JVI Figure table search 04
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bragg, J. N.
Right arrow Articles by Jackson, A. O.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bragg, J. N.
Right arrow Articles by Jackson, A. O.
Journal of Virology, July 2004, p. 7379-7391, Vol. 78, No. 14
0022-538X/04/$08.00+0     DOI: 10.1128/JVI.78.14.7379-7391.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.

The N-Terminal 85 Amino Acids of the Barley Stripe Mosaic Virus {gamma}b Pathogenesis Protein Contain Three Zinc-Binding Motifs

Jennifer N. Bragg, Diane M. Lawrence, and Andrew O. Jackson*

Department of Plant and Microbial Biology, University of California, Berkeley, California 94720

Received 19 December 2003/ Accepted 18 February 2004


   ABSTRACT
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Barley stripe mosaic virus RNA{gamma} encodes {gamma}b, a cysteine-rich protein that affects pathogenesis. Nine of the eleven cysteines are concentrated in two clusters, designated C1 (residues 1 to 23) and C2 (residues 60 to 85), that are arranged in zinc finger-like motifs. A basic motif (BM) rich in lysine and arginine (residues 19 to 47) resides between the C1 and C2 clusters. We have demonstrated that {gamma}b binds zinc and that the C1, BM, and C2 motifs have independent zinc-binding activities. To evaluate the requirements for binding, mutations were introduced into each region. Cysteine residues at positions 7, 9, 10, 19, and 23 in the C1 motif were replaced with serines. In the BM, asparagines were substituted for lysines at positions 26 and 35, glutamine for arginine at position 25, and glycines for arginines at positions 33 and 36. The C2 mutations included cysteine replacements with serines at positions 60, 64, 71, and 81, and a histidine-to-leucine change at position 85. These mutations destroyed zinc-binding activity in each of the isolated motifs. {gamma}b derivatives containing mutations in only two of the motifs retained the ability to bind zinc, whereas a {gamma}b derivative containing mutations inactivating all three motifs destroyed the ability to bind zinc. Plants inoculated with transcripts containing combinations of the C1, BM, and C2 mutations elicited a "null" phenotype in barley characteristic of {gamma}b deletion mutants and also delayed the appearance and reduced the size of local lesions in Chenopodium amaranticolor. These results show that zinc binding of each of the motifs is critical for the biological activity of {gamma}b.


   INTRODUCTION
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Barley stripe mosaic virus (BSMV) encodes a 17-kDa cysteine-rich protein (CRP) protein, designated {gamma}b, that is expressed from the 3'-proximal cistron of RNA{gamma} (Fig. 1A). The {gamma}b protein is expressed early and remains at elevated levels throughout the course of BSMV infection (6). Although {gamma}b is dispensable for both viral replication and movement in the ND18 strain of the virus, in the {gamma}{Delta}2.1 deletion mutant, which removes the bulk of the {gamma}b gene, symptom onset is delayed in barley, and symptoms are attenuated, resulting in a null phenotype characterized by an erratic mosaic pattern (21). Further analysis of this mutant suggests that {gamma}b has differential effects on RNA accumulation and protein expression and that it may regulate the synthesis of proteins encoded on RNAß. Interestingly, in the type strain of BSMV, the {gamma}b protein is essential for systemic symptom development (21), and the requirement for {gamma}b is linked to a 372-nucleotide (nt) direct repeat in the 5' region of RNA{gamma} that overlaps the start codon and increases the size of the {gamma}a gene. This gene encodes the polymerase component of the viral replicase and therefore, this strain-specific difference suggests that subtle effects on replication contribute to the phenotypic effects of mutations in the {gamma}b protein.



View larger version (22K):
[in this window]
[in a new window]
  		FIG. 1. The tripartite genome of BSMV and subgenomic RNAs (sgRNAs) used for expression of the seven proteins encoded by the RNAs and motifs located within {gamma}b. (A) The filled circles represent the 5' cap structure present on all of the BSMV genomic and subgenomic RNAs. Each 3'-proximal ORF terminates with a UAA codon, followed by an internal 4- to 40-nt polyadenylate sequence (An) that precedes the 238-nt tRNA-like 3' terminus (solid rectangles). The three gRNAs, whose ORFs are illustrated by rectangular blocks, are designated {alpha}, ß, and {gamma}. RNA{alpha} encodes the {alpha}a protein, which is required for replication. The {alpha}a protein contains helicase (Hel) and methyltransferase (Mt) domains and forms the "helicase subunit" of the viral replicase complex. RNAß encodes five proteins. The coat protein, ßa, is translated directly from the gRNA. The overlapping triple gene block (TGB) proteins TGB1, TGB2, and TGB3 are each required for virus movement and are expressed from two subgenomic RNAs: sgRNAß1 and sgRNAß2. TGB1 contains a helicase (Hel) domain, whereas TGB2 and TGB3 are small hydrophobic proteins. TGB2' is a minor translational readthrough protein that is dispensable for infection. RNA{gamma} is bicistronic. The {gamma}a protein, which is translated from the gRNA{gamma}, contains the GDD domain and is the polymerase subunit of the replicase. The cysteine-rich {gamma}b protein, which is expressed from the capped sgRNA{gamma}, is involved in pathogenesis. (B) The {gamma}b protein is 152 amino acids in length and contains two cysteine-rich regions, C1 and C2, between amino acids 7 to 23 and 60 to 85, respectively. A BM, rich in lysine and arginine residues, lies between amino acids 19 to 47. Six heptad repeats that form a coiled-coil structure are predicted between amino acids 95 and 140.

 
Nine of the eleven cysteine residues in the {gamma}b protein are concentrated in two clusters toward the N terminus of the protein. The C1 (amino acids 1 to 23) and C2 (amino acids 60 to 85) clusters are arranged in zinc finger-like motifs (with one histidine residue included in C2 at position 85). The {gamma}b protein also contains a basic motif (BM) between amino acids 19 to 47 and six heptad repeats of a coiled-coil motif between residues 95 to 140 (Fig. 1B). Extensive analyses of mutations introduced into the C1, BM, and C2 regions of {gamma}b have demonstrated that these motifs each have important virulence functions (4). Subcellular fractionation of infected barley extracts has revealed that the site-specific mutations in these regions also have striking effects on the solubility of the protein and that mutant derivatives are often shifted from the soluble fraction to the cell wall and membrane-containing fractions (4). Furthermore, single-stranded RNA-binding activity has been demonstrated for a recombinant GST-{gamma}b fusion protein in in vitro gel shift studies. Lysine and arginine residues within the BM were demonstrated to mediate this binding, whereas cysteine and histidine substitutions in the C1 and C2 regions had little effect on in vitro RNA-binding activity (5).

Small cysteine-rich proteins (CRPs) are also encoded by the 3'-terminal open reading frame (ORFs) in a number of other plant viruses including members of the Tobravirus, Carlavirus, Furovirus, and Pecluvirus genera (17). Although these CRPs do not share a substantial amount of sequence identity with {gamma}b, they do share limited regions of sequence and predicted structural similarity. For example, amino acids 26 to 134 of the BSMV {gamma}b protein, which encompass the majority of the BM, C2, and putative coiled-coil regions of the protein, can be aligned with the corresponding amino acids of the CRPs of the other hordeiviruses, Poa semilatent virus (PSLV) and Lychnis ringspot virus, and also with the CRPs of Soilborne cereal mosaic virus (SBCMV), Soilborne wheat mosaic virus (SBWMV), Peanut clump virus (PCV), Indian peanut clump virus, Chinese wheat mosaic virus (CWMV), Sorghum chlorotic spot virus (SCSV), and Oat golden stripe virus (OGSV) (10). One predicted zinc binding motif (CCCH) is conserved among all of the proteins of this family, as is the arrangement of the basic, zinc-binding, and coiled-coil motifs. Three additional CRPs that affect replication (Beet necrotic yellow vein virus (BNYVV) p14, Tobacco rattle virus (TRV) p16, and Pea early browning virus (PEBV) p12) each contain cysteine and histidine residues arranged in putative zinc-binding motifs adjacent to a region of clustered basic arginine and lysine residues (15, 18) but lack a predicted coiled-coil region.

