|Home | About | Journals | Submit | Contact Us | Français|
The RNase A ribonucleases are complex group of functionally diverse secretory proteins with conserved enzymatic activity. We have identified novel RNase 1 genes from four species of squirrel (order Rodentia, family Sciuridae). Squirrel RNase 1 genes encode typical RNase A ribonucleases, each with eight cysteines, a conserved CKXXNTF signature motif, and a canonical His12-Lys41-His119 catalytic triad. Two alleles encode Callosciurus prevostii RNase 1, which include a Ser18↔Pro, analogous to the sequence polymorphisms found among the RNase 1 duplications in the genome of Rattus exulans. Interestingly, although the squirrel RNase 1 genes are closely related to one another (77 to 95% amino acid sequence identity), the cluster as a whole is distinct and divergent from the clusters including RNase 1 genes from other rodent species. We examined the specific sites at which Sciuridae RNase 1s diverge from Muridae / Cricetidae RNase 1s, and determined that the divergent sites are located on the external surface, with complete sparing of the catalytic crevice. The full significance of these findings awaits a more complete understanding of biological role of mammalian RNase 1s.
RNase 1 is the most thoroughly characterized lineage of the RNase A ribonucleases, which is an extensive family of secretory enzymes characterized in vertebrate species (Beintema and Kleineidam, 1994; Dubois et al., 1999; Dyer and Rosenberg 2006; Rosenberg 2008; Marshall et al., 2008; Cho et al., 2005; Cho and Zhang, 2006; Yu and Zhang 2006). Bovine RNase A, the RNase 1 ortholog initially isolated from pancreatic tissue, has served as a prototype “master protein” and was the source of many of the earliest discoveries on primary structure, crystallographic ultrastructure, protein folding and catalytic mechanism (Hirs et al., 1956; Anfinsen, 1957; Aqvist and Anfinsen, 1959; Scheraga and Rupley, 1962). However, despite the complex biochemical characterization, our understanding of the biology of RNase 1 remains rudimentary. Given its abundance in bovine pancreas, localization as a secretory protein, and overall catalytic efficiency, RNase A was presumed to function as a digestive enzyme, serving to promote catabolism of ingested ribonucleic acid polymers. However, more recent studies have shown that RNase 1 expression is substantially less prominent in pancreatic tissue of non-ruminant mammals (Beintema et al., 1973), and is not limited to the gastrointestinal tract (Futami et al. 1997). Similarly, the RNase 1 lineage was considered to be relatively stable from an evolutionary perspective, at least compared to extensive divergence and duplication reported for the eosinophil ribonucleases (RNase 2 / RNase 3; Rosenberg et al., 1995; Zhang et al., 2000). However, recently Dubois and colleagues (2002) identified multiple RNase 1 duplications in the genomes of rat species (family Muridae), suggesting that RNase 1 genes, at least among rodents, may be responding to more complex evolutionary constraints than may have been previously appreciated.
As part of a larger exploration of the RNase 1 lineage in rodent species, we have isolated and characterized full coding sequences from genomic DNAs from four distinct species of squirrels (order Rodentia, family Sciuridae). Squirrels are widely distributed, morphologically primitive rodents that have diversified in concert with major changes in geography and climate (Mercer and Roth, 2003); an analysis of nuclear genes c-myc and RAG-1 has confirmed five major squirrel lineages (Steppan et al., 2004). In this work, we explore the sequences of the Sciuridae RNase 1s, note the similaries and differences from their counterparts among species from rodent families, and evaluate the divergent sites within the context of the tertiary protein structure.
The Callosciurus notatus fibroblast cell line NZP-46 was obtained from American Type Culture Collection (CRL-1926, ATCC, Manassas, VA). Genomic DNA was isolated from cells by lysis in 10 mM Tris-HCl, pH 8.0, with 100 mM NaCl, 25 mM EDTA, 0.5% SDS and 0.1 mg/mL proteinase K at 56°C overnight followed by phenol: chloroform extraction. All other genomic DNAs were prepared from animal tissue as described (Mercer and Roth, 2003).
