A WHEP Domain Regulates the Dynamic Structure and Activity of Caenorhabditis elegans Glycyl-tRNA Synthetase*

WHEP domains exist in certain eukaryotic aminoacyl-tRNA synthetases and play roles in tRNA or protein binding. We present evidence herein that cytoplasmic and mitochondrial forms of Caenorhabditis elegans glycyl-tRNA synthetase (CeGlyRS) are encoded by the same gene (CeGRS1) through alternative initiation of translation. The cytoplasmic form possessed an N-terminal WHEP domain, whereas its mitochondrial isoform possessed an extra N-terminal sequence consisting of an mitochondrial targeting signal and an appended domain. Cross-species complementation assays showed that CeGRS1 effectively rescued the cytoplasmic and mitochondrial defects of a yeast GRS1 knock-out strain. Although both forms of CeGlyRS efficiently charged the cytoplasmic tRNAsGly of C. elegans, the mitochondrial form was much more efficient than its cytoplasmic counterpart in charging the mitochondrial tRNAGly isoacceptor, which carries a defective TψC hairpin. Despite the WHEP domain per se lacking tRNA binding activity, deletion of this domain reduced the catalytic efficiency of the enzyme. Most interestingly, the deletion mutant possessed a higher thermal stability and a somewhat lower structural flexibility. Our study suggests a role for the WHEP domain as a regulator of the dynamic structure and activity of the enzyme.

Aminoacyl-tRNA synthetases (aaRSs) 3 belong to a ubiquitous and ancient family of enzymes that establish their genetic codes by attaching specific amino acids to their cognate tRNAs. The resultant aminoacyl-tRNAs are then delivered to ribosomes to decipher mRNA codons through base pairing with the anticodon of the aminoacyl-tRNA (1). Because protein translation takes place in both the cytoplasm and mitochondria in eukaryotes, two distinct sets of aaRSs are required: one functioning in the cytoplasm and the other functioning in mitochondria (1)(2)(3)(4). In most cases, the cytoplasmic and mitochondrial forms of an aaRS are encoded by two different nuclear genes. Occasionally, both isoforms of an aaRS are encoded by the same nuclear gene through alternative initiation of translation, examples of which include genes encoding yeast alanyl-, glycyl-, histidyl-, and valyl-tRNA synthetases (5)(6)(7)(8)(9). As a result, cytoplasmic and mitochondrial forms of an aaRS, for example yeast glycyl-tRNA synthetase (GlyRS), possess essentially the same polypeptide sequence, except for a cleavable mitochondrial targeting signal (MTS) attached at the N terminus of the mitochondrial precursor form.
GlyRS is one of the most intriguing aaRSs because of its divergent quaternary structure and evolutionary origin. Two distinct oligomeric structures of GlyRS exist: one with an ␣ 2 structure and the other with an ␣ 2 ␤ 2 structure (10,11). These two forms are divergent not only in subunit composition but also in molecular size and protein sequence (12)(13)(14). Even so, they possess the same signature motifs and are thus assigned to the same class (class II). To date, ␣ 2 ␤ 2 enzymes exist only in bacteria and chloroplasts, but ␣ 2 enzymes are spread over all three domains of life. The major identity elements of tRNA Gly reside in the discriminator base (N73), the top three base pairs of the acceptor stem (1:72, 2:71, and 3:70), and C35/C36 in the anticodon loop (4). Among these identity elements, the discriminator base is of particular interest. It is a U in bacteria and an A in eukaryotic cytoplasm. In general, U73-containing tRNA Gly pairs with an ␣ 2 ␤ 2 -type GlyRS enzyme, whereas A73containing tRNA Gly pairs with an ␣ 2 -type GlyRS enzyme (15). However, this once-tight rule was broken by the discovery that Thermus thermophilus possesses a U73/␣ 2 pair (15).
GlyRS has attracted enormous attention over the past decade because of its implication in Charcot-Marie-Tooth (CMT) disease, one of the most common inherited neurological disorders (16). To date, 13 missense mutations of human GlyRS were shown to cause a dominant axonal form of CMT, also known as CMT type 2D (17,18). However, there is no direct causal relationship between loss of the primary function of the enzyme and CMT disease, because not all CMT-causing mutants possess impaired aminoacylation activity (19). Recent studies suggested that GlyRS mutations also cause a CMT-like syndrome in other animals such as mice (20) and flies (21).
