(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Combinatorial interactions between viral proteins expand the potential functional landscape of the tomato yellow leaf curl virus proteome [1]

['Liping Wang', 'Shanghai Center For Plant Stress Biology', 'Center For Excellence In Molecular Plant Sciences', 'Chinese Academy Of Sciences', 'Shanghai', 'University Of The Chinese Academy Of Sciences', 'Beijing', 'Huang Tan', 'Department Of Plant Biochemistry', 'Center For Plant Molecular Biology']

Date: 2022-11 

As shown in Fig 2E and 2F , only certain pairwise viral protein-protein interactions could be detected by all used approaches (Rep-Rep, Rep-C3, C2-C4, C3-C4, C3-V2, and V2-V2). The differences in the detected outcomes when using different techniques might be due to specific requirements of each of the assays (e.g. strength or stability of the interaction necessary for this to be detected), or to the effect of the tags used (nature and position) on the interaction.

A summary of all detected pairwise interactions between viral proteins is shown in Fig 2E and 2F ; all viral proteins were found to interact with one another, including self-interactions, by at least two independent methods. Importantly, some of these interactions could also be detected in unbiased affinity purification followed by mass spectrometry (AP-MS) experiments with C-terminal GFP-tagged versions of the viral proteins expressed in infected N. benthamiana cells [ 40 ] when these datasets were re-analyzed to search for the viral proteins, indicating that viral proteins physically associate with one another in the context of the infection in their native state ( S1 Table ).

(A) Viral protein-protein interactions detected in yeast two-hybrid. The minimal synthetic defined (SD) medium without leucine (Leu), tryptophan (Trp), histidine (His), and adenine (Ade) was used to select positive interactions; SD without Leu and Trp was used to select co-transformants ( S1 Fig ). The interaction between the SV40 large T antigen (T) and the murine tumor suppressor p53 is a positive control. AD: activation domain; BD: binding domain. This experiment was repeated twice with similar results. (B) Summary of viral protein-protein interactions detected by co-immunoprecipitation (co-IP) in the absence (-) or presence (+) of TYLCV. These experiments were repeated at least three times; the colour scale represents the percentage of positive interaction results among all replicates, with 1 = 100%. The original co-IP blots are shown in S2 Fig (in the absence of TYLCV) and S3 Fig (in the presence of TYLCV). An interaction between two viral proteins was considered as positive if at least two replicates showed positive interactions either in the absence or presence of TYLCV. (C) Viral protein-protein interactions detected by bimolecular fluorescence complementation (BiFC) in N. benthamiana leaves. nYFP: N-terminal half of the YFP; cYFP: C-terminal half of the YFP. Images were taken at 2 days post-infiltration (dpi). Scale bar = 10 μm. This experiment was repeated at least four times; combination with Hoechst staining and negative controls can be found in S4 Fig . Additional images are shown in S5 Fig . (D) Viral protein-protein interactions detected by split-luciferase assay in N. benthamiana leaves. nLuc: N-terminal part of the luciferase protein; cLuc: C-terminal part of the luciferase protein. Images were taken at 2 dpi. The colour scale represents the intensity of the interaction in counts per second (CPS). This experiment was repeated three times with similar results. (E) Summary of the intra-viral protein-protein interactions identified in (A-D) . Different colours represent different methods, as indicated; circle size indicates the number of the methods in which a positive interaction was detected. (F) Network of intra-viral protein-protein interactions. The colored lines indicate the positive interactions detected by Y2H, Co-IP, BiFC, split-luciferase assay, or AP-MS. See also S1 – S5 Figs; S1 Table .

