J Virol 91(12): e00367-17

PMC: PMC5446643

PMID: 28381573

Potato Virus Y HCPro Suppression of Antiviral Silencing in Nicotiana benthamiana Plants Correlates with Its Ability To Bind In Vivo to 21- and 22-Nucleotide Small RNAs of Viral Sequence

INTRODUCTION

Potyvirus HCPro is a multifunctional protein that has been studied since the 1980s, initially because of its involvement in the horizontal transmission of these economically important viruses by their insect vectors, hence its original name “helper component” (1), and later on because of its functions in the posttranslational processing of the viral polyprotein through its protease activity (2), as a pathogenicity determinant that can enhance the virulence of other viruses such as Potato virus X (PVX) in synergistic infections (3), and as a suppressor of antiviral defenses (4,–7) and because of other functions (8). This protein of around 50 kDa has three differentiated domains: an N-terminal domain, associated with aphid transmission (9, 10) and interaction with the proteasome (11, 12); a central domain associated with the suppression-of-silencing function (5, 13,–16); and a C-terminal domain responsible for its protease activity (3). In plants, HCPro is a cytoplasmic protein that forms homodimers and larger soluble aggregates as well as large cytoplasmic amorphous inclusions, the latter perhaps being end products of a translational strategy that produces equimolar amounts of each viral protein (17, 18). Inside cells, the subcellular whereabouts of HCPro are complex and dynamic: HCPro from Turnip mosaic virus (TuMV) tagged with green fluorescent protein (GFP) was found to localize to the cytoplasm of epidermal Nicotiana benthamiana cells as filaments, around the nucleus, and in what could be the endoplasmic reticulum (ER) (19). More recently, live confocal microscopy with tagged proteins visualized the interaction of Potato virus A (PVA) HCPro with a protein associated with the microtubule cytoskeleton (20) as well as the association of Potato virus Y (PVY) HCPro with small, regular, dot-like structures associated with the ER and its coating of the microtubule cytoskeletons under some stress conditions (21). The relationships between these observations and protein functions in silencing suppression or transmission by vectors remain to be clarified.

The precise molecular mechanisms by which HCPro interferes with host antiviral gene silencing defenses are not well understood either. It is not known whether HCPro inhibits the silencing defense at one level or, like some other viral suppressors, can operate at several levels. The known interactome of HCPro is extensive and includes a host tobacco calmodulin-related protein (rgs-CaM) factor that directs HCPro to degradation through the autophagy pathway (22, 23) and the Arabidopsis thaliana transcription factor RAV2, whose expression appears to modulate silencing and its suppression (24). HCPro also binds to components of the proteasome, a structure potentially involved in antiviral defense, modulating and inhibiting it (12, 25, 26), as well as to translation initiation factors (27) and to chloroplast factors (11, 28). To date, interactions of HCPro with protein components of the silencing machinery are few: interactions between Zucchini yellow mosaic virus (ZYMV) HCPro and the RNA methyltransferase HEN1 in vitro (29) and, as reported recently (7), between PVA HCPro and Argonaute1 (AGO1), which has been proposed to be related to the relief of viral RNA translational repression by ribosome-associated multiprotein complexes containing HCPro, have been demonstrated. On the other hand, HCPro binds to RNAs in vitro, and this property could also be related to its silencing suppression function: PVY HCPro interacted in vitro with 200-nucleotide (nt)-long RNAs (30, 31), and both Tobacco etch virus (TEV) and ZYMV HCPro proteins interacted in vitro with synthetic small RNAs (sRNAs) (15, 32). In vitro, Papaya ringspot virus (PRSV) HCPro bound synthetic sRNAs in a temperature-dependent manner (33), and it has been shown that a plant-purified TEV HCPro protein bound sRNAs in a size-dependent manner (34), with a preference for sRNAs of 21 nt with 3′-end overhangs and requiring the presence of an unknown host cofactor. One feature common to all of these HCPro-RNA interactions in vitro is that they required protein-to-RNA molar ratios much higher than the 2:1 ratio characterized for other suppressors proposed to interfere with silencing through sRNA sequestration, such as the P19 and 2b protein suppressors of Tomato bushy stunt virus (TBSV) and Cucumber mosaic virus (CMV) (see references 35 and 36, respectively), thus leading to the question of whether HCPro could be operating in this way. However, the use of next-generation sequencing (NGS) demonstrated that TuMV HCPro binds in vivo to sRNAs of viral sequences (viral sRNAs [vsRNAs]) of 21 and 22 nt in AGO2-deficient infected arabidopsis plants, supporting that the sequestration of sRNAs away from the silencing machinery by HCPro is a means of suppression of silencing (37). On the other hand, TuMV HCPro was shown to interfere with the biogenesis and action of microRNAs (miRNAs), but no direct binding to them was observed (37, 38).

In this work, we have used NGS to characterize short RNAs of <500 nt and sRNAs (18 to 30 nt) that were bound to PVY HCPro expressed from a PVX vector and purified under nondenaturing conditions from wild-type N. benthamiana plants. Our results show that PVY HCPro binds to sRNAs in vivo in this host and that this property correlates with the suppression-of-silencing activity of this protein. We also present data on specific properties of the sRNAs bound by HCPro.

RESULTS

Characterization of the silencing suppression activities and steady-state levels of accumulation of PVY HCPro mutants in N. benthamiana leaves.

We tested whether we could obtain a PVY HCPro mutant that lacked suppression-of-silencing activity but could accumulate sufficiently in plants to allow it to be purified and used as a control in our analysis. We created two alanine substitution mutants (6×-HCPro mutant A [mutA] and 6×-HCPro mutB) and a deletion mutant (6×-HCPro mutC). These mutations affected amino acidic motifs through the central domain of the protein (Fig. 1A) that had been related to HCPro silencing suppression or pathogenicity determinant properties in the literature (5, 13, 14, 16). None of the three mutants could suppress the silencing of a GFP reporter in agropatch assays (Fig. 1B, left). Mutants A and C failed to accumulate in infiltrated patches when transiently expressed in the presence of GFP (Fig. 1B, right) and accumulated weakly in the presence of the heterologous CMV 2b protein suppressor but not in the presence of the PVX p25 suppressor (Fig. 1B and data not shown). Only HCPro mutB accumulated in infiltrated patches even in the absence of heterologous suppressors as well as when expressed from a PVX vector (P1-6×-HCPro mutB construct) (Fig. 1B, right, last two lanes), making it amenable for purification from infected tissue when expressed from this vector.

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630001.jpg

FIG 1

Characterization of the activities of silencing suppression of three PVY HCPro mutants (mutants A, B, and C) generated for this study and of their steady-state levels of accumulation in Nicotiana benthamiana plants. (A) Schematic representation of the amino acid substitutions or deletions that affect each of the three mutants. (B) Assessment of the activities of the three mutants in the suppression of silencing of a GFP reporter transiently expressed in infiltrated leaf patches by agroinfiltration of the binary vectors that express them (agropatch assay). The GFP reporter was expressed in either the absence (−) or the presence of PVY 6×-HCPro (HCPro wild type [HC wt]), 6×-HCPro mutant A (HC mutA), 6×-HCPro mutant B (HC mutB), or 6×-HCPro mutant C (HC mutC) or the Cucumber mosaic virus 2b protein, indicated with labels in the corresponding leaf patches. The steady-state levels of accumulation of the suppressors in the patches were assessed by the Western blot analyses (right). The rightmost two lanes of the middle Western blot panel also shows the accumulation of 6×-HCPro mutB in systemic tissue when expressed from a PVX vector (construct PVX-P1-6×-HCPro mutB). M, molecular mass marker (in kilodaltons).

Purification of HCPro, HCPro mutB, and control samples under nondenaturing conditions and isolation of associated RNAs.

We performed two purifications of the histidine-tagged 6×-HCPro protein under nondenaturing conditions using N. benthamiana plants infected with PVX-P1-6×-HCPro as an input and plants infected with the empty PVX vector in the first purification event (Fig. 2A) or plants infected with PVX-P1-6×-HCPro mutB that expressed the suppressor-activity-deficient HCPro mutB in the second purification event (Fig. 2B) as parallel purified controls. The P1 protein detaches itself posttranslationally from 6×-HCPro via proteolysis (39). Samples from each purification step and elution fractions were resolved by SDS-PAGE (Fig. 2). The presence of the ∼50-kDa 6×-HCPro protein band was visualized by Coomassie blue staining of the SDS-PAGE-resolved purified preparations and confirmed serologically. Other minor protein components with affinities for nickel resin under nondenaturing conditions could also be detected in the Coomassie-stained gels. Nonlinear silver staining of gels made apparent the presence of a common background composed of proteins of all sizes with affinities for nickel resin in the HCPro and control samples (Fig. 2A, bottom, and B, bottom). Densitometry analysis of the Coomassie blue-stained gel lanes corresponding to eluted fractions used for the isolation of bound RNAs determined that the percentage of 6×-HCPro or of 6×-HCPro mutB over the total protein content was between 54 and 64% (Fig. 2). RNAs associated with purified fractions were isolated as described in Materials and Methods.

