Mismatch cleavage by single-strand specific nucleases
http://www.100md.com
《核酸研究医学期刊》
1 Basic Sciences Division and 3 Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, WA 98109, USA and 2 Department of Biology, Box 355325, University of Washington, Seattle, WA 98195, USA
*To whom correspondence should be addressed. Tel: +1 206 667 4515; Fax: +1 206 667 5889; Email: steveh@fhcrc.org
ABSTRACT
We have investigated the ability of single-strand specific (sss) nucleases from different sources to cleave single base pair mismatches in heteroduplex DNA templates used for mutation and single-nucleotide polymorphism analysis. The TILLING (Targeting Induced Local Lesions IN Genomes) mismatch cleavage protocol was used with the LI-COR gel detection system to assay cleavage of amplified heteroduplexes derived from a variety of induced mutations and naturally occurring polymorphisms. We found that purified nucleases derived from celery (CEL I), mung bean sprouts and Aspergillus (S1) were able to specifically cleave nearly all single base pair mismatches tested. Optimal nicking of heteroduplexes for mismatch detection was achieved using higher pH, temperature and divalent cation conditions than are routinely used for digestion of single-stranded DNA. Surprisingly, crude plant extracts performed as well as the highly purified preparations for this application. These observations suggest that diverse members of the S1 family of sss nucleases act similarly in cleaving non-specifically at bulges in heteroduplexes, and single-base mismatches are the least accessible because they present the smallest single-stranded region for enzyme binding. We conclude that a variety of sss nucleases and extracts can be effectively used for high-throughput mutation and polymorphism discovery.
INTRODUCTION
Single-strand specific (sss) nucleases related to S1 nuclease from Aspergillus oryzae are among the most versatile tools in molecular biology (1). They have been used in several applications to remove single-stranded regions from DNA or RNA while leaving duplex regions intact (1–5). The most widely used sss nucleases are extracellular glycoproteins that have been purified from several sources, including S1 and P1 nucleases from fungi (6–8) and mung bean and CEL I nucleases from plants (9,10). S1 and mung bean nucleases have been popular for more than two decades for a variety of applications, such as removing single-strand overhangs from resected DNA without significantly degrading duplexed regions (11). Sequencing of genes encoding many of these enzymes has revealed them to be homologous (1), with common active site motifs that coordinate three zinc ions that are essential for catalysis (12). Therefore, it would be expected that these enzymes will behave similarly in acting on nucleic acid templates.
Despite sharing key residues and having similar predicted structures, sss nucleases have been reported to vary widely in their ability to cleave single base pair mismatches (2,13–17). This application is of practical importance given the potential for using sss nucleases to detect single nucleotide polymorphisms (SNPs) and point mutations in heteroduplexes between mutant and wild-type templates (18,19). Although some early studies reported successes in using sss nucleases to cleave mismatched heteroduplexes (2,13,20–22), there have been reports of failures and sequence preferences that discourage the general use of these sss nucleases for SNP and mutation screening applications (10,14,16,18,23). An important advance was made by Yeung and colleagues, who showed that the CEL I enzyme purified from celery was able to cleave all single base pair mismatches, albeit with different sensitivities (15). This finding encouraged us to use CEL I for high-throughput detection of chemically induced point mutations (‘TILLING’: Targeting Induced Local Lesions IN Genomes) and natural polymorphisms (‘Ecotilling’) with the LI-COR gel analysis system (5,24,25). Currently, several efforts are underway to use this technique for mutation and natural polymorphism detection in a variety of organisms, including Arabidopsis and other plants and zebrafish and other animals (19,26,27).
Yeung and colleagues also showed that they could detect activities similar to CEL I from a variety of plant extracts (15). They argued that these CEL I-like activities are different from previously characterized enzymes belonging to the S1 nuclease family, such as mung bean nuclease, which in their hands was unable to cleave single base pair mismatches (10). They concluded that members of the CEL I group of sss nucleases constitute a distinct class of sss nucleases that are specific for single base pair mismatches. However, the sporadic successes and failures in cleaving single-base mismatches using S1, mung bean and some other sss nucleases led us to wonder whether the failures resulted from the use of insensitive assays or inadequate reaction conditions. We have tested this possibility and have found that all preparations tested, whether purified enzymes or unpurified extracts, can be optimized to cleave single-base mismatches in heteroduplexes when assayed by TILLING. Our results have general implications for the mechanism of action of sss nucleases and for the optimization of sss nucleases for genomic applications.
MATERIALS AND METHODS
Preparation of sss nucleases
CEL I nuclease was purified as described (28). To prepare celery juice extract (CJE), only the extraction, salting out and dialysis steps of the purification protocol were performed. Specifically, 0.5 kg of store-bought celery was juiced at 4°C, adjusted to 0.1 M Tris–HCl, pH 7.7, 100 μM PMSF, and spun for 20 min at 2600 g to pellet debris. The supernatant was brought to 25% saturation in (NH4)2SO4, mixed for 30 min at 4°C and spun at 13 000–16 200 g at 4°C for 40 min. The resulting supernatant was adjusted to 80% (NH4)2SO4, mixed for 30 min at 4°C and spun at 13 000–16 200 g for 1.5 h. The pellet was suspended in 0.1 M Tris–HCl, pH 7.7, 100 μM PMSF (1/10 starting volume). The suspension was transferred to a dialysis tube (Spectra/Por? 10 000 MW cut-off) and dialyzed against a total of 32 l of the same buffer with four changes over 4 h. Aliquots of extract were stored at –20°C.
Mung bean extract (MBE) was prepared by following the first steps of a standard protocol for mung bean nuclease purification (9). Briefly, 0.35 kg of fresh store-bought mung bean sprouts was ground in a Waring blender at 4°C together with 100 ml cold water, filtered through cheesecloth, brought to 20% saturation in (NH4)2SO4 and stirred for 20 min at 4°C. Ten grams of Celite 545 (Fisher) was added to the mixture, stirred briefly and filtered through Whatman? 4 filter paper using a Buchner funnel. The solution was brought to 50% saturation in (NH4)2SO4, stirred for 20 min and filtered in the presence of Celite 545 as before. This solution was brought to 80% saturation in (NH4)2SO4 and filtered using a Whatman? 52 hardened filter disc in the presence of 2 g of Celite 545. Precipitate was suspended in 45 ml of 30 mM ammonium acetate buffer pH 6.0 and dialyzed against the same buffer for 20 h with two buffer changes (8 l total). The solution was centrifuged at 2600 g for 20 min at 4°C. The supernatant was adjusted to 0.1 M NaCl, cooled to 0°C, and brought to 35% (v/v) ethanol by addition of 95% ethanol at 0°C. The solution was stored at 4°C for 20 min and centrifuged for 20 min at 4°C at 2600 g. The supernatant was cooled to –20°C, adjusted to 55% (v/v) ethanol by addition of 95% ethanol at –20°C, and centrifuged for 90 min at –17°C at 13 000–16 200 g. The pellet was suspended in cold water, centrifuged for 30 min at 27 000 g at 0°C, and dialyzed against 0.02 M ammonium acetate pH 6.0. Dialyzed aliquots were stored at –20°C.
SurveyorTM CEL I was obtained from Transgenomic, mung bean nuclease from New England Biolabs (Catalog M0250L, Batches 21 and 22 yielded similar results) and S1 nuclease from Fermentas (Catalog EN0321, lot 1312). Other commercial sources of mung bean and S1 nucleases were tested, but no qualitative differences were detected.
Units of nuclease activities have previously been based on the ability to digest single-stranded nucleic acids under standard conditions. Because this reaction is minimized under conditions that may favor mismatch detection, we define a TILLING unit (1 U) as the amount of enzyme activity per ml that provides best mismatch cleavage detection under conditions optimized for purified enzyme. We measured 1 TILLING U in 1.07 μl CEL I stock, 1.65 μl SurveyorTM Lot 1, 1.65 μl CJE, 2.7 μl mung bean nuclease (New England Biolabs Batch 22, 10 U/μl), 2.8 μl MBE and 0.75 μl S1 (Fermentas, 100 U/μl).
Genomic DNA templates and primer design
Arabidopsis thaliana genomic DNAs with confirmed ethylmethanesulfonate (EMS)-induced point mutations in the OXI1 gene (At3g25250, NM113431) were obtained from individuals represented in the DNA library used by the Arabidopsis TILLING Project (24), a public service for reverse genetics. The 992 bp target fragment was chosen as a representative example that showed good PCR amplification among the gene targets whose TILLING was requested by users of the service. The following PCR primers were used: TACGCGGC GGAGCTTGTATTAGCA (left) and CCATTTGAAGC CAAGGTCCAACGA (right). Each DNA sample used as template was derived from a single Arabidopsis M2 plant and harbors a single G:CA:T point mutation in the target region (Table 1).
Table 1. Sequence context of mutations and polymorphisms
Arabidopsis thaliana genomic DNA templates containing natural polymorphisms in the PIF2 gene (At5g24500, AY530749 ) were obtained from the previously described Ecotilling library of Arabidopsis ecotype DNA (25). The 1002 bp target fragment was chosen as being representative of those used for Ecotilling in that study. The following PCR primers were used: CAACCGACGATGACGATGCTTCTG (left) and CCTTCGGCTGACATTGCTGCTTTC (right). Seven of the eight tested DNA samples used in this study contain one or more polymorphisms in the target region, and one contains a G:CA:T EMS-induced mutation (Table 1).
