RoXaN, a Novel Cellular Protein Containing TPR ... - Journal of Virology

2 downloads 0 Views 863KB Size Report
Aug 22, 2003 - Finger Motifs, Forms a Ternary Complex with Eukaryotic. Initiation ... initiation factor eIF4G I. To further our understanding of the role of NSP3 in rotavirus replication, we looked for ... larly important for efficient and accurate translation initiation ... amino-terminal domain (aa 1 to 150) is required for sequence-.
JOURNAL OF VIROLOGY, Apr. 2004, p. 3851–3862 0022-538X/04/$08.00⫹0 DOI: 10.1128/JVI.78.8.3851–3862.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Vol. 78, No. 8

RoXaN, a Novel Cellular Protein Containing TPR, LD, and Zinc Finger Motifs, Forms a Ternary Complex with Eukaryotic Initiation Factor 4G and Rotavirus NSP3 Damien Vitour, Pierre Lindenbaum,† Patrice Vende, Michelle M. Becker, and Didier Poncet* Virologie Mole´culaire et Structurale, Unite´ Mixte de Recherche, CNRS-INRA, 91198 Gif-sur-Yvette, France Received 22 August 2003/Accepted 23 December 2003

Rotavirus mRNAs are capped but not polyadenylated, and viral proteins are translated by the cellular translation machinery. This is accomplished through the action of the viral nonstructural protein NSP3, which specifically binds the 3ⴕ consensus sequence of viral mRNAs and interacts with the eukaryotic translation initiation factor eIF4G I. To further our understanding of the role of NSP3 in rotavirus replication, we looked for other cellular proteins capable of interacting with this viral protein. Using the yeast two-hybrid assay, we identified a novel cellular protein-binding partner for rotavirus NSP3. This 110-kDa cellular protein, named RoXaN (rotavirus X protein associated with NSP3), contains a minimum of three regions predicted to be involved in protein-protein or nucleic acid-protein interactions. A tetratricopeptide repeat region, a proteinprotein interaction domain most often found in multiprotein complexes, is present in the amino-terminal region. In the carboxy terminus, at least five zinc finger motifs are observed, further suggesting the capacity of RoXaN to bind other proteins or nucleic acids. Between these two regions exists a paxillin leucine-aspartate repeat (LD) motif which is involved in protein-protein interactions. RoXaN is capable of interacting with NSP3 in vivo and during rotavirus infection. Domains of interaction were mapped and correspond to the dimerization domain of NSP3 (amino acids 163 to 237) and the LD domain of RoXaN (amino acids 244 to 341). The interaction between NSP3 and RoXaN does not impair the interaction between NSP3 and eIF4G I, and a ternary complex made of NSP3, RoXaN, and eIF4G I can be detected in rotavirus-infected cells, implicating RoXaN in translation regulation. factor eIF4E, and the 3⬘ poly(A) tail is bound by the poly(A)binding protein (PABP) (52, 53). These two mRNA elements and the proteins bound to them act synergistically to enhance translation (18, 27). eIF4Gs (eIF4G I and II) are scaffold proteins that bring together eIF4E and PABP (14, 19) and cause the circularization of the mRNA (62). As a result of these interactions, the eIF3 protein complex is recruited to the mRNA, followed by the small subunit of the ribosome, which initiates scanning of the 5⬘ untranslated region for a start codon (43, 50). Circularization of the mRNA seems particularly important for efficient and accurate translation initiation when competition exists between mRNAs (49) or when ribosomes or initiation factors are in limited supply (38, 51). Rotavirus mRNAs contain a 5⬘ cap structure and a 3⬘ consensus sequence, both of which are important for efficient translation. In the case of group A rotaviruses, the 3⬘ end consensus sequence UGACC is highly conserved among the 11 genes, with the exception (UGAACC) of gene 5 of rotavirus strain SA11 (42). We have previously shown that the function of rotavirus NSP3 in viral translation parallels that of PABP in cellular poly(A)-dependent translation. In rotavirus-infected cells, NSP3 binds the 3⬘ end of rotavirus mRNAs (46) and coimmunoprecipitates with eIF4G (45). NSP3 interacts with the same region of eIF4G as does PABP (22, 29, 45), but with a higher affinity (22). Therefore, during rotavirus infection, NSP3 evicts PABP from eIF4G, leading to the enhancement of translation of rotavirus mRNAs and the concomitant impairment of translation of cellular mRNAs (45, 60). In vivo, NSP3

Rotaviruses, the major cause of diarrhea in young animals and children, are involved in the death of more than 400,000 children under the age of five each year (40, 41). Rotaviruses are members of the Reoviridae family, and their genomes are composed of 11 segments of double-stranded RNA which encode six structural proteins and five or six nonstructural proteins. The viral replication cycle occurs entirely in the cytoplasm. Upon virus entry, the viral transcriptase synthesizes capped, but nonpolyadenylated, mRNAs (16, 28). These molecules are competent to serve as templates for either translation or replication. A subset of viral mRNAs are packaged into core particles and replicated to form the genomic doublestranded RNAs of progeny virions. Morphogenesis continues and viral proteins are added to generate the middle and outer layers surrounding the core particle. Mature virions, which consist of triple-layered particles, are released from the cell by lysis (reviewed in reference 16). Cellular translation involves a myriad of proteins and elements found at both the 5⬘ and 3⬘ termini of the majority of cellular mRNAs in a highly regulated cascade, controlled by specific protein-RNA and protein-protein interactions. In the steps leading to translational initiation, the cap structure at the 5⬘ terminus of an mRNA is bound by the translation initiation

* Corresponding author. Mailing address: UMR, CNRS-INRA, Virologie Mole´culaire et Structurale, 1 avenue de la Terrasse, Baˆtiment 14B, 91198 Gif-sur-Yvette Cedex, France. Phone: 33-(0)1 69823835. Fax: 33-(0)1 69824308. E-mail: [email protected]. † Present address: Integragen, 91000 Evry, France. 3851

3852

VITOUR ET AL.

stimulates the translation of viral-like mRNAs in synergy with the cap structure (9, 61). Three functional and structural domains have been identified along the 313-amino-acid (aa) sequence of NSP3. The amino-terminal domain (aa 1 to 150) is required for sequencespecific RNA binding (13, 44). Our laboratory has shown that a dimer of NSP3 binds a single RNA molecule (44), and the three-dimensional structure of the amino terminus of NSP3 confirms this, showing a heart-shaped dimer bound to one RNA molecule (13). The central domain of NSP3 (aa 150 to 241) is predicted to form a coiled-coil structure, allowing NSP3 dimerization (44). The carboxy-terminal 110 aa of NSP3 are required for binding to eIF4G I in a two-hybrid assay and for coimmunoprecipitation analysis (44, 45). The three-dimensional structure of the carboxy terminus of NSP3 (aa 206 to 313) also shows that NSP3 is a dimer when bound to eIF4G I (aa 132 to 160), but contacts between NSP3 and eIF4G are limited to the carboxy-terminal 50 aa of NSP3 (22). NSP3mediated stimulation of translation requires the simultaneous binding of NSP3 to mRNA and eIF4G, since deletion of either the amino- or carboxy-terminal domain abolishes translation enhancement (61). In this report, we describe the identification and characterization of a novel cellular protein named RoXaN (rotavirus X protein associated with NSP3) which contains a unique combination of protein-protein interaction and nucleic acid-binding domains. We have determined that mammalian cells contain at least two related proteins, termed RoXaN I and II; however, most of the work in this article describes characteristics of RoXaN I. We show that RoXaN I interacts with the dimerization domain of NSP3 and forms a ternary complex with eIF4G I during rotavirus infection. Our data suggest that RoXaN I is a novel cellular protein involved in translation regulation. MATERIALS AND METHODS Cells and viruses. African green monkey kidney COS-7 cells were grown in Dulbecco’s modified Eagle’s medium (BioWhittaker). Adenovirus type 5-transformed human embryonic kidney (HEK 293) cells containing the recombinant vectors required for the Tet-Off expression system (Clontech) were maintained in Eagle’s minimum essential medium (BioWhittaker). Both culture media were supplemented with 10% fetal bovine serum (HyClone), 100 IU of penicillin (BioValley)/ml, and 100 ␮g of streptomycin (BioValley)/ml. The bovine RF strain of group A rotaviruses was used to infect COS-7 cells. Viral infectivity was determined by a plaque assay with embryonic rhesus monkey kidney MA104 cells, as described previously (47). Infections were performed at a multiplicity of infection (MOI) of 10 PFU/cell in Dulbecco’s modified Eagle’s medium in the presence of trypsin (0.44-␮g/ml type IX trypsin; Sigma) and antibiotics, but without serum. Cells were harvested 6 h after infection. Yeast two-hybrid and “sandwich”-hybrid assays. A yeast two-hybrid assay using Saccharomyces cerevisiae (Matchmaker; Clontech) was used to detect interactions between proteins, as previously described (5, 17). All procedures described below were carried out according to the manufacturer’s instructions. The yeast two-hybrid cloning vector pGBT9, containing the GAL4 DNAbinding domain (DBD) [GAL4(1–147) DBD TRP1 Ampr] (Clontech), was used for bait constructs, and vectors pGAD424 and pGAD10, containing the GAL4 transcriptional activation domain (AD) [GAL4 (768–881) AD LEU2 Ampr] (Clontech), were used for prey constructs. The budding yeast two-hybrid strain HF7c (Clontech), with the HIS3 and lacZ reporter genes downstream of heterologous GAL4-responsive promoter elements and with trp1 and leu2 genes as selection markers, was used for transformation. The HF7c yeast strain was transformed with a pGBT9-NSP3:88-313 construct previously described as NSP3A-met88 (45) and then with a cDNA library from monkey kidney CV-1 cells in the pGAD10 vector (Clontech) by use of lithium acetate. Cells were plated onto