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630002.jpg

FIG 2

Purification of HCPro and control samples under nondenaturing conditions. (A, top) Coomassie blue-stained SDS-PAGE gels from different steps in the purification of the control (left) and HCPro (right) samples from PVX- and PVX-P1-6×-HCPro-infected plants, respectively, and Western blot detection of the 6×-HCPro protein (panel below the right gel). The rightmost lanes in each gel (second elutions) correspond to the eluted control and purified HCPro samples, respectively. The value of 64% indicates the minimum content of 6×-HCPro relative to the total protein content in the HCPro sample. (Bottom) Silver-stained SDS-PAGE gel of the purified samples. The band corresponding to 6×-HCPro is marked by an arrow. Many other bands corresponding to proteins of different sizes with affinities for nickel resin under nondenaturing conditions are visible, and their patterns are similar for both samples. (B, top) Coomassie blue-stained SDS-PAGE gels from different steps in the purification of 6×-HCPro mutB (left) and 6×-HCPro (right) samples from PVX-P1-6×-HCPro mutB- and PVX-P1-6×-HCPro-infected plants, respectively, and Western blot detection (panels below the gels). Values of 64 and 54% indicate the minimum content of 6×-HCPro of the total protein in the HCPro mutB and the HCPro samples, respectively. (Bottom) Silver-stained SDS-PAGE gels of the purified samples. The bands corresponding to 6×-HCPro mutB and 6×-HCPro are marked by arrows. All nonlabeled lanes in gels are unloaded lanes. SN, supernatant. The lane labeled M shows RNA molecular mass markers, indicated in kilodaltons at the sides of gels.

Characterization of the sRNA populations present in input plants.

We also determined the properties of the sRNA populations present in input N. benthamiana plants infected with either PVX or PVX-P1-6×-HCPro. It was not previously known how the expression of biologically active PVY HCPro from a PVX vector could affect the relative ratios of viral to plant sRNAs or their composition. Our NGS analysis found rather similar proportions of sRNAs (18 to 30 nt) in both types of infected plants with regard to their origin (for viral versus plant sequences, 30.72% versus 69.28% in PVX-infected plants and 33.70% versus 66.30% in PVX-P1-6×-HCPro-infected plants) (Table 1 and Fig. 1A). These proportions were inverted when considering only 21- and 22-nt sRNAs (for viral versus plant sequences, 65.63% versus 34.37% in PVX-infected plants and 61.05% versus 38.95% in PVX-P1-6×-HCPro-infected plants) (Table 1). The distributions of reads by sRNA size (18 to 30 nt) were also similar in both types of infected plants (Fig. 1A). The distribution of 21- to 22-nt vsRNAs throughout the viral genomes and their frequencies remained remarkably similar (Fig. 1B), as were their diversities (Fig. 1C). Given the limited effect of the expression of P1-6×-HCPro on the sRNAs of plants infected by a PVX vector, we did not test the case of PVX-P1-6×-HCPro mutB. Data on the properties of sRNA populations in infected plants are summarized in Table 1.

TABLE 1

Summary of the properties of reads of 18 to 30 nt obtained by deep sequencing of libraries produced from sRNAs isolated from input Nicotiana benthamiana plants infected with either a PVX vector or with PVX expressing a modified P1-6-HCPro bicistron from the potyvirus Potato virus Y^a

Parameter	Value for group
Parameter	Plants infected with PVX	Plants infected with PVX-P1-6×-HCPro
Total no. of reads with rRNAs (10⁾	27.77	32.40
Total no. (%) of reads without rRNAs (10^{) of:}	27.09	31.27
Plant sequence	18.77 (69.28)	20.73 (66.30)
Viral sequence	8.32 (30.72)	10.54 (33.70)
21 and 22 nt of:	10.53 (38.87)	15.51 (49.60)
Plant sequence	3.62 (34.37)	6.04 (38.95)
Viral sequence	6.91 (65.63)	9.47 (61.05)
Total no. of reads of microRNAs
Matching pre-miRNAs	26,655	39,376
Matching mature miRNAs	854	1,254

^{Numbers in total reads are given as millions of reads, except for those matching pre-miRNAs or mature miRNAs, which appear as total reads. After subtraction of the number of reads corresponding to rRNAs, percentages of total reads that match}Nicotiana benthamiana or viral sequences are also given (in parentheses).

Characterization of the sRNAs associated with purified HCPro and control samples.

Properties of the different RNA species associated with the purified HCPro and control samples were determined by high-throughput deep sequencing. The results are summarized in Table 2. Reads that corresponded to rRNAs were discarded, leaving for analysis 10.66 million and 16.54 million reads of 18 to >120 nt from the control and HCPro samples of the first purification event, respectively, and 34.84 million and 40.48 million reads of 18 to 30 nt from the HCPro mutB and HCPro samples of the second purification event, respectively (Table 2). In both experiments, reads that corresponded to sRNAs of 21 and 22 nt were the most abundant ones. Very low numbers of reads, in contrast, corresponded to pre-miRNA or mature miRNA sequences, and these numbers even decreased in the 6×-HCPro sample with regard to the controls (Table 2).

TABLE 2

Summary of properties of the reads obtained by deep sequencing of RNAs associated with purified control and HCPro samples in the first purification experiment (reads of 18 to >120 nt) or 6×-HCPro mutB and 6×HCPro samples in the second experiment (reads of 18 to 30 nt)^a

Parameter	Value for group
	Expt 1		Expt 2
	Control	6×-HCPro	6×-HCPro mutB	6×-HCPro
Total no. of reads with rRNAs (10⁾	21.11	26.06	35.23	40.65
Total no. (%) of reads without rRNAs (10^{) of:}	10.66	16.54	34.84	40.48
Plant sequence	10.50 (98.50)	15.17 (91.72)	34.33 (98.53)	37.87 (93.55)
Viral sequence	0.15 (1.41)	1.37 (8.28)	0.51 (1.47)	2.60 (6.45)
18 to 30 nt of:	7.50 (70.36)	11.40 (68.92)	34.84 (100.00)	40.48 (100.00)
Plant sequence	7.38 (98.37)	10.14 (89.00)	34.33 (98.53)	37.87 (93.55)
Viral sequence	0.12 (1.63)	1.25 (11.00)	0.51 (1.47)	2.6 (6.45)
21 and 22 nt of:	3.94 (36.96)	7.86 (47.52)	26.42 (75.83)	32.32 (92.76)
Plant sequence	3.85 (97.54)	6.65 (84.66)	25.94 (98.19)	29.87 (92.41)
Viral sequence	0.10 (2.46)	1.20 (15.34)	0.47 (1.81)	2.45 (7.59)
Total no. of reads of microRNAs
Matching pre-miRNAs	31,666	16,859	25	15
Matching mature miRNAs	400	131	2	0

^{Numbers of total reads are given as millions of reads, except for those matching pre-miRNAs or mature miRNAs, which appear as total reads. After subtraction of the number of reads corresponding to rRNAs, percentages of total reads that match}Nicotiana benthamiana or viral sequences are also given in parentheses.

In the first purification event, 98.37% of the reads of 18- to 30-nt sRNAs in the control sample were of plant sequences, and 1.63% were viral reads. Considering only 21- and 22-nt sRNAs, these proportions were 97.54% for plant reads and 2.46% for viral reads, far below the 34.37% plant reads and 65.63% viral reads of the infected plant input (Table 2 versus Table 1 and see Fig. 4A). This remarkable result shows that one or more proteins from the background bind to sRNAs but with little or no affinities for vsRNAs. In the HCPro sample, the proportions of vsRNA reads of 18 to 30 nt rose to 11.00%, and considering only those of 21 and 22 nt, the proportions increased from 6.77 to 27.20% and from 0.88 to 5.72% of the total, respectively (see Fig. 4A). In the second purification event, the proportions of plant versus viral sRNA reads of 21 and 22 nt in the HCPro mutB sample were also very low (1.81%) (Table 2 and see Fig. 4B). The presence of 6×-HCPro in the HCPro sample correlated with a >4-fold increase in the proportion of vsRNAs to 7.59% (Table 2 and see Fig. 4B), and in the case of those of 21 and 22 nt, the proportions increased from 11.76 to 43.38% and from 0.61 to 2.42%, respectively. The control and HCPro mutB samples from both experiments were similar in their near absence of vsRNAs despite the presence of the 6×-HCPro mutB protein in the latter.

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630004.jpg

FIG 4

Reads of sRNAs of plant or viral sequences present in the HCPro and control samples from the first (A) and second (B) purification experiments. The percentages of reads of viral and plant sequences are shown: top pie graphs show sRNAs of 18 to 30 nt. Middle and bottom pie graphs show only 21- and 22-nt sRNAs, respectively. The numbers of reads per million for each sRNA size are shown in the graphs at the bottom for plant and viral sequences combined (top), only plant sequences (middle), or only viral sequences (bottom). Fold change values for the viral sRNAs of 21 and 22 nt are also indicated.