PCR and nuclease reactions
For each trial, PCR was performed in a 10 μl volume with 0.075 ng genomic DNA as previously described (28). For samples from homozygotes, DNA was mixed 1:1 with Col-er (sample G) or Col-0 (samples J–P) wild-type reference DNA to generate heteroduplexes for polymorphism detection.
Reactions using CEL I nuclease, CJE and SurveyorTM nuclease were carried out as described (5,28). Trials with other nucleases followed the standard TILLING protocol, varying conditions one at a time to systematically explore parameter space. Digestions were performed by addition of 20 μl of an ice-cold solution containing 1.5x enzyme–buffer cocktail, supplemented with 0.003% Triton X100 and 0.0003 mg/ml BSA, to 10 μl PCR product in 96-well microtiter plates on ice, mixed and incubated in a heating block for digestion. Enzyme–buffer cocktails contained varying concentrations of enzyme, buffer (HEPES, Bis–Tris or potassium acetate), salt, divalent cation and tetramethylene sulfoxide (TMSO) added as a helix destabilizing agent (29). Reactions were stopped by the addition of 5 μl of 225 mM EDTA, desalted using Sephadex G50, heated at 90°C to reduce the volume and robotically loaded onto a 100-tooth membrane comb for LI-COR gel analysis as previously described (5).
RESULTS
Fungal and plant sss nucleases related to S1 share common features
Reports that S1 and mung bean nucleases fail to cleave single base pair mismatches led us to scrutinize the sequence relationships among sss nucleases belonging to the S1 nuclease family. We searched the GenBank non-redundant protein sequence databank with BLASTP (http://www.ncbi.nlm.nih.gov/blast/ January, 2004) using the 287 amino acid sequence of S1 nuclease (BAA08310 as query and detected related fungal, plant and bacterial sequences 300 amino acids long that globally align with S1 nuclease. A phylogeny was derived from a global alignment of these sequences (Fig. 1A). Although no sequence for mung bean nuclease was found in our PSI-BLAST search, several plant proteins, including CEL I, were detected. Single enzymes from five different fungal genomes were found, and these form a single clade, as do the plant (and the bacterial) genes. Therefore, the fungal nucleases appear to be orthologous to one another, and to share a single common ancestor with all of the plant genes.
Figure 1. (A) Phylogenetic tree of fungal (boxed), bacterial (underlined) and plant members of the S1 family of sss nucleases. ClustalW (35) was used to produce a global alignment and a neighbor-joining bootstrap tree (excluding gapped regions), which was displayed using TreeView (http://taxonomy.zoology. gla.ac.uk/rod/rod.html). Bootstrap percentages >70% are shown, based on 1000 trials. Rice A–C and A.thaliana BFN2-4 sequence numbers are arbitrary. (B) Alignment of S1 and CEL I nucleases with the two catalytic regions in P1 nuclease, with identical residues shaded. Based on the P1 structure (12), critical active site residues that coordinate zinc ions are indicated by carets. GenBank accession numbers for sequences used are: CEL I, AAF42954 A.thaliana BFN1, NP_172585 ; BFN2, NP_567631 .1; BFN3, PIR_T05167; BFN4, PIR_T05168; ZEN1, BAA28948 1; ZEN2, AAD00694 1; ZEN3, AAD00695 1; rice A, BAB03377 1; rice B, CAE04161 3; rice C, NP_909099 .1; barley, BAA82696 1; rice B, CAE04161 3; Neurospora, XP_331586 .1; Magnaporthe, EAA47610 1; P1, NUP1_PENCI; S1, NUS1_ASPOR; Mushroom, PIR_JC7275; Xanthomonas, NP_638544 ; Chromobacterium, NP_899730 .
Plant genomes encode multiple members of the S1 nuclease family, which fall into at least two clades (30), each of which includes proteins from both monocots and dicots (Fig. 1A). For the clade that includes CEL I (Apium graveolens CEL I, Zinnia elegans ZEN1, Arabidopsis thaliana BFN1 and Rice A), increasing evolutionary distances are found to conform with the known phylogeny, indicating orthology. Perez-Amador et al. (30) have shown that the BFN1 and ZEN1 nucleases are induced and secreted during senescence, consistent with roles in degradation of RNA and DNA for nutrient salvage following cell death. The ZEN2 and ZEN3 genes, which are, respectively, only 49% and 44% identical to ZEN1, are also induced during senescence (30). The likelihood that these genes are functionally redundant is consistent with the lack of phenotype observed for three different T-DNA insertions into A.thaliana BFN-1 (http://arabidopsis.org, locus AT1G11190). These observations suggest that diverse plant enzymes, including CEL I, BFN1, ZEN1-3 and mung bean nucleases, have common functions in senescence. Therefore, we find no evidence from phylogeny or function that CEL I belongs to a different class of enzymes from other plant or fungal nucleases.
Similarities among these nucleases are especially apparent at the active site, where critical catalytic residues are identical between CEL I, S1 and P1, whose 3-D structure is known (Fig. 1B). In addition, CEL I shares with its fungal and plant counterparts four critical cysteine residues involved in disulfide bond formation and two of the asparagine residues that are glycosylated in P1. So despite the fact that CEL I is only 25–26% identical overall to S1 and P1, the pattern of divergence is as expected for enzymes with similar basic properties.
A sensitive system for assaying cleavage of single base pair mismatches by sss nucleases
A possible explanation for the observed differences in substrate preferences for various members of the S1 nuclease family is that they were not tested using a sufficiently sensitive assay. Our high-throughput TILLING method is a robust assay for single base pair mismatches, having been shown to reliably detect all such mismatches regardless of sequence context, even in pools (31). Using this system and varying enzyme concentrations and reaction conditions, we expected to be able to determine whether previous failures resulted from basic differences between enzymes in their substrate preferences or from limitations of the assays used. One limitation is that suitable reaction conditions might substantially differ for these enzymes considering their divergent amino acid sequences.
To assay a wide range of single base pair mismatches and sequence contexts, we used two types of targets. One contains only EMS-induced mutations, which are G:CA:T transitions, presumably the most challenging class of mismatches to detect. We chose a representative region of the Arabidopsis genome where we had previously identified a large number of mismatches within a 1-kb fragment amplified from the OXI1 gene (Table 1). A second target, a 1-kb fragment within the PIF2 gene, was amplified from induced and naturally occurring variants of different types. After PCR amplification, we subjected these target heteroduplex fragments to cleavage by CEL 1, mung bean and S1 nucleases, varying the concentration and reaction conditions to optimize for mutation detection. A typical concentration series for CEL 1 nuclease using the PIF2 test fragment is shown in Figure 2.
Figure 2. CEL I concentration dependence for detection of single base pair mismatches. Increasing concentrations of CEL I nuclease were added to three PCR-amplified and heteroduplexed PIF2 test samples (J, K and L in Table 1) using standard TILLING conditions and LI-COR-based detection. For each CEL I concentration tested, all possible single base pair mismatches were evaluated (see Table 1). Panels from left to right are no enzyme, 0.1, 0.33, 1, 3 and 10 U/ml. Both IRD700 (left) and IRD800 (right) channels are shown to display the size distribution of bands for each labeled strand. Bands resulting from confirmed mismatch cleavages are marked (arrows). Cleavages occur on either strand at a mismatched base, yielding product sizes that add up to that of the full-length band (992 bp, marked by circles). A prominent example of a sporadic mispriming product is seen in the first lane of the 0.33 U/ml sample, identified as comigrating bands seen in both channels (marked by asterisks). Some contamination of adjacent lanes is not uncommon, such as the first lane in the 1 U/ml sample that contaminates the lane to the right.
LI-COR gel images from untreated samples reveal a dense background of fragments smaller than full-length product, evidently resulting from incomplete polymerase extension and/or mispriming (Fig. 2, 0 U samples). Treatment with a low concentration of CEL I has only a minor effect on the background and on the amount of full-length product. Background bands likely result from incomplete polymerase extension during PCR; these vary between templates both in pattern and intensity (5). At a low level of digestion, mismatch cleavage is easily detectable (Fig. 2, 0.1 U samples). As the concentration of CEL I is raised, background bands and full-length product decrease in intensity while mismatch cleavage products accumulate (Fig. 2, 0.33 and 1 U samples). A slow migrating smear sharply decreases in intensity as enzyme concentrations increase, presumably resulting from digestion of PCR product networks that had remained annealed despite the denaturing conditions used for gel electrophoresis. When an optimal CEL I concentration is reached (Fig. 2, 1 U samples), the contrast between mismatch cleavage signal and background is maximal. As CEL I concentration is further increased, all products begin to fade (Fig. 2, 3 and 10 U samples). The fact that the size distribution of fragments does not change indicates that fading results from the loss of end label as opposed to internal cleavage. Importantly, we detected only minor intensity variations in cleavage of all different mismatches, as though CEL I shows little if any mismatch preference in our assay.