J. VIROL. synthetic dropout medium lacking tryptophan and leucine to monitor the transformation efficiency and also onto dropout medium lacking tryptophan, leucine, and histidine to select for interacting proteins. Plates were incubated at 30°C for 3 days. Plasmids encoding cellular p53 and nuclear lamin genes in frame with the GAL4 AD (Clontech) were used as negative controls. The yeast sandwich-hybrid system developed by Ozenberger and Young (39) was used to analyze interactions between three proteins. Yeast strain CY770, a gift of Kathleen H. Young, was transformed with DBD and AD fusion proteins and a third nonchimeric protein expressed from the pYX212 plasmid vector (Ingenius). Uracil was omitted from the medium to maintain selection for the pYX212 plasmid. In addition, the medium was supplemented to contain 20 mM 3-amino-triazole (Sigma), a competitive inhibitor of the constitutively expressed yeast HIS3 protein His3p, to suppress background growth of yeast colonies. Cells were plated onto synthetic dropout medium lacking tryptophan, leucine, and uracil to monitor the transformation efficiency. Colonies capable of growth on this dropout medium were resuspended and diluted 1:10 in water, and 5-␮l drops were plated on dropout medium lacking tryptophan, leucine, uracil, and histidine to select for interacting proteins. Plates were incubated at 30°C for 3 days. Plasmid constructions. In-frame deletion and cloning strategies were determined with the CloneIt program (34; http://genome.jouy.inra.fr/cgi-bin/CloneIt /CloneIt). Deletion mutants of NSP3 used in the two-hybrid assays were described previously (44). The pGBT9-NSP3:88-313 plasmid (45) was digested with restriction endonucleases Ecl16II and SalI. The fragment released was ligated with the pYX212 plasmid vector digested with SmaI and XhoI, using T4 DNA ligase (Euromedex), to generate pYX212-NSP3:88-313. The pBS RF7 plasmid, which contains the full-length NSP3 open reading frame (ORF) (GenBank accession no. Z21639), was used as a template for PCR amplification with Pfu DNA polymerase (Stratagene) of the fragment encoding aa 150 to 313 of NSP3, using oligonucleotides 1A and 1B (Table 1, which contains all of the oligonucleotides mentioned below), and the fragment encoding aa 206 to 313, using oligonucleotides 2A and 1B appended with 5⬘ BglII and 3⬘ EcoRI restriction sites. PCR products were digested with BglII and EcoRI and ligated with a pEGFP-C1 vector (Clontech) digested with the same restriction endonucleases to generate pEGFP-NSP3:150-313 and pEGFP-NSP3:206-313, respectively. Site-directed mutagenesis (QuikChange; Stratagene) was used to generate an EcoRI restriction site in both pEGFP-NSP3:150-313 and pEGFP-NSP3:206-313, using complementary oligonucleotides 3A and 3B. After PCR amplification, both plasmids were digested with EcoRI, and the ends of the vectors were religated to generate pEGFP-NSP3:150-238 and pEGFP-NSP3:206-238, respectively. The plasmid recovered from six independent yeast clones after the yeast two-hybrid screen using pGBT9-NSP3:88-313 as bait and the CV-1 cDNA library as prey was called pGAD-RoXaN-N1 (Fig. 1). It encodes aa 1 to 387 of RoXaN I and is now called pGAD-RoXaN I:1-387. The DNA fragment generated by the digestion of pGAD-RoXaN I:1-387 with EcoRI was ligated to pGBT9 which had been digested with the same enzyme to generate pGBT9-RoXaN I:1-387. For the generation of deletion mutants, pGAD-RoXaN I:1-387 was digested with StuI and religated to generate pGAD-RoXaN I:1-29,136-387 and was digested with Eco109I and religated to generate pGAD-RoXaN I:1-162,344-387. pGAD-RoXaN I:1-387 was digested with SacII and StuI and then religated to generate pGAD-RoXaN I:135-387, which was digested with SmaI and religated to generate pGAD-RoXaN I:135-342. DNA fragments were amplified by PCR using pGAD-RoXaN I:1-387 as a template, with different pairs of oligonucleotides appended with 5⬘ EcoRI and 3⬘ PstI restriction sites. Oligonucleotide 4A was used in conjunction with oligonucleotides 4B, 4C, and 4D. The resulting DNA fragments were digested with EcoRI and PstI and ligated with pGAD424 digested with the same restriction endonucleases to generate pGAD-RoXaN I:144-341, pGAD-RoXaN I:144-265, and pGAD-RoXaN I:144-174, respectively. PCR amplification of a portion of pGAD-RoXaN I:1-387 was performed, using oligonucleotides 5A and 4D. The resulting DNA fragment was digested with EcoRI and PstI and ligated with pGAD424 digested with the same restriction endonucleases to generate pGADRoXaN I:1-174. Oligonucleotide 6A was used in conjunction with oligonucleotide 4B to amplify the nucleotides encoding aa 244 to 341 of RoXaN I by PCR. The resulting DNA fragment was digested with EcoRI and PstI and ligated with pGAD424 digested with the same restriction endonucleases to generate pGADRoXaN I:244-341. The DNA fragment encoding RoXaN I aa 1 to 766 recovered after digestion of the RoXaN I⫺N2 plasmid (GenBank accession no. AF188530) with XhoI and EcoRI restriction endonucleases was ligated with the pGAD424 vector digested with the same restriction endonucleases to generate pGAD-RoXaN I:1-766.

VOL. 78, 2004

RoXaN-NSP3-eIF4G TERNARY COMPLEX

3853

TABLE 1. Oligonucleotides used for plasmid construction Oligonucleotide

Sequence (5⬘33⬘)a

Orientation

1A 1B 2A 3A 3B 4A 4B 4C 4D 5A 6A 7A 7B 8A 8B 9A 9B 10A 10B 11A 11B 12A 12B 13A 13B 14A 14B 15A 15B 16A 16B

AAGATTagatctGGTGAAGTCGAAGTGGATG AAGCTTgaattcTATTCATAGCTATATTCA AAGATTagatctATGTACTCTCTCCAAAACG CAAAATAAAATTAGTTgaattcTGTCTTCAGTCG CGACTGAAGACAgaattcAACTAATTTTATTTTG CCTCgaattcCCCCACGATGAAAGC GATTggatccTCAATCGAGGGCATCCAG CCGGTggatccTCATGGACCAAAGACATC ATTAAggatccTTTACTTATATGCCTTGCGAA CGGAGgaattcATGGAGAGGCAGAAACGGAA CAGGgaattcCCCAGCACCGAC CTAGCCACCATGGACTACAAGGACGACGATGACAAGgcggccgcG GATCCgcggccgcCTTGTCATCGTCGTCCTTGTAGTCCATGGTGG TAGCCACCATGGCCAGTATGACTGGTGGACAGCAAATGGGTgc ggccgcACCCATTTGCTGTCCACCAGTCATACTGGCCATGGT GAATTCgcggccgcCATGGAGAGGCAGAAACGGAA GAATTCgcggccgcCTACTCCCCAGTGGTGGCGGT caTATAAGAGGCCCgcggccgcGGAAACCTTTTCtc gaGAAAAGGTTTCCgcggccgcGGGCCTCTTATAtg CTCATCAAGAACCCCTAGaagcttACCCACGAGTTCAAGC GCTTGAACTCGTGGGTaagcttCTAGGGGTTCTTGATGAG GGATGACTTCTCAGCGGCCGCGGTCTTTGGCCCAG CTGGGCCAAAGACCGCGGCCGCTGAGAAGTCATCC CAGAGCTGGACACCGCGGCCGCGTCGCTGTCTCTGGTC GACCAGAGACAGCGACGCGGCCGCGGTGTCCAGCTCTG CTGGATGACTTCTGaagcttGGATGTCTTTGGC GCCAAAGACATCCaagcttCAGAAGTCATCCAG CAGGGTGGCCTGTAGaagcttGGCGTGCCTAGTGAGTTGC GCAACTCACTAGGCACGCCaagcttCTACAGGCCACCCTG AATTCGccatggCActcgag GATCctcgagTGccatggCG

Forward Reverse Forward Forward Reverse Forward Reverse Reverse Reverse Forward Forward Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse

a

The restriction endonuclease recognition sites are indicated in lowercase letters.