For each sRNA size class from 18 to 30 nt, we analyzed its relative enrichment in HCPro versus control or HCPro mutB samples. To do this, we used reads per million for each particular sRNA size class in the samples. Reads per million for RNAs of 18 to 30 nt of either plant or viral sequences are shown in Fig. 4 (bottom). In the first purification event, the number of reads corresponding to 21-nt sRNAs of plant sequences was 2.60 times higher in the HCPro sample than in the control sample, whereas those of viral sequences were up 13.35 times higher. For the 22-nt sRNAs, these numbers were 1.43 and 9.78 times higher, respectively (see Fig. 4, left). In the second purification event, 21- and 22-nt plant reads did not change, and only the numbers of reads of vsRNAs of 21 and 22 nt were increased in the HCPro sample relative to the HCPro mutB sample, with fold change values of 4.5 and 4.1, respectively (see Fig. 4, bottom right). These fold change increases in 21- and 22-nt reads of viral sequences relative to the controls show that 6×-HCPro binds at least to vsRNAs.

Given than in input plants, one-third of the 21- and 22-nt sRNA populations are of plant sequences and two-thirds are of viral sequences (Fig. 3), the observed increase in the proportions of 21- and 22-nt vsRNAs in the purified HCPro versus control samples (Fig. 4, right versus left) could be compatible with 6×-HCPro providing 21- and 22-nt sRNAs of only the viral sequence, of viral and plant sequences indistinctly, or somewhere in between. In the former case, the contribution of 6×-HCPro to the pool of 21- and 22-nt sRNAs would be smaller than in the other cases, to explain the fold change increases in vsRNA reads in the HCPro sample (Fig. 4). We calculated mathematically the contribution of 6×-HCPro to the pool of 21- and 22-nt sRNAs associated with the purified HCPro samples under three scenarios of preference for binding to sRNAs of plant or of viral sequences, including no preference for binding viral over plant sequences (50:50) or some preference for (75:25) or binding to only (100:0) viral sequences, and found that for 21-nt sRNAs, this value ranged from a maximum of 58% if it bound to sRNAs of viral or plant sequences indistinctly to a minimum of 22% if it bound only to sRNAs of viral sequences. The contribution was much lower for 22-nt-long sRNAs (Fig. 5).

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630003.jpg

FIG 3

Analysis of sRNAs isolated from Nicotiana benthamiana plants infected with either a PVX vector or PVX expressing a modified P1-6×-HCPro bicistron from the potyvirus Potato virus Y. (A) Percentages of reads of viral and of plant sequences in 18- to 30-nt or only 21- and 22-nt sRNAs. The graphs at the bottom show reads per million for each RNA size of plant and viral sequences (top), only plant sequences (middle), or only viral sequences (bottom). (B) Distribution in the viral genome of 21- and 22-nt reads in reads per million of vsRNAs. At the top, the viral genomic RNA used to match the sequences and also the subgenomic RNAs are represented schematically. Positive-sense RNAs are represented at the top of the charts (dark gray), whereas negative-polarity RNAs are represented at the bottom (light gray). The top scale was capped at 150 reads/million. (C) Venn diagram of unique 21- and 22-nt vsRNA reads of either polarity obtained from PVX-infected plants (clear circle at the left) or from PVX-P1-6×-HCPro-infected plants (dotted circle at the right). Numbers associated with each circle represent unique reads matching PVX sequences. The number inside the streaked overlapping area between both circles shows the number of unique reads common to both samples.

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630005.jpg

FIG 5

Estimation of the percent contribution of 6×-HCPro to the population of 21- and 22-nt sRNAs associated with purified HCPro samples. This contribution was calculated under three different assumptions of binding preference of 6×-HCPro for sRNAs of viral over plant sequences: no preference (50% viral and 50% plant sequences), a preference for 75% viral and 25% plant sequences, or an exclusive preference for viral over plant sequences (100% viral and 0% plant sequences). The contribution of 6×-HCPro was obtained with the algorithm indicated below the chart. Detailed information on the mathematical formula for this analysis can be found in Fig. S1 in the supplemental material.

In both purification events, reads of viral sequences were distributed throughout the whole viral genome, although there were a few hot spots (Fig. 6A). Considering only the vsRNAs of 21 and 22 nt and only those of the PVX sequence (excluding those with sequences corresponding to the P1-6×-HCPro or P1-6×-HCPro mutB insert), HCPro samples were enriched not only in vsRNA reads but also in their diversity relative to the control or HCPro mutB samples (Fig. 6A and andB).B). RNA reads of viral sequences were largely of positive (sense) polarity, although for some vsRNA sizes, negative-sense reads were also abundant, at around 40% (Fig. 6C).

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630006.jpg

FIG 6

Distribution of reads of viral sequences throughout the viral genomes obtained from HCPro and control samples from the first (left) and second (right) purification experiments. (A) Distribution of reads for RNAs of all sizes (18 to >120 nt in first experiment and 18 to 30 nt in second experiment) and of 21 and 22 nt. Values are shown as reads per million, and the scale of the charts is capped at 150 reads/million. Positive-sense RNAs are represented on top side of the charts (dark gray), whereas negative-polarity RNAs are represented on the bottom side (light gray). (B) Venn diagram of unique 21- and 22-nt vsRNA reads of either polarity. Numbers associated with each circle represent unique reads matching PVX sequences. The number inside the streaked overlapping area between both circles is the number of unique reads common to both samples. (C) Percentages of positive and negative polarities in the reads of viral sequences in the samples, distributed by size.

To test whether 6×-HCPro could show a preference for binding to sRNAs with a particular 5′-end nucleotide, we analyzed the 5′-end nucleotide compositions of the RNA reads in purified samples and calculated the fold changes in the frequencies of each of the four nucleotides at the 5′ ends for every RNA size in the HCPro versus control and in the HCPro versus HCPro mutB purification events (Fig. 7A and andB,B, respectively). We found that the numbers of 21- and 22-nt-long vsRNA reads that had an adenine at their 5′ ends increased in the HCPro samples in comparison to the reads in the control or HCPro mutB samples (Fig. 7, arrows). Fold changes for 5′-end adenines in the 21-nt vsRNAs of the HCPro samples of the first and second purifications (11.72- and 4.50-fold, respectively) were absent in the corresponding sRNAs of plant sequences as well as in the vsRNA populations of infected input plants (Fig. 7). An additional 5′-end preference for cytosines was also observed in reads of viral RNAs in the HCPro sample of 79 to 84 nt, but reads corresponded mainly to the hot spot in the PVX 166K replicase gene, and a large proportion of these reads start at their 5′ ends with a cytosine at position 2042 in the PVX genome; therefore, they are not indicative of a bias toward binding RNAs with this particular starting nucleotide.

An external file that holds a picture, illustration, etc.
Object name is zjv9991826630007.jpg

FIG 7

The 5′-end nucleotide prevalence in sRNA reads of either plant (top) or viral (bottom) sequences in purified samples. (A, left) Percent 5′-end nucleotide prevalences in reads of RNAs of 18 to 30 nt in the control (top) and HCPro (bottom) samples from the first purification experiment. (Right) Chart showing the corresponding fold change values in the 5′-end nucleotides for all RNA sizes (18 to >120 nt). (B, left) Percentages of 5′-end nucleotides in RNA reads ranging from 18 to 30 nt from the HCPro mutB (top) and HCPro (bottom) samples. (Right) Chart showing the corresponding fold change values in the HCPro sample versus the HCPro mutB sample for sizes ranging from 18 to 30 nt. The percentages of 5′-end nucleotides and fold change variations in input plants infected with PVX or PVX-P1-6×-HCPro appear at the bottom.

Characterization of short RNAs of >30 nt that potentially bound to HCPro.

In the first purification event, we also analyzed reads of >30 nt, and we studied whether enrichments for some of these reads in the HCPro versus control samples occurred. We found some enrichments, but their read numbers were very small. These fold change increases were clustered mainly into two size ranges, the first one in reads of RNAs of 73 to 84 nt and the second one in reads of 98 to 99 nt (see Fig. S2 in the supplemental material). Reads in the HCPro sample of 73 to 84 nt were 20 times more abundant than in the control sample, and this was mainly attributable to three hot spots of sequences with more than 1,000 reads each but especially to one of them, found only in the HCPro sample, that corresponded to a sequence in the PVX 166K gene from nt 2038 to 2227 and was composed of 13,000 reads with sense polarity (Fig. S2). This viral region did not correlate with a noncanonical open reading frame (ORF) and did not form any stable secondary structure that could help explain its origin (analyzed by using Mfold [40]) (data not shown). Reads of 98 and 99 nt were also more abundant in the HCPro sample, and they formed another hot spot from nt 6264 to 6362 located within the potyviral P1 cistron and therefore were not present in the control sample (Fig. S2). The existence of this hot spot with 1,693 reads could not be explained either by a noncanonical ORF or by the formation of secondary structure in this viral region (analyzed by Mfold) (data not shown), and its cause also remains unexplained.

Characterization of the silencing suppression activities and steady-state levels of accumulation of PVY HCPro mutants in N. benthamiana leaves.

FIG 1

Purification of HCPro, HCPro mutB, and control samples under nondenaturing conditions and isolation of associated RNAs.

FIG 2

Characterization of the sRNA populations present in input plants.