Efficient cleavage of single base pair mismatches by sss nucleases and crude extracts
In addition to the CEL I that we purified from celery, we tested commercially prepared CEL I (SurveyorTM), mung bean and S1 nucleases and crude extracts from juiced celery (CJE) and from ground mung bean sprouts (MBE) for their ability to recognize and cleave single base pair G:CA:T mismatches. We used standard TILLING assays to find optimal concentrations for these enzyme preparations (Table 2 and data not shown) and to explore different digestion conditions. To optimize the efficiency of cleavage by mung bean nuclease, we tested different pHs, divalent cations, temperatures, ionic strengths, anions, additives and reaction times at varying concentrations. We also performed limited optimization tests on S1 nuclease. For both enzymes, we found conditions that allowed detection of G:CA:T mismatches in the heteroduplexed OXI1 test fragment. Using standard conditions established previously for TILLING, CEL I, SurveyorTM and CJE consistently detected all eight mismatches with comparable efficiency (Fig. 3). Commercial mung bean nuclease provided similar detection efficiencies, showing band and background intensities that appeared comparable to those for CJE, despite being incubated at different pHs, temperatures and divalent cation concentrations. All eight bands were also detected using MBE under conditions optimized for mung bean nuclease, but the efficiency was reduced relative to that of commercial mung bean nuclease. With S1 nuclease, six of eight bands were detected, usually at lower efficiency than for the other enzymes. We conclude that all enzyme preparations are capable of detecting mismatches in heteroduplexes consisting of only single-base transition mutations within 1-kb fragments.
Table 2. Activities of sss nucleases under different reaction conditions
Figure 3. Comparison of sss nuclease preparations for detection of G:CA:T transitions. OXI1 heteroduplexes (A–H in Table 1) were digested and products displayed as in Figure 2, except that optimized enzyme–buffer cocktails and incubation conditions were used. Enzymes are present at 1 U/ml. From left to right: CEL I in CEL I buffer (10 mM KCl, 10 mM MgSO4, 10 mM HEPES pH 7.5, 0.002% Triton X100 and 0.0002 mg/ml BSA), 45°C, 15 min; SurveyorTM in CEL I buffer, 45°C, 15 min; CJE in CEL I buffer, 45°C, 15 min; MBE in Bis–Tris buffer (10 mM MgSO4, 0.2 mM ZnSO4, 20 mM Bis–Tris pH 6.5, 0.002% Triton X100 and 0.0002 mg/ml BSA), 60°C, 30 min; mung bean nuclease in Bis–Tris buffer, 60°C, 30 min; S1 nuclease in 2 mM ZnCl2, 10 mM potassium acetate pH 5.5, 10 mM MgSO4, 0.002% Triton X100 and 0.0002 mg/ml BSA, 45°C, 30 min. Top, IRD700; bottom, IRD800, where the image gain was increased overall for clarity.
Efficient cleavage of different mismatches by sss nucleases
To determine if the enzyme preparations could be used to detect other types of nucleotide polymorphisms, we tested seven samples that together contained all different single base pair mismatches and three different deletions in a 1-kb amplicon (Table 1). We used the same digestion conditions as for detection of single G:CA:T mismatches in the OXI1 amplicon and followed the standard TILLING protocol. Overall results for PIF2 were similar to what we found with OXI1 in comparing the different enzyme preparations. Polymorphisms were easily detected in the seven test samples for all CEL I and mung bean nuclease preparations (Fig. 4). In the case of S1 nuclease, detection was generally weaker, although 10 of 11 tested polymorphisms were consistently detected in both channels. Interestingly, the single failure was for a CG transversion (C:C and G:G mismatches) embedded in a stretch of 10 consecutive G/C base pairs, which is by far the most G/C-rich stretch tested (Table 1). It is possible that under the low Mg++ conditions that we used, S1 nuclease was less able to gain access to either strand for cleavage, and that optimization of S1 nuclease for cleavage of this extremely G/C-rich mismatch will improve its robustness.
Figure 4. Comparison of sss nuclease preparations for detection of mismatches and loop-outs. All different mismatches (A:C and G:T in lanes 1, 2 and 7, T:C and C:A in lanes 2 and 7, C:C and G:G in lane 3, T:T and A:A in lane 4), three deletions (9 bp in lane 5, 3 bp in lane 6 and 2 bp in lane 7) and >30 polymorphisms (lane 8), are represented in the two channels. PIF2 heteroduplexes (I–P in Table 1) were digested and products displayed as in Figures 2 and 3, with the full-length product at the top. Conditions are the same as for Figure 3. Top, IRD700; bottom, IRD800, where the image gain was increased overall for clarity.
One of the samples in our assay was of a rare PIF2 allele that showed an anomalous degree of polymorphism. In our initial Ecotilling study of this allele, we detected only a subset of polymorphisms using purified CEL I (25). Interestingly, digestion with CJE, MBE and mung bean nuclease all led to improved detection. We attribute these improvements to the lower overall degree of digestion by these sss nuclease preparations, as evident from the higher level of full-length product that remains. Such detection differences emphasize the trade-off that was made when optimizing conditions for single transition mutations. We found that signal-over-noise was best for CEL I under conditions in which the amount of full-length product was strongly reduced relative to undigested product (Fig. 2), but for mung bean nuclease, signal-over-noise was found to be best when more of the full-length product remained intact. This signal-over-noise trade-off probably reflects competing reactions in which sss nucleases cleave mismatched duplexes while simultaneously eroding their ends (32). Maximizing signal-over-noise in face of competing reactions is challenging, and encourages further exploration of sss nuclease digestion conditions for Ecotilling and TILLING applications.
DISCUSSION
We have shown that different members of the S1 family of sss nucleases can be used to detect single base pair mismatches in heteroduplexes between wild-type and mutant or polymorphic amplicons. Robust and sensitive detection using the TILLING system did not appear to depend on the sequence context or the identity of the mismatch, but depended strongly on the digestion conditions used. For example, we found robust mismatch cleavage by mung bean nuclease to occur over a pH range centered around pH 6.5 at 60°C, whereas Yeung and co-workers found that CEL I cleaves mismatches throughout a pH range centered around pH 8 at 37–45°C (10,15). These differences in pH and temperature optima for the two enzymes could account for the inability to detect mismatch cleavage by mung bean or S1 nucleases using conditions that are optimal for CEL I (10,23), or for failures to detect mismatches using mung bean nuclease at pH 5.5 (10) and S1 nuclease at pH 4.5 (14). When optimized for mismatch cleavage, mung bean nuclease behaved similarly to CEL I nuclease on all tested heteroduplexes. Although S1 nuclease performed less well than the two plant enzymes under tested conditions (at pH 5.5 in low Mg++), it nevertheless detected 13 of the 16 single-base mismatches in the different heteroduplex fragments tested. Such differences in pH and other optima of the kind that we encountered have been documented in previous studies of the action of sss nucleases on various templates (1,16,33). We conclude that there is no fundamental difference between CEL I and other members of the S1 nuclease family in their action on single-base mismatches.
We found that crude extracts from celery and mung bean sprouts performed as well as the purified enzymes from these sources for TILLING. In fact, we have been using the same crude extract from celery exclusively for more than a year in our high-throughput service, resulting in the discovery of more than 1500 EMS-induced mutations (http://tilling.fhcrc.org:9366/). Such robustness is surprising insofar as plant extracts have been reported to contain several presumably undesirable activities, including other endo- and exo-nucleases. There are at least two potential explanations for the effectiveness of crude extracts in mismatch detection. It is possible that the very high concentration and mismatch cleavage activity of sss nuclease in these extracts relative to undesirable activities permits the direct use of extracts without purification. Alternatively, some feature of our assay system might favor mismatch cleavage over other forms of digestion. These two possibilities are not mutually exclusive.
One possible way that our assay system might favor mismatch cleavage over confounding activities of sss nuclease preparations is in the protection of ends from being ‘nibbled’, a feature of all tested sss nucleases. The loss of end-labeled templates during digestion appears to limit mismatch detection in our system. It is possible that the hydrophobic IRD dyes at each 5' end perform an inadvertent end-blocking function, stabilizing perfectly base-paired heteroduplexes to end-cleavage activities, whereas imperfectly base-paired PCR failure products that constitute the background are not as well stabilized and so are digested. Yeung and colleagues had noted that addition of Taq polymerase improved mismatch detection (15). No such improvement was seen when using CEL I for TILLING, perhaps because binding of Taq polymerase to ends of DNA duplexes acts to protect them, and the presence of Taq in a standard TILLING reaction carried over from the amplification step may suffice. Nibbling would not be a limiting factor when sss nucleases are used to cleave heteroduplexes that are detected on agarose electrophoretic gels by ethidium bromide staining (16,17), because the slight shortening that removes the terminal 5' base will be almost imperceptible. In such cases, higher concentrations of enzyme are needed to cleave both strands simultaneously, which increases both the rate of nibbling and the number of ends that would be attacked during the reaction. We have found that celery extracts are also capable of double-strand cleavage when assayed by agarose gel analysis (data not shown), an activity that has been documented for the SurveyorTM preparation of CEL I. This is consistent with the abundance of sss nuclease activities in plant extracts that are capable of efficient mismatch cleavage.
The consistent action of sss nucleases on base pair mismatches and their lack of sequence preference provides support for a model of their action that was first put forward in 1976. Kroeker et al. (32) proposed that the active site of mung bean nuclease cannot accommodate a DNA strand that is perfectly base paired, but transient disruptions of base pairing will allow occasional binding and cleavage. Such disruptions would occur regularly at the ends of duplexes, leading to nibbling of 5' and 3' ends in alternation. Their model can also account for nicking of supercoiled plasmids, for cleavage at sites of Y junctions and damaged base pairs and for digestion of looped-out regions by these enzymes (1). Single base pair mismatches are the smallest bulged-out regions and are the least susceptible to cleavage (13). Unless conditions are optimal for sss nucleases to bind and cleave these small bulges, nibbling of end-labeled duplexes predominates, decreasing sensitivity. The acid pH conditions that are optimal for all known relatives of S1 nuclease to digest single-stranded DNA favor duplex degradation, but by raising the pH, duplexes are stabilized (32,34). Thus, more neutral pH conditions that stabilize duplexes but still allow small bulges to be cleaved are likely to be best for mismatch detection. S1, mung bean and CEL I nucleases may be sufficiently different in pH optima and other properties that they differ in the conditions needed to obtain an optimal balance between cleavage of small bulges and nibbling of duplexes. The relative lack of sequence preference is understandable if only steric hindrance of binding by base pairing limits mismatch cleavage.