There is only one GlyRS gene in the nuclear genome of Caenorhabditis elegans, namely, CeGRS1. Because mitochondrial tRNAs in this organism typically possess a defective TC hairpin, the question arose as to how enzyme(s) encoded by this gene can efficiently charge such a tRNA species. Our study shows that CeGlyRS acquired two functional domains, a mitochondrial appended domain (MAD) and a WHEP domain, during evolution. The MAD enabled the enzyme to charge its cognate tRNA (containing a defective TC hairpin) with a higher efficiency, whereas the WHEP domain contributed to regulat-ing the dynamic structure of the enzyme. As a result, deletion of the WHEP domain altered the thermal stability, structural flexibility, and catalytic rate of the enzyme. These results provide new insights into understanding the structure-function relationships of CeGlyRS and its homologues.

Results
C. elegans GRS1 Is a Dual Functional Gene-Analysis of the ORF of CeGRS1 suggested that this gene is dual functional, with ATG1 and ATG65 being the respective initiator codons of the mitochondrial and cytoplasmic forms of CeGlyRS (Fig. 1A). The mitochondrial form of CeGlyRS (CeGlyRS m ) possessed a 64-residue polypeptide extension with a mitochondrial matrixprocessing peptidase cleavage site in between N-terminal resi-dues 20 and 21. The sequence containing N-terminal residues 1-20 was rich in positively charged and hydroxylated residues and was devoid of acidic residues, a feature characteristic of an MTS, and the sequence containing N-terminal residues 21-64 was specific to the mitochondrial form; this sequence is herein referred to as the MAD. The MAD was absent from all other eukaryotic GlyRS homologues shown herein. In contrast, the N-terminal domain (containing residues 1-66) of the cytoplasmic form of CeGlyRS (CeGlyRS c ) shared high sequence identity (57-70% identity) with the WHEP domains of GlyRSs of Drosophila melanogaster, Bombyx mori, and Homo sapiens (Fig. 1B).
Because only one GlyRS gene exists in the genome of C. elegans, one would normally assume that the identity elements must be strictly conserved between its nuclear and mito- chondrial encoded tRNAs Gly (CetRNA n Gly and CetRNA m Gly , respectively) so that CeGlyRS can efficiently charge both tRNA isoacceptors. As expected, the discriminator base A73 and C35/ C36 in the anticodon loop were conserved in yeast and C. elegans tRNA n Gly and tRNA m Gly (Fig. 1C). However, contrary to our anticipation, the first three base pairs in the acceptor stem highly diverged among these tRNA isoacceptors, suggesting that they are not critical for recognition in these two eukaryotic organisms. In addition to differences identified in the acceptor stem, CetRNA m Gly possessed a defective TC hairpin, which prompted us to ask how CeGlyRS can effectively charge this tRNA.
C. elegans GRS1 Can Rescue Growth Defects of a Yeast GRS1 Knock-out Strain-To test the functional potential of CeGRS1, we cloned the ORF of this gene or its derivatives into pADH and tested the ability of the resultant construct to rescue the growth defects of a yeast grs1 Ϫ strain. As shown in Fig. 2, CeGlyRS m conferred a positive growth phenotype to the null allele on both 5-fluoroorotic acid (5-FOA) and YPG, suggesting that this construct can functionally substitute for the yeast GRS1 gene (row 3 in Fig. 2, A and B). This result also implied that a minor portion of the CeGlyRS m protein was retained in the cytoplasm for functioning, a scenario often seen in yeast aaRS complementation (22). Deletion of the MTS alone or both the MTS and MAD specifically impaired its mitochondrial rescue activity; MTS-deleted mutants restored the growth phenotype of the knock-out strain on 5-FOA but not YPG (rows 4 and 5). Thus, the MTS was required for mitochondrial targeting of CeG-lyRS m , and the MAD was dispensable for efficient aminoacylation of yeast tRNA n Gly by CeGlyRS. Further deletion of the WHEP domain impaired both the cytoplasmic and mitochondrial rescue activities (rows 6 and 8). As expected, attaching a heterologous MTS to CeGlyRS c yielded a fusion construct that could restore the growth phenotype of the knock-out strain on YPG (row 7), lending further support to the observation that the MAD is not required for efficient aminoacylation of yeast tRNA m Gly by CeGlyRS. A Western blot analysis using an anti-His 6 tag antibody showed that except for