TABLE 1

Parameter	Value for group
Parameter	Plants infected with PVX	Plants infected with PVX-P1-6×-HCPro
Total no. of reads with rRNAs (10⁾	27.77	32.40
Total no. (%) of reads without rRNAs (10^{) of:}	27.09	31.27
Plant sequence	18.77 (69.28)	20.73 (66.30)
Viral sequence	8.32 (30.72)	10.54 (33.70)
21 and 22 nt of:	10.53 (38.87)	15.51 (49.60)
Plant sequence	3.62 (34.37)	6.04 (38.95)
Viral sequence	6.91 (65.63)	9.47 (61.05)
Total no. of reads of microRNAs
Matching pre-miRNAs	26,655	39,376
Matching mature miRNAs	854	1,254

Characterization of the sRNAs associated with purified HCPro and control samples.

TABLE 2

Parameter	Value for group
	Expt 1		Expt 2
	Control	6×-HCPro	6×-HCPro mutB	6×-HCPro
Total no. of reads with rRNAs (10⁾	21.11	26.06	35.23	40.65
Total no. (%) of reads without rRNAs (10^{) of:}	10.66	16.54	34.84	40.48
Plant sequence	10.50 (98.50)	15.17 (91.72)	34.33 (98.53)	37.87 (93.55)
Viral sequence	0.15 (1.41)	1.37 (8.28)	0.51 (1.47)	2.60 (6.45)
18 to 30 nt of:	7.50 (70.36)	11.40 (68.92)	34.84 (100.00)	40.48 (100.00)
Plant sequence	7.38 (98.37)	10.14 (89.00)	34.33 (98.53)	37.87 (93.55)
Viral sequence	0.12 (1.63)	1.25 (11.00)	0.51 (1.47)	2.6 (6.45)
21 and 22 nt of:	3.94 (36.96)	7.86 (47.52)	26.42 (75.83)	32.32 (92.76)
Plant sequence	3.85 (97.54)	6.65 (84.66)	25.94 (98.19)	29.87 (92.41)
Viral sequence	0.10 (2.46)	1.20 (15.34)	0.47 (1.81)	2.45 (7.59)
Total no. of reads of microRNAs
Matching pre-miRNAs	31,666	16,859	25	15
Matching mature miRNAs	400	131	2	0

Characterization of short RNAs of >30 nt that potentially bound to HCPro.

DISCUSSION

The silencing suppression activity of the HCPro component from members of the genus Potyvirus was discovered soon after RNA-based gene silencing as a mechanism of defense against plant viruses (4, 5), but the modes by which it interferes with silencing are still under investigation. On the one hand, HCPro has been shown to interact with several factors from the host, such as HEN1 in vitro (29), whose methylation of sRNAs is required for their incorporation into RNA-induced silencing complexes. Perhaps related to this, it has been observed that in a transgenic HCPro background, the small interfering RNA (siRNA) population derived from infection of tobacco plants by CMV appeared to be less methylated at their 3′ ends than in nontransgenic plants (41). Recently, HCPro from PVA has been proposed to suppress silencing through two mechanisms: by its interaction with components of the methionine cycle that would deprive HEN1 of its substrate, preventing sRNA methylation and therefore antiviral silencing, or by relieving viral translational repression through its interaction with AGO1 (7). An alternative mode by which HCPro could suppress silencing would be through its binding of sRNAs, as has been shown for some viruses in vitro by using synthetic sRNAs (15, 31,–34) and in vivo for 21- and 22-nt sRNAs by TuMV HCPro in infected AGO2-deficient arabidopsis plants (37). HCPro could thus be suppressing silencing by any of these mechanisms or by a combination of some or all of them.

To test whether PVY HCPro binds RNAs in planta, investigate the properties of these RNAs, and explore their potential relationship to its suppressor function, we expressed hexahistidine-tagged HCPro from a PVX vector in the compatible host N. benthamiana, purified it under nondenaturing conditions using Ni^{-nitrilotriacetic acid (NTA) resin, and analyzed RNAs associated with the purified sample, using high-throughput sequencing. To follow this approach, we needed to analyze and compare the compositions of the sRNA populations found in input infected plants and in the purified HCPro and control samples. During the course of this work, we had some interesting observations.}

The profiles of sRNAs in plants infected with PVX and those infected with PVX expressing PVY HCPro are very similar.

It was known that the expression of PVY HCPro from PVX vectors reduced viral titers and the level of genomic viral RNAs (42) and that the expression of HCPro from other potyviruses increased the accumulation of PVX subgenomic RNAs in plants relative to that of the genomic RNA (43). However, when we determined the properties of the sRNA populations present in input N. benthamiana plants infected with either PVX or PVX-P1-6×-HCPro, we found very similar profiles for 18- to 30-nt-long sRNAs with regard to their distribution by sizes or throughout the viral genome or in their ratios of viral versus plant sequences (Fig. 3).

Preparations purified by using Ni-NTA resin under nondenaturing conditions also contain a protein background that binds to specific sRNAs.

Plant proteins of different sizes with metal affinities can bind to Ni^{-NTA resin under nondenaturing extraction and purification conditions, as was observed previously for the purification of PVY HCPro (}39) or the CMV replicase complex (44). We found similar protein backgrounds with and without 6×-HCPro in purified control and HCPro samples (Fig. 2, bottom). This was not surprising, as the maximum capacity of the resin to bind hexahistidine-tagged proteins is 5 to 10 mg/ml of resin (QIAExpressionist handbook; Qiagen GmbH, Hilden, Germany), and we purified at most 0.6 mg of 6×-HCPro/ml of resin, leaving over 90% of the sites still available for less specific binding by other factors with metal affinities. In the purified HCPro samples, 6×-HCPro constituted at least 54% of the total protein content (Fig. 2). Sequencing of associated RNAs in the control sample from the first purification event produced millions of reads in both cases, indicating that binding to sRNAs by unidentified components of the protein background had taken place. Analysis of reads in the control sample showed that the proportion of reads of vsRNAs was extremely low (1.63%) (Fig. 4A). This is a remarkable fact, as it implies that the protein background has very little or no affinity for vsRNAs, and its binding to plant sRNAs must take place under conditions (i.e., in their timing or because of physical barriers through subcellular compartmentalization) that prevent binding to vsRNAs. The nature of the sRNAs of mainly plant sequences associated with the protein background and the protein(s) responsible remain to be investigated. However, the most abundant sRNA sizes of 21 and 22 nt (Fig. 4A) had mainly 5′-end cytosines and adenines, respectively (Fig. 7B), suggesting that binding by the protein background is specific.

During infection, PVY HCPro binds in vivo to vsRNAs of 21 nt and, to a much lesser extent, of 22 nt, and the binding ability correlates with suppressor activity.

Analysis of the RNA reads associated with the purified suppressor-deficient HCPro mutB sample in the second purification produced a profile almost identical to that of the control sample in the previous purification (Fig. 4). In contrast, HCPro samples from both purification experiments showed marked differential enrichments for vsRNAs of 21 nt and 22 nt with regard to the controls but far less or none of the same-sized sRNAs of plant sequences (Fig. 4, bottom). This increase in the proportion of vsRNAs of 21 and 22 nt in the HCPro sample relative to the controls demonstrates that 6×-HCPro contributes, at least with vsRNAs of 21 and 22 nt, to the pool of sRNAs associated with the purified sample, although these data would also be compatible with 6×-HCPro also contributing same-sized sRNAs of plant sequences (Fig. 5). The relative contribution of 6×-HCPro to the sRNA pool found to be associated with the HCPro sample could be calculated using different binding preference scenarios based on their being of viral or plant sequences: in the second experiment, it contributed at most 58% of the 21-nt sRNAs in the case of no preference (Fig. 5), even though 6×-HCPro constituted over 54% of the protein in the purified sample (Fig. 2). This indicates that the binding of 6×-HCPro to these sRNAs was not stronger than that of the unidentified component(s) of the protein background either in planta or under the extraction and purification conditions for the samples.

Our analysis of the 5′-end nucleotide prevalence in sRNAs associated with the HCPro sample found an enrichment of those sRNAs containing adenines at their 5′ ends specific to vsRNAs of 21 and 22 nt that was not observed in the sRNAs of plant sequences of the same size (Fig. 7). These data could be explained if 6×-HCPro bound with a preference for 21- and 22-nt vsRNAs with 5′-end adenines. Interestingly, the fold changes of 5′-end adenines in 21-nt vsRNAs (Fig. 7, middle) were similar to those found in the enrichment of 21-nt vsRNAs in HCPro versus control samples (Fig. 4, bottom). However, in both the control and HCPro mutB samples against which these fold changes were calculated, the total numbers of reads were comparatively low (Table 2), and the vsRNAs of 21 and 22 nt with 5′-end adenines were quite underrepresented (Fig. 7) for unknown reasons. Therefore, these data should be interpreted with caution. Nevertheless, preferences in vivo for sRNAs with specific 5′-end nucleotides are known for some sRNA-binding proteins, such as AGO proteins (37, 45,–49), and also for viral suppressors, such as the 2b protein of CMV (50). Our PVY HCPro purification approach produced samples enriched in 6×-HCPro from wild-type N. benthamiana plants (Fig. 2) that were biologically active as a helper factor in virus transmission assays (not shown). This purification approach is different from the immunoprecipitation procedure used to obtain purified TuMV HCPro from infected arabidopsis plants with AGO2 mutant backgrounds (37), but nevertheless, analysis of raw data from this work indicates that some enrichment in 5′-end adenines could also have occurred in the purified native versus mutant TuMV HCPro samples (37).