The ability of a variety of sss nucleases to robustly cleave mismatches in our TILLING assay has implications for genetic technologies. At present, high-throughput TILLING appears to be the method of choice for mutation scanning, and Ecotilling has potential for SNP discovery applications (19). By default, our TILLING project uses 8-fold pooling and obtains consistent detection of heterozygous bands present at 1/16th the level of wild-type (31); higher levels of pooling would increase efficiency and lower the cost of the overall process. Higher efficiency might be achieved in part from continuing improvements in PCR technology that would reduce the intensity of the nuclease-resistant background banding pattern (Fig. 2). As we have demonstrated here, optimization of sss nuclease conditions can also improve detection by increasing mismatch cleavage relative to end erosion. The availability of several different sources of nucleases that act similarly, but under different reaction conditions, increases opportunities to improve detection efficiency and robustness. For example, our finding that crude celery extract and mung bean nuclease appear to work better than purified CEL I when there are large numbers of mismatches makes these the sss nuclease preparations of choice for TILLING and Ecotilling of highly polymorphic regions. We anticipate that our growing understanding of the mechanism by which these enzymes discriminate between mismatches and duplexes will result in improved genomic methodologies.
ACKNOWLEDGEMENTS
We thank Faith Hassinger, Christine Codomo and other members of the Seattle TILLING Project team for input and technical assistance. Our work is supported by grants from the Plant Genome Research Program and the Arabidopsis 2010 Project of the National Science Foundation.
REFERENCES
Desai,N.A. and Shankar,V. (2003) Single-strand-specific nucleases. FEMS Microbiol. Rev., 26, 457–491.
Shenk,T.E., Rhodes,C., Rigby,P.W. and Berg,P. (1975) Biochemical method for mapping mutational alterations in DNA with S1 nuclease: the location of deletions and temperature-sensitive mutations in simian virus 40. Proc. Natl Acad. Sci. USA, 72, 989–993.
Wiegand,R.C., Godson,N. and Radding,C.M. (1975) Specificity of the S1 nuclease from Aspergillus oryzae. J. Biol. Chem., 250, 8848–8855.
Berk,A.J. (1989) Characterization of RNA molecules by S1 nuclease analysis. Methods Enzymol., 180, 334–347.
Colbert,T., Till,B.J., Tompa,R., Reynolds,S., Steine,M.N., Yeung,A.T., McCallum,C.M., Comai,L. and Henikoff,S. (2001) High-throughput screening for induced point mutations. Plant Physiol., 126, 480–484.
Vogt,V.M. (1973) Purification and further properties of single-strand-specific nuclease from Aspergillus oryzae. Eur. J. Biochem., 33, 192–200.
Fujimoto,M., Kuninaka,A. and Yoshino,H. (1974) Purification of a nuclease from Penicillium citrinum. Agric. Biol. Chem., 38, 777–783.
Vogt,V.M. (1980) Purification and properties of S1 nuclease from Aspergillus. Methods Enzymol., 65, 248–255.
Laskowski,M.,Sr (1980) Purification and properties of the mung bean nuclease. Methods Enzymol., 65, 263–276.
Yang,B., Wen,X., Kodali,N.S., Oleykowski,C.A., Miller,C.G., Kulinski,J., Besack,D., Yeung,J.A., Kowalski,D. and Yeung,A.T. (2000) Purification, cloning and characterization of the CEL I nuclease. Biochemistry, 39, 3533–3541.
Henikoff,S. (1984) Unidirectional digestion with Exonuclease III creates targeted breakpoints for DNA sequencing. Gene, 28, 351–359.
Volbeda,A., Lahm,A., Sakiyama,F. and Suck,D. (1991) Crystal structure of Penicillium citrinum P1 nuclease at 2.8? resolution. EMBO J., 10, 1607–1618.
Dodgson,J.B. and Wells,R.D. (1977) Action of single-strand specific nucleases on model DNA heteroduplexes of defined size and sequence. Biochemistry, 16, 2374–2379.
Silber,J.R. and Loeb,L.A. (1981) S1 nuclease does not cleave DNA at single-base mismatches. Biochim. Biophys. Acta, 656, 256–264.
Oleykowski,C.A., Bronson Mullins,C.R., Godwin,A.K. and Yeung,A.T. (1998) Mutation detection using a novel plant endonuclease. Nucleic Acids Res., 26, 4597–4602.
Howard,J.T., Ward,J., Watson,J.N. and Roux,K.H. (1999) Heteroduplex cleavage analysis using S1 nuclease. Biotechniques, 27, 18–19.
Sokurenko,E.V., Tchesnokova,V., Yeung,A.T., Oleykowski,C.A., Trintchina,E., Hughes,K.T., Rashid,R.A., Brint,J.M., Moseley,S.L. and Lory,S. (2001) Detection of simple mutations and polymorphisms in large genomic regions. Nucleic Acids Res., 29, e111.
Taylor,G.R. and Deeble,J. (1999) Enzymatic methods for mutation scanning. Genet. Anal., 14, 181–186.
Henikoff,S. and Comai,L. (2003) Single-nucleotide mutations for plant functional genomics. Annu. Rev. Plant Biol., 54, 375–401.
Ahmad,A., Holloman,W.K. and Holliday,R. (1975) Nuclease that preferentially inactivates DNA containing mismatched bases. Nature, 258, 54–56.
Pukkila,P.J. (1978) The recognition of mismatched base pairs in DNA by DNase I from Ustilago maydis. Mol. Gen. Genet., 161, 245–250.
Bartok,K., Fraser,M.J. and Fareed,G.C. (1974) Detection of sequence heterology by use of the N. crassa nucleases. Biochem. Biophys. Res. Commun., 60, 507–514.
Oleykowski,C.A., Bronson Mullins,C.R., Chang,D.W. and Yeung,A.T. (1999) Incision at nucleotide insertions/deletions and base pair mismatches by the SP nuclease of spinach. Biochemistry, 38, 2200–2205.
Till,B.J., Reynolds,S.H., Greene,E.A., Codomo,C.A., Enns,L.C., Johnson,J.E., Burtner,C., Odden,A.R., Young,K., Taylor,N.E., Henikoff,J.G., Comai,L. and Henikoff,S. (2003) Large-scale discovery of induced point mutations with high throughput TILLING. Genome Res., 13, 524–530.
Comai,L., Young,K., Till,B.J., Reynolds,S.H., Greene,E.A., Codomo,C.A., Enns,L.C., Johnson,J.E., Burtner,C., Odden,A.R. and Henikoff,S. (2004) Efficient discovery of DNA polymorphisms in natural populations by ecotilling. Plant J., 37, 778–786.
Wienholds,E., van Eeden,F.J., Kosters,M., Mudde,J., Plasterk,R.H. and Cuppen,E. (2003) Efficient target-selected mutagenesis in zebrafish. Genome Res., 13, 2700–2707.
Smits,B.M. G., Mudde,J., Plasterk,R.H. A. and Cuppen,E. (2004) Target-selected mutagenesis of the rat. Genomics, 83, 332–334.
Till,B.J., Colbert,T., Tompa,R., Enns,L.C., Codomo,C.A., Johnson,J.E., Reynolds,S.H., Henikoff,J.G., Greene,E.A., Stein,M.N., Comai,L. and Henikoff,S. (2003) High-throughput TILLING for functional genomics. In Grotewald,E. (ed.), Plant Functional Genomics: Methods and Protocols. Humana Press, Clifton, NJ, pp. 205–220.
Chakrabarti,R. and Schutt,C.E. (2002) Novel sulfoxides facilitate GC-rich template amplification. Biotechniques, 32, 866–874.
Perez-Amador,M.A., Abler,M.L., De Rocher,E.J., Thompson,D.M., van Hoof,A., LeBrasseur,N.D., Lers,A. and Green,P.J. (2000) Identification of BFN1, a bifunctional nuclease induced during leaf and stem senescence in Arabidopsis. Plant Physiol., 122, 169–180.
Greene,E.A., Codomo,C.A., Taylor,N.E., Henikoff,J.G., Till,B.J., Reynolds,S.H., Enns,L.C., Burtner,C., Johnson,J.E., Odden,A.R., Comai,L. and Henikoff,S. (2003) Spectrum of chemically induced mutations from a large-scale reverse-genetic screen in Arabidopsis. Genetics, 164, 731–740.
Kroeker,W.D., Kowalski,D. and Laskowski,M.,Sr (1976) Mung bean nuclease I. Terminally directed hydrolysis of native DNA. Biochemistry, 15, 4463–4467.
Esteban,J.A., Salas,M. and Blanco,L. (1992) Activation of S1 nuclease at neutral pH. Nucleic Acids Res., 20, 4932.
Kowalski,D. and Sanford,J.P. (1982) Action of mung bean nuclease on supercoiled PM2 DNA. J. Biol. Chem., 257, 7820–7825.