MTS-⌬WHEP, which was poorly expressed, all other CeGlyRS constructs used for cross-species rescue assays were properly expressed in yeast (rows 3-8 in Fig.  2C). Thus, ⌬WHEP was much more stable in the cytoplasm than in mitochondria. We also noted that the processed CeG-lyRS m was somewhat larger in size than its cytoplasmic counterpart CeGlyRS c (rows [3][4][5], suggesting that the MAD was not cleaved away from CeGlyRS m following its import into mitochondria. Deletion of the WHEP Domain Has Little Effect on the Halflife of CeGlyRS c -To check whether deletion of the WHEP domain affects the half-life of CeGlyRS c , we performed a cycloheximide (CHX) chase assay in yeast as previously described (23). Genes encoding CeGlyRS c and ⌬WHEP were respectively cloned into pGAL1, and the resultant constructs were transformed into a S. cerevisiae strain, INVSc1. Cultures of the transformants were induced with galactose for 4 h, followed by the addition of CHX to terminate protein synthesis. CHX-treated cells were harvested at various intervals (0ϳ16 h) following induction. Protein extracts were prepared for Western blot analyses using an anti-His 6 tag antibody. As shown in Fig. 2D, both proteins were fairly stable; less than 20% of the proteins were degraded throughout the time period tested. Thus, deletion of the WHEP domain did not impair the stability of the protein in vivo.
CeGlyRS m and CeGlyRS c Are Respectively Localized in the Mitochondria and Cytoplasm-To map the subcellular localization of CeGlyRS and its derivatives in yeast, we carried out a fluorescence microscopic analysis. A DNA sequence that encodes the GFP was amplified by a PCR and inserted in-frame into the 3Ј end of CeGlyRS m , CeGlyRS m (⌬MTS), and CeG-lyRS c . We then transformed these GFP fusion constructs into the yeast knock-out strain and examined their subcellular distributions via fluorescence microscopy. As shown in Fig. 3, CeGlyRS m was predominantly confined to mitochondria, whereas CeGlyRS c was evenly distributed in the cytoplasm.
Deletion of the MTS from CeGlyRS m redirected the protein to the cytoplasm. Because of the imaging technology resolution constraints, we could not rule out the possibility that a minor portion of the GFP fusion proteins targeted other unintended cellular compartments.
The WHEP Domain-deleted CeGlyRS Variant Possesses Higher Thermal Stability-Because deletion of the WHEP domain impaired the cross-species rescue activity of CeGlyRS c (Fig. 2), we next investigated whether and to what extent this deletion alters the protein's thermal stability in vitro. Pursuant to this objective, genes encoding CeGlyRS variants were transformed into Escherichia coli, and His 6 -tagged CeGlyRS proteins were purified to homogeneity through nickel-nitrilotriacetic acid affinity chromatography (Fig. 4A). Purified CeGlyRS c and ⌬WHEP proteins were then subjected to CD spectrometry at 200 -250 nm. As shown in Fig. 4B, CeGlyRS c retained most of its secondary structure (with high molar ellipticity values at 222 and 208 nm) at temperatures below 40°C but lost much of its secondary structure when the temperature reached 50°C or higher. In contrast, ⌬WHEP retained most of its secondary structure even when the temperature reached 50°C. The deletion mutant did not lose much of its secondary structure until the temperature reached 60°C or higher (Fig.  4B). As a result, CeGlyRS c and ⌬WHEP. respectively. had melting temperatures of ϳ60 and ϳ68°C (Fig. 4C). Thus, deletion of the WHEP domain increased the thermal stability of the protein.
The WHEP Domain-deleted CeGlyRS Variant Possesses Slightly Lower Structural Flexibility-To further study the effect of the deletion on the structural flexibility of CeGlyRS c , limited proteolysis was carried out with CeGlyRS c /trypsin in a ratio of 100:1. Limited proteolysis is often used to probe the structure and dynamics of proteins. Exposed regions such as loops and other flexible regions are more susceptible to the prolific protease. As shown in Fig. 4D, ⌬WHEP was somewhat more resistant to the protease than was the WT. Approximately 80% of the WT protein was hydrolyzed after 2 h of protease treatment, and only ϳ10% protein remained after 4 h of treatment. In contrast, ϳ40% of the ⌬WHEP protein remained after 4 h of treatment. This result suggests that the WT protein possessed a slightly more-flexible structure than did the deletion mutant.