The pRoXaN plasmid, which contains the full-length RoXaN I ORF, was generated by ligation of the fragment resulting from XhoI digestion of pGADRoXaN I:1-387 into the KIAA1031 plasmid (human cDNA) digested with the same enzyme. The deduced amino acid sequence of RoXaN I in pGAD-RoXaN I:1-387 from the monkey CV-1 cDNA library shows only one amino acid change compared to human RoXaN I, at position 371 (human, glycine; monkey, valine); however, the XhoI fragment from pGAD-RoXaN I:1-387 used to construct pRoXaN does not include this change. Therefore, the amino acid sequence of RoXaN I in pRoXaN is identical to the human sequence. The complementary primers 7A and 7B, containing the nucleotide sequence coding for the Flag epitope tag (54), were annealed and ligated with pcDNA3.1/ Hygro(⫺) vector (Invitrogen) digested with NheI and BamHI to generate pcDNA3.1-Flag. Similarly, complementary oligonucleotides 8A and 8B, containing the nucleotide sequence coding for the T7tag epitope tag (54), were annealed and ligated with pcDNA3.1/Hygro(⫺) vector digested with NheI and BamHI to generate pcDNA3.1-T7tag. Oligonucleotides 7A, 7B, 8A, and 8B also contain a NotI restriction site, which was used to insert RoXaN I fragments in frame with either the Flag or T7tag epitope tag in pcDNA3.1-Flag and pcDNA3.1-T7tag. The full-length RoXaN I ORF was amplified by PCR with Pfu DNA polymerase (Stratagene), using pRoXaN as a template and oligonucleotides 9A and 9B appended with 5⬘ and 3⬘ NotI sites. PCR products were digested by NotI and ligated to pcDNA3.1-T7tag vectors digested with the same restriction endonuclease to generate pcDNA3.1-T7tag-RoXaN I:1-977. Site-directed mutagenesis was used to generate a NotI restriction site in pcDNA3.1-T7tag-RoXaN I:1-977, using complementary oligonucleotides 10A and 10B. The plasmid was then digested partially with the NotI restriction endonuclease and religated to generate pcDNA3.1-T7tag-RoXaN I:1-173 and pcDNA3.1-T7tag-RoXaN I:177-977. A stop codon followed by a HindIII restriction site was introduced into pcDNA3.1-T7tag-RoXaN I:177-977 by site-directed mutagenesis with complementary oligonucleotides 11A and 11B to generate pcDNA3.1-T7tag-RoXaN I:177-414. Triplet amino acid changes were generated by site-directed mutagenesis of pcDNA3.1-T7tag-RoXaN I:177-414, using complementary oligonucleotides 12A

and 12B to generate pcDNA3.1-T7tag-RoXaN I:177-414-DGD:257/259:AAA and oligonucleotides 13A and 13B to generate pcDNA3.1-T7tag-RoXaN I:177414-LLD:268/270:AAA. Site-directed mutagenesis with complementary oligonucleotides 14A and 14B was used to introduce a stop codon followed by a HindIII restriction site in pcDNA3.1-T7tag-RoXaN I:1-977. After PCR amplification, the vector was digested with HindIII and then religated to generate pcDNA3.1-T7tag-RoXaN I:1-255. The same strategy was used to generate pcDNA3.1-T7tag-RoXaN I:1279, using complementary oligonucleotides 15A and 15B. The pFASTbac-Flag-eIF4G I plasmid, which contains the full-length human eIF4G I cDNA, was generated by the ligation of two fragments, one recovered from the pHFC1 plasmid (64) digested with EcoRI restriction endonuclease and the other from plasmid p8/1 (GenBank accession no. AAC78442) digested with XmaI and EcoRI restriction endonucleases, with the pFASTbac-Flag plasmid (8) digested with XmaI and EcoRI. This construct was digested completely with XhoI and partially with BamHI, and the fragment corresponding to the ORF of eIF4G I was ligated with pcDNA3 (Clontech) digested with the same enzymes to generate pcDNA3-Flag-eIF4G I. The complementary oligonucleotides 16A and 16B, containing NcoI and XhoI restriction sites, were annealed and ligated into the EcoRI and BamHI restriction sites of pGBT9 to generate pGBT9-NX. Next, the full-length ORF of human PABP, recovered after digestion of the pGEM-hPABP plasmid (21) with restriction endonucleases NcoI and XhoI, was ligated with pGBT9-NX which had been digested with NcoI and SalI restriction endonucleases to generate pGBT9hPABP. Rotavirus strain RF proteins NSP1, NSP2, and NSP5 in yeast two-hybrid bait and prey vectors have been described previously (44, 48). Antibodies. Mouse monoclonal antibodies (MAbs) specific for rotavirus RF nonstructural proteins NSP1, NSP3, and NSP5, expressed in Spodoptera frugiperda cells infected with recombinant baculoviruses, have been described previously (3, 4, 37, 48). Green fluorescent protein (GFP)- and Flag M5-specific MAbs were purchased from Sigma, and a T7tag-specific MAb was purchased from Novagen. An eIF4G I:157-637-specific polyclonal rabbit antiserum was a gift of Nahum Sonenberg (McGill University, Montreal, Canada) and has

3854

VITOUR ET AL.

J. VIROL.

FIG. 1. RoXaN cDNA and protein. (A) Reconstitution of RoXaN I cDNA. The different clones or sequences used to reconstitute the complete RoXaN I cDNA, shown at bottom, are indicated. RoXaN I:1-387 (RoXaN-N1) is the clone obtained from the yeast two-hybrid analysis. RoXaN I:1-764 (RoXaN-N2) corresponds to the sequence of GenBank accession no. AF188530, and KIAA1031 has been described previously (57). The position of the stop codon upstream of the RoXaN I start codon, found in several ESTs, is indicated along with one representative EST (GenBank accession no. AI927120). (B) Schematic representation of the structural motifs of RoXaN I. The motifs and regions found within the deduced amino acid sequence of RoXaN I are indicated, along with their positions. C3H1 and C2H2 zinc finger motifs, numbered from 1 to 5, are indicated in light gray and striped boxes, respectively. Black boxes indicate the presence of other histidine and cysteine residues conserved between RoXaN I and II which have the potential to form zinc finger-like motifs. The predicted coiled-coil region is indicated by a checkerboard pattern.

been described previously (20). Secondary antibodies goat anti-mouse immunoglobulin G (IgG), donkey anti-goat IgG (Jackson Laboratory), and goat anti-rabbit IgG (Sigma) conjugated to horseradish peroxidase were used as described below. Transfections. HEK 293 cells (2 ⫻ 106) in 60-mm-diameter poly-L-lysinecoated plates (Biocoat) and COS-7 cells (3 ⫻ 105) in 9.5-cm2 plates were seeded 24 to 36 h prior to transfection. Transfections were performed with Lipofectamine (Gibco BRL) used according to the manufacturer’s instructions. Cells were washed and processed for infection or coimmunoprecipitation analysis 36 or 48 h after transfection. Immunoprecipitations. Cells were either transfected for 36 or 48 h, infected for 6 h, or both, and then they were washed once with 1 ml of cold (4°C) phosphate-buffered saline (PBS) supplemented with protease inhibitor cocktail (one tablet per 50 ml) (Complete; Roche) and lysed in 1 ml of TMGK buffer (20 mM Tris-HCl [pH 8], 20 mM MgCl2, 110 mM KCl, 1% Triton X-100, Complete protease inhibitor cocktail). Cell debris was removed by centrifugation, supernatants were transferred to 1.5-ml Eppendorf tubes, and 30 ␮l of lysate was removed prior to further treatment. Immunoprecipitation was performed by adding 1 ␮l of MAb or polyclonal rabbit antiserum to the whole-cell lysate supernatant and incubating it overnight at 4°C with endover-end rotation. Next, 50 ␮l of a 50% suspension of protein A-Sepharose beads in TMGK buffer was added and the incubation continued for 1 h at 4°C with end-over-end rotation. Protein A-Sepharose beads were then subjected to centrifugation (13,000 ⫻ g, 30 s) and washed three times with 400 ␮l of TMGK buffer. Immunoblots. After immunoprecipitation, proteins complexed with protein A-Sepharose beads were boiled in loading buffer (10 mM Tris-HCl [pH 6.8], 2% sodium dodecyl sulfate [SDS], 10% glycerol, 150 mM ␤-mercaptoethanol),

resolved by SDS–10% polyacrylamide gel electrophoresis (PAGE) in Laemmli buffer (33), and then transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore) by transverse electrophoresis in 10 mM CAPS [pH 11]–10% methanol. Next, the membrane was incubated for 30 min at room temperature in PBS containing 0.1% Tween 20–5% dry milk or in SuperBlock (Pierce) blocking reagent solution, washed three times with PBS– 0.1% Tween 20, and incubated for 1 h at room temperature with a primary antibody (1:10,000 dilution for NSP3-, T7tag-, and Flag M5-specific MAbs, 1:4,000 dilution for eIF4G I-specific polyclonal rabbit antiserum, and 1:2,500 dilution for GFP-specific MAb). The membrane was washed three times in PBS containing 0.1% Tween 20, incubated for 1 h at room temperature with a secondary antibody (1:20,000 dilution for goat anti-mouse IgG and 1:100,000 dilution for goat anti-rabbit IgG or donkey anti-goat IgG) conjugated to horseradish peroxidase, washed again three times in PBS containing 0.1% Tween 20, incubated with peroxidase substrate (SuperSignal; Pierce), and exposed to film for various times. Sequence analyses. RoXaN I cDNA nucleotide sequence homology searches were done at the National Center for Biotechnology Information (NCBI) Web site (http://www.ncbi.nlm.nih.gov/BLAST/) with the basic local alignment search tool (BLAST) program. Peptide sequence analyses (Prosite motif scan) were carried out at the ISREC (Swiss Institute for Experimental Cancer Research) Web site (http://www.isrec.isb-sib.ch), and the pFam domain homology search was conducted through a program at the Washington University in St. Louis Web site (http://pfam.wustl.edu). Protein alignments by the ClustalW program were performed at the Lyon I Universite´ website (http://pbil.univ-lyon1.fr/). Coiled-coil regions were identified with the Lupas algorithm on the same server.