Primers were designed from the 5’ and 3’ ends of an RNase 1 fragment originally reported for Oryctolagus cuniculus (Dubois et al., 2003; [Table 1]). The 338 bp product resulting from amplification using these primers and C. notatus genomic DNA as a template (standard conditions, 35 cycles) was cloned into PCR2.1, sequenced (ABI 3730xl Genetic Analyzer) and the information used to design internal unidirectional primers in order to isolate the complete C. notatus RNase 1 gene and to identify additional squirrel RNase 1s, as described in the following sections. Primers used to clone other C. notatus genes in order to confirm the identity of this cell line are also listed in Table 1.
Full length squirrel RNase 1 sequences were generated using the Genome Walker Universal kit (BD Biosciences / Clontech, Mountainview, CA). Two nested 5’-directed and two nested 3’-directed primers were designed based on sequence from the original C. notatus RNase 1 fragment [Figure 1]. Manufacturer’s protocols were followed, save for the use of 0.25 µg of genomic DNA to generate each linker-based library rather than the recommended 2.5 µg. Sequences were amplified using Advantage 2 PCR enzyme (Clontech, Mountain View, CA), cycling parameters including a hold 95°C for 2 minutes, followed by 25 cycles of 95°C for 20 seconds, 50°C for 20 seconds and 72°C for 30 seconds, followed by a 5 minute hold at 72°C. Amplification products to be evaluated were cloned into the PCR2.1 sequencing vector. Full length sequence information for each RNase 1 was initially determined from overlapping clones from the genome walk libraries, and then full length sequences were re-isolated de novo from fresh samples of unmanipulated genomic DNA, utilizing primers designed from sequence situated outside the designated open reading frames. All final clones were isolated from at least two independent amplification reactions and sequenced completely in two directions.
Genomic Southern blot was performed by Lofstrand Labs Limited (Gaithersburg, MD) with genomic DNA and unlabelled cDNA probes supplied by our group. Genomic DNA was isolated as described above and subjected to restriction digestion and gel electrophoresis (1 µg / lane), acid / base denaturation, neutralization, and transfer to nylon membrane by standard methods. DNA was UV linked to the membrane, and the membrane subjected to prehybridization (68°C, 3 hrs in 6X SSC, 5X Denhardt’s solution, 0.5% SDS) followed by hybridization for 25 hrs at 68°C with random primed probes. Filters were washed at low stringency in 2X SSC with 0.01% SDS at 68°C with three buffer changes over 60 minutes and exposed to film for 2.5 days with an intensifier screen at −80°C.
Primary sequence analysis was determined with the assistance of Sequencher analysis software (GeneCodes, Ann Arbor, MI). Amino acid sequence similarities were determined via algorithms within the Wisconsin Genetics Computer Group software package available online at the National Institutes of Health. Isoelectric points of encoded proteins were determined by EMBOSS at http://isoelectric.ovh.org/ . Amino acid and nucleotide sequence alignments were generated via ClustalW (http://clustalw.genome.jp/). Phylogenetic analysis was conducted using MEGA version 4.0 (Tamura et al., 2007). The ribbon diagram of human RNase 1 (pdb 2k11) was from the RCSB protein data bank (http://www.rcsb.org/pdb/home/home.do), and was manipulated via the Java-based MBT protein workshop algorithms. Nucleotide and amino acid sequences of all RNase 1s other than those identified in this study were obtained via direct accession number searches and via BLAST searches at the NCBI / National Library of Medicine portal at http://www.ncbi.nlm.nih.gov/sites/entrez.
The strategy used to isolate complete open reading frames of RNase 1 genes from squirrel genomic DNAs is shown in Figure 1. The NZP-46 C. notatus fibroblast cell line was used as an initial source of genomic DNA. We confirmed this cell line as derived from C. notatus by cloning and sequencing a ~400 bp fragment of the 16S ribosomal RNA gene reported for this species (GenBank acc no. AY227453). As shown in Figure 1A, we generated a primary 338 bp fragment encoding C. notatus RNase 1 from genomic DNA isolated from the NZP-46 cell line using 24 bp 5’ and 3’ primers derived from sequence originally reported to encode RNase 1 from the European rabbit (O. cunniculus, see Table 1 for these and all primer sequences). The published rabbit sequence (listed as Genbank AJ535681 in Dubois et al., 2003) was remarkably similar to squirrel sequences isolated in this report; this sequence file been changed to AJ535681.1, now denoting RNase 1 from Sciurus vulgaris. We utilized sequence from this primary C. notatus fragment to design nested primer pairs (3A, 3B, 5A, 5B) for two rounds of unidirectional PCR amplification reactions with restriction fragmented, linkered libraries created from multiple sources of squirrel genomic DNA as templates (Figure 1B, and as described in Methods). Amplification products were sequenced, overlapping contigs were aligned, and the complete open reading frames were ultimately re-isolated from unmanipulated genomic DNA using primers derived from genomic sequence located 5’ to the start ATG and 3’ to the stop codon [Figure 1C]. Each gene sequence reported was isolated from multiple independent amplification trials.