The MAD and WHEP Domain Respectively Play Roles in Aminoacylation of CetRNA m Gly and CetRNA n Gly -As mentioned above, CeGlyRS possesses two distinctive domains at its N terminus: MAD and WHEP. We were prompted to ask whether these two domains are involved in tRNA aminoacylation. Aminoacylation reactions were carried out at ambient temperature with 20 or 200 nM of enzyme and 5 M of in vitro transcribed tRNA Gly . Note that purified CeGlyRS m represents a mature form of CeGlyRS m (without MTS). As shown in Fig. 5A, CeGlyRS m and CeGlyRS c charged CetRNA n Gly with a similarly high efficiency, suggesting that the MAD is dispensable for aminoacylation of CetRNA n Gly . On the other hand, deletion of the WHEP domain from CeGlyRS c reduced its glycylation activity ϳ3-fold, suggesting that the WHEP domain is somehow involved in this reaction (Fig. 5A). Although in vitro transcribed CetRNA m Gly (lacking the entire T-arm) was a very poor sub- strate for glycylation (presumably because of a lack of modification at position 9 with 1-methyladenosine) (24), CeGlyRS m was more efficient (at least 5-fold) than CeGlyRS c (and ⌬WHEP) in charging this T-armless tRNA (Fig. 5B). These results together suggest that the MAD enhances aminoacylation of CetRNA m Gly by CeGlyRS m and that the WHEP domain contributes to aminoacylation of CetRNA n Gly by CeGlyRS c . Relative aminoacylation activities of CeGlyRS c and ⌬WHEP at various temperatures were determined and shown in Fig. 5C.
To take a closer look at the role of the WHEP domain in tRNA aminoacylation, kinetic parameters for aminoacylation of CetRNA n Gly by CeGlyRS c and ⌬WHEP were determined. As shown in Table 1, CeGlyRS c had a K m value of 0.24 M for CetRNA n Gly and a k cat value of 1.73 s Ϫ1 , whereas ⌬WHEP had a K m value of 0.21 M for CetRNA n Gly and a k cat value of 0.62 s Ϫ1 . Thus, deletion of the WHEP domain had almost no effect on the apparent affinity of the enzyme for CetRNA n Gly but reduced its catalytic rate 2.8 times. That is, CeGlyRS c possessed a catalytic efficiency (k cat /K m ) 2.4 times higher than that of ⌬WHEP. Similar results were obtained using yeast tRNA n Gly as the substrate ( Table 1).
The WHEP Domain per se Does Not Bind tRNA-To test whether the WHEP domain itself binds tRNA, a DNA sequence encoding the WHEP domain was cloned into pET21, and the resultant construct was transformed into an E. coli strain, BL21(DE3), for expression. The recombinant WHEP domain (with a C-terminal His 6 tag) was purified through nickel-nitrilotriacetic acid affinity chromatography, and an affinity coelectrophoresis assay was carried out using CetRNA n Gly as the ligand. Affinity co-electrophoresis is particularly useful for determining weak protein-nucleic acid interactions. As shown in Fig. 5D, whereas CeGlyRS c effectively bound tRNA n Gly (with a K d value of ϳ4 M), the purified WHEP domain failed to bind the tRNA even at high protein concentrations (up to 64 M). This result is in agreement with the findings of the kinetic study shown in Table 1.
Overexpression of the WHEP Domain-deleted CeGlyRS Variant Can Rescue the Yeast Knock-out Strain-Because deletion of the WHEP domain only reduced the aminoacylation efficiency (k cat /K m ) of CeGlyRS for yeast tRNA Gly 3.5-fold (Table  1), we wondered whether the deletion mutant could be made a functional yeast enzyme in vivo by overexpression from a stronger promoter. To this end, a DNA sequence encoding ⌬WHEP was cloned into pTEF1 (with a strong TEF1 promoter), and the ability of the resultant construct to rescue the growth defects of the yeast GRS1 knock-out strain was tested. As shown in Fig.  6A, ⌬WHEP, overexpressed from a TEF1 promoter, successfully restored the growth phenotype of the knock-out strain on 5-FOA, but not on YPG (row 3). Attaching a heterologous MTS to ⌬WHEP failed to yield a functional yeast mitochondrial enzyme (row 4). Western blotting results showed that MTS-⌬WHEP had an expression level much lower than that of its counterpart without an MTS, i.e. ⌬WHEP, in yeast (Fig. 6B, left  panel), regardless of the promoters used. This may have been the reason why MTS-⌬WHEP failed to provide sufficient glycylation activity in yeast mitochondria (Figs. 2, 6). We also found that the TEF1 promoter used was ϳ12-fold stronger than the ADH promoter used (Fig. 6B, right panel).