VOL. 78, 2004

RoXaN-NSP3-eIF4G TERNARY COMPLEX

RESULTS Discovery of RoXaN by interaction with NSP3 in a yeast two-hybrid assay. To investigate the possible interaction of rotavirus NSP3 with cellular proteins, we employed the yeast two-hybrid assay originally described by Fields and Song (17). The transformation of yeast with the plasmid containing the entire NSP3 coding sequence (NSP3:1-313), fused in frame with the DBD of GAL4, activated the reporter gene. Deletion of the first 88 aa of NSP3 (NSP3:88-313) abolished transactivation in yeast but maintained the NSP3 properties of multimerization and interaction with eIF4G I (45). Epithelial monkey kidney cells are susceptible to rotavirus infection, so we used NSP3:88-313 as bait to screen a cDNA library from monkey kidney CV-1 cells. Of the 35 yeast colonies originally obtained, two different cDNA clones were isolated repeatedly. One cDNA was isolated from 27 yeast colonies, encoded aa 1 to 926 of the human translation factor eIF4G I, and has been further characterized (45). The other cDNA, isolated from six yeast clones, contained a 1,274-bp insert which encoded a single large ORF in frame with the GAL4 AD. This ORF, which did not include in-frame stop codons along the entire insert, encoded a previously unidentified cellular protein that we named RoXaN. The cDNA fragment found in the screen was called RoXaN-N1 and the protein it encodes is now referred to as RoXaN I:1-387 (Fig. 1A). RoXaN I:1-387 was tested for interactions with unrelated cellular proteins, such as p53, nuclear lamin, human eIF4G I, and PABP, and with rotavirus proteins, including NSP1, NSP2, and NSP5, fused to the GAL4 AD. No interaction was observed between RoXaN I:1-387 and these proteins when either the HIS3 or ␤-galactosidase reporter gene was used (data not shown); thus, according to the yeast two-hybrid assay standards to eliminate artifactual interactions and false positives (5), RoXaN I is a true interactor with NSP3. Complete cDNA of RoXaN. The 1.2-kb RoXaN I cDNA recovered from the yeast two-hybrid screen was a partial cDNA, as it contained a potential initiation codon but no stop codon at the 3⬘ end (Fig. 1A). Furthermore, a Northern blot of CV-1 and MA104 poly(A) RNA with this cDNA as a probe indicated that RoXaN I was encoded by an mRNA of approximately 6 kb (data not shown). In addition, we used the NCBI BLAST software (1) to compare the nucleotide sequence of RoXaN I to other sequences deposited in GenBank and identified a human expressed sequence tag (EST; GenBank accession no. AI028132) with partial sequence identity to RoXaN I:1-387. To ascertain the complete RoXaN I coding sequence, we determined the nucleotide sequence of this EST, which contains a 2,398-bp insert (RoXaN-N2; GenBank accession no. AF188530) but does not include any stop codons 5⬘ to or 3⬘ in frame with the RoXaN I ORF. This cDNA clone is now referred to as RoXaN I:1-764. During our effort to obtain the full-length cDNA of RoXaN I, the nucleotide sequence of a human genomic DNA clone derived from chromosome 22 was made available (GenBank accession no. AL035659.22) and included sequences corresponding to RoXaN I:1-387 and RoXaN I:1-764. Soon after, the nucleotide sequence of a large cDNA clone (KIAA1031; GenBank accession no. AB028954) was determined by the Kazusa DNA Research Institute (57).

3855

The KIAA1031 insert overlaps RoXaN I:1-387 and RoXaN I:1-764 sequences and contains a stop codon at the 3⬘ end of the RoXaN I ORF. We determined that the ATG codon found near the 5⬘ end of RoXaN I:1-387 is the genuine RoXaN I start codon by identifying stop codons upstream of the RoXaN I ORF in the sequence of human ESTs (for example, GenBank accession no. AI927120) and in the RoXaN I genomic sequence (GenBank accession no. AL035659). Figure 1A shows the schematic reconstitution of the complete coding sequence of RoXaN I from these different sources. The human RoXaN I protein consists of 977 aa and has a predicted molecular mass of 110 kDa (Fig. 1B). The availability of virtually the complete nucleotide sequence of the human genome and ESTs allowed the identification of a second gene, located on chromosome 16 (GenBank accession no. AC009048), encoding a protein of 971 aa which is 45% identical to RoXaN I. This protein is also known as HSPC055 and is encoded by a previously identified partial cDNA (65) (GenBank accession no. NM_014153.2). Therefore, the protein from chromosome 22 is called RoXaN I, the protein from chromosome 16 is called RoXaN II, and the family of proteins they form together is called RoXaN. RoXaN amino acid sequence analyses. To gain insight into the function of RoXaN, we investigated the presence of conserved domains and motifs in RoXaN I and II. (Unless otherwise stated, amino acid numbering will refer to the RoXaN I sequence.) pFam (http://hits.isb-sib.ch/cgi-bin/PFSCAN) analysis of the amino acid sequence of RoXaN I showed the presence of two tetratricopeptide repeats (TPRs) near the amino terminus (aa 82 to 115 and 116 to 149; Fig. 1B). A third TPR was predicted at aa 37 to 71 only by the NCBI Conserved Domain Search program (http://www.ncbi.nlm.nih.gov/Structure /cdd/wrpsb.cgi), and a fourth TPR was predicted at aa 1 to 45 when the E (expected) value of the program was lowered to 1. TPRs are short, highly degenerate motifs repeated at least three times (35) and are involved in protein-protein interactions (7). In addition, four Cys-X8-Cys-X5-Cys-X3-His (C3H1) and one C2H2-type zinc finger motif was predicted for the carboxyterminal half of RoXaN; both types of zinc finger motifs are found to be involved in protein-nucleic acid binding (32, 36). Other cysteine and histidine residues, not included in predicted zinc finger motifs, are conserved between RoXaN I and RoXaN II. The carboxy-terminal region of RoXaN (aa 904 to 941of RoXaN I and 920 to 954 of RoXaN II) exhibits a high probability to form a coiled-coil structure, as predicted by the Lupas algorithm (http://npsa-pbil.ibcp.fr/cgi-bin/primanal_lupas .pl). RoXaN I and II exhibit 49 and 54% identity at the amino acid level in the TPR and zinc finger regions, respectively, but these domains are separated by a less well-conserved, prolinerich region (aa 175 to 414) with only 18% identity. However, a cluster of amino acids (aa 257 to 271) found within this region are well conserved between RoXaN I and II. In RoXaN I, this cluster (eLDtLLdsL) matches perfectly with the protein-protein interaction motif LDxLLxxL, or LD motif, defined by Turner and colleagues from peptide repeats found in proteins of the paxillin superfamily (58, 59). In RoXaN II, the last leucine of the LD motif is changed to alanine (eLDdLLdsA). RoXaN mRNAs are expressed in many cell types. Several observations indicate that RoXaN is probably a family of ubiq-

3856

VITOUR ET AL.

uitously expressed proteins. RoXaN I:1-387 was isolated from a monkey CV-1 cell cDNA library (see above). Hybridization of poly(A) mRNAs extracted from MA104 and CV-1 cells by use of a RoXaN I:1-387 insert as a probe showed that RoXaN I mRNA was present in these two cell lines, which are both epithelial monkey kidney cell lines and fully support rotavirus replication. The use of a polyclonal serum specific for a peptide of RoXaN I indicated that the protein is present in MA104 and HEK 293 cells (M. M. Becker, unpublished results). A systematic reverse transcription-PCR study from the Kazusa DNA Research Institute, using primers designed from the sequence of the KIAA1031 clone, found RoXaN I mRNA in all the human cell types and tissues tested (http://zearth.kazusa .or.jp/huge/gfpage/KIAA1031/). ESTs containing RoXaN I or II sequences are found in cDNA libraries from all human tissue types (http://www.ncbi.nlm.nih.gov/UniGene/clust.cgi ?ORG⫽Hs&CID⫽25347 and http://www.ncbi.nlm.nih.gov /UniGene/clust.cgi?ORG⫽Hs&CID⫽179898). RoXaN I mRNA has also been identified by RNA differential display as ScRG3 (GenBank accession no. AJ223207), a cellular mRNA downregulated in the brains of scrapie-infected mice (11). This 6.5-kb mRNA was found in all the adult mouse tissues tested, with a lower expression in the liver and spleen (11). As RoXaN I mRNA is found in many types of tissues tested, RoXaN I is most likely present in many, if not all, cell types, irrespective of their susceptibility to rotavirus infection. Evolutionary conservation of RoXaN. As a result of searches of nucleotide sequences present in GenBank, clear homologues of RoXaN were found in mammals, fish (Danio rerio; GenBank accession no. AAH48884), amphibians (Xenopus laevis; GenBank accession no. BU916757), birds (Gallus gallus; GenBank accession no. BG625273), and the invertebrate chordate sea squirt (Ciona intestinalis; GenBank accession no. TC28509). Regions similar to the TPR domains and to zinc finger motifs can be found in other experimental model organisms (Saccharomyces, Drosophila, and Caenorhabditis) (data not shown) but are not present in the same protein. RoXaN proteins thus appear to be specific for chordates. Localization of NSP3 and RoXaN I interaction domains by yeast two-hybrid assay. We identified two different potential protein-protein interaction motifs in RoXaN I; therefore, it was important to establish whether the region of RoXaN INSP3 interaction included one of these motifs. Several deletion mutants of RoXaN I were constructed in the pGAD424 and pGBT9 vectors. These constructs were utilized in the yeast two-hybrid assay in conjunction with NSP3:88-313 in the corresponding vector (Fig. 2A). Deletion of the TPR region of RoXaN I (aa 1 to 175) did not abolish the interaction of RoXaN I with NSP3 (Fig. 2A). A fragment containing aa 244 to 341 of RoXaN I was still capable of interacting with NSP3: 88-313 (Fig. 2A). Thus, NSP3 interacts with the proline-rich region of RoXaN I between the TPR and the zinc finger regions. Deletion mutants of NSP3 constructed previously to identify the eIF4G I-interacting domain of NSP3 (44) were utilized in the two-hybrid assay to determine the region of NSP3 capable of interacting with RoXaN I (Fig. 2B). The results indicate that the region of NSP3 between aa 163 and 240 is necessary and sufficient for interaction with RoXaN I in a yeast two-hybrid assay. This region is located within the predicted coiled-coil