In conclusion, our work demonstrates that PVY HCPro expressed from a virus vector binds in planta to sRNAs with viral sequences of 21 nt and also, to a much lesser extent, of 22 nt and that this binding ability correlates with its suppression-of-silencing activity (Fig. 1 and and4).4). The 21-nt vsRNAs bound to HCPro were also found to be differentially enriched in 5′-end adenines relative to purified controls (Fig. 7). These data support physical binding of HCPro to vsRNAs of 21 and 22 nt as a means to interfere with antiviral silencing. In addition to this, our data would be compatible with the possibility that PVY HCPro binds in vivo to 21- and 22-nt-long sRNAs of viral or plant sequences indistinctly and with any 5′ ends but also with the possibility that during infection, HCPro binds mainly or even only to vsRNAs of 21 and 22 nt with adenines at their 5′ ends (Fig. 5), thus specifically targeting vsRNAs that could become loaded into AGO2.

The profiles of sRNAs in plants infected with PVX and those infected with PVX expressing PVY HCPro are very similar.

Preparations purified by using Ni-NTA resin under nondenaturing conditions also contain a protein background that binds to specific sRNAs.

During infection, PVY HCPro binds in vivo to vsRNAs of 21 nt and, to a much lesser extent, of 22 nt, and the binding ability correlates with suppressor activity.

MATERIALS AND METHODS

Plants and viruses.

The N. benthamiana plants used in this study were kept in controlled growth chambers with temperatures of 25°C during the day and 20°C at night, with a 16-h/8-h day/night photoperiod and ∼2,500 lx of daylight intensity. Four- to five-week-old plants were agroinoculated with Agrobacterium tumefaciens containing the appropriate binary constructs expressing three infectious PVX vectors: an “empty” PVX vector was expressed from binary construct pgR107 that originated from D. C. Baulcombe's group (University of Cambridge, UK), which expresses an infectious PVX that contains an additional coat protein (CP) promoter and a polylinker for the insertion and expression of foreign genes (51); a binary construct, PVX-P1-6×-HCPro, that was derived from vector pgR107, modified to express HCPro with a six-histidine tag at its N terminus (6×-HCPro) from a PVY P1-6×-HCPro-modified bicistron through an additional subgenomic RNA (Fig. 1B), as described previously (39); and a binary construct, PVX-P1-6×-HCPro mutB, that was obtained by amplification by PCR of the entire P1-6×-HCPro mutB bicistron from a pROK2-based binary construct, P1-6×-HCPro mutB (see below), with appropriate oligonucleotides and insertion of the amplified fragment after digestion into a ClaI- and SmaI-linearized pgR107 PVX vector. For transient expression in plants, a binary construct expressing a free GFP reporter was used (39), along with viral suppressor genes expressed from pROK2-based binary constructs expressing a modified Cucumber mosaic virus 2b suppressor (construct 2b-HA, where HA is human influenza virus hemagglutinin) (52); the p25 suppressor from PVX (construct pCAM-P25) (43); the modified PVY P1-6×-HCPro bicistron (construct P1-6×-HCPro) (39); as well as its mutant variants mutA, mutB, and mutC. Construct P1-6×-HCPro mutA contains the triplets for amino acids (aa) 236 to 238 of PVY HCPro (Arg-Lys-His; CGC-AAG-CAT) replaced with three alanine triplets (GCC-GCG-GCT). This mutant is the equivalent of mutant AS9 in Tobacco etch virus HCPro (5) and was obtained by amplifying two PCR fragments using appropriate oligonucleotides, ligating these fragments overnight, and performing PCR amplification of the ligated fragments. The fusion PCR fragment thus obtained was cloned after digestion into XhoI- and SacI-linearized construct P1-6×-HCPro. Construct P1-6×-HCPro mutB contains the triplets for aa 131 and 132 of PVY HCPro (Asn-Leu; AAT-TTA) replaced by two alanines (GCT-GCA). This mutant is the equivalent of mutant HCL134H in Plum pox virus HCPro (13). P1-6x-HCPro mutB was obtained by PCR amplification of a fragment containing this sequence modification using appropriate oligonucleotides and cloning after digestion into an XhoI- and SwaI-linearized construct, P1-6×-HCPro. Construct P1-6×-HCPro mutC contains a deletion of 20 aa between leucine 113 and asparagine 133 and a new glutamine residue (Leu113-Δ20 aa-Glu-Asn133). This deletion in mutC encompasses the motif of mutB and a Glu-Lys-Gln motif upstream. It was obtained by digestion of construct P1-6×-HCPro with XhoI and SwaI, filling of the XhoI 3′ recess with the Klenow fragment of DNA polymerase I, phosphorylation, and religation. All PCRs were performed by using Phusion DNA polymerase (Finnzymes, Finland), and mutant constructs were confirmed by sequencing.

Purification of HCPro from infected plants.

Our protocol for the purification of HCPro tagged with six histidines (6×-HCPro) is based on methods described previously (18), with some modifications: fully developed infected leaves were collected 7 days after agroinoculation with PVX-P1-6×-HCPro or with either the PVX-P1-6×-HCPro mutB vector or the empty PVX vector to obtain suppressor-deficient HCPro mutB or the empty controls, respectively. In all cases, 100 g of systemically infected tissue was used as the starting material. Every step in the purification of the HCPro and control samples was performed in parallel, at between 0°C and 4°C, with prechilled buffers and laboratory equipment. Leaves were homogenized in extraction buffer (3.3 ml of buffer/g of leaf tissue [100 mM Tris-HCl {pH 8.5}, 20 mM MgSO₄, 500 mM NaCl, 0.5 mM EGTA, and 20% sucrose supplemented with 0.2% Na₂SO₃, 0.1% polyvinylpyrrolidone 40, and 5 mM 2-mercaptoethanol]). After filtration, centrifugation, and ultracentrifugation, total soluble proteins were precipitated with 40% (NH₄)₂SO₄. Proteins were resuspended in extraction buffer and incubated with 1 ml of Ni^{-NTA agarose (Qiagen GmbH, Hilden, Germany). Several rinses of the resin were performed with extraction buffer, and after this, two consecutive steps of elution were done by using 4 ml of extraction buffer supplemented with 400 mM EGTA. Elution samples originating from plants infected with PVX, PVX-P1-6×-HCPro, or PVX-P1-6×-HCPro mutB constituted the purified preparations used in this work (here the control, HCPro, and HCPro mutB samples, respectively).}

Analysis of protein content during purification.

Samples from the different steps in the purification procedure were mixed 1:1 with 2× Laemmli buffer (100 mM Tris-HCl [pH 6.8], 4% SDS, 20% [vol/vol] glycerol, 0.01% bromophenol blue, 2% 2-mercaptoethanol), boiled, and fractionated by 10% SDS-PAGE. Gels were stained with a solution of 0.27% (wt/vol) Coomassie blue in 4:4:1 (vol/vol/vol) ethanol-water-glacial acetic acid to visualize protein bands. Quantification of the protein band content was done by densitometry analysis with Image J software (v. 1.48) (https://imagej.nih.gov/ij/). Silver staining of SDS-PAGE gels was performed as described previously (44). In addition, HCPro was also detected in samples by Western blot analysis: SDS-PAGE-resolved proteins were wet blotted onto Hybond-P polyvinylidene difluoride (PVDF) membranes (Amersham, GE Healthcare, Buckinghamshire, UK). Detection of HCPro was performed by using a mouse monoclonal antibody to PVY HCPro (antibody 1A11) (53) followed by commercial alkaline phosphatase-linked secondary antibodies and SigmaFast 5-bromo-4-chloro-3-indolylphosphate (BCIP)–nitroblue tetrazolium (NBT) substrate tablets (Sigma-Aldrich, St. Louis, MO, USA). Detection of the 2b protein of CMV was performed with a mouse monoclonal antibody to this protein (52).

RNA isolation and high-throughput sequencing.