Understanding the functions of these proteins during virus infections is of particular interest because, like {gamma}b, a number of them have been demonstrated to be viral pathogenicity determinants and are suggested to play regulatory roles during infection (4, 7, 12, 13, 18). Deletion of BSMV {gamma}b (5, 21) and TRV p16 (18), both of which have RNA-binding activity, or PCV p15 (7) causes a dramatic decreases in viral RNA levels that can be rescued by supplying the appropriate CRP in trans. Furthermore, the TRV p16 protein can be functionally replaced by the BSMV {gamma}b, SBWMV p19, or Cucumber mosaic virus (CMV) 2b proteins, suggesting that these proteins may share some common biological functions (18). The TRV p16, BSMV {gamma}b, and SBWMV p19 proteins are all CRPs but, interestingly, the CMV 2b protein does not encode a cysteine-rich region. The 2b protein is most commonly recognized as a suppressor of RNA silencing (20). The PCV CRP, p15, also acts as a suppressor of RNA silencing (8) and more recently, a similar role has been suggested for the BSMV and PSLV {gamma}b proteins (30). In these experiments, Nicotiana benthamiana plants were inoculated with a TMV derivative (TMV-GFP) that expresses green fluorescent protein (GFP). RNA silencing, preventing systemic movement and fluorescence of TMV-GFP, was induced by coinoculation with a Potato virus X (PVX) vector encoding a truncated GFP sequence that does not fluoresce (PVX-GF). However, when TMV-GFP was coinoculated with PVX-GF that also expresses either the BSMV or PSLV {gamma}b proteins, systemic movement of TMV-GFP was rescued, and GFP fluorescence was detected in the systemically infected leaves (30).

The C1 and C2 motifs of {gamma}b have long been predicted to be zinc-binding regions and are comparable in size (20 to 40 amino acids) to the majority of identified zinc-binding motifs (1). The {gamma}b C1 and C2 motifs have the patterns CX2/3CX8CX3C and CX3CX16CX3H (or CX3CX6CX9CX3H), respectively, and whereas the spacings are unique, the identity and arrangement of the ligands are characteristic of structural sites in which the metal stabilizes the protein and is required for its proper folding (1, 2).

To provide more insight into the role the {gamma}b protein plays in BSMV infections, we analyzed the zinc-binding activity of wild-type (wt) {gamma}b obtained from infected plants, as well as that of recombinant {gamma}b fusion protein derivatives containing either large deletions or site-specific mutations within the C1, BM, and C2 regions. Mutations that affected {gamma}b zinc-binding activity were incorporated into biologically active cDNA clones and were used to evaluate the phenotypic effects of the substitutions, as well as their effects on RNA and protein accumulation during BSMV infections.


   MATERIALS AND METHODS
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Oligonucleotides and recombinant plasmids. Standard molecular biology techniques were used for all recombinant plasmid manipulations (25). A summary of constructs generated for zinc binding experiments is shown in Table 1. The sequences of the oligonucleotides used for introduction of mutations can be found in Table 2. Appropriate regions of each construct were sequenced to ensure that the desired mutations had been introduced.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Constructs used in zinc-binding experimentsa

 

View this table:
[in this window]
[in a new window]
  		TABLE 2. Oligonucleotides used to construct clones for zinc-binding studies

 
{gamma}b derivatives used to assess zinc binding. The constructs used in 65ZnCl2-binding experiments were generated by using the pMAL protein fusion and purification system from New England Biolabs (NEB), Beverly, Mass. (catalog no. E8000S). This system was used to generate fusions of the maltose-binding protein (MBP) to the N terminus of either wt {gamma}b or mutant {gamma}b derivatives (Table 1).

MBP fusions suitable for affinity purification were constructed with the wt {gamma}b (pMalC2X{gamma}b) and a {gamma}b mutant containing cysteine (C)-to-serine (S) mutations at amino acid positions 7, 9, 10, 19, 23, 60, 64, 71, and 81 and a histidine (H)-to-leucine (L) mutation at amino acid 85 [pMalC2X{gamma}b(–)C1C2]. For this purpose, the {gamma}b gene was amplified from the {gamma}42 cDNA clone of BSMV RNA{gamma} (23) or RNA{gamma}(–)C1C2 by using the primers BamMal2 and gb3'Pst1. The amplification introduced a BamHI site directly before the start codon of the {gamma}b gene and a PstI site at the 3' end of the amplified region. The amplified fragments were then introduced into the BamHI and PstI sites of pMalC2X, a vector designed to create MBP fusions to the N terminus of proteins of interest.

A mutant designed to express a MBP fusion to the N-terminal 85 amino acids of {gamma}b (pMalC2X{gamma}b1-85) was constructed by digesting pMalC2X{gamma}b with SphI and PstI, blunting with Klenow, and religating the vector. To generate a clone (pMalC2X{gamma}b86-152) to express MBP fused to the C-terminal 67 amino acids of {gamma}b, the primers C2X{gamma}b3' and {gamma}bXba3' were used to amplify amino acids 86 to 152 of {gamma}b from the {gamma}42 clone. These primers introduced a BamHI site directly before codon 86 of the {gamma}b gene and an XbaI site at the 3' end of the amplified region to permit cloning of the fragment into the BamHI and XbaI sites of pMalC2X.

For investigations of the contribution of the C1 region to zinc-binding activity, C-to-S mutations were introduced in the C1 motif of {gamma}b at positions 7, 9, 10, 19, and 23 to generate the pMalC2X{gamma}b(–)C1 clone. C-to-S mutations were introduced in the C2 motif of {gamma}b at positions 60, 64, 71, and 81, and an H-to-L mutation was introduced at position 85 to construct the pMalC2X{gamma}b(–)C2 plasmid. These plasmids were constructed by overlap PCR (28), whereby the pMalC2X{gamma}b and pMalC2X{gamma}b(–)C1C2 constructs were each used as templates for two PCRs. One reaction used the primers NcoMalE5' and g2234R to amplify the C1 region of {gamma}b, and the second reaction used the primers g2216F and MalERev to amplify the C2 region of {gamma}b. To generate the pMalC2X{gamma}b(–)C1 clone, the C1 region from pMalC2X{gamma}b(–)C1C2 and the C2 region from pMalC2X{gamma}b were mixed, annealed, extended, and then amplified by using the NcoMalE5' and MalERev primers. The pMalC2X{gamma}b(–)C2 plasmid was constructed from the C1 region of pMalC2X{gamma}b and the C2 region of pMalC2X{gamma}b(–)C1C2 by the same method.

The plasmids pMalC2XC1 and pMalC2XC2 were designed for expression of MBP fusions to {gamma}b amino acid residues 1 to 23 (C1) and 60 to 85 (C2), respectively. To generate these plasmids, regions of the {gamma}42 clone were amplified by using the primers BamMal2 and C1R (C1 region) or C2F and C2R (C2 region), and the products were introduced into the BamHI and PstI sites of pMALC2X. Similarly, mutant plasmids pMalC2X(–)C1 (containing C-to-S substitutions at positions 7, 9, 10, 19, and 23) and pMalC2X(–)C2 (containing C-to-S mutations at positions 60, 64, 71, and 81 and an H-to-L mutation at position 85) were generated by PCR amplification of the cDNA of the RNA{gamma}(–)C1C2 mutant by priming with (–)C1F and (–)C1R or (–)C2F and (–)C2R, respectively. These PCR products were cloned into the BamHI and PstI sites of the pMalC2X vector.

Four plasmids were generated for expression of MBP fusions to wt and mutant sequences spanning the BM region of {gamma}b (amino acids 21 to 49). Fragments for construction of the plasmids pMalC2XBM (containing wt sequence) and pMalC2XBM29 (containing arginine [R]-to-glycine [G] mutations at amino acids 33 and 36 and a lysine [K]-to-asparagine [N] substitution at amino acid 35) were amplified from the {gamma}42 clone or the BM29 mutant (4), respectively, by using the primers BMF and BMR. The fragments used to produce pMalC2XBM26 (with an R-to-glutamine [Q] change at position 25 and a K-to-N substitution at amino acid 26) and pMalC2X(–)BM (which includes the five mutations from the BM26 and BM29 constructs) were derived from the mutants BM26 or BM29 (4), respectively, by amplifying with the oligonucleotides BM26F and BMR. The four fragments were then introduced into the BamHI and PstI sites of pMalC2X.