J. VIROL.

FIG. 2. Mapping of the interacting domains of RoXaN I and NSP3 by yeast two-hybrid analysis. Interactions were determined by the capacity of yeast to grow (⫹) or not (⫺) on medium lacking leucine, tryptophan, and histidine. (A) Interactions between NSP3 and RoXaN deletion mutants were tested. b, NSP3:88-313 as bait, in pGBT9 constructs; p, NSP3:88-313 as prey, in pGAD424 constructs; b,p, both bait and prey constructs were tested. The TPR domain of RoXaN I is indicated by a striped box, zinc finger motifs and conserved cysteine and histidine clusters are indicated by black boxes (for detailed mapping, see Fig. 1B), and the LD domain is indicated by a stippled box. (B) Interactions of RoXaN with NSP3 deletion mutants were tested. b, RoXaN I:1-387 as bait, in pGBT9 constructs; p, RoXaN I:1-387 as prey, in pGAD424 constructs; b,p, both bait and prey constructs were tested. The RNA-binding domain of NSP3 is indicated by a white box, the dimerization domain is indicated by a striped box, and the eIF4G I-binding domain is indicated by a gray box.

structure of the viral protein, between the RNA- and eIF4Gbinding domains (44). RoXaN I interacts with NSP3 during rotavirus infection. To determine whether the interaction between NSP3 and RoXaN I occurs during rotavirus infection, we transfected COS-7 cells with RoXaN I:1-387 fused to an amino-terminal Flag epitope tag (Flag-RoXaN I:1-387). Transfected cells were then infected with rotavirus RF at an MOI of 10 PFU/cell. Cells were lysed 6 h after infection and immunoprecipitated with a Flagspecific MAb, and then NPS3 was detected by immunoblotting with the NSP3-specific MAb ID3 (Fig. 3A). Cells transfected with RoXaN I cDNA constructs did not look different from mock-transfected cells, as judged by light microscopy. The infection protocol used for these experiments was similar to that used previously to detect NSP3-eIF4G interactions (45).

VOL. 78, 2004

FIG. 3. Interaction of RoXaN with NSP3 during rotavirus infection. (A) Coimmunoprecipitation of NSP3 with RoXaN I:1-387. (B) Coimmunoprecipitation of RoXaN I:1-387 with NSP3. (C) Interaction of full-length RoXaN I with NSP3. COS-7 cells were transfected (TF) with Flag-RoXaN I:1-387 (A and B) or T7tag-RoXaN I:1-977 (C), either mock infected or infected (IF) with rotavirus RF at an MOI of 10 PFU/cell for 6 h, and then lysed. Cell lysates were immunoprecipitated (IP) with specific MAbs as indicated. Immunoblotting (IB) was performed with an NSP3 (A), Flag (B), or T7tag (C) specific MAb. As a negative control, cell lysates were immunoprecipitated with an irrelevant NSP1- or NSP5-specific MAb. The positions of NSP3, RoXaN I, and immunoglobulin heavy (IgH) and light (IgL) chains are indicated.

Therefore, these experiments likely reveal relevant proteinprotein interactions. NSP3 coimmunoprecipitated with FlagRoXaN I:1-387 from infected and transfected cells (Fig. 3A, lane 2), but not from cells that were only infected or transfected (Fig. 3A, lanes 1 and 4). NSP3 was not detected when transfected and infected cell lysates were immunoprecipitated with a control rotavirus NSP1-specific MAb (Fig. 3A, lane 5). The NSP3-RoXaN I interaction was further confirmed by reversing the order of the antibodies used (Fig. 3B). Transfected and infected cell lysates were immunoprecipitated with either an NSP1- or NSP3-specific MAb, and Flag-RoXaN I:1-387 was detected by immunoblotting using the Flag-specific MAb (Fig. 3B). Flag-RoXaN I:1-387 coimmunoprecipitated with NSP3 (Fig. 3B, lane 7), but not with NSP1 (Fig. 3B, lanes 1 and 2), from cells that were both transfected and infected. For confirmation of the interaction between NSP3 and fulllength RoXaN I, a complete RoXaN I cDNA was generated and appended with a T7tag or Flag epitope tag. COS-7 cells

RoXaN-NSP3-eIF4G TERNARY COMPLEX

3857

were transfected with T7tag-RoXaN I:1-977 and infected. Cell lysates were immunoprecipitated with an NSP3-specific MAb, and RoXaN I was detected by immunoblotting using a T7tagspecific MAb (Fig. 3C). In this experiment, a 110-kDa protein was detected (Fig. 3C), which exhibits the expected molecular mass of full-length RoXaN I. Next, cells were transfected with T7tag-RoXaN I:1-977 and infected with rotavirus RF, and cell lysates were treated as described above (Fig. 3C). T7tagRoXaN I:1-977 was not detected after the immunoprecipitation of mock-infected cell lysate or when the NSP3-specific MAb was replaced with an NSP5-specific MAb for immunoprecipitation. The absence of coimmunoprecipitation of RoXaN I with the viral proteins NSP1 and NSP5 shows that overexpression of RoXaN I did not induce artifactual interactions, even with RNA-binding proteins, and underscores the specificity of the NSP3-RoXaN I interaction. Together, these data indicate that the interaction between NSP3 and RoXaN I, first detected by yeast two-hybrid analysis with truncated proteins, does occur with full-length RoXaN I in cells infected with rotavirus. NSP3 interacts with the LD domain of RoXaN I. The yeast two-hybrid assay showed that aa 244 to 341 of RoXaN I are capable of interacting with aa 163 to 240 of NSP3 in the absence of other viral proteins (Fig. 2). Interestingly, this region of RoXaN I is within aa 177 to 414, in which little similarity exists between RoXaN I and RoXaN II, with the exception of a cluster of conserved residues at positions 257 to 272. To determine whether this conserved region is required for the interaction of RoXaN I with NSP3, we introduced stop mutations on either side of this region of T7tag-RoXaN I:1-977 to express proteins truncated at positions 174, 256, or 280. We also generated a plasmid capable of expressing GFP fused to the amino-terminal half of NSP3 (pEGFP-NSP3:150-313). HEK 293 cells were transfected with pEGFP-NSP3:150-313 and one of the truncated T7tag-RoXaN I constructs (Fig. 4). After immunoprecipitation with the T7tag-specific MAb, the GFP-NSP3:150-313 fusion protein was detected by immunoblotting using a GFP-specific MAb. GFP-NSP3:150-313 coimmunoprecipitated with RoXaN I when cells were cotransfected with a plasmid expressing full-length RoXaN I (Fig. 4, lane 1), indicating that the presence of GFP does not perturb the interaction between NSP3 and RoXaN I. GFP-NSP3:150-313 also coimmunoprecipitated with RoXaN I:1-279, but not with RoXaN I:1-255 or RoXaN I:1-173 (Fig. 4, lanes 2, 3, and 4, respectively), suggesting that aa 256 to 279 of RoXaN I are important for the interaction with NSP3. Three amino acid triplets (aa 257 to 259 [DGD], 264 to 266 [ELD], and 268 to 270 [LLD]) are conserved between RoXaN I and II, within a region exhibiting little similarity between the two proteins. Two of the triplets (264 to 266 [ELD] and 268 to 270 [LLD]) are found within the paxillin LD motif. To assess the importance of the LD motif for NSP3-RoXaN I interaction, aa 257 to 259 (DGD), outside the LD motif, and aa 268 to 270 (LLD), within the LD motif, were changed to triple alanines by site-directed mutagenesis of T7tag-RoXaN I:177414. The wild type and the two mutant proteins were then each tested for an interaction with GFP-NSP3:150-313 (Fig. 4, lanes 5, 6, and 7) after transfection in HEK 293 cells. GFP-NSP3: 150-313 coimmunoprecipitated with wild-type RoXaN I:177414 and, although with slightly lower efficiency, with the DGD:

3858

VITOUR ET AL.

FIG. 4. NSP3 interacts with the LD domain of RoXaN I. HEK 293 cells were cotransfected with the carboxy-terminal half of NSP3 fused to GFP (GFP-NSP3:150-313) and with the full length (lane 1), deletion mutants (lanes 2 to 5), or point mutants (lanes 6 and 7) of T7tagRoXaN I, incubated for 36 h, and then lysed. Cell lysates were immunoprecipitated with a T7tag-specific MAb. Immunoblotting was performed with a GFP-specific MAb. The amino acid positions of the RoXaN I deletion mutants are indicated above each lane (lanes 2 to 5). The point mutations introduced into T7tag-RoXaN I:177-414 are indicated as 257/259 and 268/270 (lanes 6 and 7, respectively).

