A nuclease isolated from spinach, called SP, is a particularly intriguing CEL I ortholog, showing properties intermediate between CEL I and S1. It has a strong preference for AT-rich regions, yet is able to cut single-base mismatches and has a neutral pH optimum. Unlike CEL I, it is unable to recognize mismatches with guanine bases at the mismatched site 30. We cloned this CEL I ortholog from spinach mRNA. We called this putative nuclease SP I to distinguish the properties of this clonally purified form from the native SP nuclease preparations that may contain more than one homolog that are very difficult to separate during native enzyme purification. The SP I mRNA sequence was deposited to GenBank under accession no. [GenBank:EF032908].

SP I and CEL I amino acid sequences are 71.1% identical. Amino acid sequence alignment of SP I with other known S1-like nucleases reveals that all residues identified by structural studies 28 as crucial for binding of the three Zn²⁺ atoms and for catalysis are preserved in SP I. Interestingly, the nucleotide binding site shows less sequence conservation (Figure 1A). A significant divergence was observed in a fragment within a loop located close to the ligand, comprising residues 134–139 in SP I and 127–132 in P1, respectively. The H135 residue is one of the most prominent sequence features within this fragment of the SP I nuclease, compared with CEL I and P1. To test if H135 is important for the SP I mismatch-specific nuclease activity, we produced a recombinant virus expressing a H135A mutant of the SP I nuclease.

Infection of a Sf9 cell culture with recombinant viruses containing the CEL I or SP I genes under control of a constitutive promoter resulted in accumulation of a single-strand specific nuclease activity both in the culture media and cell extract (Figure 1C). This activity adhered to a Ni²⁺ affinity column and was eluted with 150 mM imidazole. A single major nuclease band was observed when the partially purified nuclease preparations were separated on a SDS PAGE, in-gel refolded, and stained for single-strand specific nuclease activity (Figure 1B). This activity co-migrated with a native CEL I control (purified from celery), implying that the recombinant enzyme contains a similar quantity of N-linked glycans. We also detected the recombinant protein by a Western blot experiment with an anti-hexahistidine monoclonal antibody (not shown). The Ni²⁺ affinity column-purified nucleases were stable on ice for at least a week and infinitely stable when stored in 50% glycerol at -20°C. The enzyme activities were reduced by freezing/thawing cycles, decreasing by roughly 50% after each cycle (data not shown).

To test the efficacy of our recombinant nuclease preparations in mutation detection, a highly polymorphic section of exon 11 of BRCA1 was used as the substrate 8. A PCR product derived from one patient contained three single base pair polymorphisms as revealed by a control experiment with CEL I purified from celery. This substrate is challenging because of multiple single-base substitutions in close proximity to each other, a quality that would render many mutation detection techniques ineffective 9. While the mismatches 2196 G → A and 2430 C → T were well detected by all our recombinant nuclease preparations (Figure 2), little cutting of the nucleotide substitution T → C at position 2201 of BRCA1 was observed, reflected by the low signal from the 300 nt long fragment (Figure 2). We also observed an additional 303–304 nt peak which may have originated from processing of the 305 nt peak by CEL I and SP I exonuclease activities. Control experiments have indicated efficient cutting of T → C at position 2201 of BRCA1 by native CEL I. Since this result was reproduced in all our recombinant SP I and CEL I preparations, it indicates that the mismatch sequence preference and possibly the balance between the exo- and endonuclease activities of the expressed enzymes is slightly different from their native counterparts purified from plant tissues. Surprisingly, unlike its native counterpart 30, the recombinant SP I was capable of introducing nicks specifically 3' of an extrahelical G nucleotide (Figure 3). This result was reproduced on several preparations of SP I and confirmed by mass spectrometry analysis of the incised heteroduplex substrates to exclude the possibility of a non-specific action of SP I on its substrate (Figure 4), further indicating that mismatch preferences of CEL I orthologs can be modified by recombinant expression.

SP I ^H135A was an active nuclease, with mismatch recognition properties similar to those of the wild type SP I (Figures 2 and 3), indicating that H135 is dispensable for the mismatch nuclease activity of SP I.

Total RNA was prepared from store-bought fresh spinach (Spinacia oleracea Melody hybrid) leaves using the phenol SDS procedure for plant RNA extraction as described 32. Stratagene's Pro-Star First Strand RT-PCR kit was used to synthesize first-strand cDNA. We used the CEL I nuclease amino acid sequence [GenBank:AAF42954] 5 to construct two pairs of degenerate primers that allowed amplification of SP I cDNA in two segments. The resulting products were cloned in a TA vector using the TA Cloning^® Kit (Invitrogen), and sequenced with the use of vector-specific primers. By using 5' and 3' RACE technology (Stratagene), sequences of the 3' and 5' SP I mRNA coding regions were obtained. A pair of primers (sequences 5' TTTCAATGTCGCGTTCTACT and 5' AGTCCTAAACATTGGAAGCC) and Pfu DNA polymerase were used to amplify the entire protein-coding region of SP I cDNA which was cloned in the pCR^®2.1 TA vector (Invitrogen), yielding the pSP plasmid. The entire insert in the pSP plasmid was sequenced using vector-specific primers and the SP I cDNA sequence was deposited to GenBank under accession No. [GenBank:EF032908].