The three sets of mutations described for the C1, BM, and C2 motifs were introduced into the full-length {gamma}b protein in four combinations. The pMalC2X(–)C1C2 mutant containing the 10 mutations within the (–)C1 and (–)C2 derivatives was described above. The plasmid pMalC2X{gamma}b(–)C1BM, incorporating the 10 mutations of the (–)C1 and (–)BM derivatives, was generated by using the primers 23F2 and gbXba3' to amplify a portion of the {gamma}b BM29 mutant (4). This fragment was cloned into the pCR-BluntII TOPO vector and then digested with BglII to create an intermediate designated (–)C1BM/BglII. A BglII fragment excised from pMalC2X{gamma}b(–)C1 was next cloned into the (–)C1BM/BglII intermediate, and the resulting clone was digested with BamHI and XbaI. The BamHI/XbaI fragment was ligated into the corresponding sites of pMalC2X. The plasmid pMalC2X{gamma}b(–)BMC2, which contains the 10 mutations present of the (–)BM and (–)C2 mutants, was constructed with the primers BamMal2 and BsmIR to amplify a region of pMalC2X{gamma}b(–)BM. The PCR product was cloned into the pCR-BluntII TOPO vector and then a three-part ligation was performed by using the BamHI/BsmI fragment from this product, the BsmI/PstI fragment of pMalC2X(–)C2, and the pMalC2X vector that was digested with BamHI/PstI.

The plasmid pMalC2X{gamma}b(–)C1BMC2, which contains all 15 mutations introduced into the (–)C1, (–)BM, and (–)C2 derivatives, was constructed in multiple steps. First, the BM29 mutant was amplified with the primers 232526F and BsmIR to create a fragment spanning amino acids 23 to 62 of {gamma}b. The BglII/BsmI fragment from this product, containing the five desired BM mutations, was introduced into a pCR-BluntII TOPO plasmid containing a {gamma}b mutant with C-to-S mutations at amino acids 7, 9, 10, 19, 23, 60, 64, 71, and 81 to create intermediate A. This intermediate spanned amino acids 23 to 152 of {gamma}b and was used as a template for two PCRs whose products were each cloned into the pCR-BluntII TOPO vector. One reaction amplified amino acids 23 to 152 of intermediate A by using the primers 23F2 and BSMV3' to produce intermediate B. The second reaction amplified amino acids 81 to 152 of intermediate A with the primers 81/85F to introduce the H85L mutation and BSMV3' to generate intermediate C. The MscI fragment from intermediate B (containing sequence encoding amino acids 82 to 152 of {gamma}b and 3' sequence derived from vector) was cloned into the MscI site of intermediate C to produce intermediate D, which contained amino acids 23 to 152 of {gamma}b with all 10 of the BM and C2 mutations. The BglII/PstI fragment from intermediate D was then introduced into the corresponding sites of the TOPO {gamma}b(–)C1C2 construct to produce intermediate E, which consisted of the full-length {gamma}b sequence containing all 15 C1, BM, and C2 mutations. The BamHI/PstI fragment from intermediate E was next introduced into the corresponding sites of pMalC2X to produce the final product, pMalC2X{gamma}b(–)C1BMC2.

Incorporating {gamma}b zinc-binding mutations into the RNA{gamma} cDNA clone. To evaluate the disease phenotypes elicited by the amino acid substitutions, four BSMV mutants were generated to incorporate the mutations described for the C1, BM, and C2 motifs of {gamma}b into the RNA{gamma} plasmid to produce RNA{gamma}(–)C1C2, RNA{gamma}(–)C1BM, RNA{gamma}(–)BMC2, and RNA{gamma}(–)C1BMC2 (Table 1). The RNA{gamma}(–)C1C2 construct contains C-to-S mutations at amino acids 7, 9, 10, 19, 23, 60, 64, 71, and 81 and a H-to-L mutation at amino acid 85. To engineer {gamma}b(–)C1C2, two blunt-ended fragments were amplified by PCR. The first reaction used the primers –1875 and C19/23R with the C1(9,10,19) mutant template (4), and the second reaction used the primers g2150F and BSMV3' with the C2(60,64) mutant template (4). These fragments were ligated, and the resulting product was amplified with the primers g1629F and C71/81R to produce intermediate F. The KpnI/MscI fragment of intermediate F was introduced into the corresponding sites of the C1(7,9,10) mutant (4). Next, the histidine to leucine mutation at position 85 was introduced by site-directed mutagenesis with the primers 85F and 85R with a QuikChange site-directed mutagenesis kit (catalog no. 200518; Stratagene, La Jolla, Calif.). The pCR-BluntII TOPO intermediates used to construct MBP fusion clones containing the mutants {gamma}b(–)C1BMC2, {gamma}b(–)C1BM, and {gamma}b(–)BMC2 were digested with KpnI and HpaI. These fragments were introduced into the corresponding sites of the {gamma}b(–)C1C2 mutant for {gamma}b(–)C1BMC2 and {gamma}b(–)C1BM or into the wt RNA{gamma} background for {gamma}b(–)BMC2 (Table 1). The resulting mutants were analyzed in two hosts, barley and Chenopodium amaranticolor, to evaluate the phenotypic effects resulting from the abrogation of zinc-binding activity in each of the mutants.

Zinc affinity chromatography. Infectious RNA transcripts were generated from linearized cDNA templates by using T7 RNA polymerase as described previously (22). RNA{alpha} and RNA{gamma} were linearized with MluI, and an RNAß derivative (B7), in which the coat protein is not expressed, was linearized with SpeI. The B7 mutant was used because larger amounts of {gamma}b protein were recovered from infected tissue in the absence of the coat protein. Seven- to ten-day-old seedlings of the barley cultivar "black hulless" were inoculated with RNA transcripts in GKP buffer (50 mM glycine, 30 mM KHPO4 [pH 9.2], 1% bentonite, 1% celite) (22). Inoculated plants were grown in a growth chamber with a 14-h light regimen and a diurnal temperature fluctuating between 18 and 26°C and were then harvested 7 days postinoculation (dpi). After the infected tissue was weighed, all steps were carried out at 4°C. Leaves (1 to 2 g) were ground with a mortar and pestle in 5 ml of cold zinc column buffer (100 mM NaPO4 [pH 8.0], 150 mM NaCl) to which 0.1% 2-mercaptoethanol (ß-ME) and a cocktail of protease inhibitors (1 µg/ml each of leupeptin, pepstatin, and aprotinin and 50 µM phenylmethylsulfonyl fluoride [PMSF]) were added just before grinding. The extract was filtered through cheesecloth into a 15-ml conical tube and centrifuged at 2,000 rpm (500 x g) for 10 min at 4°C in a SS-34 rotor to eliminate most of the cellular debris. The supernatant from this step was transferred to ultracentrifuge tubes and recentrifuged at 17,000 rpm (30,000 x g) for 35 min at 4°C in a Beckman 70.1 Ti rotor to pellet nuclei, chloroplasts, membrane fragments, and higher-molecular-weight polymers. The supernatant (S30) from this step (5 ml) was used as the protein extract for chromatography.

A zinc affinity column was constructed for chromatography by using 3 ml of chelating Sepharose fast-flow resin (catalog no. 17-0575-01; Amersham, Piscataway, N.J.) that had been charged with zinc for immobilized metal affinity chromatography as described by the manufacturer. First, the resin was equilibrated in deionized water and then mixed with 150 mM ZnSO4 at room temperature for 30 min. The resin was washed twice with deionized water, exchanged into zinc column buffer, and pipetted into a 5-ml syringe plugged with glass wool. After the resin settled, the column was washed at 4°C with 5 volumes of cold zinc column buffer, and then the protein extract (5 ml) was applied to the column at the rate of 4 drops/min (~200 µl/min). The column was washed first with 10 ml of the zinc column binding buffer and second with 10 ml of the high-salt zinc wash buffer (50 mM Tris-HCl [pH 8.0], 700 mM NaCl). Protein was eluted by using a pH step gradient in 0.5-U intervals from pH 7.5 to 3.5 (1 ml per elution buffer step, 100 mM NaPO4, variable pH, 700 mM NaCl). All washing and elution steps were performed at a flow rate of 6 to 10 drops/min, and 500-µl fractions were collected. For the control experiments, the barley extracts were applied to a column that contained chelating Sepharose fast-flow resin that had not been charged with zinc.