257/259:AAA mutant (Fig. 4, lanes 5 and 6), but not with the LLD:268/270:AAA mutant (Fig. 4, lane 7), despite similar levels of expression. Thus, alteration of the LLD:268-270 triplet, but not the DGD:257-259 triplet, was sufficient to totally abolish the interaction between NSP3 and RoXaN I, indicating that these residues, which are conserved between RoXaN I and II and included in the LD motif, are required for the interaction with NSP3. Furthermore, the fact that RoXaN I containing limited mutations or deletions and expressed under

J. VIROL.

control of the same cytomegalovirus promoter and with the same sequence tag as full-length RoXaN I is unable to interact with NSP3 is a very strong argument against a nonspecific interaction due to the overexpression of RoXaN I. NSP3 dimers interact simultaneously with RoXaN I and eIF4G I. The region of NSP3 that is necessary for its interaction with RoXaN I includes a portion of the NSP3 dimerization domain and is close to the eIF4G-binding domain (44). Dimerization of NSP3 is required for its eIF4G- and RNAbinding properties (13, 22, 44), and consequently, for its biological properties (61). By interacting with NSP3, RoXaN I could impair the binding of eIF4G to NSP3, either by disrupting NSP3 dimers or by concealing its eIF4G-binding site. Therefore, it was essential to determine whether NSP3 is capable of interacting simultaneously with RoXaN I and eIF4G. To answer this question, we used the yeast sandwich-hybrid assay described by Ozenberger and Young (39). In this assay, a third, nonhybrid protein partner is expressed, allowing reconstitution of a fully functional GAL4 transactivator by bridging the two fusion proteins that are incapable of direct interaction. First, we expressed eIF4G I:120-330 as a DBD fusion protein along with NSP3:88-313 to confirm that these two proteins, which interact in the yeast two-hybrid assay, do not activate the reporter gene in the absence of the GAL4 AD (Fig. 5, lane 4). In addition, we determined that DBD-eIF4G I:120-330 and AD-RoXaN I:1-387 did not activate the reporter gene when expressed together (Fig. 5, lane 5), but that NSP3 is capable of dimerization under these conditions (Fig. 5, lane 2). When NSP3:88-313 was expressed as the bridging protein in combination with DBD-eIF4G I:120-330 and AD-RoXaN I:1387, the transformed yeast colonies were able to grow in the absence of histidine (Fig. 5, lane 3), demonstrating that NSP3, eIF4G I, and RoXaN I can form a multiprotein complex. Deletion of the NSP3/PABP-binding domain of eIF4G I (eIF4G I:142-548) impaired the formation of this complex, and

FIG. 5. RoXaN I, NSP3, and eIF4G I form a ternary complex. Yeast strain CY770 was transformed with combinations of GAL4 DBD and GAL4 AD hybrid constructs and pYX212 nonhybrid constructs as indicated on the left. Each combination of triple transformants was selected on medium lacking leucine, uracil, and tryptophan, and six colonies were replicated on medium lacking leucine, uracil, tryptophan, and histidine to detect interactions. NSP3, NSP3:88-313; RoXaN, RoXaN I:1-387; eIF4G, eIF4G I:120-330; eIF4G⌬, eIF4G I:142-548; ⫺, empty vector.

VOL. 78, 2004

RoXaN-NSP3-eIF4G TERNARY COMPLEX

3859

FIG. 6. Interaction of RoXaN I, NSP3, and eIF4G I during rotavirus infection. COS-7 cells were transfected with T7tag-RoXaN I:1977, either mock infected (lanes 1 and 3) or infected with rotavirus RF at an MOI of 10 PFU/cell (lanes 2 and 4) for 6 h, and then lysed. Cell lysates were either immunoprecipitated with an eIF4G I-specific antiserum (lanes 1 and 2) or directly resolved by SDS-PAGE (lanes 3 and 4). Immunoblotting was performed with a T7tag-specific MAb. Total cell lysates from mock-infected (lane 3) or infected (lane 4) cells are shown.

yeast colonies were unable to grow on selective medium (Fig. 5, lane 6). With the controls described above, these results demonstrate that the interaction of RoXaN I with NSP3 does not impair the binding of NSP3 to eIF4G I and that the NSP3 dimer can physically interact simultaneously with these two cellular proteins. RoXaN I, NSP3, and eIF4G I form a ternary complex in rotavirus-infected cells. The results obtained with the yeast sandwich-hybrid assay, which uses truncated fusion proteins, do not exclude the possibility that, in mammalian cells, binding of full-length RoXaN I to NSP3 may impair the formation of an NSP3-eIF4G I complex. If this were the case, RoXaN I would not be detected after immunoprecipitation of endogenous eIF4G I from infected cell lysates. To investigate the presence of a RoXaN I-NSP3-eIF4G I complex during rotavirus infection, COS-7 cells were transfected with T7tag-RoXaN I and infected with rotavirus RF, and the cell lysates were immunoprecipitated with an eIF4G I-specific antiserum (Fig. 6). RoXaN I was detected by immunoblotting using the T7tagspecific MAb and was only seen in lysates from cells that were both infected and transfected (Fig. 6, lane 2). These results indicate that a complex including RoXaN I, eIF4G I, and NSP3 is also formed during rotavirus infection. RoXaN I and eIF4G I associate only in the presence of NSP3. The presence of RoXaN I and eIF4G I in the same protein complex during rotavirus infection suggests that RoXaN I may also be involved in cellular translation and that eIF4G I and RoXaN I may interact in uninfected cells. The fact that T7tag-RoXaN I:1-977 did not coimmunoprecipitate with eIF4G I (Fig. 6, lane 1) may indicate that this interaction is transient, as would be expected for the stepwise assembly of a macromolecular complex on an mRNA molecule during translational initiation (19). Therefore, we examined the possibility that a small fraction of eIF4G I and RoXaN I interacts in uninfected cells and that overexpression of the two proteins allows detection of this interaction in the absence of rotavirus infection. We transfected HEK 293 cells with Flag-eIF4G I, T7tag-RoXaN I:1-977, and one of several constructs contain-

FIG. 7. Interaction of RoXaN I with eIF4G I is dependent on the two protein-binding domains of NSP3. HEK 293 cells were transfected with T7tag-RoXaN I:1-977, Flag-eIF4G I, and various deletion mutants of NSP3 fused to GFP, incubated for 36 h, and then lysed. Cell lysates were immunoprecipitated with a T7tag-specific MAb. Immunoblotting was performed with a GFP-specific or Flag-specific MAb, as indicated. Total cell lysates from transfected cells are shown. The portion of NSP3 present in the GFP fusion proteins used for transfection is shown above each lane. Proteins migrating faster than GFP fusion proteins are likely degradation products of the protein and are often seen with this antibody (D. Vitour, unpublished observations).

ing the different NSP3 protein-binding domains fused to GFP (Fig. 7). The construct GFP-NSP3:150-313 expresses the RoXaN I- and eIF4G-binding domains of NSP3. GFP-NSP3: 206-313 and GFP-NSP3:150-238 express the eIF4G- and RoXaN I-binding domains, respectively. GFP-NSP3:206-238 expresses a fragment of NSP3 which is not known to bind either protein or RNA and is used as a negative control. Cell lysates were immunoprecipitated with a T7tag-specific MAb to isolate RoXaN I and proteins bound to it, followed by immunoblotting with a rabbit antiserum to detect eIF4G I. Similar to what was seen previously, even when eIF4G I and RoXaN I were abundant, an interaction between the two proteins (Fig. 7, lane 4) was not detectable. In addition, the expression of only one of the two protein-binding domains of NSP3 did not allow detection of the interaction between eIF4G I and RoXaN I (Fig. 7, lanes 2 and 3). Immunoblotting with a GFPspecific MAb confirmed that GFP-NSP3:150-313 and GFPNSP3:150-238, which both contain the RoXaN I-binding domain, coimmunoprecipitated with RoXaN I (Fig. 7, lanes 1 and 2, third row). However, neither GFP-NSP3:206-313 nor GFPNSP3:206-238, which do not contain the RoXaN I-binding domain, immunoprecipitated with RoXaN I (Fig. 7, lanes 3 and 4, third row). Thus, overexpression of both cellular proteins or the expression of a deletion mutant of NSP3 did not reveal a detectable interaction between RoXaN I and eIF4G I.

3860

VITOUR ET AL.

