Hantavirus nucleocapsid protein oligomerization

Abstract

Hantaviruses are enveloped, negative-strand RNA viruses which can be lethal to humans, causing either a hemorrhagic fever with renal syndrome or a hantaviral pulmonary syndrome. The viral genomes consist of three RNA segments: the L segment encodes the viral polymerase, the M segment encodes the viral surface glycoproteins G1 and G2, and the S segment encodes the nucleocapsid (N) protein. The N protein is a 420- to 430-residue, 50-kDa protein which appears to direct hantavirus assembly, although mechanisms of N protein oligomerization, RNA encapsidation, budding, and release are poorly understood. We have undertaken a biochemical and genetic analysis of N protein oligomerization. Bacterially expressed N proteins were found by gradient fractionation to associate not only as large multimers or aggregates but also as dimers or trimers. Chemical cross-linking of hantavirus particles yielded N protein cross-link products with molecular masses of 140 to 150 kDa, consistent with the size of an N trimer. We also employed a genetic, yeast two-hybrid method for monitoring N protein interactions. Analyses showed that the C-terminal half of the N protein plus the N-terminal 40 residues permitted association with a full-length N protein fusion. These N-terminal 40 residues of seven different hantavirus strains were predicted to form trimeric coiled coils. Our results suggest that coiled-coil motifs contribute to N protein trimerization and that nucleocapsid protein trimers are hantavirus particle assembly intermediates.

FIG. 1

Hantavirus N protein. Shown at the top is the map of the coding region of the N protein of SNV from residues 1 to 428. Vertical lines are indicative of the number of charged residues per five residue segments: thin lines depict one charged residue per segment while progressively thicker lines depict two, three, and four charged residues per segment. Cysteine (C) residues are denoted above the map line, while positions of restriction endonuclease sites in the coding region are shown below the line. Beneath the map, the SNV N protein sequence (3) is provided, along with a comparison to the sequence of PHV (20). Amino acids are depicted in standard single letter code, while restriction enzyme abbreviations are as follows: Hpa, HpaI; Eco, EcoRI; Sn, SnaB1; Sma, SmaI; Ssp, SspI; E57, Eco571; Bam, BamHI; EcoV, EcoRV.

FIG. 2

Sucrose gradient fractionation of hantavirus N proteins. The His-tagged hantavirus (SNV) N protein (5) was expressed in Escherichia coli cells, solubilized in 100 mM HEPES (pH 7.0)–1 M MgSO₄, and applied to a 15 to 30% sucrose gradient in 50 mM Tris (pH 7.4)–100 mM NaCl–0.1 mM EDTA in parallel with high-molecular-weight marker proteins (Amersham Pharmacia Biotech). Gradients were centrifuged at 4°C at 243,000 × g for 18 h, after which 25 0.2-ml fractions were collected from the gradient top to bottom, as indicated. Fractions were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and detected by Coomassie blue staining for markers (19) or by N protein immunoblotting (11, 17, 18), using anti-hantaviral N protein monoclonal antibody Hy11E5EF6CE7 as the primary antibody and an alkaline phosphatase-conjugated secondary antibody. After detection, protein levels were quantitated densitometrically and are displayed as percentages of the highest detected protein level across all gradient fractions for a given protein. Gradient profiles correspond to the SNV N protein (50 kDa) (A), bovine serum albumin (67 kDa) (B), lactate dehydrogenase tetramers (140 kDa) (C), and catalase tetramers (232 kDa) (D).

FIG. 3

Cross-linking of hantavirus N proteins. Virus particles (lanes A and B) and cell pellets (lanes E and F) from PHV-infected Vero E6 (African green monkey kidney) cells were collected and either mock treated (lanes A and E) or treated with a 1 mM concentration of the cross-linking agent BMH (11, 17, 18) (lanes B and F). After treatments, samples were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis along with size markers (lanes C and F), and N proteins were detected by immunoblotting, using the anti-N protein monoclonal antibody Hy12A6CF6 as the primary antibody and an alkaline phosphatase-conjugated anti-mouse secondary antibody (Promega). N protein and 200-, 97-, and 68-kDa size marker bands are as shown. Cross-link products are indicated by arrowheads. The calculated sizes of cross-link products were 146 kDa (lane B), 142 kDa (lane F), and 135 kDa (lane F), while the predicted size of the N protein is 50 kDa.