Thompson,J.D., Higgins,D.G. and Gibson,T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res., 22, 4673–4680.(Bradley J. Till1, Chris Burtner2, Luca C)
*To whom correspondence should be addressed. Tel: +1 206 667 4515; Fax: +1 206 667 5889; Email: steveh@fhcrc.org
ABSTRACT
We have investigated the ability of single-strand specific (sss) nucleases from different sources to cleave single base pair mismatches in heteroduplex DNA templates used for mutation and single-nucleotide polymorphism analysis. The TILLING (Targeting Induced Local Lesions IN Genomes) mismatch cleavage protocol was used with the LI-COR gel detection system to assay cleavage of amplified heteroduplexes derived from a variety of induced mutations and naturally occurring polymorphisms. We found that purified nucleases derived from celery (CEL I), mung bean sprouts and Aspergillus (S1) were able to specifically cleave nearly all single base pair mismatches tested. Optimal nicking of heteroduplexes for mismatch detection was achieved using higher pH, temperature and divalent cation conditions than are routinely used for digestion of single-stranded DNA. Surprisingly, crude plant extracts performed as well as the highly purified preparations for this application. These observations suggest that diverse members of the S1 family of sss nucleases act similarly in cleaving non-specifically at bulges in heteroduplexes, and single-base mismatches are the least accessible because they present the smallest single-stranded region for enzyme binding. We conclude that a variety of sss nucleases and extracts can be effectively used for high-throughput mutation and polymorphism discovery.
INTRODUCTION
Single-strand specific (sss) nucleases related to S1 nuclease from Aspergillus oryzae are among the most versatile tools in molecular biology (1). They have been used in several applications to remove single-stranded regions from DNA or RNA while leaving duplex regions intact (1–5). The most widely used sss nucleases are extracellular glycoproteins that have been purified from several sources, including S1 and P1 nucleases from fungi (6–8) and mung bean and CEL I nucleases from plants (9,10). S1 and mung bean nucleases have been popular for more than two decades for a variety of applications, such as removing single-strand overhangs from resected DNA without significantly degrading duplexed regions (11). Sequencing of genes encoding many of these enzymes has revealed them to be homologous (1), with common active site motifs that coordinate three zinc ions that are essential for catalysis (12). Therefore, it would be expected that these enzymes will behave similarly in acting on nucleic acid templates.
Despite sharing key residues and having similar predicted structures, sss nucleases have been reported to vary widely in their ability to cleave single base pair mismatches (2,13–17). This application is of practical importance given the potential for using sss nucleases to detect single nucleotide polymorphisms (SNPs) and point mutations in heteroduplexes between mutant and wild-type templates (18,19). Although some early studies reported successes in using sss nucleases to cleave mismatched heteroduplexes (2,13,20–22), there have been reports of failures and sequence preferences that discourage the general use of these sss nucleases for SNP and mutation screening applications (10,14,16,18,23). An important advance was made by Yeung and colleagues, who showed that the CEL I enzyme purified from celery was able to cleave all single base pair mismatches, albeit with different sensitivities (15). This finding encouraged us to use CEL I for high-throughput detection of chemically induced point mutations (‘TILLING’: Targeting Induced Local Lesions IN Genomes) and natural polymorphisms (‘Ecotilling’) with the LI-COR gel analysis system (5,24,25). Currently, several efforts are underway to use this technique for mutation and natural polymorphism detection in a variety of organisms, including Arabidopsis and other plants and zebrafish and other animals (19,26,27).
Yeung and colleagues also showed that they could detect activities similar to CEL I from a variety of plant extracts (15). They argued that these CEL I-like activities are different from previously characterized enzymes belonging to the S1 nuclease family, such as mung bean nuclease, which in their hands was unable to cleave single base pair mismatches (10). They concluded that members of the CEL I group of sss nucleases constitute a distinct class of sss nucleases that are specific for single base pair mismatches. However, the sporadic successes and failures in cleaving single-base mismatches using S1, mung bean and some other sss nucleases led us to wonder whether the failures resulted from the use of insensitive assays or inadequate reaction conditions. We have tested this possibility and have found that all preparations tested, whether purified enzymes or unpurified extracts, can be optimized to cleave single-base mismatches in heteroduplexes when assayed by TILLING. Our results have general implications for the mechanism of action of sss nucleases and for the optimization of sss nucleases for genomic applications.
MATERIALS AND METHODS
Preparation of sss nucleases
CEL I nuclease was purified as described (28). To prepare celery juice extract (CJE), only the extraction, salting out and dialysis steps of the purification protocol were performed. Specifically, 0.5 kg of store-bought celery was juiced at 4°C, adjusted to 0.1 M Tris–HCl, pH 7.7, 100 μM PMSF, and spun for 20 min at 2600 g to pellet debris. The supernatant was brought to 25% saturation in (NH4)2SO4, mixed for 30 min at 4°C and spun at 13 000–16 200 g at 4°C for 40 min. The resulting supernatant was adjusted to 80% (NH4)2SO4, mixed for 30 min at 4°C and spun at 13 000–16 200 g for 1.5 h. The pellet was suspended in 0.1 M Tris–HCl, pH 7.7, 100 μM PMSF (1/10 starting volume). The suspension was transferred to a dialysis tube (Spectra/Por? 10 000 MW cut-off) and dialyzed against a total of 32 l of the same buffer with four changes over 4 h. Aliquots of extract were stored at –20°C.
Mung bean extract (MBE) was prepared by following the first steps of a standard protocol for mung bean nuclease purification (9). Briefly, 0.35 kg of fresh store-bought mung bean sprouts was ground in a Waring blender at 4°C together with 100 ml cold water, filtered through cheesecloth, brought to 20% saturation in (NH4)2SO4 and stirred for 20 min at 4°C. Ten grams of Celite 545 (Fisher) was added to the mixture, stirred briefly and filtered through Whatman? 4 filter paper using a Buchner funnel. The solution was brought to 50% saturation in (NH4)2SO4, stirred for 20 min and filtered in the presence of Celite 545 as before. This solution was brought to 80% saturation in (NH4)2SO4 and filtered using a Whatman? 52 hardened filter disc in the presence of 2 g of Celite 545. Precipitate was suspended in 45 ml of 30 mM ammonium acetate buffer pH 6.0 and dialyzed against the same buffer for 20 h with two buffer changes (8 l total). The solution was centrifuged at 2600 g for 20 min at 4°C. The supernatant was adjusted to 0.1 M NaCl, cooled to 0°C, and brought to 35% (v/v) ethanol by addition of 95% ethanol at 0°C. The solution was stored at 4°C for 20 min and centrifuged for 20 min at 4°C at 2600 g. The supernatant was cooled to –20°C, adjusted to 55% (v/v) ethanol by addition of 95% ethanol at –20°C, and centrifuged for 90 min at –17°C at 13 000–16 200 g. The pellet was suspended in cold water, centrifuged for 30 min at 27 000 g at 0°C, and dialyzed against 0.02 M ammonium acetate pH 6.0. Dialyzed aliquots were stored at –20°C.
SurveyorTM CEL I was obtained from Transgenomic, mung bean nuclease from New England Biolabs (Catalog M0250L, Batches 21 and 22 yielded similar results) and S1 nuclease from Fermentas (Catalog EN0321, lot 1312). Other commercial sources of mung bean and S1 nucleases were tested, but no qualitative differences were detected.
Units of nuclease activities have previously been based on the ability to digest single-stranded nucleic acids under standard conditions. Because this reaction is minimized under conditions that may favor mismatch detection, we define a TILLING unit (1 U) as the amount of enzyme activity per ml that provides best mismatch cleavage detection under conditions optimized for purified enzyme. We measured 1 TILLING U in 1.07 μl CEL I stock, 1.65 μl SurveyorTM Lot 1, 1.65 μl CJE, 2.7 μl mung bean nuclease (New England Biolabs Batch 22, 10 U/μl), 2.8 μl MBE and 0.75 μl S1 (Fermentas, 100 U/μl).
Genomic DNA templates and primer design
Arabidopsis thaliana genomic DNAs with confirmed ethylmethanesulfonate (EMS)-induced point mutations in the OXI1 gene (At3g25250, NM113431) were obtained from individuals represented in the DNA library used by the Arabidopsis TILLING Project (24), a public service for reverse genetics. The 992 bp target fragment was chosen as a representative example that showed good PCR amplification among the gene targets whose TILLING was requested by users of the service. The following PCR primers were used: TACGCGGC GGAGCTTGTATTAGCA (left) and CCATTTGAAGC CAAGGTCCAACGA (right). Each DNA sample used as template was derived from a single Arabidopsis M2 plant and harbors a single G:CA:T point mutation in the target region (Table 1).
Table 1. Sequence context of mutations and polymorphisms
Arabidopsis thaliana genomic DNA templates containing natural polymorphisms in the PIF2 gene (At5g24500, AY530749 ) were obtained from the previously described Ecotilling library of Arabidopsis ecotype DNA (25). The 1002 bp target fragment was chosen as being representative of those used for Ecotilling in that study. The following PCR primers were used: CAACCGACGATGACGATGCTTCTG (left) and CCTTCGGCTGACATTGCTGCTTTC (right). Seven of the eight tested DNA samples used in this study contain one or more polymorphisms in the target region, and one contains a G:CA:T EMS-induced mutation (Table 1).