Subcellular fractionation of plant extracts. Barley tissue (1 to 2 g) was ground in a mortar and pestle in 5 ml of cold nickel column buffer (100 mM Tris [pH 8.0], 200 mM NaCl, 30 mM MgCl2, 0.1% Triton X-100) containing a cocktail of protease inhibitors (1 µg/ml each of leupeptin, pepstatin, and aprotinin and 50 µM PMSF) and 0.1% ß-ME. The extract was passed through cheesecloth, and the liquid was squeezed into a 15-ml conical tube. The residue in the cheesecloth was rinsed and mixed with Laemmli loading buffer (100 mM Tris-Cl [pH 6.8], 4% sodium dodecyl sulfate [SDS], 0.2% bromophenol blue, 20% glycerol, 200 mM ß-ME) to obtain a crude cell debris fraction of unbroken cells and fibrous materials. The liquid recovered after filtration through cheesecloth was centrifuged at 4°C for 10 min at 3,000 rpm (1,000 x g) in a SS-34 rotor to pellet nuclei and chloroplasts (fraction P1). The resulting supernatant (S1) was centrifuged at 4°C for 35 min at 17,000 rpm (30,000 x g) to separate membranes (pellet, fraction P30) from the soluble proteins (supernatant, fraction S30). Samples of these fractions were separated on 10% SDS-polyacrylamide gels, blotted onto nitrocellulose, and detected with {gamma}b and goat anti-mouse horseradish peroxidase (GAM-HRP; catalog no. 170-6516; Bio-Rad, Hercules, Calif.) antibodies.

Escherichia coli protein expression and amylose affinity chromatography. Zinc-binding constructs were transformed into E. coli strain TB1 (catalog no. E4122S; NEB) for the expression of fusion proteins. Cultures were grown to an optical density at 600 nm of 0.5 at 37°C in 80 ml of rich medium (10 g of tryptone, 5 g of yeast extract, 5 g of NaCl, and 2 g of glucose/liter) containing 100 µg of ampicillin/ml, and protein expression was induced for 2 h by the addition of 300 µM IPTG (isopropyl-ß-D-thiogalactopyranoside). Cells were pelleted at 5,000 rpm (4,000 x g) for 10 min at room temperature in a GSA rotor and resuspended in 10 ml of amylose column buffer (20 mM Tris-HCl [pH 7.5], 200 mM NaCl, 1 mM EDTA). Resuspended cells were either frozen overnight at –20°C or subjected to two rounds of freezing in a dry ice-ethanol bath and thawing in cool water. After thawing, 0.1% ß-ME and a cocktail of protease inhibitors were added (1 µg/ml each of leupeptin, pepstatin, and aprotinin and 50 µM PMSF), and the cells were sonicated on ice for 2 min (four 30-s bursts). The extract was centrifuged at 8,700 rpm (9,000 x g) for 20 min at 4°C in an SS-34 rotor. The supernatant containing the crude protein extract was divided into aliquots and frozen at –80°C in 10% glycerol for future use. To prepare amylose resin (NEB catalog no. E8021S) for affinity purification, a 50/50 slurry of resin and amylose column buffer were mixed. Affinity purification was conducted by a batch method in a 1.5-ml microfuge tube in which 100 µl of resin was added to 1 ml of crude protein extract for 30 min on ice, followed by occasional mixing. The resin was washed twice with 1 ml of amylose column buffer, and protein was either eluted at 4°C by using 200 µl to 1.0 ml of amylose column buffer containing 10 mM maltose or directly solubilized into the same volume of Laemmli loading buffer.

Zinc blotting. Proteins were separated on SDS-8% polyacrylamide gels and electroblotted onto nitrocellulose membranes. The membranes were washed for 1 h at room temperature in 20 ml of renaturing buffer (100 mM Tris-HCl [pH 6.8], 50 mM NaCl, 10 mM dithiothreitol), rinsed twice with zinc-binding buffer (100 mM Tris-HCl [pH 6.8], 50 mM NaCl), and incubated for 30 min in 20 ml of zinc-binding buffer containing 80 µCi of 65ZnCl2 (~10 µM) (Perkin-Elmer [Boston, Mass.] catalog no. NEZ111). The membrane was then rinsed twice with zinc binding buffer and exposed to X-ray film. All buffers were degassed to prevent oxidation of the proteins. Alcohol dehydrogenase (ADH; 10 µg) was used as a positive control in some experiments (Sigma [St. Louis, Mo.] catalog no. A-7011). In competition experiments, divalent metal ions were included in the zinc-binding buffer at a concentration of 10 mM (~1,000 times the concentration of 65ZnCl2) during incubation with the radioisotope.


   RESULTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
The {gamma}b protein from infected barley binds zinc in affinity chromatography experiments. In order to determine whether the {gamma}b protein has the ability to bind zinc, native protein was extracted from BSMV-infected tissue at 7 dpi, and the extracts were subjected to zinc affinity chromatography (24). The {gamma}b protein was abundant in the initial extract (Fig. 2A, lane 1, and B, lane 1), and when the extract was applied to the column, nearly all of the {gamma}b protein bound to and was retained on the zinc matrix (M), even after low- and high-salt washes (Fig. 2A, lane 2, and 2B, lanes 5 and 8). A control column was constructed with resin that had not been charged with zinc. When the infected barley extract was passed over this matrix, {gamma}b was observed only in the initial extract and in the effluent and wash fractions (Fig. 2A, lanes 1, 6, and 7), indicating that {gamma}b binds specifically to the zinc ligand during affinity chromatography. As the pH decreases, the interactions of cysteine and histidine residues with zinc become less stable and therefore a pH step gradient from 7.5 to 3.5 was used to elute proteins from the zinc column. The {gamma}b protein was detected in fractions eluting between pH 6.5 and 4.5 with the peak release observed at pH 5.5 (Fig. 2C). Interestingly, a substantial amount of {gamma}b (estimated to be ca. 50% of the bound protein) remained associated with the zinc resin after the pH gradient elution steps, and the matrix required boiling in Laemmli loading buffer for release of the residual protein (Fig. 2B, lanes 5 and 8). The {gamma}b protein often became insoluble during purification manipulations, and therefore the tenacious binding to the zinc matrix may have resulted from insoluble aggregates of {gamma}b that formed during chromatography.



View larger version (51K):
[in this window]
[in a new window]
  		FIG. 2. Purification of {gamma}b by zinc affinity chromatography. Barley was inoculated with a derivative of BSMV in which the coat protein was not expressed, and infected tissue was harvested and extracted at 7 dpi. Protein samples were separated on 10% polyacrylamide gels, blotted onto nitrocellulose, and detected with {gamma}b and GAM-HRP antibodies. (A) The initial extract (I) that contains {gamma}b (lane 1) was subjected to zinc affinity chromatography. Lane 2 shows the {gamma}b protein that bound to the zinc affinity matrix (M), and lane 3 shows the effluent (E) fraction containing {gamma}b protein that failed to bind to the matrix. After application of the barley protein extract, the column was washed (W) with a low-salt buffer (lane 4). Lanes 5 to 7 show the {gamma}b protein present in the matrix, effluent, and wash fractions after the initial extract was applied to Sepharose matrix that was not charged with zinc. (B) Lane 1 shows the {gamma}b protein in the initial extract (I) from infected barley. Lanes 2 to 5 show the {gamma}b protein present in the effluent (E1), low-salt wash (W1), high-salt wash (H1), and matrix bound (M1) fractions after the first application to the zinc affinity matrix. Lanes 6 to 8 show the effluent (E2), low-salt wash (W2), and matrix-bound (M2) fractions of {gamma}b protein that were recovered from the E1 fraction after application to a fresh batch of zinc affinity matrix. (C) The {gamma}b protein bound to the zinc affinity matrix was eluted with a pH step gradient from pH 7.5 to 3.0.

 
To determine whether the cysteine and histidine residues in the C1 and C2 regions of {gamma}b are involved in interactions between {gamma}b and the zinc affinity matrix, the {gamma}b(–)C1C2 mutant was constructed. In this mutant, cysteine residues at positions 7, 9, 11, 19, 23, 60, 64, 71, and 81 were changed to serines, and the histidine at position 85 was changed to a leucine. The {gamma}b(–)C1C2 protein was detected in infected barley plants by Western blotting; however, after extraction for affinity chromatography, the mutant protein (Fig. 3B) behaved quite differently from the wt protein (Fig. 3A). When infected tissue was extracted by grinding leaves directly in Laemmli loading buffer, the wt protein resolved almost entirely as a monomer after SDS-polyacrylamide gel electrophoresis (PAGE) and Western blot analysis (Fig. 3B, lane 1). In contrast, the majority of the {gamma}b(–)C1C2 protein was present as multimers or as high-molecular-weight aggregates with essentially no protein migrating at the size expected for the {gamma}b monomer (Fig. 3B, lanes 2 to 6). Interestingly, the aggregates were not observed when the {gamma}b(–)C1C2 protein was produced by in vitro translation (Fig. 3C, lane 2), so it is likely that insoluble aggregates were produced during infection or during extraction from plants.



