The Sireviruses, a plant-specific lineage of the Ty1/copia retrotransposons, interact with a family of proteins related to dynein light chain 8

Abstract

Plant genomes are rich in long terminal repeat retrotransposons, and here we describe a plant-specific lineage of Ty1/copia elements called the Sireviruses. The Sireviruses vary greatly in their genomic organization, and many have acquired additional coding information in the form of an envelope-like open reading frame and an extended gag gene. Two-hybrid screens were conducted with the novel domain of Gag (the Gag extension) encoded by a representative Sirevirus from maize (Zea mays) called Hopie. The Hopie Gag extension interacts with a protein related to dynein light chain 8 (LC8). LC8 also interacts with the Gag extension from a Hopie homolog from rice (Oryza sativa). Amino acid motifs were identified in both Hopie Gag and LC8 that are responsible for the interaction. Two amino acids critical for Gag recognition map within the predicted LC8-binding cleft. Two-hybrid screens were also conducted with the Gag extension encoded by the soybean (Glycine max) SIRE1 element, and an interaction was found with light chain 6 (LC6), a member of the LC8 protein family. LC8 and LC6 proteins are components of the dynein microtubule motor, with LC8 being a versatile adapter that can bind many unrelated cellular proteins and viruses. Plant LC8 and LC6 genes are abundant and divergent, yet flowering plants do not encode other components of the dynein motor. Although, to our knowledge, no cellular roles for plant LC8 family members have been proposed, we hypothesize that binding of LC8 proteins to Gag aids in the movement of retrotransposon virus-like particles within the plant cell or possibly induces important conformational changes in the Gag protein.

Figure 1.

Genomic organization of the Sireviruses. A phylogenetic analysis of the plant Pseudoviridae (Ty1/copia elements) was previously described (Gao et al., 2003). The neighbor-joining tree resulting from this analysis revealed two major groups, the classical Ty1/copia elements and the Sireviruses. The Sirevirus clade contains elements with unusually large Gag proteins and varied genomic organization with respect to env-like ORFs and breaks in translation between gag and pol.

Figure 2.

Cartoon depicting the structural features of the Gag and Env-like proteins of individual Sirevirus elements. Boxes represent ORFs followed by their approximate length in amino acids. Arrows represent primers that delimit the regions of the Gag extension that were cloned and used for two-hybrid assays. The asterisks identify those ORFs predicted from a nucleotide consensus. ZK, CX₂CX₃HX₄C zinc knuckle.

Figure 3.

Interaction of some Sireviruses with LC8 family members. A, Hopie has all features associated with Sireviruses, namely an extended gag gene, a break in reading frame between gag and pol, and an env-like ORF. Primers (represented by arrows) amplify the gag extension, which was fused to the LexA DNA-binding domain for use in a two-hybrid screen. Boxed arrows represent LTRs; open rectangles represent coding regions. B, Ten-fold serial dilutions of yeast cells expressing either the Hopie (maize) or Osr10 (rice) Gag extensions as bait and ZmLC8 as the prey. Yeast grown on −Leu −Trp serve as a control for cell number. Yeast grown on media lacking His identify two-hybrid interactions. C, ZmLC8 and the Hopie Gag extension interact in vitro. A Coomassie-stained polyacrylamide gel shows purification of GST-LC8. The autoradiograph demonstrates that GST-LC8 can immunoprecipitate radiolabeled Gag translated in vitro. D, Ten-fold serial dilutions of yeast cells expressing the SIRE1-4 Gag extension as a bait and GmLC6 as the prey. The plasmid containing the B42p transcription activation domain is under the control of the galactose (Gal) promoter. Yeast grown on −Leu −Trp 2% dextrose (Dex) serve as a loading control. Yeast grown on −Leu −Trp 2% Gal 1% raffinose (Raf) show a growth defect when both the Gag and GmLC6 proteins are present.

Figure 4.

Mutations in Gly and Cys residues of the Hopie and Osr10 Sirevirus Gag extension abrogate binding to ZmLC8. A, The first section depicts 10-fold serial dilutions of yeast cells expressing the Hopie Gag or the G85R or C88R mutants along with ZmLC8. The second section depicts cells expressing either the cloned Osr10 Gag or the G81E or C84E mutants. B, Hopie Gag mutants made by site-directed mutagenesis. Yeast grown on −Leu −Trp serve as a loading control. All Gag mutant constructs are fused to the LexA DNA-binding domain, and two-hybrid interactions are measured against GAD-ZmLC8. C, ClustalX alignment of maize and rice Sirevirus gag sequences. RiceA is the sequence used as bait. RiceCon is the Osr10 consensus sequence, whereas RiceB is another Osr10 sequence cloned during our experiments. MaizeA is the Hopie sequence used for these experiments, whereas MaizeB is from the sequence present in GenBank for the first-described maize Sirevirus (Hopie) element. The LC8 interaction domain is located between a predicted coiled-coil domain and a CX₂CX₃HX₄C zinc knuckle.

Figure 5.

ZmLC8 mutants that abrogate binding to the Hopie Gag extension. A, Ten-fold serial dilutions are shown of cells expressing three of the four ZmLC8 mutants: S93P, Y94N, and V112D. All of these mutants fail to grow on media lacking His when either the Hopie or Osr10 Sirevirus Gag extension is expressed as a GAD fusion. B, Alignment of some LC8 protein sequences from animals, fungi, and plants. Arrows represent amino acid positions, which, when mutated, abrogate binding of ZmLC8 to the Hopie and Osr10 Gag extensions. C, The amino acids important in binding the maize and rice Gag extensions were mapped onto the crystal structure of human LC8 bound to a peptide. LC8 was crystallized as a dimer, and the blue and pink domains represent a monomer of that dimer. The green and brown residues represent the bound peptide. Yellow amino acids were identified in this study as being important for binding the Gag extension.

Figure 6.

Neighbor-joining phylogenetic tree of LC8 (represented by bold-line branches) and LC6 sequences (represented by thin-line branches) with emphasis placed on those genes found in plants. Arrows identify the LC8/LC6 sequences discussed in this study. Numbers represent bootstrap values (100 replicates). GenBank or PlantGDB identifiers follow each entry. PlantGDB identifiers contain “tuc” and can be retrieved from www.plantgdb.org. For entries found in GenBank, the genus initial followed by the species name is given. For plant sequences, hosts can be identified by the following abbreviations: (AT, Arabidopsis; BV, Beta vulgaris; CS, Citrus sinesis; GA, Gossypium arboretum; GM, soybean; HV, Hordeum vulgare; LE, tomato; LJ, Lotus japonicus; MT, Medicago truncatula; OS, rice; SB, Sorghum bicolor; SC, Secale cereale; ST, Solanum tuberosum; TA, Triticum aestivum; ZM, maize).

Figure 7.

Interactions between the various Arabidopsis LC8/LC6 proteins and either the Hopie Gag extension (A) or the SIRE1-4 Gag extension (B). Ten-fold serial dilutions were plated onto either control media or media lacking His to identify protein-protein interactions. In these experiments, the baits were the different Arabidopsis dynein light chains expressed as LexA fusions. In B, the LexA plasmid is under the control of the GAL promoter. The preys are indicated above each section. The last sections in both A and B serve as negative controls for autoactivation of the HIS3 marker gene by the bait. Cell number controls for these sections were consistent with the other sections (data not shown).