PCR and nuclease reactions
For each trial, PCR was performed in a 10 μl volume with 0.075 ng genomic DNA as previously described (28). For samples from homozygotes, DNA was mixed 1:1 with Col-er (sample G) or Col-0 (samples J–P) wild-type reference DNA to generate heteroduplexes for polymorphism detection.
Reactions using CEL I nuclease, CJE and SurveyorTM nuclease were carried out as described (5,28). Trials with other nucleases followed the standard TILLING protocol, varying conditions one at a time to systematically explore parameter space. Digestions were performed by addition of 20 μl of an ice-cold solution containing 1.5x enzyme–buffer cocktail, supplemented with 0.003% Triton X100 and 0.0003 mg/ml BSA, to 10 μl PCR product in 96-well microtiter plates on ice, mixed and incubated in a heating block for digestion. Enzyme–buffer cocktails contained varying concentrations of enzyme, buffer (HEPES, Bis–Tris or potassium acetate), salt, divalent cation and tetramethylene sulfoxide (TMSO) added as a helix destabilizing agent (29). Reactions were stopped by the addition of 5 μl of 225 mM EDTA, desalted using Sephadex G50, heated at 90°C to reduce the volume and robotically loaded onto a 100-tooth membrane comb for LI-COR gel analysis as previously described (5).
RESULTS
Fungal and plant sss nucleases related to S1 share common features
Reports that S1 and mung bean nucleases fail to cleave single base pair mismatches led us to scrutinize the sequence relationships among sss nucleases belonging to the S1 nuclease family. We searched the GenBank non-redundant protein sequence databank with BLASTP (http://www.ncbi.nlm.nih.gov/blast/ January, 2004) using the 287 amino acid sequence of S1 nuclease (BAA08310 as query and detected related fungal, plant and bacterial sequences 300 amino acids long that globally align with S1 nuclease. A phylogeny was derived from a global alignment of these sequences (Fig. 1A). Although no sequence for mung bean nuclease was found in our PSI-BLAST search, several plant proteins, including CEL I, were detected. Single enzymes from five different fungal genomes were found, and these form a single clade, as do the plant (and the bacterial) genes. Therefore, the fungal nucleases appear to be orthologous to one another, and to share a single common ancestor with all of the plant genes.
Figure 1. (A) Phylogenetic tree of fungal (boxed), bacterial (underlined) and plant members of the S1 family of sss nucleases. ClustalW (35) was used to produce a global alignment and a neighbor-joining bootstrap tree (excluding gapped regions), which was displayed using TreeView (http://taxonomy.zoology. gla.ac.uk/rod/rod.html). Bootstrap percentages >70% are shown, based on 1000 trials. Rice A–C and A.thaliana BFN2-4 sequence numbers are arbitrary. (B) Alignment of S1 and CEL I nucleases with the two catalytic regions in P1 nuclease, with identical residues shaded. Based on the P1 structure (12), critical active site residues that coordinate zinc ions are indicated by carets. GenBank accession numbers for sequences used are: CEL I, AAF42954 A.thaliana BFN1, NP_172585 ; BFN2, NP_567631 .1; BFN3, PIR_T05167; BFN4, PIR_T05168; ZEN1, BAA28948 1; ZEN2, AAD00694 1; ZEN3, AAD00695 1; rice A, BAB03377 1; rice B, CAE04161 3; rice C, NP_909099 .1; barley, BAA82696 1; rice B, CAE04161 3; Neurospora, XP_331586 .1; Magnaporthe, EAA47610 1; P1, NUP1_PENCI; S1, NUS1_ASPOR; Mushroom, PIR_JC7275; Xanthomonas, NP_638544 ; Chromobacterium, NP_899730 .
Plant genomes encode multiple members of the S1 nuclease family, which fall into at least two clades (30), each of which includes proteins from both monocots and dicots (Fig. 1A). For the clade that includes CEL I (Apium graveolens CEL I, Zinnia elegans ZEN1, Arabidopsis thaliana BFN1 and Rice A), increasing evolutionary distances are found to conform with the known phylogeny, indicating orthology. Perez-Amador et al. (30) have shown that the BFN1 and ZEN1 nucleases are induced and secreted during senescence, consistent with roles in degradation of RNA and DNA for nutrient salvage following cell death. The ZEN2 and ZEN3 genes, which are, respectively, only 49% and 44% identical to ZEN1, are also induced during senescence (30). The likelihood that these genes are functionally redundant is consistent with the lack of phenotype observed for three different T-DNA insertions into A.thaliana BFN-1 (http://arabidopsis.org, locus AT1G11190). These observations suggest that diverse plant enzymes, including CEL I, BFN1, ZEN1-3 and mung bean nucleases, have common functions in senescence. Therefore, we find no evidence from phylogeny or function that CEL I belongs to a different class of enzymes from other plant or fungal nucleases.
Similarities among these nucleases are especially apparent at the active site, where critical catalytic residues are identical between CEL I, S1 and P1, whose 3-D structure is known (Fig. 1B). In addition, CEL I shares with its fungal and plant counterparts four critical cysteine residues involved in disulfide bond formation and two of the asparagine residues that are glycosylated in P1. So despite the fact that CEL I is only 25–26% identical overall to S1 and P1, the pattern of divergence is as expected for enzymes with similar basic properties.
A sensitive system for assaying cleavage of single base pair mismatches by sss nucleases
A possible explanation for the observed differences in substrate preferences for various members of the S1 nuclease family is that they were not tested using a sufficiently sensitive assay. Our high-throughput TILLING method is a robust assay for single base pair mismatches, having been shown to reliably detect all such mismatches regardless of sequence context, even in pools (31). Using this system and varying enzyme concentrations and reaction conditions, we expected to be able to determine whether previous failures resulted from basic differences between enzymes in their substrate preferences or from limitations of the assays used. One limitation is that suitable reaction conditions might substantially differ for these enzymes considering their divergent amino acid sequences.
To assay a wide range of single base pair mismatches and sequence contexts, we used two types of targets. One contains only EMS-induced mutations, which are G:CA:T transitions, presumably the most challenging class of mismatches to detect. We chose a representative region of the Arabidopsis genome where we had previously identified a large number of mismatches within a 1-kb fragment amplified from the OXI1 gene (Table 1). A second target, a 1-kb fragment within the PIF2 gene, was amplified from induced and naturally occurring variants of different types. After PCR amplification, we subjected these target heteroduplex fragments to cleavage by CEL 1, mung bean and S1 nucleases, varying the concentration and reaction conditions to optimize for mutation detection. A typical concentration series for CEL 1 nuclease using the PIF2 test fragment is shown in Figure 2.
Figure 2. CEL I concentration dependence for detection of single base pair mismatches. Increasing concentrations of CEL I nuclease were added to three PCR-amplified and heteroduplexed PIF2 test samples (J, K and L in Table 1) using standard TILLING conditions and LI-COR-based detection. For each CEL I concentration tested, all possible single base pair mismatches were evaluated (see Table 1). Panels from left to right are no enzyme, 0.1, 0.33, 1, 3 and 10 U/ml. Both IRD700 (left) and IRD800 (right) channels are shown to display the size distribution of bands for each labeled strand. Bands resulting from confirmed mismatch cleavages are marked (arrows). Cleavages occur on either strand at a mismatched base, yielding product sizes that add up to that of the full-length band (992 bp, marked by circles). A prominent example of a sporadic mispriming product is seen in the first lane of the 0.33 U/ml sample, identified as comigrating bands seen in both channels (marked by asterisks). Some contamination of adjacent lanes is not uncommon, such as the first lane in the 1 U/ml sample that contaminates the lane to the right.
LI-COR gel images from untreated samples reveal a dense background of fragments smaller than full-length product, evidently resulting from incomplete polymerase extension and/or mispriming (Fig. 2, 0 U samples). Treatment with a low concentration of CEL I has only a minor effect on the background and on the amount of full-length product. Background bands likely result from incomplete polymerase extension during PCR; these vary between templates both in pattern and intensity (5). At a low level of digestion, mismatch cleavage is easily detectable (Fig. 2, 0.1 U samples). As the concentration of CEL I is raised, background bands and full-length product decrease in intensity while mismatch cleavage products accumulate (Fig. 2, 0.33 and 1 U samples). A slow migrating smear sharply decreases in intensity as enzyme concentrations increase, presumably resulting from digestion of PCR product networks that had remained annealed despite the denaturing conditions used for gel electrophoresis. When an optimal CEL I concentration is reached (Fig. 2, 1 U samples), the contrast between mismatch cleavage signal and background is maximal. As CEL I concentration is further increased, all products begin to fade (Fig. 2, 3 and 10 U samples). The fact that the size distribution of fragments does not change indicates that fading results from the loss of end label as opposed to internal cleavage. Importantly, we detected only minor intensity variations in cleavage of all different mismatches, as though CEL I shows little if any mismatch preference in our assay.