Single intronless open reading frames were identified in three of the four sources of genomic DNA except for C. prevostii, in which two highly homologous RNase 1 sequences were identified. The squirrel sequences encode proteins that are typical of the RNase 1 lineage [Figure 2]. All encoded RNase 1s have a 28 amino acid hydrophobic signal sequence preceding an amino terminus that was inferred by homology to human and mouse RNase 1s (Mizuta et al., 1990; Lenstra and Beintema, 1979). All polypeptides maintain the His-Lys-His catalytic triad, eight appropriately spaced cysteines, and the RNase A family signature CKXXNTF motif. The polypeptides are 128 amino acids in length, and have isoelectric points ranging from 8.40 to 8.58.
As noted above, two RNase 1 sequences were identified in the DNA sample from C. prevostii; each unique sequence was identified in multiple amplification trials. We performed Southern blot analysis, which revealed single bands in each of three restriction-digested C. notatus DNA samples probed with radiolabelled C. notatus RNase 1, and likewise a single band in the lane containing restriction-digested C. prevostii DNA probed with radiolabelled C. prevostii RNase 1 [Figure 3]. These findings suggest that the two C. prevostii RNase 1 sequences are alleles rather than gene duplications. Of the four divergent sites within this allelic pair, two are in the signal sequence, and two are within the mature protein, at positions Ser18↔Pro and Arg55↔Gln [Figure 2]. Interestingly, Ser18↔Pro was identified in the gene duplication of the Polynesian rat (Rattus exulans), both the comparison of RNase 1α ↔ RNase 1β, and also RNase 1α ↔ RNase 1δ (Dubois et al., 2002).
An unrooted neighbor-joining tree displaying relationships among mammalian RNase 1 DNA sequences is shown in Figures 4A. The Sciuridae RNase 1s form a cluster that is distinct from those of other mammalian species, and that is divergent from the cluster that includes RNase 1s from Muridae and Cricetidae species. Sciridae RNase 1s in relation to all rodent sequences, including those of the rodent infraorder Hystricognathi (guinea pig, capybara, cuis), for which DNA sequence is unavailable, are shown in the tree in Figure 4B. Average distances within and between clusters were determined [Table 2]. Clearly, the rodent Sciuridae RNase 1s are no more similar to rodent Muridae / Cricetidae RNase 1s than they are to any of the other phylogenetically more distant mammalian species. Interestingly, the same is true when this type of analysis is applied to the data shown in Figure 4B; the Sciuridae RNase 1 sequences are no more similar to those of the Hystricognathi RNase 1s than they are to the Muridae RNase 1s (data not shown).
We identified 18 sites that demarcate specific divergence between the squirrel RNases and those in the Muridae cluster [Table 3]. There are no linear regions or domains with a specific propensity for divergence. However, when the divergent sites are mapped onto the three dimensional ribbon diagram of the homologous human RNase 1 (pdb #2k11; Figure 5), they localize universally on the exterior surface of the protein. The segments that interact with the incoming ribonucleotide, those demarcated by and encompassing the His12 - Lys41 - His119 catalytic triad, are remarkably conserved.