View larger version (63K):
[in this window]
[in a new window]
  		FIG. 3. Expression and subcellular distribution of wt {gamma}b and the {gamma}b(–)C1C2 mutant in barley. (A and B) Barley was inoculated with RNA{alpha}, the RNAß B7 derivative in which the coat protein was not expressed, and an RNA{gamma} encoding either the wt {gamma}b protein or the {gamma}b(–)C1C2 mutant and infected tissue was harvested at 7 dpi. (A) wt {gamma}b was extracted from infected tissue and fractionated. Lane 1, debris containing cellular and other fibrous materials that was retained after squeezing through cheese cloth; lane 2, the 1,000 x g P1 pellet fraction from expressed sap; lane 3, the 30,000 x g S30 soluble fraction; lane 4, the 30,000 x g P30 pellet fraction. (B) Lane 1, total wt {gamma}b protein recovered after grinding infected tissue directly in Laemmli buffer; lane 2, total {gamma}b(–)C1C2 protein extracted after grinding in Laemmli buffer; lane 3, {gamma}b(–)C1C2 cell debris retained by cheese cloth; lane 4, the {gamma}b(–)C1C2 1000 x g P1 pellet fraction from the cheesecloth filtrate; lane 5, the {gamma}b(–)C1C2 30,000 x g, S30 soluble fraction; lane 6, the {gamma}b(–)C1C2 30,000 x g, P30 pellet fraction. All fractions were separated on 10% SDS-polyacrylamide gels and blotted onto nitrocellulose. Proteins were detected with {gamma}b and GAM-HRP antibodies. Species of {gamma}b corresponding to sizes expected for monomers (M), dimers (D), tetramers (T), and aggregates (A) are designated by arrows. (C) Lane 1 (wt {gamma}b) and lane 2 [{gamma}b(–)C1C2] show [35S]methionine-labeled proteins recovered after translation in vitro in a rabbit reticulocyte system. Proteins were separated on a SDS-10% polyacrylamide gel, and the gel was dried and exposed to X-ray film.

 
To compare subcellular localization of wt {gamma}b and the {gamma}b(–)C1C2 derivative, infected tissue was ground in extraction buffer and fractionated. Approximately half of the wt protein appeared in the cell debris fraction (Fig. 3A, lane 1), and the remainder was present in the P1 pellet and the soluble (S30) fractions (Fig. 3A, lanes 2 and 3). In each fraction, the wt {gamma}b protein resolved as a monomer after electrophoresis. In contrast, almost all of the {gamma}b(–)C1C2 protein shifted from the soluble S30 fraction into the insoluble cell debris and P30 fractions (Fig. 3B, lane 3 and 6), as was previously reported for the {gamma}b mutants characterized by Donald and Jackson (4). Moreover, none of the {gamma}b(–)C1C2 protein aggregate from the cell debris fraction was able to enter the gel, whereas the {gamma}b(–)C1C2 protein in the P30 fraction had the size expected for a {gamma}b tetramer. An attempt was made to extract the {gamma}b(–)C1C2 protein from the P30 fraction by using buffers containing from 0.1% to 1% Triton X-100, but the amounts of protein recovered were not sufficient for use in zinc chromatography experiments (data not shown).

An MBP-{gamma}b fusion protein expressed from E. coli binds a 65Zn(II) radioisotope. Putative zinc-binding regions of {gamma}b were assayed for zinc-binding activity by blotting protein derivatives onto nitrocellulose and probing them with radiolabeled 65ZnCl2 (26). This assay has been used previously to demonstrate and map the zinc-binding ability of a number of different viral proteins (3, 9, 19, 27). Since the {gamma}b(–)C1C2 mutant could not be purified from infected barley in amounts sufficient for use in zinc affinity chromatography, we assessed the zinc binding of wt {gamma}b and {gamma}b mutants expressed as MBP fusions in E. coli. In all cases, the induced MBP-{gamma}b fusion derivatives and a MBP-lacZ control were expressed to high levels (Fig. 4, lanes 1 to 3). After amylose affinity separation, the MBP-lacZ and {gamma}b fusions were the major proteins present in the sample (data not shown). Although a Factor Xa protease cleavage recognition site had been inserted between the MBP sequence and the {gamma}b sequence for each construct, the protease failed to release {gamma}b from the fusion proteins even after extensive digestion under a variety of conditions (data not shown). A similar problem was encountered with attempted thrombin digestion of the GST-{gamma}b fusion protein used in the in vitro RNA-binding studies performed by Donald and Jackson (5). Due to the difficulties in releasing {gamma}b from MBP, the uncleaved fusion protein was used in the zinc blotting experiments, and an MBP-lacZ fusion protein was used as a negative control to ensure that positive results were not attributable to the MBP. In addition, a known zinc-binding protein, ADH, was used as a positive control in early experiments. The purified fusion proteins were separated by SDS-PAGE, blotted onto nitrocellulose, renatured on the membrane, and incubated with radiolabeled 65ZnCl2 to evaluate metal-binding activities.



View larger version (29K):
[in this window]
[in a new window]
  		FIG. 4. 65Zn(II)-binding activity of ADH and MBP-{gamma}b fusion proteins. Purified ADH and MBP fusion proteins separated on SDS-8% polyacrylamide gels were either stained with Coomassie brilliant blue and dried or blotted onto nitrocellulose, incubated with radiolabeled 65Zn(II), and exposed to X-ray film to assess zinc binding. (A) The ADH positive control, MBP-lacZ negative control, and MBP-{gamma}b proteins were tested for 65Zn(II) binding. The numbers at the bottom of the gel are for reference in the text. (B) Replicate samples of ADH and MBP-{gamma}b proteins were incubated either with 65ZnCl2 alone or with 65ZnCl2 and a divalent metal ion competitor present in a 1,000x molar concentration over the radioisotope.

 
In the initial experiments, the zinc-binding activities of ADH, MBP-lacZ, and MBP-{gamma}b were analyzed. As expected, the positive control, ADH, exhibited strong zinc binding (Fig. 4A, lane 4). The negative control, MPB-lacZ, failed to bind zinc (Fig. 4A, lane 5), and 65ZnCl2 binding was not detected even after protracted exposure of the blot (data not shown). MBP-{gamma}b exhibited a strong signal, indicating high zinc-binding capacity (Fig. 4A, lane 6). In addition, zinc-binding activity was investigated in the presence of divalent metal cation competitors (Fig. 4B). Replicate samples of ADH and MBP-{gamma}b were separated by SDS-PAGE, blotted onto nitrocellulose, and used in zinc-binding experiments. During the 65ZnCl2 binding step, Mg(II), Mn(II), Co(II), Cu(II), and Zn(II) chloride salts were included at 10 mM, ~1,000x greater concentration than the radioisotope. Unlabeled Zn(II) and Cu(II) were highly efficient competitors of radiolabeled 65Zn(II) and reduced the signal associated with the ADH and MBP-{gamma}b proteins to levels that could be detected only after lengthy exposures (i.e., 5- to 10-fold) of the blots (Fig. 4B and data not shown). Co(II) was also an effective competitor but had substantially lower affinity for the proteins than Zn(II) and Cu(II) (Fig. 4B). The signal observed after Mn(II) competition was only slightly lower than that observed for the negative control that lacked a competitor ion, indicating that Mn(II) competes very inefficiently with 65Zn(II) (Fig. 4B). Finally, Mg(II) was unable to compete with 65Zn(II) for metal-binding sites in the ADH and {gamma}b proteins (Fig. 4B). In summary, unlabeled Zn(II) and Cu(II) appear to compete most efficiently with 65Zn(II) for binding to the ADH and {gamma}b proteins. Co(II) has a lower binding affinity than Zn(II) and Cu(II), and Mn(II) and Mg(II) appear to compete poorly for the metal-binding sites in these two proteins. These observations are consistent with the results of other studies demonstrating that only metallic cations, such as Cu(II) or Co(II), that substitute structurally or functionally for Zn(II) can effectively compete with 65Zn(II) for protein binding (26).