Efficient cleavage of single base pair mismatches by sss nucleases and crude extracts
In addition to the CEL I that we purified from celery, we tested commercially prepared CEL I (SurveyorTM), mung bean and S1 nucleases and crude extracts from juiced celery (CJE) and from ground mung bean sprouts (MBE) for their ability to recognize and cleave single base pair G:CA:T mismatches. We used standard TILLING assays to find optimal concentrations for these enzyme preparations (Table 2 and data not shown) and to explore different digestion conditions. To optimize the efficiency of cleavage by mung bean nuclease, we tested different pHs, divalent cations, temperatures, ionic strengths, anions, additives and reaction times at varying concentrations. We also performed limited optimization tests on S1 nuclease. For both enzymes, we found conditions that allowed detection of G:CA:T mismatches in the heteroduplexed OXI1 test fragment. Using standard conditions established previously for TILLING, CEL I, SurveyorTM and CJE consistently detected all eight mismatches with comparable efficiency (Fig. 3). Commercial mung bean nuclease provided similar detection efficiencies, showing band and background intensities that appeared comparable to those for CJE, despite being incubated at different pHs, temperatures and divalent cation concentrations. All eight bands were also detected using MBE under conditions optimized for mung bean nuclease, but the efficiency was reduced relative to that of commercial mung bean nuclease. With S1 nuclease, six of eight bands were detected, usually at lower efficiency than for the other enzymes. We conclude that all enzyme preparations are capable of detecting mismatches in heteroduplexes consisting of only single-base transition mutations within 1-kb fragments.
Table 2. Activities of sss nucleases under different reaction conditions
Figure 3. Comparison of sss nuclease preparations for detection of G:CA:T transitions. OXI1 heteroduplexes (A–H in Table 1) were digested and products displayed as in Figure 2, except that optimized enzyme–buffer cocktails and incubation conditions were used. Enzymes are present at 1 U/ml. From left to right: CEL I in CEL I buffer (10 mM KCl, 10 mM MgSO4, 10 mM HEPES pH 7.5, 0.002% Triton X100 and 0.0002 mg/ml BSA), 45°C, 15 min; SurveyorTM in CEL I buffer, 45°C, 15 min; CJE in CEL I buffer, 45°C, 15 min; MBE in Bis–Tris buffer (10 mM MgSO4, 0.2 mM ZnSO4, 20 mM Bis–Tris pH 6.5, 0.002% Triton X100 and 0.0002 mg/ml BSA), 60°C, 30 min; mung bean nuclease in Bis–Tris buffer, 60°C, 30 min; S1 nuclease in 2 mM ZnCl2, 10 mM potassium acetate pH 5.5, 10 mM MgSO4, 0.002% Triton X100 and 0.0002 mg/ml BSA, 45°C, 30 min. Top, IRD700; bottom, IRD800, where the image gain was increased overall for clarity.
Efficient cleavage of different mismatches by sss nucleases
To determine if the enzyme preparations could be used to detect other types of nucleotide polymorphisms, we tested seven samples that together contained all different single base pair mismatches and three different deletions in a 1-kb amplicon (Table 1). We used the same digestion conditions as for detection of single G:CA:T mismatches in the OXI1 amplicon and followed the standard TILLING protocol. Overall results for PIF2 were similar to what we found with OXI1 in comparing the different enzyme preparations. Polymorphisms were easily detected in the seven test samples for all CEL I and mung bean nuclease preparations (Fig. 4). In the case of S1 nuclease, detection was generally weaker, although 10 of 11 tested polymorphisms were consistently detected in both channels. Interestingly, the single failure was for a CG transversion (C:C and G:G mismatches) embedded in a stretch of 10 consecutive G/C base pairs, which is by far the most G/C-rich stretch tested (Table 1). It is possible that under the low Mg++ conditions that we used, S1 nuclease was less able to gain access to either strand for cleavage, and that optimization of S1 nuclease for cleavage of this extremely G/C-rich mismatch will improve its robustness.
Figure 4. Comparison of sss nuclease preparations for detection of mismatches and loop-outs. All different mismatches (A:C and G:T in lanes 1, 2 and 7, T:C and C:A in lanes 2 and 7, C:C and G:G in lane 3, T:T and A:A in lane 4), three deletions (9 bp in lane 5, 3 bp in lane 6 and 2 bp in lane 7) and >30 polymorphisms (lane 8), are represented in the two channels. PIF2 heteroduplexes (I–P in Table 1) were digested and products displayed as in Figures 2 and 3, with the full-length product at the top. Conditions are the same as for Figure 3. Top, IRD700; bottom, IRD800, where the image gain was increased overall for clarity.
One of the samples in our assay was of a rare PIF2 allele that showed an anomalous degree of polymorphism. In our initial Ecotilling study of this allele, we detected only a subset of polymorphisms using purified CEL I (25). Interestingly, digestion with CJE, MBE and mung bean nuclease all led to improved detection. We attribute these improvements to the lower overall degree of digestion by these sss nuclease preparations, as evident from the higher level of full-length product that remains. Such detection differences emphasize the trade-off that was made when optimizing conditions for single transition mutations. We found that signal-over-noise was best for CEL I under conditions in which the amount of full-length product was strongly reduced relative to undigested product (Fig. 2), but for mung bean nuclease, signal-over-noise was found to be best when more of the full-length product remained intact. This signal-over-noise trade-off probably reflects competing reactions in which sss nucleases cleave mismatched duplexes while simultaneously eroding their ends (32). Maximizing signal-over-noise in face of competing reactions is challenging, and encourages further exploration of sss nuclease digestion conditions for Ecotilling and TILLING applications.
DISCUSSION
We have shown that different members of the S1 family of sss nucleases can be used to detect single base pair mismatches in heteroduplexes between wild-type and mutant or polymorphic amplicons. Robust and sensitive detection using the TILLING system did not appear to depend on the sequence context or the identity of the mismatch, but depended strongly on the digestion conditions used. For example, we found robust mismatch cleavage by mung bean nuclease to occur over a pH range centered around pH 6.5 at 60°C, whereas Yeung and co-workers found that CEL I cleaves mismatches throughout a pH range centered around pH 8 at 37–45°C (10,15). These differences in pH and temperature optima for the two enzymes could account for the inability to detect mismatch cleavage by mung bean or S1 nucleases using conditions that are optimal for CEL I (10,23), or for failures to detect mismatches using mung bean nuclease at pH 5.5 (10) and S1 nuclease at pH 4.5 (14). When optimized for mismatch cleavage, mung bean nuclease behaved similarly to CEL I nuclease on all tested heteroduplexes. Although S1 nuclease performed less well than the two plant enzymes under tested conditions (at pH 5.5 in low Mg++), it nevertheless detected 13 of the 16 single-base mismatches in the different heteroduplex fragments tested. Such differences in pH and other optima of the kind that we encountered have been documented in previous studies of the action of sss nucleases on various templates (1,16,33). We conclude that there is no fundamental difference between CEL I and other members of the S1 nuclease family in their action on single-base mismatches.
We found that crude extracts from celery and mung bean sprouts performed as well as the purified enzymes from these sources for TILLING. In fact, we have been using the same crude extract from celery exclusively for more than a year in our high-throughput service, resulting in the discovery of more than 1500 EMS-induced mutations (http://tilling.fhcrc.org:9366/). Such robustness is surprising insofar as plant extracts have been reported to contain several presumably undesirable activities, including other endo- and exo-nucleases. There are at least two potential explanations for the effectiveness of crude extracts in mismatch detection. It is possible that the very high concentration and mismatch cleavage activity of sss nuclease in these extracts relative to undesirable activities permits the direct use of extracts without purification. Alternatively, some feature of our assay system might favor mismatch cleavage over other forms of digestion. These two possibilities are not mutually exclusive.
One possible way that our assay system might favor mismatch cleavage over confounding activities of sss nuclease preparations is in the protection of ends from being ‘nibbled’, a feature of all tested sss nucleases. The loss of end-labeled templates during digestion appears to limit mismatch detection in our system. It is possible that the hydrophobic IRD dyes at each 5' end perform an inadvertent end-blocking function, stabilizing perfectly base-paired heteroduplexes to end-cleavage activities, whereas imperfectly base-paired PCR failure products that constitute the background are not as well stabilized and so are digested. Yeung and colleagues had noted that addition of Taq polymerase improved mismatch detection (15). No such improvement was seen when using CEL I for TILLING, perhaps because binding of Taq polymerase to ends of DNA duplexes acts to protect them, and the presence of Taq in a standard TILLING reaction carried over from the amplification step may suffice. Nibbling would not be a limiting factor when sss nucleases are used to cleave heteroduplexes that are detected on agarose electrophoretic gels by ethidium bromide staining (16,17), because the slight shortening that removes the terminal 5' base will be almost imperceptible. In such cases, higher concentrations of enzyme are needed to cleave both strands simultaneously, which increases both the rate of nibbling and the number of ends that would be attacked during the reaction. We have found that celery extracts are also capable of double-strand cleavage when assayed by agarose gel analysis (data not shown), an activity that has been documented for the SurveyorTM preparation of CEL I. This is consistent with the abundance of sss nuclease activities in plant extracts that are capable of efficient mismatch cleavage.
The consistent action of sss nucleases on base pair mismatches and their lack of sequence preference provides support for a model of their action that was first put forward in 1976. Kroeker et al. (32) proposed that the active site of mung bean nuclease cannot accommodate a DNA strand that is perfectly base paired, but transient disruptions of base pairing will allow occasional binding and cleavage. Such disruptions would occur regularly at the ends of duplexes, leading to nibbling of 5' and 3' ends in alternation. Their model can also account for nicking of supercoiled plasmids, for cleavage at sites of Y junctions and damaged base pairs and for digestion of looped-out regions by these enzymes (1). Single base pair mismatches are the smallest bulged-out regions and are the least susceptible to cleavage (13). Unless conditions are optimal for sss nucleases to bind and cleave these small bulges, nibbling of end-labeled duplexes predominates, decreasing sensitivity. The acid pH conditions that are optimal for all known relatives of S1 nuclease to digest single-stranded DNA favor duplex degradation, but by raising the pH, duplexes are stabilized (32,34). Thus, more neutral pH conditions that stabilize duplexes but still allow small bulges to be cleaved are likely to be best for mismatch detection. S1, mung bean and CEL I nucleases may be sufficiently different in pH optima and other properties that they differ in the conditions needed to obtain an optimal balance between cleavage of small bulges and nibbling of duplexes. The relative lack of sequence preference is understandable if only steric hindrance of binding by base pairing limits mismatch cleavage.