In this work, we isolated and characterized four novel RNase 1 genes from squirrel genomic DNAs via a two-tiered unidirectional PCR approach. We used a variant of this method in a previous study in which we described the isolation of a unique RNase A ribonuclease from the amphibian species Iguana iguana (Nitto et al., 2005). This approach is ideally suited for isolation of novel RNase A ribonucleases, which have relatively small, intronless open reading frames with characterized sequence motifs and specific regions of predictable cross-species homologies. Here, we chose to re-isolate the complete open-reading frames de novo from unmanipulated genomic DNA once the complete sequences were assembled from genomic fragments. In this way, we were able to assess the sequence of each gene directly, and we were able to eliminate the possibility that our isolated sequences contained unrelated fragments that became ligated to one another during library assembly.
Among the interesting findings, we observed that rodent RNase 1s from squirrels and from other rodent families emerged as distinct clusters. The constraints leading to divergence among the rodent RNase 1s remain obscure for several reasons. First, despite the historical focus on bovine pancreatic RNase A, the full biological role of RNase 1 in non-ruminant species remains obscure. Futami and colleagues (1997) have demonstrated that RNase 1s are expressed widely in tissues outside the gastrointestinal tract, and Beintema and colleagues (1973) have demonstrated that there is little to no actual enzymatic activity in the pancreas other than in mammals with ruminant digestive systems, and even less in pancreatic tissue of squirrels than in those of other rodents, all findings that lead one to question the role of this protein as purely a digestive enzyme. Yet few other hypotheses have emerged in its place.
Although extensive gene duplication has been described for the rodent RNase 2/3s and the Rana ribonucleases (Rosenberg et al., 2001), there were only a few duplications characterized among RNase 1 genes until quite recently; the best known of these are those among ruminants, the bovine brain and bovine seminal RNases (D’Alessio et al., 1991). More recently, Zhang and colleagues (2002, 2006) have described duplications of RNase 1s of African and Asian langurs, and Dubois and colleagues (2002) have identified extensive RNase 1 duplication among several rat species. Duplication of RNase 1 in the Polynesian rat (Rattus exulans) includes a nucleotide polymorphism resulting in the conversion of Pro18↔Ser, similar to that observed between C. prevostii RNase 1 alleles (Ser18↔Pro). Interestingly, Ser18 is a component of a crucial hydrogen bond in the native structure of bovine RNase A and provides stability for the amino terminal segment of the protein (Richards and Wycoff, 1971); conversion to proline likely eliminates this bond, which could release the amino terminal segment from this constrained position, permitting novel interactions to occur.
Despite our inability to come to any universal conclusions regarding function, information obtained from cross-species evaluations provides us with some clues regarding functionally crucial vs. flexible sites. We evaluated individual sites and found that, despite the divergence calculated for the rodent Sciuridae vs. rodent Muridae / Cricetidae sequences, the essential nature of the RNase 1 proteins remained unchanged. In other words, none of the RNase A family signature features were lost, there was no prominent change in net charge, nor were there any systematic alterations in hydrophilicity / hydrophobicity resulting from the changes at the 18 individual divergent sites. However, when these sites were examined in three-dimensional perspective, it was clear that mutational change was all focused at the external surface of the protein. Not only were the catalytic amino acids left intact (His12, Lys41, His119), but the entire catalytic crevice remained free of any mutational change (Figure 5). Furthermore, only one of the 18 total sites (Q69) was among those recognized as a contact site for interaction with intracellular ribonuclease inhibitor [RI; (Kobe and Deisenhofer, 1995)]. This finding suggests that maintaining strong interactions with RI is likewise an important evolutionary constraint.
In summary, we have identified four novel RNase 1 genes from the rodent family Sciuridae. The sequences encode proteins with properties typical of the RNase A ribonuclease family, including the signal sequence, His-Lys-His catalytic triad and eight canonically spaced cysteines. The most interesting feature of the squirrel sequences is their nature of their divergence from the rodent Muridae/Cricetidae RNase 1 counterparts. Three-dimensional projection indicates that the divergent sites are located on the external surface of RNase 1, while the catalytic crevice remains unchanged. The significance of this finding awaits a clearer understanding of the full spectrum of activities of the mammalian RNase 1 lineage.
The work presented in this manuscript was supported by NIAID DIR funding to HFR and National Science Foundation #DEB 9726855 to JMM and VLR.
Sequence data from this article have been deposited with the DDBJ/EMBL/GenBank Data Libraries under Accession Nos. FJ659109 – FJ659113