The N-terminal 85 residues of {gamma}b contain three independent zinc-binding regions. To determine the contributions of the different motifs to metal binding, a number of mutant {gamma}b derivatives were tested for zinc affinity. The purified fusion proteins were loaded onto a polyacrylamide gel in approximately equal amounts and formed the major species in each lane (Fig. 5A and B, see Coomassie blue-stained panels). However, all lanes loaded with proteins extracted from E. coli also contained a host-derived protein band migrating just above the MBP fusions. This protein was not visible after Coomassie blue staining, but a minor zinc-binding signal was observed (Fig. 5A and B). Interestingly, the intensity of zinc binding observed for this host protein varied in different protein extracts, but the differences in signal did not correlate with the zinc-binding activity observed for the MBP fusion proteins. Thus, it is likely that variable amounts of the E. coli protein were present in the different samples.



View larger version (44K):
[in this window]
[in a new window]
  		FIG. 5. 65Zn(II)-binding activity of ADH and MBP-{gamma}b fusion proteins. (A and B) MBP fusion proteins separated on SDS-8% polyacrylamide gels were either stained with Coomassie brilliant blue and dried or blotted onto nitrocellulose, incubated with radiolabeled 65Zn(II), and exposed to X-ray film to assess zinc binding. The MBP fusion partner derivatives are designated at the top of each lane, and the numbers at the bottom are for reference in the text. The asterisk indicates the location of an E. coli-derived protein that copurifies with the MBP fusions. This protein is not visible in the Coomassie blue-stained gels but displays a minor zinc-binding signal.

 
To evaluate the individual contributions of the cysteine-rich regions to the zinc-binding activity of {gamma}b, three {gamma}b mutants, MBP-{gamma}b(–)C1, MBP-{gamma}b(–)C2, and MBP-{gamma}b(–)C1C2, were designed. These derivatives contain cysteine-to-serine or histidine-to-leucine substitutions in the C1, C2, or C1 and C2 regions, respectively. Both the MBP-{gamma}b(–)C1 and MBP-{gamma}b(–)C2 mutants demonstrated strong zinc-binding activity (Fig. 5A, lanes 3 and 4). Surprisingly, the MBP-{gamma}b(–)C1C2 mutant, which contained all 10 cysteine-to-serine and histidine-to-leucine mutations, still bound zinc with an intensity that appeared to be similar to that of the wt protein (Fig. 5A, lanes 2 and 5, and B, lanes 2 and 6). These results indicate that amino acids other than the cysteine and histidine residues in the C1 and C2 motifs contribute to zinc binding and that the {gamma}b protein could contain multiple zinc-binding regions, including the C1 and/or C2 motifs.

To define these regions, a series of additional mutants were designed to map the zinc-binding sites within the {gamma}b protein. Because the MBP-{gamma}b(–)C1C2 mutant retained considerable zinc-binding activity (Fig. 5A, lane 5), the {gamma}b protein was first divided into two components, MBP-{gamma}b1-85 and MBP-{gamma}b86-152. MBP-{gamma}b1-85 expresses the N-terminal 85 amino acids of {gamma}b fused to MBP and contains the C1, BM, and C2 motifs. The MBP-{gamma}b86-152 derivative encompasses the C-terminal 67 amino acids of {gamma}b fused to MBP and contains the associated coiled-coil motif. When tested for zinc binding, the MBP-{gamma}b1-85 mutant produced a strong signal that was similar in intensity to the signal from the full-length protein (Fig. 5B, lane 3). However, binding of 65Zn by the MBP-{gamma}b86-152 mutant was not detected (Fig. 5B, lane 4), even after lengthy exposure of the blot (data not shown). These results clearly demonstrate that the vast majority of the zinc-binding activity resides within the 85 N-terminal residues of {gamma}b.

The N-terminal {gamma}b residues were analyzed in more detail in order to provide a more precise map of the zinc-binding regions of the protein. To separately evaluate the C1, BM, and C2 regions, three MBP fusion clones—MBP-C1, MBP-BM, and MBP-C2—were constructed. These clones express {gamma}b amino acids 1 to 23 (C1), 21 to 49 (BM), and 60 to 85 (C2), respectively. When purified from E. coli by amylose affinity chromatography and tested in the zinc blotting assay, all three of the motifs bound zinc independently (Fig. 5B, lanes 11, 13, and 15). The signal intensities detected from the MBP-C1 and MBP-BM proteins were both strong, but the signal from the MBP-C2 protein was less intense. These results suggest that the C1 and BM motifs bind zinc with a substantially higher affinity than the C2 motif.

To determine which residues within the C1 and C2 regions mediate zinc binding, each of the cysteine residues in C1 were changed to serines, and the cysteine and histidine residues in C2 were changed to serines and to a leucine, respectively. When tested in the zinc blotting assay, neither the MBP(–)C1 or the MBP(–)C2 mutants retained the ability to bind zinc (Fig. 5B, lanes 12 and 14). These results demonstrate that the cysteine and histidine residues in the C1 and C2 motifs are required for coordination of zinc binding. Furthermore, these observations suggest that the disease phenotypes observed in previous studies of {gamma}b C1 and C2 mutants (4) could be the result of compromised zinc-binding activity.

The amino acids targeted in the BM were previously found to be important for in vitro RNA-binding activity (5). Substitution of arginine and lysine residues within BM resulted in diverse disease phenotypes when the mutant RNA{gamma} derivatives were inoculated onto plants (4, 5). Two BM mutants, BM26 (containing mutations R25Q and K26N) and BM29 (containing mutations R33G, K35N, and R36G) were also compromised in their ability to bind RNA. These changes were incorporated into the MBP-BM construct to create MBP-BM26 and MBP-BM29, respectively. When tested for zinc binding, both mutants appeared to have an affinity similar to the MBP-BM derivative (Fig. 5B, lanes 17 to 19). Because the mutations in the BM26 and BM29 derivatives compromised RNA-binding affinity without disrupting metal binding, these results suggest that the two activities can be separated. However, when substitutions of these five residues in the BM were incorporated into a single construct to produce MBP(–)BM, zinc binding was impaired significantly, and only a small amount of residual binding was detected upon long exposures of the blot (Fig. 5B, lane 16, and data not shown). These results indicate that the BM residues (R25, K26, R33, K35, and R36) that are required for metal binding are also involved in RNA binding.

To investigate zinc binding in the context of the full-length protein, mutations destroying zinc binding in each of the three isolated motifs were introduced into the MBP-{gamma}b protein in four combinations. The MBP-{gamma}b(–)C1C2 construct (C7,9,10,19,23,60,64,71,81S and H85L) that had been made previously bound zinc with a high affinity (Fig. 5A, lane 5). Three additional mutant derivatives were also constructed. Two of these, MBP-{gamma}b(–)C1BM (C7,9,10,19,23S, R25Q, R33,36G, and K26,35N) and MBP-{gamma}b(–)BMC2 (R25Q, R33,36G, K26,35N, C60,64,71,81S, and H85L) contained mutations that destroyed zinc-binding activity in two of the three identified motifs and left one zinc-binding region with its wt sequence. In contrast, the MBP-{gamma}b(–)C1BMC2 derivative contained all 15 mutations (C7,9,10,19,23,60,64,71,81S, H85L, R25Q, R33,36G, and K26,35N). The MBP-{gamma}b(–)C1BM and MBP-{gamma}b(–)BMC2 mutant combinations each retained binding activity (Fig. 5B, lanes 7 and 8). However, the signal detected from the MBP-(–)C1/BM mutant, which contained the wt sequence for the C2 region, was substantially lower than the signal for the MBP-{gamma}b(–)C1C2 and MBP-{gamma}b(–)BMC2 mutants. This binding affinity was similar to that of the three zinc-binding regions expressed individually outside of the context of the full-length protein. In both cases, the C2 region displayed considerably lower zinc-binding activity than either the C1 or BM regions (Fig. 5B, lanes 11, 13, and 15). In contrast, the mutant MBP-{gamma}b(–)C1BMC2, with substitutions in all three motifs, failed to show detectable activity (Fig. 5B, lane 5). Only long exposures of the blot revealed a residual amount of zinc signal associated with MBP-{gamma}b(–)C1BMC2, and this intensity was comparable to that detected for the MBP-(–)BM mutant (Fig. 5B, lane 16). These results thus demonstrate that the {gamma}b protein contains three independent zinc-binding regions whose activities must each be disrupted to generate a full-length {gamma}b protein lacking metal-binding activity.