RoXaN I can be recovered along with eIF4G only when it is tethered to NSP3. DISCUSSION We have identified RoXaN I as a novel cellular partner for rotavirus NSP3. The interaction between RoXaN I and NSP3 does not require other viral or cellular components, since the interaction was initially detected in yeast with proteins lacking their RNA-binding domains. Thus, it appears likely that the interaction between RoXaN I and NSP3 consists of direct protein-protein contact. Interestingly, the region of RoXaN I determined to interact with NSP3 involves a domain conserved between RoXaN I and II which is embedded in a larger region showing the greatest dissimilarity between the two proteins. Limited mutations in this conserved domain abolish the capacity of RoXaN I to coimmunoprecipitate with NSP3, emphasizing the specificity of the interaction between NSP3 and RoXaN I. We have shown that the interaction initially detected with truncated RoXaN I and NSP3 fusion proteins in yeast also occurs with full-length RoXaN I during rotavirus infection. RoXaN I is present in CV-1 and MA104 cells (M. M. Becker, unpublished results), which are susceptible to rotavirus infection; thus, the interaction we observed is relevant for rotavirus replication. An analysis of the predicted amino acid sequence of RoXaN I reveals the presence of several functional domains involved in protein-protein and protein-nucleic acid interactions, and particularly, in multiprotein complexes. First, a TPR region is present at the amino terminus of RoXaN. TPRs are alphahelical domains in which repetition forms a groove that specifically fits the structure of a protein partner (12, 23, 24). TPR proteins were first identified in cell cycle studies of yeast (25, 56), and TPRs have since been observed in a growing number of unrelated proteins (2, 7, 23). As a consequence, the presence of TPRs is not predictive of a particular function, but rather reveals a potential for interacting with other proteins. The identification of other cellular proteins capable of interaction with the TPR region of RoXaN I will lead to a more precise definition of the cellular function of RoXaN I. Four zinc finger motifs of the C3H1 type and one of the C2H2 type are predicted for the carboxy-terminal half of RoXaN. These two kinds of zinc finger motifs are seen in RNA- and DNAbinding proteins, respectively (32). Proteins containing C2H2 zinc finger motifs were first identified in DNA-binding proteins involved in transcriptional regulation, such as transcription factor TFIIIA (10). The hypothesis that RoXaN I acts as a transcription factor cannot be ruled out, but a single C2H2 zinc finger is generally unable to bind DNA, and the recognition of the DNA sequence in target genes usually requires a series of three or more C2H2 zinc fingers (36). The C3H1 zinc fingers are present in sequence-specific RNA-binding proteins, such as tristetraprolin. Tristetraprolin specifically binds an AU-rich sequence present at the 3⬘ end of cytokine mRNAs, including tumor necrosis factor alpha mRNA, and promotes their degradation (6). As RoXaN I is present in cells before viral protein synthesis begins, it can be envisioned that RoXaN I interacts with the viral RNAs and facilitates a first round of translation prior to the production of NSP3. Because the role of NSP3 in rotavirus mRNA translation is

J. VIROL.

well established, the interaction between RoXaN I and NSP3 suggests a role for RoXaN I in translation. To support this hypothesis, we have shown, both in yeast and during rotavirus infection, that NSP3, eIF4G I, and RoXaN I form a ternary complex. Since the domains involved in the interaction between RoXaN I and NSP3 are not RNA-binding domains, it is likely that mRNA molecules, either viral or cellular, also are present in NSP3-RoXaN I-eIF4G I ternary complexes. Moreover, we have previously shown that rotavirus NSP3 mimics the function of PABP in translation in that it interacts with the 3⬘ end of viral mRNAs and binds eIF4G (45). The functional analogy between NSP3 and PABP is reinforced by the recent observation of an interaction between PABP and paxillin (63). Paxillin is a cytoskeletal adapter protein and the founder of the superfamily of paxillin proteins, whose members contain repeated LD motifs according to the consensus sequence LDxLLxxL (55, 58, 59). A paxillin-PABP interaction requires the amino-terminal half of paxillin, which contains five LD motifs. Interestingly, NSP3 binds to the LD motif of RoXaN I and has been described as a cytoskeleton-associated protein (26, 37). Numerous links exist between the translational machinery and the cytoskeleton (15, 30, 31). For example, the PABP-paxillin interaction has been proposed to target the delivery of mRNAs to the leading edge of migrating cells (63). The lack of a detectable interaction between RoXaN I and eIF4G I in the absence of NSP3 could be the result of a weak, transient, or indirect interaction between the two cellular proteins. We can also envision that, in uninfected cells, eIF4G I and RoXaN I are functionally linked through a series of regulated protein-protein or protein-RNA interactions. Rotaviruses could circumvent this regulation as a result of the continuous interaction of NSP3 with both RoXaN I and eIF4G I. The results presented in this study indicate that rotavirus NSP3 interacts with a novel cellular protein, RoXaN I, which contains a unique combination of structural domains. This interaction is currently the only known function of RoXaN I and the sole tool available to decipher the role of this protein in uninfected and infected cells. ACKNOWLEDGMENTS We acknowledge the skillful technical assistance of Nathalie Castagne´. We thank Kathleen H. Young (Neuroscience Discovery Research, Wyeth Research, Princeton, N.J.) for the gift of the CY770 yeast strain used in the yeast sandwich-hybrid assay, Nahum Sonenberg (McGill University, Montreal, Canada) for the gift of eIF4G I-specific antiserum, Takahiro Nagase (Kazusa DNA Research Institute, Chiba, Japan) for the KIAA1031 clone, Robert E. Rhoads (Louisiana State University, Shreveport) for human eIF4G I clone HFC1, and Thierry Grange (Institut Jacques Monod, Universite´s Paris 6-7, Paris, France) for the human PABP cDNA. This work was supported in part by the Progamme de Recherche Fondamentale en Microbiologie, Maladie Infectieuse et Parasitologie. D.V. was supported by a fellowship from le Ministe`re de la Recherche et des Nouvelles Technologies. P.L. was supported by a fellowship from Ministe`re de l’Education Nationale de la Recherche et de la Technologie. M.M.B. was supported by European Union grant number QLK2-CT-2002-01249. P.V. and D.P. are members of INRA staff. REFERENCES 1. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389–3402.

VOL. 78, 2004 2. Andrade, M. A., C. Perez-Iratxeta, and C. P. Ponting. 2001. Protein repeats: structures, functions, and evolution. J. Struct. Biol. 134:117–131. 3. Aponte, C., N. M. Mattion, M. K. Estes, A. Charpilienne, and J. Cohen. 1993. Expression of two bovine rotavirus non-structural proteins (NSP2, NSP3) in the baculovirus system and production of monoclonal antibodies directed against the expressed proteins. Arch. Virol. 133:85–95. 4. Aponte, C., D. Poncet, and J. Cohen. 1996. Recovery and characterization of a replicase complex in rotavirus-infected cells by using a monoclonal antibody against NSP2. J. Virol. 70:985–991. 5. Bartel, P., C. T. Chien, R. Sternglanz, and S. Fields. 1993. Elimination of false positives that arise in using the two-hybrid system. BioTechniques 14:920–924. 6. Blackshear, P. J. 2002. Tristetraprolin and other CCCH tandem zinc-finger proteins in the regulation of mRNA turnover. Biochem. Soc. Trans. 30:945– 952. 7. Blatch, G. L., and M. Lassle. 1999. The tetratricopeptide repeat: a structural motif mediating protein-protein interactions. Bioessays 21:932–939. 8. Bushell, M., D. Poncet, W. E. Marissen, H. Flotow, R. E. Lloyd, M. J. Clemens, and S. J. Morley. 2000. Cleavage of polypeptide chain initiation factor eIF4GI during apoptosis in lymphoma cells: characterisation of an internal fragment generated by caspase-3-mediated cleavage. Cell Death Differ. 7:628–636. 9. Chizhikov, V., and J. T. Patton. 2000. A four-nucleotide translation enhancer in the 3⬘-terminal consensus sequence of the nonpolyadenylated mRNAs of rotavirus. RNA 6:814–825. 10. Clemens, K. R., V. Wolf, S. J. McBryant, P. Zhang, X. Liao, P. E. Wright, and J. M. Gottesfeld. 1993. Molecular basis for specific recognition of both RNA and DNA by a zinc finger protein. Science 260:530–533. 11. Dandoy-Dron, F., F. Guillo, L. Benboudjema, J. P. Deslys, C. Lasmezas, D. Dormont, M. G. Tovey, and M. Dron. 1998. Gene expression in scrapie. Cloning of a new scrapie-responsive gene and the identification of increased levels of seven other mRNA transcripts. J. Biol. Chem. 273:7691–7697. 12. Das, A. K., P. W. Cohen, and D. Barford. 1998. The structure of the tetratricopeptide repeats of protein phosphatase 5: implications for TPR-mediated protein-protein interactions. EMBO J. 17:1192–1199. 13. Deo, R. C., C. M. Groft, K. R. Rajashankar, and S. K. Burley. 2002. Recognition of the rotavirus mRNA 3⬘ consensus by an asymmetric NSP3 homodimer. Cell 108:71–81. 14. Dever, T. E. 1999. Translation initiation: adept at adapting. Trends Biochem. Sci. 24:398–403. 15. Dreyfuss, G., V. N. Kim, and N. Kataoka. 2002. Messenger-RNA-binding proteins and the messages they carry. Nat. Rev. Mol. Cell Biol. 3:195–205. 16. Estes, M. K. 2001. Rotaviruses and their replication, p. 1747–1785. In D. M. Knipe, P. M. Howley, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. E. Straus (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa. 17. Fields, S., and O. Song. 1989. A novel genetic system to detect proteinprotein interactions. Nature 340:245–246. 18. Gallie, D. R. 1991. The cap and poly(A) tail function synergistically to regulate mRNA translational efficiency. Genes Dev. 5:2108–2116. 19. Gallie, D. R. 2002. Protein-protein interactions required during translation. Plant Mol. Biol. 50:949–970. 20. Gradi, A., H. Imataka, Y. V. Svitkin, E. Rom, B. Raught, S. Morino, and N. Sonenberg. 1998. A novel functional human eukaryotic translation initiation factor 4G. Mol. Cell. Biol. 18:334–342. 21. Grange, T., C. M. de Sa, J. Oddos, and R. Pictet. 1987. Human mRNA polyadenylate binding protein: evolutionary conservation of a nucleic acid binding motif. Nucleic Acids Res. 15:4771–4787. 22. Groft, C. M., and S. K. Burley. 2002. Recognition of eIF4G by rotavirus NSP3 reveals a basis for mRNA circularization. Mol. Cell. Biol. 9:1273–1283. 23. Groves, M. R., and D. Barford. 1999. Topological characteristics of helical repeat proteins. Curr. Opin. Struct. Biol. 9:383–389. 24. Groves, M. R., N. Hanlon, P. Turowski, B. A. Hemmings, and D. Barford. 1999. The structure of the protein phosphatase 2A PR65/A subunit reveals the conformation of its 15 tandemly repeated HEAT motifs. Cell 96:99–110. 25. Hirano, T., N. Kinoshita, K. Morikawa, and M. Yanagida. 1990. Snap helix with knob and hole: essential repeats in S. pombe nuclear protein nuc2⫹. Cell 60:319–328. 26. Hua, J., X. Chen, and J. T. Patton. 1994. Deletion mapping of the rotavirus metalloprotein NS53 (NSP1): the conserved cysteine-rich region is essential for virus-specific RNA binding. J. Virol. 68:3990–4000. 27. Iizuka, N., L. Najita, A. Franzusoff, and P. Sarnow. 1994. Cap-dependent and cap-independent translation by internal initiation of mRNAs in cell extracts prepared from Saccharomyces cerevisiae. Mol. Cell. Biol. 14:7322– 7330. 28. Imai, M., K. Akatani, N. Ikegami, and Y. Furuichi. 1983. Capped and conserved terminal structures in human rotavirus genome double-stranded RNA segments. J. Virol. 47:125–136. 29. Imataka, H., A. Gradi, and N. Sonenberg. 1998. A newly identified Nterminal amino acid sequence of human eIF4G binds poly(A)-binding protein and functions in poly(A)-dependent translation. EMBO J. 17:7480– 7489.