The ability of a variety of sss nucleases to robustly cleave mismatches in our TILLING assay has implications for genetic technologies. At present, high-throughput TILLING appears to be the method of choice for mutation scanning, and Ecotilling has potential for SNP discovery applications (19). By default, our TILLING project uses 8-fold pooling and obtains consistent detection of heterozygous bands present at 1/16th the level of wild-type (31); higher levels of pooling would increase efficiency and lower the cost of the overall process. Higher efficiency might be achieved in part from continuing improvements in PCR technology that would reduce the intensity of the nuclease-resistant background banding pattern (Fig. 2). As we have demonstrated here, optimization of sss nuclease conditions can also improve detection by increasing mismatch cleavage relative to end erosion. The availability of several different sources of nucleases that act similarly, but under different reaction conditions, increases opportunities to improve detection efficiency and robustness. For example, our finding that crude celery extract and mung bean nuclease appear to work better than purified CEL I when there are large numbers of mismatches makes these the sss nuclease preparations of choice for TILLING and Ecotilling of highly polymorphic regions. We anticipate that our growing understanding of the mechanism by which these enzymes discriminate between mismatches and duplexes will result in improved genomic methodologies.
ACKNOWLEDGEMENTS
We thank Faith Hassinger, Christine Codomo and other members of the Seattle TILLING Project team for input and technical assistance. Our work is supported by grants from the Plant Genome Research Program and the Arabidopsis 2010 Project of the National Science Foundation.
REFERENCES
Desai,N.A. and Shankar,V. (2003) Single-strand-specific nucleases. FEMS Microbiol. Rev., 26, 457–491.
Shenk,T.E., Rhodes,C., Rigby,P.W. and Berg,P. (1975) Biochemical method for mapping mutational alterations in DNA with S1 nuclease: the location of deletions and temperature-sensitive mutations in simian virus 40. Proc. Natl Acad. Sci. USA, 72, 989–993.
Wiegand,R.C., Godson,N. and Radding,C.M. (1975) Specificity of the S1 nuclease from Aspergillus oryzae. J. Biol. Chem., 250, 8848–8855.
Berk,A.J. (1989) Characterization of RNA molecules by S1 nuclease analysis. Methods Enzymol., 180, 334–347.
Colbert,T., Till,B.J., Tompa,R., Reynolds,S., Steine,M.N., Yeung,A.T., McCallum,C.M., Comai,L. and Henikoff,S. (2001) High-throughput screening for induced point mutations. Plant Physiol., 126, 480–484.
Vogt,V.M. (1973) Purification and further properties of single-strand-specific nuclease from Aspergillus oryzae. Eur. J. Biochem., 33, 192–200.
Fujimoto,M., Kuninaka,A. and Yoshino,H. (1974) Purification of a nuclease from Penicillium citrinum. Agric. Biol. Chem., 38, 777–783.
Vogt,V.M. (1980) Purification and properties of S1 nuclease from Aspergillus. Methods Enzymol., 65, 248–255.
Laskowski,M.,Sr (1980) Purification and properties of the mung bean nuclease. Methods Enzymol., 65, 263–276.
Yang,B., Wen,X., Kodali,N.S., Oleykowski,C.A., Miller,C.G., Kulinski,J., Besack,D., Yeung,J.A., Kowalski,D. and Yeung,A.T. (2000) Purification, cloning and characterization of the CEL I nuclease. Biochemistry, 39, 3533–3541.
Henikoff,S. (1984) Unidirectional digestion with Exonuclease III creates targeted breakpoints for DNA sequencing. Gene, 28, 351–359.
Volbeda,A., Lahm,A., Sakiyama,F. and Suck,D. (1991) Crystal structure of Penicillium citrinum P1 nuclease at 2.8? resolution. EMBO J., 10, 1607–1618.
Dodgson,J.B. and Wells,R.D. (1977) Action of single-strand specific nucleases on model DNA heteroduplexes of defined size and sequence. Biochemistry, 16, 2374–2379.
Silber,J.R. and Loeb,L.A. (1981) S1 nuclease does not cleave DNA at single-base mismatches. Biochim. Biophys. Acta, 656, 256–264.
Oleykowski,C.A., Bronson Mullins,C.R., Godwin,A.K. and Yeung,A.T. (1998) Mutation detection using a novel plant endonuclease. Nucleic Acids Res., 26, 4597–4602.
Howard,J.T., Ward,J., Watson,J.N. and Roux,K.H. (1999) Heteroduplex cleavage analysis using S1 nuclease. Biotechniques, 27, 18–19.
Sokurenko,E.V., Tchesnokova,V., Yeung,A.T., Oleykowski,C.A., Trintchina,E., Hughes,K.T., Rashid,R.A., Brint,J.M., Moseley,S.L. and Lory,S. (2001) Detection of simple mutations and polymorphisms in large genomic regions. Nucleic Acids Res., 29, e111.
Taylor,G.R. and Deeble,J. (1999) Enzymatic methods for mutation scanning. Genet. Anal., 14, 181–186.
Henikoff,S. and Comai,L. (2003) Single-nucleotide mutations for plant functional genomics. Annu. Rev. Plant Biol., 54, 375–401.
Ahmad,A., Holloman,W.K. and Holliday,R. (1975) Nuclease that preferentially inactivates DNA containing mismatched bases. Nature, 258, 54–56.
Pukkila,P.J. (1978) The recognition of mismatched base pairs in DNA by DNase I from Ustilago maydis. Mol. Gen. Genet., 161, 245–250.
Bartok,K., Fraser,M.J. and Fareed,G.C. (1974) Detection of sequence heterology by use of the N. crassa nucleases. Biochem. Biophys. Res. Commun., 60, 507–514.
Oleykowski,C.A., Bronson Mullins,C.R., Chang,D.W. and Yeung,A.T. (1999) Incision at nucleotide insertions/deletions and base pair mismatches by the SP nuclease of spinach. Biochemistry, 38, 2200–2205.
Till,B.J., Reynolds,S.H., Greene,E.A., Codomo,C.A., Enns,L.C., Johnson,J.E., Burtner,C., Odden,A.R., Young,K., Taylor,N.E., Henikoff,J.G., Comai,L. and Henikoff,S. (2003) Large-scale discovery of induced point mutations with high throughput TILLING. Genome Res., 13, 524–530.
Comai,L., Young,K., Till,B.J., Reynolds,S.H., Greene,E.A., Codomo,C.A., Enns,L.C., Johnson,J.E., Burtner,C., Odden,A.R. and Henikoff,S. (2004) Efficient discovery of DNA polymorphisms in natural populations by ecotilling. Plant J., 37, 778–786.
Wienholds,E., van Eeden,F.J., Kosters,M., Mudde,J., Plasterk,R.H. and Cuppen,E. (2003) Efficient target-selected mutagenesis in zebrafish. Genome Res., 13, 2700–2707.
Smits,B.M. G., Mudde,J., Plasterk,R.H. A. and Cuppen,E. (2004) Target-selected mutagenesis of the rat. Genomics, 83, 332–334.
Till,B.J., Colbert,T., Tompa,R., Enns,L.C., Codomo,C.A., Johnson,J.E., Reynolds,S.H., Henikoff,J.G., Greene,E.A., Stein,M.N., Comai,L. and Henikoff,S. (2003) High-throughput TILLING for functional genomics. In Grotewald,E. (ed.), Plant Functional Genomics: Methods and Protocols. Humana Press, Clifton, NJ, pp. 205–220.
Chakrabarti,R. and Schutt,C.E. (2002) Novel sulfoxides facilitate GC-rich template amplification. Biotechniques, 32, 866–874.
Perez-Amador,M.A., Abler,M.L., De Rocher,E.J., Thompson,D.M., van Hoof,A., LeBrasseur,N.D., Lers,A. and Green,P.J. (2000) Identification of BFN1, a bifunctional nuclease induced during leaf and stem senescence in Arabidopsis. Plant Physiol., 122, 169–180.
Greene,E.A., Codomo,C.A., Taylor,N.E., Henikoff,J.G., Till,B.J., Reynolds,S.H., Enns,L.C., Burtner,C., Johnson,J.E., Odden,A.R., Comai,L. and Henikoff,S. (2003) Spectrum of chemically induced mutations from a large-scale reverse-genetic screen in Arabidopsis. Genetics, 164, 731–740.
Kroeker,W.D., Kowalski,D. and Laskowski,M.,Sr (1976) Mung bean nuclease I. Terminally directed hydrolysis of native DNA. Biochemistry, 15, 4463–4467.
Esteban,J.A., Salas,M. and Blanco,L. (1992) Activation of S1 nuclease at neutral pH. Nucleic Acids Res., 20, 4932.
Kowalski,D. and Sanford,J.P. (1982) Action of mung bean nuclease on supercoiled PM2 DNA. J. Biol. Chem., 257, 7820–7825.
Thompson,J.D., Higgins,D.G. and Gibson,T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res., 22, 4673–4680.(Bradley J. Till1, Chris Burtner2, Luca C)