To analyze sRNA populations, three separate deep sequencing events were performed. The first one analyzed the sRNAs present in plants infected with either PVX or PVX-P1-6×-HCPro, used as the input for purification. For this, total RNAs were extracted from ∼25 g of infected tissue in both cases. Thirty micrograms of total RNA was sent immediately for sequencing of sRNAs (18 to 30 nt). Total RNAs were isolated with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The second sequencing event analyzed short RNAs of 18 to >120 nt that were present in the purified control and HCPro samples from the first purification experiment. For this, 1.5-ml aliquots from the purified elution samples obtained from tissues infected with either PVX or PVX-P1-6×-HCPro were used to obtain 9.61 and 13.38 ng of gel-eluted RNAs from the control and HCPro samples, respectively. Total RNAs from the purified samples were fractioned by electrophoresis in 10% PAGE gels containing 8 M urea. The gel region resolving ∼12 nt to ∼500 nt was sliced with a razor, RNAs were eluted from the gel with 0.3 M NaCl, and the resulting elution samples were precipitated with 1 μl of glycogen (20 μg/μl) (Roche, Basel, Switzerland) and 1 volume of isopropanol. The third sequencing event analyzed sRNAs of 18 to 30 nt present in the Ni^{-NTA column-purified HCPro mutB and HCPro samples from the second purification experiment. For this, 1.5-ml aliquots from the purified elution samples obtained from tissues infected with either PVX-P1-6×-HCPro mutB or PVX-P1-6×-HCPro were used to obtain 597 and 339 ng of total RNAs for sequencing, respectively.}

RNA sequencing was performed by Fasteris SA (Chemin du Pont-du-Centenaire, Geneva, Switzerland). Briefly, RNA samples were subjected to 5′ cap removal with tobacco acid pyrophosphatase treatment and end-repair reactions, followed by single-stranded ligation of 3′- and 5′-indexed adapters. Afterwards, reverse transcription and PCR amplification were performed to generate the cDNA libraries that were multiplexed and sequenced in one lane of 50-bp single-end runs for the first and third deep sequencing events, or 125-bp single-end runs for the second deep-sequencing event, by using an Illumina HiSeq 2500 sequencer (Illumina, San Diego, CA, USA).

Analysis of RNA sequences.

Sequence reads from the RNA libraries with their adapters removed were computationally processed after discarding any insert reads of <18 nt, using our own Perl scripts. Reads were then mapped by using Bowtie (for reads between 18 and 49 nt) (54) or Bowtie2 (for reads of 50 nt and higher) (55), considering only perfect matches. Reads mapping to plant rRNAs, including mitochondrial and chloroplast ones from the NCBI database, were removed from the analysis. The remaining reads were then compared with (i) the genomes of the viral constructs, (ii) the N. benthamiana genome, and (iii) known miRNA sequences. The genomic sequence of PVX was that in the pgR107 binary construct (GenBank accession number {"type":"entrez-nucleotide","attrs":{"text":"AY297842.1","term_id":"31747248","term_text":"AY297842.1"}}AY297842.1), and those of PVX-P1-6×-HCPro and PVX-P1-6×-HCPro mutB were obtained by inserting the sequences of P1-6×-HCPro or P1-6×-HCPro mutB into the PVX vector polylinker and confirmed by Sanger sequencing. The genome sequence of N. benthamiana was obtained from the University of Sydney (http://sydney.edu.au/) (version 0.5) (56). For miRNAs, we used sequences in the Plant Noncoding RNA Database (PNRD) (57), including mature miRNA as well as pre-miRNA sequences. Sorting of sequences by length, counting of total and unique sequences, and determinations of percentages of nucleotide identities at the RNA 5′ ends were all achieved by employing different Perl scripts developed in our laboratory. To measure how many times an RNA type became enriched or depleted in the HCPro versus the control or HCPro mutB samples, we used fold change values. These values were obtained by dividing the values for the HCPro sample by the corresponding ones for the control sample. When the fold change value of the number of reads (x) was <1, the negative of its inverse (−1/x) was used in the charts. To calculate the contribution of 6×-HCPro to the sRNA population present in the purified HCPro samples, we used the mathematical formula shown in Fig. 5 (developed in Fig. S1 in the supplemental material).

Accession number(s).

High-throughput sequencing data from this article have been uploaded to the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo) with accession number {"type":"entrez-geo","attrs":{"text":"GSE71921","term_id":"71921"}}GSE71921.

Plants and viruses.

Purification of HCPro from infected plants.

Analysis of protein content during purification.

RNA isolation and high-throughput sequencing.

Analysis of RNA sequences.

Accession number(s).

University of Maryland, College Park;

^{Departamento de Biología Medioambiental, Centro de Investigaciones Biológicas, CSIC, Madrid, Spain}

^{Universidad Politécnica de Madrid, Centro de Biotecnología y Genómica de Plantas, Campus de Montegancedo, Madrid, Spain}

^{National Institute of Horticultural & Herbal Science, Agricultural Research Center for Climate Change, Wanju, Republic of Korea}

^{Corresponding author.}

Address correspondence to Francisco J. del Toro, se.cisc.bic@stdjf, or Tomás Canto, se.cisc.bic@otnac.samot.

Citation del Toro FJ, Donaire L, Aguilar E, Chung B-N, Tenllado F, Canto T. 2017. Potato virus Y HCPro suppression of antiviral silencing in Nicotiana benthamiana plants correlates with its ability to bind in vivo to 21- and 22-nucleotide small RNAs of viral sequence. J Virol 91:e00367-17. https://doi.org/10.1128/JVI.00367-17.

Received 2017 Mar 7; Accepted 2017 Mar 30.

ABSTRACT

We have investigated short and small RNAs (sRNAs) that were bound to a biologically active hexahistidine-tagged Potato virus Y (PVY) HCPro suppressor of silencing, expressed from a heterologous virus vector in Nicotiana benthamiana plants, and purified under nondenaturing conditions. We found that RNAs in purified preparations were differentially enriched in 21-nucleotide (nt) and, to a much lesser extent, 22-nt sRNAs of viral sequences (viral sRNAs [vsRNAs]) compared to those found in a control plant protein background bound to nickel resin in the absence of HCPro or in a purified HCPro alanine substitution mutant (HCPro mutB) control that lacked suppressor-of-silencing activity. In both controls, sRNAs were composed almost entirely of molecules of plant sequence, indicating that the resin-bound protein background had no affinity for vsRNAs and also that HCPro mutB failed to bind to vsRNAs. Therefore, PVY HCPro suppressor activity correlated with its ability to bind to 21- and 22-nt vsRNAs. HCPro constituted at least 54% of the total protein content in purified preparations, and we were able to calculate its contribution to the 21- and the 22-nt pools of sRNAs present in the purified samples and its binding strength relative to the background. We also found that in the 21-nt vsRNAs of the HCPro preparation, 5′-terminal adenines were overrepresented relative to the controls, but this was not observed in vsRNAs of other sizes or of plant sequences.

IMPORTANCE It was previously shown that HCPro can bind to long RNAs and small RNAs (sRNAs) in vitro and, in the case of Turnip mosaic virus HCPro, also in vivo in arabidopsis AGO2-deficient plants. Our data show that PVY HCPro binds in vivo to sRNAs during infection in wild-type Nicotiana benthamiana plants when expressed from a heterologous virus vector. Using a suppression-of-silencing-deficient HCPro mutant that can accumulate in this host when expressed from a virus vector, we also show that sRNA binding correlates with silencing suppression activity. We demonstrate that HCPro binds at least to sRNAs with viral sequences of 21 nucleotides (nt) and, to a much lesser extent, of 22 nt, which were are also differentially enriched in 5′-end adenines relative to the purified controls. Together, our results support the physical binding of HCPro to vsRNAs of 21 and 22 nt as a means to interfere with antiviral silencing.

KEYWORDS: HCPro, sRNAs, suppressor of silencing, potyvirus, antiviral silencing

ABSTRACT

Click here to view.

ACKNOWLEDGMENTS

We thank Peter Palukaitis for his critical reading of the manuscript.

E.A. is the recipient of an FPU fellowship from the former Spanish Ministry of Education and Sport. This work was supported by the Spanish Ministry of Economy and Competitiveness (grant references BIO2013-47940-R and BIO2016-75619-R [AEI/FEDER, European Union]) and by the Rural Development Administration (RDA) of the Republic of Korea (grant reference PJ00946102).

ACKNOWLEDGMENTS

Footnotes

Supplemental material for this article may be found at https://doi.org/10.1128/JVI.00367-17.