RoXaN-NSP3-eIF4G TERNARY COMPLEX

3861

30. Jansen, R. P. 1999. RNA-cytoskeletal associations. FASEB J. 13:455–466. 31. Kloc, M., N. R. Zearfoss, and L. D. Etkin. 2002. Mechanisms of subcellular mRNA localization. Cell 108:533–544. 32. Krishna, S. S., I. Majumdar, and N. V. Grishin. 2003. Structural classification of zinc fingers: survey and summary. Nucleic Acids Res. 31:532–550. 33. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. 34. Lindenbaum, P. 1998. CloneIt: finding cloning strategies, in-frame deletions and frameshifts. Bioinformatics 14:465–466. 35. Main, E. R., Y. Xiong, M. J. Cocco, L. D’Andrea, and L. Regan. 2003. Design of stable alpha-helical arrays from an idealized TPR motif. Structure (Camb.) 11:497–508. 36. Matthews, J. M., and M. Sunde. 2002. Zinc fingers—folds for many occasions. IUBMB Life 54:351–355. 37. Mattion, N. M., J. Cohen, C. Aponte, and M. K. Estes. 1992. Characterization of an oligomerization domain and RNA-binding properties on rotavirus nonstructural protein NS34. Virology 190:68–83. 38. Michel, Y. M., A. M. Borman, S. Paulous, and K. M. Kean. 2001. Eukaryotic initiation factor 4G-poly(A) binding protein interaction is required for poly(A) tail-mediated stimulation of picornavirus internal ribosome entry segment-driven translation but not for X-mediated stimulation of hepatitis C virus translation. Mol. Cell. Biol. 21:4097–4109. 39. Ozenberger, B. A., and K. H. Young. 1995. Functional interaction of ligands and receptors of the hematopoietic superfamily in yeast. Mol. Endocrinol. 9:1321–1329. (Erratum, 10:936, 1996.) 40. Parashar, U. D., J. S. Bresee, J. R. Gentsch, and R. I. Glass. 1998. Rotavirus. Emerg. Infect. Dis. 4:561–570. 41. Parashar, U. D., E. G. Hummelman, J. S. Bresee, M. A. Miller, and R. I. Glass. 2003. Global illness and deaths caused by rotavirus disease in children. Emerg. Infect. Dis. 9:565–572. 42. Patton, J. T., Z. Taraporewala, D. Chen, V. Chizhikov, M. Jones, A. Elhelu, M. Collins, K. Kearney, M. Wagner, Y. Hoshino, and V. Gouvea. 2001. Effect of intragenic rearrangement and changes in the 3⬘ consensus sequence on NSP1 expression and rotavirus replication. J. Virol. 75:2076–2086. 43. Pestova, T. V., and C. U. Hellen. 1999. Ribosome recruitment and scanning: what’s new? Trends Biochem. Sci. 24:85–87. 44. Piron, M., T. Delaunay, J. Grosclaude, and D. Poncet. 1999. Identification of the RNA-binding, dimerization, and eIF4GI-binding domains of rotavirus nonstructural protein NSP3. J. Virol. 73:5411–5421. 45. Piron, M., P. Vende, J. Cohen, and D. Poncet. 1998. Rotavirus RNA-binding protein NSP3 interacts with eIF4GI and evicts the poly(A) binding protein from eIF4F. EMBO J. 17:5811–5821. 46. Poncet, D., C. Aponte, and J. Cohen. 1993. Rotavirus protein NSP3 (NS34) is bound to the 3⬘ end consensus sequence of viral mRNAs in infected cells. J. Virol. 67:3159–3165. 47. Poncet, D., and J. Cohen. 1989. A plaque hybridization assay for rotaviruses. J. Virol. Methods 26:27–33. 48. Poncet, D., P. Lindenbaum, R. L’Haridon, and J. Cohen. 1997. In vivo and in vitro phosphorylation of rotavirus NSP5 correlates with its localization in viroplasms. J. Virol. 71:34–41. 49. Preiss, T., and M. W. Hentze. 1998. Dual function of the messenger RNA cap structure in poly(A)-tail-promoted translation in yeast. Nature 392:516–520. 50. Preiss, T., and M. W. Hentze. 1999. From factors to mechanisms: translation and translational control in eukaryotes. Curr. Opin. Genet. Dev. 9:515–521. 51. Proweller, A., and J. S. Butler. 1997. Ribosome concentration contributes to discrimination against poly(A)- mRNA during translation initiation in Saccharomyces cerevisiae. J. Biol. Chem. 272:6004–6010. 52. Sachs, A. B., P. Sarnow, and M. W. Hentze. 1997. Starting at the beginning, middle, and end: translation initiation in eukaryotes. Cell 89:831–838. 53. Sachs, A. B., and G. Varani. 2000. Eukaryotic translation initiation: there are (at least) two sides to every story. Nat. Struct. Biol. 7:356–361. 54. Sambrook, J., and D. Russell. 2001. Molecular cloning: a laboratory manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 55. Schaller, M. D. 2001. Paxillin: a focal adhesion-associated adaptor protein. Oncogene 20:6459–6472. 56. Sikorski, R. S., M. S. Boguski, M. Goebl, and P. Hieter. 1990. A repeating amino acid motif in CDC23 defines a family of proteins and a new relationship among genes required for mitosis and RNA synthesis. Cell 60:307–317. 57. Suyama, M., T. Nagase, and O. Ohara. 1999. HUGE: a database for human large proteins identified by Kazusa cDNA sequencing project. Nucleic Acids Res. 27:338–339. 58. Tumbarello, D. A., M. C. Brown, and C. E. Turner. 2002. The paxillin LD motifs. FEBS Lett. 513:114–118. 59. Turner, C. E. 2000. Paxillin interactions. J. Cell Sci. 113:4139–4140. 60. Varani, G., and F. H. Allain. 2002. How a rotavirus hijacks the human protein synthesis machinery. Nat. Struct. Biol. 9:158–160. 61. Vende, P., M. Piron, N. Castagne, and D. Poncet. 2000. Efficient translation of rotavirus mRNA requires simultaneous interaction of NSP3 with the eukaryotic translation initiation factor eIF4G and the mRNA 3⬘ end. J. Virol. 74:7064–7071.

3862

VITOUR ET AL.

62. Wells, S. E., P. E. Hillner, R. D. Vale, and A. B. Sachs. 1998. Circularization of mRNA by eukaryotic translation initiation factors. Mol. Cell 2:135–140. 63. Woods, A. J., M. S. Roberts, J. Choudhary, S. T. Barry, Y. Mazaki, H. Sabe, S. J. Morley, D. R. Critchley, and J. C. Norman. 2002. Paxillin associates with poly(A)-binding protein 1 at the dense endoplasmic reticulum and the leading edge of migrating cells. J. Biol. Chem. 277:6428–6437. 64. Yan, R., W. Rychlik, D. Etchison, and R. E. Rhoads. 1992. Amino acid

J. VIROL. sequence of the human protein synthesis initiation factor eIF-4 gamma. J. Biol. Chem. 267:23226–23231. 65. Zhang, Q. H., M. Ye, X. Y. Wu, S. X. Ren, M. Zhao, C. J. Zhao, G. Fu, Y. Shen, H. Y. Fan, G. Lu, M. Zhong, X. R. Xu, Z. G. Han, J. W. Zhang, J. Tao, Q. H. Huang, J. Zhou, G. X. Hu, J. Gu, S. J. Chen, and Z. Chen. 2000. Cloning and functional analysis of cDNAs with open reading frames for 300 previously undefined genes expressed in CD34⫹ hematopoietic stem/progenitor cells. Genome Res. 10:1546–1560.