Discussion
Based on the modular organization of aaRSs, it is widely accepted that modern aaRSs descended from successive addi-tions of new domains to the catalytic core (25). According to this hypothesis, primitive aaRSs possessed only a minimal core capable of activating amino acids. New domains were later recruited to the catalytic core, forming a larger aaRS capable of recognizing an ancient tRNA with a structure mimicking the existing tRNA acceptor stem. These early tRNA-aaRS pairs continued to grow through additions of the anticodon stemloop structure and anticodon-binding domain to yield contemporary tRNAs and aaRSs. Other functional domains, such as the editing, protein-interacting, and auxiliary tRNA-binding domains, were later recruited by aaRSs to facilitate their translation efficiencies (25). The MAD and WHEP domain of CeG-lyRS are probably among such examples (Fig. 5). However, some domains recruited by aaRSs are unrelated to their primary functions. Such domains often confer noncanonical functions, such as transcriptional regulation, translational regulation, mitochondrial RNA splicing, and cytokine-like activity, to the aaRSs (26).
In addition to serving as an intracellular translation enzyme, human GlyRS also circulates in serum and acts as a component of the innate defense system against ERK-activated tumorigen-

TABLE 1 Kinetic parameters for aminoacylation of tRNA n Gly by C. elegans GlyRS variants
Each value is determined from a hyperbolic fit of three independent data sets.  (27). Moreover, mutations in human GlyRS are known to cause an axonal form of CMT disease. A recent study further demonstrated that several CMT-associated GlyRS mutants can competitively block binding of the VEGF to its membrane receptor, Nrp1, leading to impaired motor neuron migration (28). WHEP domains exist as single domains in higher eukaryotic TrpRS, HisRS, GlyRS, and MetRS and as multiple repeated domains in Glu-ProRS, but they do not exist in any non-aaRS proteins (29). Human Glu-ProRS possesses three tandem WHEP domains that participate in regulating the noncanonical function of the enzyme through interactions with specific protein factors and the 3Ј-UTRs of interferon-␥-induced inflammatory genes (30). In contrast, the single WHEP domain of CeGlyRS appears to participate in regulating the dynamic structure of the enzyme, a feature somewhat similar to its human homologue (31). Deletion of this domain had no discernible effect on the tRNA binding of the enzyme (Table 1) but substantially affected its catalytic efficiency, thermal stability, and structural flexibility (Figs. 4 and 5). Because all WHEPcontaining GlyRS enzymes possess a similar protein sequence and domain organization (Fig. 1), the WHEP domains may play a similar role in these homologues.

Experimental Procedures
Construction of Plasmids-Cloning of the CeGRS1 gene or its derivatives into pADH (a high copy number yeast shuttle vector with a constitutive ADH promoter), pTEF1 (a high copy number yeast shuttle vector with a constitutive TEF1 promoter), or pET21b (an E. coli expression vector with a T7 promoter) followed the protocol described earlier. Briefly, a cDNA sequence containing an ORF of CeGlyRS m (1-2226 bp) was amplified by RT-PCR as an EagI-NdeI fragment using a pair of gene-specific primers. The PCR-amplified fragment was digested with EagI and NdeI and then cloned into the EagI/NdeI sites of pADH. Cloning of genes encoding various truncated forms of CeG-lyRS m followed a similar protocol. To fuse a heterologous MTS to CeGlyRS c (or its derivatives), a DNA sequence encoding amino acid residues 1-46 of the mitochondrial precursor form of yeast valyl-tRNA synthetase was amplified by a PCR as an EagI-EagI fragment and was inserted at the 5Ј end of the ORF of CeGlyRS c . To fuse the GFP to CeGlyRS c (or its derivatives), a DNA sequence encoding the GFP was amplified by a PCR as an XhoI-XhoI fragment and then inserted at the 3Ј end of the ORF of CeGlyRS c .