Footnotes

REFERENCES

References

1. Pirone TP, Blanc S. 1996. Helper-dependent vector transmission of plant viruses. Annu Rev Phytopathol34:227–247. doi:10.1146/annurev.phyto.34.1.227. [] [[PubMed]
2. Carrington JC, Herndon KL. 1992. Characterization of the potyviral HCPro autoproteolytic cleavage site. Virology187:308–315. doi:10.1016/0042-6822(92)90319-K. ] [
3. Pruss G, Ge X, Shi XM, Carrington JC, Vance VB. 1997. Plant viral synergism: the potyviral genome encodes a broad-range pathogenicity enhancer that transactivates replication of heterologous viruses. Plant Cell9:859–868. doi:10.1105/tpc.9.6.859. ] [
4. Anandalakshmi R, Pruss GJ, Ge X, Marathe R, Mallory AC, Smith TH, Vance VB. 1998. A viral suppressor of gene silencing in plants. Proc Natl Acad Sci U S A95:13079–13084. doi:10.1073/pnas.95.22.13079. ] [
5. Kasschau K, Carrington JC. 1998. A counterdefensive strategy of plant viruses: suppression of posttranscriptional gene silencing. Cell95:461–470. doi:10.1016/S0092-8674(00)81614-1. [] [[PubMed]
6. Burgyán J, Havelda Z. 2011. Viral suppressors of RNA silencing. Trends Plant Sci16:265–272. doi:10.1016/j.tplants.2011.02.010. [] [[PubMed]
7. Ivanov KI, Eskelin K, Bašić M, Swarnalok D, Löhmus A, Varjosalo M, Mäkinen K. 2016. Molecular insights into the function of the viral RNA silencing suppressor HCPro. Plant J85:30–45. doi:10.1111/tpj.13088. [] [[PubMed]
8. Valli A, Gallo A, Calvo M, de Jesús Pérez J, García JA. 2014. A novel role of the potyviral helper component proteinase contributes to enhance the yield of viral particles. J Virol88:9808–9818. doi:10.1128/JVI.01010-14. ] [
9. Canto T, López-Moya JJ, Serra-Yoldi MT, Díaz-Ruíz JR, López-Abella D. 1995. Different helper component mutations associated with lack of aphid transmissibility in two isolates of Potato virus Y. Phytopathology85:1519–1524. doi:10.1094/Phyto-85-1519. [[PubMed]
10. Blanc S, López-Moya JJ, Wang R, García-Lampasona S, Thornbury DW, Pirone TP. 1997. A specific interaction between coat protein and helper component correlates with aphid transmission of a potyvirus. Virology231:141–147. doi:10.1006/viro.1997.8521. [] [[PubMed]
11. Jin Y, Ma D, Dong J, Jin J, Li D, Deng C, Jin J, Wang T. 2007. HCPro of Potato virus Y can interact with three Arabidopsis 20S proteasome subunits in planta. J Virol81:12881–12888. doi:10.1128/JVI.00913-07. ] [
12. Sahana N, Kaur H, Palukaitis P, Canto T, Praveen S. 2014. The asparagine residue in the FRNK box of potyviral helper-component protease is critical for template function and subcellular localization. J Gen Virol95:1167–1177. doi:10.1099/vir.0.060269-0. [] [[PubMed]
13. González-Jara P, Atencio FA, Martínez-García B, Barajas D, Tenllado F, Díaz-Ruíz JR. 2005. A single amino acid mutation in the plum pox virus helper component-proteinase gene abolishes both synergistic and RNA silencing suppression activities. Phytopathology95:894–901. doi:10.1094/PHYTO-95-0894. [] [[PubMed]
14. Kasschau K, Carrington JC. 2001. Long-distance movement and replication maintenance functions correlate with silencing suppression activity of potyviral HC-Pro. Virology285:71–81. doi:10.1006/viro.2001.0901. [] [[PubMed]
15. Shiboleth YM, Haronsky E, Leibman D, Arazi T, Wassenegger M, Whitham SA, Gaba V, Gal-On A. 2007. The conserved FRNK box in HC-Pro, a plant viral suppressor of gene silencing, is required for small RNA binding and mediates symptom development. J Virol81:13135–13148. doi:10.1128/JVI.01031-07. ] [
16. Torres-Barceló C, Martín S, Darós J-A, Elena SF. 2008. From hypo- to hypersuppression: effect of amino acid substitutions on the RNA-silencing suppressor activity of the tobacco etch potyvirus HC-Pro. Genetics180:1039–1049. doi:10.1534/genetics.108.091363. ] [
17. Plisson C, Drucker M, Blanc S, German-Retama S, Le Gall O, Thomas D, Bron P. 2003. Structural characterization of HC-Pro, a plant virus multifunctional protein. J Biol Chem278:23753–23761. doi:10.1074/jbc.M302512200. [] [[PubMed]
18. Ruíz-Ferrer V, Boskovic J, Alfonso C, Rivas G, Llorca O, López-Abella D, López-Moya JJ. 2005. Structural analysis of tobacco etch potyvirus HC-Pro oligomers involved in aphid transmission. J Virol79:3758–3765. doi:10.1128/JVI.79.6.3758-3765.2005. ] [
19. Zheng H, Yan F, Lu Y, Sun L, Lin L, Cai L, Hou M, Chen J. 2011. Mapping the self-interacting domains of TuMV HC-Pro and the subcellular localization of the protein. Virus Genes42:110–116. doi:10.1007/s11262-010-0538-8. [] [[PubMed]
20. Haikonen MT, Rajamaki ML, Valkonen J. 2013. Interaction of the microtubule associated host protein HIP2 with the viral helper component proteinase is important in infection with potato virus A. Mol Plant Microbe Interact26:734–744. doi:10.1094/MPMI-01-13-0023-R. [] [[PubMed]
21. Del Toro F, Tena F, Tilsner J, Wright K, Tenllado F, Chung B-N, Praveen S, Canto T. 2014. Potato virus Y HCPro localization at distinct, dynamically-related and environment-influenced structures in the cell cytoplasm. Mol Plant Microbe Interact27:1331–1343. doi:10.1094/MPMI-05-14-0155-R. [] [[PubMed]
22. Anandalakshmi R, Marathe R, Ge X, Herr JM Jr, Mau C, Mallory A, Pruss G, Bowman L, Vance VB. 2000. A calmodulin-related protein that suppresses posttranscriptional gene silencing in plants. Science290:142–144. doi:10.1126/science.290.5489.142. [] [[PubMed]
23. Nakahara KS, Masuta C. 2014. Interaction between viral RNA silencing suppressors and host factors in plant immunity. Curr Opin Plant Biol20:88–95. doi:10.1016/j.pbi.2014.05.004. [] [[PubMed]
24. Endres MW, Gregory BD, Gao Z, Foreman AW, Mlotshwa S, Ge X, Pruss GJ, Ecker JR, Bowman LH, Vance VB. 2010. Two plant viral suppressors of silencing require the ethylene-inducible host transcription factor RAV2 to block RNA silencing. PLoS Pathog6:e1000729. doi:10.1371/journal.ppat.1000729. ] [
25. Ballut L, Drucker M, Pugniére M, Cambon F, Blanc S, Roquet F, Candresse T, Schmid H-P, Nicolas P, Le Gall O, Badaoui S. 2005. HcPro, a multifunctional protein encoded by a plant RNA virus, targets the 20S proteasome and affects its enzymatic activities. J Gen Virol88:2595–2603. doi:10.1099/vir.0.81107-0. [] [[PubMed]
26. Jin Y, Ma D, Dong J, Li D, Deng C, Jin J, Wang T. 2007. The HC-Pro protein of potato virus Y interacts with NtMinD of tobacco. Mol Plant Microbe Interact20:1505–1511. doi:10.1094/MPMI-20-12-1505. [] [[PubMed]
27. Ala-Poikela M, Goytia E, Halkonen T, Rajamaki ML, Valkonen JPT. 2011. Helper component proteinase of the genus Potyvirus is an interaction partner of translation initiation factors sIF(iso)4E and eIF4E and contains a 4E binding motif. J Virol85:6784–6794. doi:10.1128/JVI.00485-11. ] [
28. Cheng Y-Q, Liu Z-M, Xu J, Zhou T, Wang M, Chen Y-T, Li H-F, Fan Z-F. 2008. HC-Pro protein of sugar cane mosaic virus interacts specifically with maize ferredoxin-5 in vitro and in planta. J Gen Virol89:2046–2054. doi:10.1099/vir.0.2008/001271-0. [] [[PubMed]
29. Jamous RM, Boonrod K, Fuellgrabe MW, Ali-Shetayen MS, Krczai G, Wassenegger M. 2011. The helper component-proteinase of the zucchini yellow mosaic virus inhibits the Hua enhancer 1 methyltransferase activity in vitro. J Gen Virol92:2222–2226. doi:10.1099/vir.0.031534-0. [] [[PubMed]
30. Maia IG, Bernardi F. 1996. Nucleic acid-binding properties of a bacterially expressed potato virus Y helper component-proteinase. J Gen Virol77:869–877. doi:10.1099/0022-1317-77-5-869. [] [[PubMed]
31. Urcuqui-Inchima S, Maia IG, Arruda P, Haenni AL, Bernardi F. 2000. Deletion mapping of the potyviral helper component-proteinase reveals two regions involved in RNA binding. Virology268:104–111. doi:10.1006/viro.1999.0156. [] [[PubMed]
32. Mérai Z, Kerényi Z, Kertész S, Magda M, Lakatos L, Silhavy D. 2006. Double-stranded RNA binding may be a general plant RNA viral strategy to suppress silencing. J Virol80:5747–5756. doi:10.1128/JVI.01963-05. ] [
33. Mangrauthia SK, Singh-Shakya VP, Jain RK, Praveen S. 2009. Ambient temperature perception in papaya for papaya ringspot virus interaction. Virus Genes38:429–434. doi:10.1007/s11262-009-0336-3. [] [[PubMed]
34. Lakatos L, Csorba T, Pantaleo V, Chapman EJ, Carrington JC, Liu YP, Dolja VV, Calvino LF, López-Moya JJ, Burgyán J. 2006. Small RNA binding is a common strategy to suppress RNA silencing by several viral suppressors. EMBO J21:2768–2780. doi:10.1038/sj.emboj.7601164. ] [
35. Vargasson JM, Szittya G, Burgyan J, Hall TM. 2003. Size selective recognition of siRNAs by an RNA silencing suppressor. Cell115:799–811. doi:10.1016/S0092-8674(03)00984-X. [] [[PubMed]
36. González I, Rakitina D, Semashko M, Taliansky M, Praveen S, Palukaitis P, Carr J, Kalinina N, Canto T. 2012. RNA binding is more critical to the suppression of silencing function of cucumber mosaic virus 2b protein than nuclear localization. RNA18:771–782. doi:10.1261/rna.031260.111. ] [
37. García-Ruíz H, Carbonell A, Hoyer JS, Fahlgren N, Gilbert KB, Takeda A, Giampetruzzi A, García-Ruíz MT, McGinn MG, Lowery N, Martinez-Baladejo MT, Carrington JC. 2015. Roles and programming of Arabidopsis ARGONAUTE proteins during turnip mosaic virus infection. PLoS Pathog11:e1004755. doi:10.1371/journal.ppat.1004755. ] [
38. Chapman EJ, Prokhnevsky AI, Gopinath K, Dolja VV, Carrington J. 2004. Viral RNA silencing suppressors inhibit the microRNA pathway at an intermediate step. Genes Dev18:1179–1186. doi:10.1101/gad.1201204. ] [
39. Tena F, González I, Doblas P, Rodríguez C, Sahana N, Kaur H, Tenllado F, Praveen S, Canto T. 2013. The influence of cis-acting P1 protein and translational elements on the expression of potato virus Y HCPro in heterologous systems and its suppression of silencing activity. Mol Plant Pathol14:530–541. doi:10.1111/mpp.12025. ] [
40. Zuker M. 2003 Mfold Web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res31:3406–3415. doi:10.1093/nar/gkg595. ] [[Google Scholar]
41. Ebhardt HA, Thi EP, Wang M-B, Unrau PJ. 2005. Extensive 3′ modification of plant small RNAs is modulated by helper component-proteinase expression. Proc Natl Acad Sci U S A102:13398–13403. doi:10.1073/pnas.0506597102. ] [
42. Del Toro FJ, Aguilar E, Hernández-Walias FJ, Tenllado F, Chung B-N, Canto T. 2015. High temperature, high ambient CO2 affect the interactions between three positive-sense RNA viruses and a compatible host differentially, but not their silencing suppression efficiency. PLOs One10:e136062. doi:10.1371/journal.pone.0136062. ] [
43. Aguilar E, Almendral D, Allende L, Pacheco R, Chung BN, Canto T, Tenllado F. 2015. The P25 protein of potato virus X (PVX) is the main pathogenicity determinant for systemic necrosis in PVX-associated synergisms. J Virol89:2090–2103. doi:10.1128/JVI.02896-14. ] [
44. Gal-On A, Canto T, Palukaitis P. 2000. Characterisation of genetically modified cucumber mosaic virus expressing histidine-tagged 1a and 2a proteins. Arch Virol145:37–50. doi:10.1007/s007050050003. [] [[PubMed]
45. Takeda A, Iwasaki S, Watanabe T, Utsumi M, Watanabe Y. 2008. The mechanism selecting the guide strand from small RNA duplexes is different among argonaute proteins. Plant Cell Physiol49:493–500. doi:10.1093/pcp/pcn043. [] [[PubMed]
46. Montgomery TA, Howell MD, Cuperus JT, Li D, Hansen JE, Alexander AL, Chapman EJ, Fahlgren N, Allen E, Carrington JC. 2008. Specificity of ARGONAUTE7-miR390 interaction and dual functionality in TAS3 trans-acting siRNA formation. Cell133:128–141. doi:10.1016/j.cell.2008.02.033. [] [[PubMed]
47. Mi S, Cai T, Hu Y, Chen Y, Hodges E, Ni F, Wu L, Li S, Zhou H, Long C, Chen S, Hannon GJ, Qi Y. 2008. Sorting of small RNAs into Arabidopsis argonaute complexes is directed by the 5′ terminal nucleotide. Cell133:116–127. doi:10.1016/j.cell.2008.02.034. ] [
48. Harvey JJ, Lewsey MG, Patel K, Westwood J, Heimstädt S, Carr JP, Baulcombe DC. 2011. An antiviral defense role of AGO2 in plants. PLoS One6:e14639. doi:10.1371/journal.pone.0014639. ] [
49. Carbonell A, Fahlgren N, Garcia-Ruiz H, Gilbert KB, Montgomery TA, Nguyen T, Cuperus JT, Carrington JC. 2012. Functional analysis of three Arabidopsis ARGONAUTES using slicer-defective mutants. Plant Cell24:3613–3629. doi:10.1105/tpc.112.099945. ] [
50. Hamera S, Song X, Su L, Chen X, Fang R. 2012. Cucumber mosaic virus suppressor 2b binds to AGO4-related small RNAs and impairs AGO4 activities. Plant J69:104–115. doi:10.1111/j.1365-313X.2011.04774.x. [] [[PubMed]
51. Lu R, Malcuit I, Moffett P, Ruíz MT, Peart J, Wu AJ, Rathjen JP, Bendahmane A, Day L, Baulcombe DC. 2003. High throughput virus-induced gene silencing implicates heat shock protein 90 in plant disease resistance. EMBO J22:5690–5699. doi:10.1093/emboj/cdg546. ] [
52. González I, Martínez L, Rakitina DV, Lewsey MG, Atienzo FA, Llave C, Kalinina NO, Carr JP, Palukaitis P, Canto T. 2010. Cucumber mosaic virus 2b protein subcellular targets and interactions: their significance to its RNA silencing suppressor activity. Mol Plant Microbe Interact23:294–303. doi:10.1094/MPMI-23-3-0294. [] [[PubMed]
53. Canto T, Ellis JP, Bowler G, López-Abella D. 1995. Production of monoclonal antibodies to potato virus Y helper component protease and their use for strain differentiation. Plant Dis79:234–237. doi:10.1094/PD-79-0234. [[PubMed]
54. Langmead B, Trapnell C, Pop M, Salzberg SL. 2009. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol10:R25. doi:10.1186/gb-2009-10-3-r25. ] [
55. Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat Methods9:357–359. doi:10.1038/nmeth.1923. ] [
56. Naim F, Nakasugi K, Crowhurst RN, Hilario E, Zwart AB, Hellens RP, Taylor JM, Waterhouse PM. 2012. Advanced engineering of lipid metabolism in Nicotiana benthamiana using a draft genome and the V2 viral silencing-suppressor protein. PLoS One7:e52717. doi:10.1371/journal.pone.0052717. ] [
57. Yi X, Zhang Z, Ling Y, Xu W, Su Z. 2015. PNRD: a plant non-coding RNA database. Nucleic Acids Res43:D982–D989. doi:10.1093/nar/gku1162. ] [