FIG. 4

Recombinant N protein constructs. Shown are maps of the coding regions for LexA-N fusion proteins which were used to assay N protein-protein interactions, in conjunction with a wt SNV N protein (3) fusion to the VP16 (12) transcriptional activation domain. LexA domains are indicated by white boxes, and SNV N protein domains are depicted by black boxes, with amino acid numbers (see Fig. 1) shown at the top. Fusion protein constructs were based on Lex-EBfill, which expresses the DNA binding domain of LexA (2, 6, 12, 16). Lex-N expresses the LexA domain, fused to the full-length, wt N protein coding region. Lex-N1-357, Lex-N1-171, and Lex-N1-172 encode Lex-N fusion proteins, which were truncated after N protein codons 357, 171, and 172, respectively. Lex-N172-428 encodes a fusion protein which has an N-terminal N protein coding region deletion, removing N protein codons 1 to 171. Lex-N172-357 encodes an N protein truncated at its amino (residues 1 to 171) and carboxy (residues 358 to 428) termini. The Lex-NΔ157-372, Lex-NΔ40-195, and LexNΔ40-214 plasmids encode Lex-N proteins with the indicated codon deletions of the N protein, while Lex-Nins156, Lex-Nins268, and Lex-Nins372 have four codon insertions after codons for the indicated amino acid residues. Exact juncture and mutation sequences are available on request.

FIG. 5

Genetic analysis of N protein interactions. Protein-protein interactions between VP16-N proteins and the indicated Lex-N fusion proteins were monitored by quantitation of β-galactosidase activities. To do so, a yeast (S. cerevisiae) L40 (2, 12) strain expressing VP16-N was constructed by lithium acetate (2, 12) transformation and selection on Leu⁻ plates. Subsequent transformation of Lex constructs into L40 and L40 VP16-N parental strains used Trp⁻ or Trp⁻ Leu⁻ selections, respectively. For assays, parental and transformant yeast colonies were transferred with toothpicks onto nitrocellulose filter squares, which were placed into precooled aluminum boats floating on liquid nitrogen for 30 s and then submerged in liquid nitrogen for 5 s. Frozen cells on filters were thawed at room temperature, after which filters were transferred onto Whatman papers in petri dishes impregnated with Z buffer (100 mM sodium phosphate, pH 7.0, 10 mM KCl, 1 mM MgSO₄) containing 0.25 mg of X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside)/ml. Experimental, positive, and negative control filters were incubated at 30°C for 16 h to allow for color development, after which cells on filters were dried at room temperature. For quantitation, dried filter images were scanned at 150 dpi with a Hewlett-Packard Scan Jet IIc scanner and densitometrically quantitated using the gel plot program of NIH Image 1.62. Activity levels correspond to those from yeast strains doubly transformed with VP16-N and the indicated Lex-N construct and were normalized to those of the average of L40 VP16-N Lex-N (N) double transformants. Values are derived from the number (n) of independent β-galactosidase measurements listed below, and background activity levels (b) for each Lex-N construct were determined from a minimum of four measurements in transformed cells lacking the VP16-N construct and are also listed: Lex-N, n = 33, b = 5%; Lex-N1-357, n = 17, b = 19%; Lex-N1-171, n = 24, b = not done; Lex-N1-172, n = 6, b = 17%; Lex-N172-428, n = 16, b = 16%; Lex-N172-357, n = 22, b = 12%; Lex-NΔ157-372, n = 8, b = 12%; Lex-NΔ40-195, n = 10, b = 13%; Lex-NΔ40-215, n = 4, b = 22%; Lex-Nins156, n = 4, b = 20%; Lex-Nins268, n = 6, b = 16%; Lex-Nins372, n = 3, b = 42%; Lex-EBfill, n = 4, b = 13%.

FIG. 6

Prediction of N protein coiled-coil domains. The N protein sequences of Sin Nombre, Prospect Hill, Hantaan, Seoul, Tula, Pulmonary, and Sapporo strains of hantavirus were subjected to the MultiCoil (29) parallel coiled-coil prediction algorithm, using a 0.5 cutoff for the maximum scoring residue. The graphs show the calculated probabilities (y axis) for trimeric (solid bars) and dimeric (hatched bars) coiled coils versus N protein residues 1 to 110 (x axis). Note that MultiCoil probabilities for coiled coils in residues 111 to 428 were below threshold values. As shown, two possible coiled coils are predicted, at residues 1 to 34 and 38 to 80. For each sequence, predicted trimer and dimer probabilities for the 1 to 34 and 38 to 80 coiled coils are provided at the top of each graph. Probabilities of 0 indicate that no residue in a region scored above the 0.5 cutoff value.