A pair of primers (5' GGG

CTCGAG

GGTACC

GTGGTGGTGGTGGTGGTG

CTCGAG

GGTACC

GTGGTGGTGGTGGTGGTG

The QuikChange mutagenesis reaction (Stratagene) to create the pAcSPmut vector expressing a H135A mutant of SP I nuclease was conducted in accordance with manufacturer's recommendations. (For convenience, amino acid numbering throughout the manuscript is given with respect to the putative mature proteins starting with N-terminal tryptophan and lacking signal peptides. H135 of putative mature SP I corresponds to H158 of the expressed sequence.) A pair of complementary oligonucleotides was used: 5' GATATTCATCAGCCAATGCATTGCGCG

GCG

CGC

All tissue culture procedures, co-transfection and virus amplification were done according to Pharmingen recommendations 33. Briefly, monolayer Sf9 cultures were co-transformed with an expression plasmid and BaculoGold Bright linearized DNA. The recombinant virus produced was amplified twice. Fluorescent microscopy was used to inspect the cells for the presence of GFP which was the marker of infection. Flow cytometry was used to assess virus titers by an end-point dilution assay. Monolayer cultures of Sf9 cells grown in TNM-FH medium were used for protein expression. In a typical experiment 5 × 10⁷ cells were infected with 6 ml of ~1 × 10⁸ pfu/ml amplified virus stock. Three days after infection the cell extract and culture medium were analyzed for plasmid nicking activity. Hexahistidine-tagged proteins were then purified on a HIS-Select Ni⁺⁺ column (Sigma) from the cell culture media. The crude medium was passed through a 0.22 μm filter (Millipore), diluted two-fold with Equilibration/Wash buffer (50 mM Tris-HCl, pH 7.6, 300 mM NaCl, and 10 μM ZnCl₂), and loaded on a column that had been equilibrated with the same buffer. After loading, the column was washed with Equilibration/Wash buffer and then with 50 mM Tris-HCl, pH 7.6, 300 mM NaCl, 10 μM ZnCl₂, 5 mM imidazole. Nucleases were eluted with 50 mM Tris-HCl, pH 7.6, 300 mM NaCl, 10 μM ZnCl₂, and 150 mM imidazole.

Human genomic DNA, purified from blood samples from patients participating in the Margaret Dyson/Family Risk Assessment Program, was obtained from the Fox Chase Cancer Center Biorepository with approval of the Institutional Review Committee (protocol #00-824). A pair of primers specific for exon 11.4 of the BRCA1 gene (sequences 5' CCTTCCCTAGAGTGCTAAC and 5' CCCACCTAATTGTACTGAA) were synthesized with Cy5 fluorescent label at the 5' end of the forward primer and Cy5.5 label at the 5' end of the reverse primer. Twenty μl PCR reactions included 2 μl 10× PCR buffer (Applied Biosystems), 5% DMSO, 2 mM MgCl₂, 0.2 mM each dNTP, 0.0375 μM each fluorescent primer, 100 ng human genomic DNA template and 0.2 U AmpliTaq Gold DNA polymerase (Applied Biosystems). The thermal cycling protocol consisted of a 5 min initial denaturation step at 94°C, followed by 35 cycles of (denaturation at 94°C for 10 s, annealing at 55°C for 20 s and elongation at 72°C for 1 min). PCR amplification products were heated to 94°C and gradually cooled to 4°C to allow formation of heteroduplexes. The resulting fluorescent substrates were incubated with recombinant nuclease preparations at 45°C for 60 min in CEL I reaction buffer (20 mM HEPES, pH 7.5, 3 mM MgCl₂, 10 mM KCl), purified using the CEQ8000 ethanol-glycogen cleanup procedure (Beckman) and separated on Beckman CEQ8000 Genetic Analysis System according to the manufacturer's protocol.

The oligonucleotides for making the mismatched substrates were synthesized in the Fox Chase Cancer Center Fannie E. Rippel Biotechnology Facility and PAGE-purified. DNA heteroduplex substrates containing A or G extrahelical loops were constructed by annealing a 5'-labeled oligonucleotide to a partially complementary cold nucleotide as shown in Figure 3B. Prior to annealing, the singe-stranded oligonucleotides were labeled at the 5'-termini with T4 polynucleotide kinase and [γ-³²P]ATP. One hundred fmol of a heteroduplex substrate was incubated with recombinant nuclease preparations in 20 μl reaction volume in CEL I reaction buffer (20 mM HEPES, pH 7.5, 10 mM KCl, 3 mM MgCl₂). Taq DNA polymerase (0.5 Units) was added to stimulate the mismatch-specific activity of CEL I and SP I 6. The reactions were performed at 45°C for 1 h, terminated with formamide and analyzed on a denaturing PAGE gel. Autoradiography was used to visualize radioactive bands.

Unlabeled heteroduplex oligonucleotide substrate was constructed as shown in Figure 4. The CEL I mismatch endonuclease assay was performed as described above. The reaction was stopped with EDTA and the reaction products were desalted using C18 ZipTip (Millipore Corporation) before mass spectral analysis. The ZipTip pre-concentration and AnchorChip (Bruker Daltonics) technique for MALDI spotting were employed. 3-HPA (3-hydroxypicolinic acid) was used as the MALDI matrix. 0.5 μl of 1% 3-HPA, 0.1% diammonium hydrogen citrate was applied onto 400 μm spot on the anchor plate and allowed to dry. 0.5 μl of desalted and pre-concentrated oligonucleotide reaction products was applied onto the matrix crystals. Mass spectra were acquired on a MALDI-TOF-MS Reflex IV instrument (Bruker Daltonics) in a linear delayed pulse ion extraction mode. The oligonucleotides were desorbed and ionized by a nitrogen pulsed laser with a wavelength of 337 nm. Internal calibration was carried out using singly and doubly charged ions from the full-length oligonucleotide substrate.