Complementation Assay for Cytoplasmic Activity-The yeast GRS1 knock-out strain RJT3/II-1 was previously described (32). We carried out a complementation assay for cytoplasmic GlyRS activity by introducing a test plasmid carrying the target gene and a LEU2 marker into RJT3/II-1 and determining the ability of the transformants to grow in the presence of 5-FOA. Starting from a cell density of 1.0 A 600 , cell cultures were 3-fold serially diluted, and 10-l aliquots of each dilution were spotted onto designated plates containing 5-FOA. The plates were incubated at 30°C for 3 days. Transformants ejected the maintenance plasmid (which carried the WT ScGRS1 gene) with a URA3 marker in the presence of 5-FOA and thus could not grow on 5-FOA plates unless the test plasmid encoded a functional cytoplasmic GlyRS.
Complementation Assay for Mitochondrial Activity-The yeast GRS1 knock-out strain RJT3/II-1 was cotransformed with a test plasmid (which carried a LEU2 marker) and a second maintenance plasmid (which carried a TRP1 marker and only expressed the cytoplasmic form of ScGlyRS). In the presence of 5-FOA, the first maintenance plasmid (carrying a URA3 marker) was evicted from the cotransformants, and the second maintenance plasmid was retained. Thus, all cotransformants could survive 5-FOA selection because of the presence of the cytoplasmic ScGlyRS derived from the second maintenance plasmid. The mitochondrial phenotypes of the cotransformants were further tested on YPG plates at 30°C, with results documented on day 3 following plating. Because a yeast cell cannot survive on glycerol (a nonfermentable carbon source) without functional mitochondria, the cotransformants did not grow on the YPG plates unless the test plasmid encoded a functional mitochondrial GlyRS. Fluorescence Microscopy-Yeast cells were grown to ϳ0.6 A 600 in SD/ϪLeu selective medium. We pretreated cells with DAPI (0.5 g/ml) or MitoTracker (300 nM) for 30 min. Fluorescence microscopy (Axio observer.A1; Carl Zeiss, Oberkochen, Germany) was then used to examine samples with a 100ϫ objective at 25°C, and images were captured with a CCD camera (Axiocam MRm; Carl Zeiss). Nuclear and mitochondrial tracks and merged images were generated with AxioVision Rel. 4.8 software and then subjected to two-dimensional deconvolution with AutoQuant X2.
Aminoacylation Assay-Aminoacylation reactions were carried out at ambient temperature in a buffer containing 50 mM HEPES (pH 7.5), 50 mM KCl, 15 mM MgCl 2 , 5 mM dithiothreitol, 10 mM ATP, 0.1 mg/ml bovine serum albumin, 5 M tRNA Gly , and 20 M glycine (2 M [ 3 H]glycine; Moravek Biochemicals, Brea, CA). The specific activity of [ 3 H]glycine used was 35.0 Ci/mmol. The final concentration of the enzymes used in the reactions was 20 nM (unless otherwise indicated). The reactions were quenched by spotting 10-l aliquots of the reaction mixture onto Whatman filters (Maidstone, Kent, UK) soaked in 5% TCA and 2 mM glycine. Filters were washed three times for 15 min each in ice-cold 5% TCA before liquid scintillation counting. The data were obtained from three independent experiments and averaged.
Kinetic parameters for aminoacylation of tRNA by purified enzymes were determined by directly fitting the data points to the Michaelis-Menten equation. Initial rates of aminoacylation were determined at 25°C with tRNA Gly concentrations ranging 1-20 M and enzyme concentrations ranging 10 -200 nM. The data were obtained from three independent experiments and averaged.
Polyacrylamide Affinity Co-electrophoresis-In vitro transcribed CetRNA n Gly was labeled with 32 P using polynucleotide kinase (New England Biolabs, Beverly, MA) after dephosphorylation with calf intestine phosphatase. The recombinant WHEP domain was 2-fold serially diluted and mixed with a 5% polyacrylamide solution, forming a mini gel matrix with a protein gradient of 1-64 M. 32 P-Labeled tRNA was loaded into each well at an estimated concentration of 1 nM in 2-l aliquots. The gel was run in buffer containing 1ϫ TBE (90 mM Tris borate and 2 mM EDTA) and 50 mM NaCl at 20°C at 50 V for 1 h. After electrophoresis, the gel was vacuum-dried and then exposed to x-ray film.