<em class="genus-species">Potato Virus Y</em> HCPro Suppression of Antiviral Silencing in <em class="genus-species">Nicotiana benthamiana</em> Plants Correlates with Its Ability To Bind <em>In Vivo</em> to 21- and 22-Nucleotide Small RNAs of Viral Sequence

INTRODUCTION

RESULTS

Characterization of the silencing suppression activities and steady-state levels of accumulation of PVY HCPro mutants in N. benthamiana leaves.

Purification of HCPro, HCPro mutB, and control samples under nondenaturing conditions and isolation of associated RNAs.

Characterization of the sRNA populations present in input plants.

TABLE 1

Characterization of the sRNAs associated with purified HCPro and control samples.

TABLE 2

Characterization of short RNAs of >30 nt that potentially bound to HCPro.

Characterization of the silencing suppression activities and steady-state levels of accumulation of PVY HCPro mutants in N. benthamiana leaves.

Purification of HCPro, HCPro mutB, and control samples under nondenaturing conditions and isolation of associated RNAs.

Characterization of the sRNA populations present in input plants.

TABLE 1

Characterization of the sRNAs associated with purified HCPro and control samples.

TABLE 2

Characterization of short RNAs of >30 nt that potentially bound to HCPro.

DISCUSSION

The profiles of sRNAs in plants infected with PVX and those infected with PVX expressing PVY HCPro are very similar.

Preparations purified by using Ni-NTA resin under nondenaturing conditions also contain a protein background that binds to specific sRNAs.

During infection, PVY HCPro binds in vivo to vsRNAs of 21 nt and, to a much lesser extent, of 22 nt, and the binding ability correlates with suppressor activity.

The profiles of sRNAs in plants infected with PVX and those infected with PVX expressing PVY HCPro are very similar.

Preparations purified by using Ni-NTA resin under nondenaturing conditions also contain a protein background that binds to specific sRNAs.

During infection, PVY HCPro binds in vivo to vsRNAs of 21 nt and, to a much lesser extent, of 22 nt, and the binding ability correlates with suppressor activity.

MATERIALS AND METHODS

Plants and viruses.

Purification of HCPro from infected plants.

Analysis of protein content during purification.

RNA isolation and high-throughput sequencing.

Analysis of RNA sequences.

Accession number(s).

Plants and viruses.

Purification of HCPro from infected plants.

Analysis of protein content during purification.

RNA isolation and high-throughput sequencing.

Analysis of RNA sequences.

Accession number(s).

ABSTRACT

ACKNOWLEDGMENTS

Footnotes

REFERENCES

References