Mutation of zinc-binding motifs affects {gamma}b antibody recognition. The in vivo changes in subcellular localization and aggregation effects noted upon mutagenesis of wt {gamma}b to the {gamma}b(–)C1C2 derivative prompted us to evaluate whether mutagenesis affected conformational changes sufficiently to alter antibody recognition of the individual motifs. Therefore, each of the MBP-{gamma}b fusion derivatives analyzed for zinc binding were also tested by Western blot analysis for recognition by the polyclonal antibody raised against {gamma}b. All of these proteins could be detected by Coomassie blue staining (Fig. 5, Coomassie blue-stained panels), and each protein reacted with the MBP antibody (Fig. 6, upper panels). The MBP fusions to the wt protein and each of the full-length {gamma}b mutants were recognized by the {gamma}b antibody, as were the {gamma}b1-85 and {gamma}b86-152 mutants and the C2 and BM motifs (Fig. 6, lanes 2 to 8, 12, and 14). In contrast, the C1 region was not recognized by the polyclonal {gamma}b antibody (Fig. 6, lane 10), even after lengthy exposure of the blots. The mutant regions (–)C1,(–)BM, and (–)C2 also failed to bind the antibody (Fig. 6, lanes 11, 13, and 15). However, after long exposures, signal was detected associated with the (–)C2 mutant, albeit to a much lower extent than for the C2 region (data not shown). These results show that the epitopes of the wt BM and C2 regions that are recognized by the {gamma}b antibody retain their antigenicity when the motifs are expressed outside the context of the full-length protein. In addition, the residues required for zinc-binding activity by the BM and C2 motifs are also essential for recognition by the {gamma}b antibody. In contrast, the wt C1 region, which failed to react with the {gamma}b antibody when expressed as an MBP fusion may contain epitopes that are not accessible for antibody recognition in the full-length protein, or the isolated motif may require flanking sequences to present recognizable epitopes. These results suggest that the structural integrity of {gamma}b depends on the zinc-binding activity of these three motifs and that the insolubility of the mutant derivatives in planta may be due to structural changes resulting from the inability to bind zinc.



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 6. Immunoblots of MBP fusion proteins. Purified MBP fusion proteins were separated on 8% SDS-polyacrylamide gels and blotted onto nitrocellulose. Proteins were detected with either MBP primary and goat anti-rabbit horseradish peroxidase secondary antibodies (top panels) or {gamma}b primary and GAM-HRP secondary antibodies (lower panels). Lane numbers are for reference in the text.

 
BSMV {gamma}b zinc-binding mutants elicit attenuated disease phenotypes. To determine whether the MBP fusions affected the functions of {gamma}b during infection, MBP-{gamma}b was substituted for wt {gamma}b in RNA{gamma} and inoculated with RNA{alpha} and RNAß onto C. amaranticolor and barley. Symptoms were not evident in barley, and although in C. amaranticolor local lesions were observed, these were smaller than those observed for the wt virus (data not shown). Other foreign sequences fused to the {gamma}b gene, including GFP and glutathione S-transferase (GST), are able to elicit systemic symptoms in barley and lesions in C. amaranticolor that are similar to lesions observed in a wt infection (data not shown). These observations show that fusion of {gamma}b to MBP affects virulence to a greater extent than if the {gamma}b ORF were not expressed or simply deleted from the genome.

To assess the biological effects of the derivatives that were compromised in vitro for zinc binding, the {gamma}b(–)C1BMC2, {gamma}b(–)C1C2, {gamma}b(–)C1BM, and {gamma}b(–)BMC2 mutants were introduced into the wt {gamma}42 clone, and transcripts containing these mutations were mixed with RNA{alpha} and RNAß and inoculated onto host plants (Fig. 7). In C. amaranticolor, the virus expressing the wt {gamma}b protein elicited chlorotic lesions in inoculated leaves by 4 dpi. Subsequently, the lesions progressively expanded and coalesced, became red in color, and turned necrotic, and by 11 dpi, the leaves had been abscised. When BSMV contained any one of the {gamma}b zinc-binding mutant derivatives, C. amaranticolor usually failed to develop lesions on the leaf, and when lesions did develop, they were far less pronounced than those resulting from wt infections. At 7 dpi, leaves inoculated with the {gamma}b(–)C1BM, {gamma}b(–)C1C2, and {gamma}b(–)C1BMC2 derivatives began to exhibit lesions, but only in regions associated with the leaf veins. By 11 dpi, these lesions had developed a red color and become necrotic, but lesion enlargement ceased. In addition, the {gamma}b(–)BMC2 mutant developed chlorotic lesions at 5 dpi that turned necrotic by 11 dpi, but these lesions did not change color or coalesce, and the leaf remained on the plant (Fig. 7A).



View larger version (62K):
[in this window]
[in a new window]
  		FIG. 7. Symptoms elicited by BSMV containing either wt {gamma}b or {gamma}b zinc-binding mutant derivatives. The local lesion host C. amaranticolor (A) and the systemic host barley (B) were inoculated with transcripts of RNA{alpha}, RNAß, and a RNA{gamma} derivative encoding either the wt or a mutant {gamma}b protein. Leaves of healthy and infected plants were photographed between 6 and 11 dpi as designated. Before they were photographed at 11 dpi, the C. amaranticolor leaves that were inoculated with the wt virus abscised and became desiccated and shrunken. (C) RNA and proteins were extracted from barley tissue infected with BSMV expressing either wt {gamma}b or {gamma}b zinc-binding mutant derivatives. The {gamma}b derivatives are designated at the top of each lane. RNA was separated on a 1% agarose gel, blotted onto Nytran, and detected with a radiolabeled probe complementary to the 3' region conserved among the BSMV RNAs. Proteins were separated on SDS-10% polyacrylamide gels, blotted onto nitrocellulose, and detected with the {gamma}b, coat protein, or TGB1 primary antibodies, respectively, and a GAM-HRP secondary antibody.

 
BSMV encoding the mutant {gamma}b derivatives also elicited aberrant symptoms in barley that were less pronounced than those observed for the wt virus (Fig. 7B). The phenotype elicited by these mutants was similar to the null phenotype observed for the {gamma}b deletion mutant (21) and the {gamma}b C2 and BM mutants described previously (4). The infected plants only developed erratic mosaic patches that failed to expand across the leaf blades. Western blot analysis of infected leaf tissue revealed the presence of the {gamma}b(–)C1BM, {gamma}b(–)C1C2, and {gamma}b(–)BMC2 mutants, but the {gamma}b(–)C1BMC2 derivative was not detected (Fig. 7C). Moreover, the levels of the {gamma}b mutant derivatives were lower than the levels of the wt protein. Although the coat protein was present in all of the samples, the amounts detected in the samples containing the {gamma}b mutants were greatly reduced from those of the wt sample (Fig. 7C). The TGB1 protein was also easily detectable in the wt {gamma}b sample but, as previously described for the null phenotype, this major movement protein could not be detected in any of the samples from plants inoculated with the mutant {gamma}b derivatives (Fig. 7C). In addition, viral accumulation was reduced in infections containing the mutant {gamma}b derivatives (Fig. 7C). Taken together, these observations suggest that each of the three zinc-binding motifs contributes to the functions of {gamma}b and that mutations affecting zinc binding in any of these motifs yield replication and movement defects similar to those observed when {gamma}b is not expressed or is deleted from the viral genome.


   DISCUSSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The {gamma}b protein plays an integral role in BSMV pathogenesis, and substantial pleiotropic effects have been observed upon its deletion from the viral genome. Symptoms are attenuated, viral RNA accumulation is decreased to 10 to 20% of wt levels, the abundance of the coat protein is reduced by 2 to 3 orders of magnitude, and the normally abundant TGB1 protein is undetectable in infections containing the {gamma}b-null mutant. To explore how individual regions of {gamma}b contribute to virulence, Donald and Jackson (4) introduced an extensive set of mutations into the C1, BM, and C2 regions of the protein. Substitutions in each of these motifs altered symptom phenotype, revealing that all three motifs are vital for the appropriate functioning of {gamma}b. Studies to evaluate RNA binding by the {gamma}b protein demonstrated that the BM mediates this activity, whereas substitutions in the C1 and C2 regions have little effect on RNA binding. In the present study, we conducted further studies to examine how the C1 and C2 regions of {gamma}b contribute mechanistically to the numerous effects on pathogenesis. The C1 and C2 motifs have long been suggested to have zinc-binding activity, but this ability had not been tested previously. We used both the {gamma}b protein purified from infected barley and recombinant MBP-{gamma}b fusion derivatives to demonstrate that the N terminus of {gamma}b functions in zinc binding. Within the N-terminal 85 amino acids of the protein, both the C1 and the C2 regions have zinc-binding activity and, surprisingly, the BM can also independently bind zinc. In addition, we have found that {gamma}b binds zinc preferentially over other divalent metal cations, and we have analyzed the biological effects of {gamma}b mutants in which zinc-binding activity has been compromised.

The experiments with the E. coli expressed MBP fusion proteins clearly demonstrate that the C1 region of {gamma}b has a high affinity for zinc. To assess the critical residues that participate in zinc binding, a series of cysteine to serine changes were introduced into the C1 motif, and substitutions at positions 7, 9, 10, 19, and 23 were found to disrupt the affinity of this region for zinc. Although the wt C1 sequence is competent to bind zinc, when this portion of the protein is expressed alone, it is not recognized by the {gamma}b antibody. The N terminus of {gamma}b is strongly hydrophobic, and this property may mask accessibility of antigenic sites in this portion of the protein. Inefficient protease cleavage of the N-terminal GST and MBP fusion partners from the GST-{gamma}b and MBP-{gamma}b proteins provides further support for this notion. The C1 motif may either be buried in the interior of the protein or obscured by interactions with other protein or RNA molecules. Furthermore, amino acid substitutions in the C1 motif may disrupt proper folding or interactions and result in exposure of the hydrophobic residues to the aqueous environment, leading to the observed insolubility of the mutant proteins.

Zinc binding of the C2 region was also disrupted by cysteine-to-serine changes at amino acids 60, 64, 71, and 81, and a histidine-to-leucine substitution at amino acid 85. Interestingly, the C2 motif displayed much lower binding than either the C1 or BM motifs. This difference is observed both in the C2 region expressed outside of the context of the full-length protein and in the {gamma}b(–)C1BM derivative, which contains substitutions disrupting the zinc-binding activity of the C1 and BM regions while preserving the wt C2 motif. However, the zinc-binding signal associated with the MBP-{gamma}b(–)C1BM mutant appears to be elevated over that observed with the isolated C2 motif. This observation suggests that sequences flanking the C2 region may contribute to its optimal activity by providing unidentified structural elements or by presenting an alternative ligand for zinc, such as the C50 residue that resides between the BM and C2 regions of {gamma}b.

The BM appears to have redundant basic amino acids functioning in zinc binding. Two mutants, BM26 and BM29 (5), that had previously been shown to be compromised for RNA binding, were each competent to bind zinc, but when substitutions of all five basic amino acids were combined into the (–)BM mutant, zinc binding was reduced to nearly undetectable levels. The minor residual signal detected for the (–)BM mutant could be attributed to the increased charge interaction between acidic residues in this region and the zinc cation, as has also been suggested for residual zinc binding observed with a mutant of the Rubella virus nonstructural protease (19). These results indicate that the basic residues that mediate RNA-binding activity are also critical for zinc binding.

To our knowledge, zinc-binding activity of a basic motif is unprecedented. Arginine is often found in the active site of zinc enzymes, but this amino acid is typically observed to participate in substrate orientation and activation or transition state stabilization rather than in direct interactions that bind metals (11, 14). However, an arginine residue that is substituted for a histidine in a mutant of the metalloenzyme carbonic anhydrase, has been shown by X-ray crystallography to coordinate with zinc in a catalytic metal-binding site (11). Within the protein environment, positively charged arginine residues can be deprotonated by a neighboring amino acid to permit entry of the resulting neutral derivative into the metal-binding region, as has been demonstrated for liver arginase (16). In addition, in binuclear zinc sites, metal binding has been observed by the amino group of lysine (14). Thus, the {gamma}b BM has a number of unusual chemical properties related to its zinc binding that appear to distinguish it from other well-characterized proteins involved in regulation.

Our compiled knowledge of its biochemical properties indicate that {gamma}b is a multifunctional protein that participates in several biological activities that require coordination of zinc binding. The differences in affinity for zinc among the three motifs suggest that each motif has a distinct role in the global function of the protein. In particular, the antigenicity and conformational shifts resulting from mutations that affect zinc interactions may contribute to the differences in RNA accumulation and disease phenotypes that were observed previously (4, 5). The in vitro RNA binding of the basic motif and the biological effects of mutations within this motif suggest that {gamma}b has an as-yet-undefined trans-acting role in the regulation of RNA interactions, particularly on RNAß. Although the biochemical roles of the C1 and C2 regions are still unclear, the distinct disease phenotypes noted after mutagenesis indicate that both motifs are critical for the function of the protein. One or both of the motifs may mediate sequence specificity of RNA binding in vivo in conjunction with the BM and contribute to interactions with host and/or viral proteins.

When {gamma}b mutant derivatives that affect zinc binding in vitro were introduced into the wt sequence in RNA{gamma}, BSMV was capable of eliciting systemic infections of barley but produced a null phenotype characterized by a mottled mosaic pattern. Several of the characteristics of the null phenotype elicited by the {gamma}b zinc-binding mutants {gamma}b(–)C1BMC2, {gamma}b(–)C1C2, {gamma}b(–)C1BM, and {gamma}b(–)BMC2 are indicative of a movement compromised virus. In plants inoculated with BSMV containing these derivatives, spread of the virus was clearly restricted, as indicated by narrow stripes and erratic mosaic patterns in barley and by lesions in C. amaranticolor that are reduced in diameter. Nevertheless, the mutant virus is capable of sufficient cell-to-cell movement to elicit detectable lesions in C. amaranticolor and to reach the vasculature in barley. Thus, even though the TGB1 movement protein is not detectable by Western blotting, sufficient amounts must be expressed from these BSMV derivatives to mediate virus spread.

A search using the pFam program to detect CRP homologs to BSMV {gamma}b revealed proteins from only two other genera of viruses. These included two pecluviruses (PCV and IPCV) and five furoviruses (SBWMV, SBCMV, OGSV, CWMV, and SCSV) (10), implying that the CRPs from these three genera have a common ancestry. Interestingly, the pFam program aligned only the BM, C2, and coiled-coil regions of {gamma}b with the other proteins and did not reveal similarities between the C1 region of {gamma}b and the CRPs of the furoviruses and pecluviruses. The BNYVV, TRV, and PEBV CRPs also contain basic and zinc binding motifs but lack a predicted coiled-coil region. The results of the pFam database search combined with this variation in structural organization suggests that the BNYVV, TRV, and PEBV proteins may be more distantly related to {gamma}b than the pecluvirus and other furovirus proteins. Thus, each region may have characteristic modular functions that were differentially acquired by a number of virus genera and conserved to varied extents during evolution. In different virus species their order may have been rearranged to accommodate overall function and yet are individually tailored to meet the needs specific to each viral infection cycle.

Although not identified by the pFam search for sequence similarity, the C2 pathogenesis protein of the geminivirus Tomato yellow leaf curl virus displays a number of similarities to {gamma}b. C2 is a basic protein that contains a cysteine-rich motif (CCCH). Like {gamma}b, C2 exhibits both zinc-binding and nucleic acid-binding activities (29). However, the DNA virus-encoded C2 protein binds preferentially to dsDNA, and cysteine residues within its CCCH motif were observed to participate in this binding. In contrast, {gamma}b preferentially binds ssRNA and ssDNA, and this activity is primarily mediated by lysine and arginine residues located within the BM, although mutations of cysteine residues within C1 also weaken the RNA-binding affinity. Finally, mutations within the CCCH motif of C2 lead to abnormal aggregation of the protein, a finding reminiscent of that observed for mutant derivatives of {gamma}b (4). The combination of RNA and zinc-binding functions exhibited by the ssRNA hordeivirus {gamma}b protein and the ssDNA geminivirus C2 protein suggests t