In order to test whether virus (TYLCV)-encoded proteins associate with one another, we employed a number of protein-protein interaction methods, namely yeast two-hybrid (Y2H), and in planta co-immunoprecipitation (co-IP), bimolecular fluorescence complementation (BiFC), and split-luciferase assays. Several viral protein-protein interactions were identified in yeast (Figs 2A and S1 ) (Rep-Rep, Rep-C3, C2-C3, C2-C4, C3-C3, C3-C4, C3-CP, C3-V2, C4-C3, V2-V2); of note, the C2-C2 self-interaction could not be evaluated, since full-length C2 fused to the GAL4 binding domain displays auto-activation ( Fig 2A ), as previously described for other geminiviral C2 proteins [ 36 – 37 ]. Next, pairwise interactions between viral proteins were tested by co-IP assays following transient expression of C-terminally tagged versions of the viral proteins in N. benthamiana. The number of associations between viral proteins found in co-IP, which detects both direct and indirect interactions, was higher (Figs 2B , S2 and S3 ). The viral infection reshapes the cell environment where the viral proteins coexist: the presence/absence of host proteins, the activation of post-translational modification pathways, or the presence of additional viral proteins might influence the outcome of the tested interactions. Therefore, co-IP assays were performed both in the presence and absence of the virus. Most reproducible interactions detected in the absence of the virus were maintained in the context of the infection (Rep-Rep, C2-C2, C4-C2, C4-C3, C4-C4, C4-V2, CP-C2, V2-C2, V2-V2), and additional interactions were detected in an infection-dependent manner (Rep-C2, Rep-C3, Rep-C4, Rep-CP, Rep-V2, CP-Rep, C4-CP, CP-CP, CP-V2, V2-C3). Viral protein-protein interactions were further tested in planta by BiFC and split-luciferase assays ( Fig 2C and 2D ). In both assays, the viral proteins were fused to one half of the protein to be reconstituted upon a positive interaction (nYFP and cYFP for YFP, or nLuc and cLuc for luciferase, respectively) and transiently expressed in N. benthamiana leaves; for the BiFC experiments, n-YFP and c-YFP were fused to the C-terminus of the viral proteins, while for split-luciferase experiments nLuc was fused to the C-terminus and cLuc to the N-terminus of the viral protein. Interestingly, BiFC indicates that most of the detected interactions occur in the nucleus, with different distribution patterns, including localization in the nucleoplasm (e.g. Rep-C2), nucleolus (e.g. C3-C3), or nuclear speckles (e.g. Rep-C3) (Figs 2C and S4 ; additional patterns of interactions observed by BiFC can be found in S5 Fig ). This nuclear prevalence of viral protein-protein interactions correlates with the nucleus hosting most of the viral cycle, including replication, transcription of viral genes, and encapsidation [ 38 ]. One notable exception is the interaction between C4 and V2, which takes place in intracellular punctate structures outside of the nucleus; this localization may be linked to the proposed role of these proteins in viral movement [ 39 ]. All viral proteins were shown to interact with one another (including self-interactions) at least in one direction by BiFC ( Fig 2C ). Similarly, all viral proteins displayed a positive interaction with each of the viral proteins by split-luciferase assays, with three exceptions (Rep-CP, C4-CP, and V2-CP) ( Fig 2D ). The identification of positive interactions in these reconstitution assays that were not detected as such by co-IP could be explained by the potential weak or transient nature of these associations, which could be overcome by artificial stabilization provided by the complementation of the split reporter.

Since, in the absence of the virus, C2-GFP appears evenly distributed in the nucleoplasm and is excluded from the nucleolus, but it gains strong nucleolar accumulation in the presence of the virus, we reasoned that C2 might perform additional functions in the context of the infection, and decided to use the C2-CP interaction as a proof-of-concept for the idea that viral proteins might have combinatorial functions.

Using transient co-expression of C2-GFP and each viral protein fused to RFP at their C-terminus in N. benthamiana leaves, we could determine that CP is sufficient to enable a strong accumulation of C2 in the nucleolus, an effect that can also be triggered by an untagged version of CP (Figs 3B and 3C ; S6 ). Of note, C2 and CP have been shown to interact in the nucleolus ( Fig 2C ). Curiously, only C2-GFP, but not GFP-C2, re-localizes to the nucleolus when in the presence of CP, likely due to a positional effect of the GFP tag ( Fig 3C ). Although full functionality of C2-GFP or GFP-C2 during the viral infection has not been demonstrated, C-terminal GFP fusions have been used in functional studies with other geminiviral C2 proteins (e.g. [ 42 – 45 ]). Local infection assays with two TYLCV mutant viruses unable to express CP, TYLCV-CPmut1 and TYLCV-CPmut2, in which early stop codons are introduced and alternative transcriptional initiation sites have been removed (for details, see Materials and Methods ), demonstrate that the presence of CP is not only sufficient, but also required for the re-localization of C2-GFP into the nucleolus in infected cells ( Fig 3D ). These mutants accumulate to wild type-like levels in the transiently transformed leaves ( S6B Fig ).

(A) Subcellular localization of the TYLCV-encoded proteins fused to GFP at their C-terminus expressed alone (+EV; co-transformed with an empty vector control) or in the context of the viral infection (+TYLCV; co-transformed with a TYLCV infectious clone) in N. benthamiana leaves at 2 days post infiltration (dpi). Scale bar = 10 μm. EV: empty vector. (B) Subcellular localization of C2-GFP co-expressed with each of the viral proteins fused to RFP in N. benthamiana leaves at 2 dpi. Scale bar = 10 μm. AF: Autofluorescence. (C) Subcellular localization of C2-GFP or GFP-C2 when expressed alone (+EV) or co-expressed with CP (+CP) in N. benthamiana leaves at 2 dpi. The accumulation of the CP transcript is shown in S6A Fig . Scale bar = 10 μm. EV: empty vector. (D) Subcellular localization of C2-GFP when expressed alone (+EV) or in the context of the infection by the WT TYLCV virus (+TYLCV) or mutated versions unable to produce CP (+TYLCV-CPmut1; +TYLCV-CPmut2) in N. benthamiana leaves at 2 dpi. Scale bar = 10 μm. EV: empty vector. Viral accumulation is shown in S6B Fig . For details on TYLCV-CPmut1 and TYLCV-CPmut2, see Materials and Methods . In ( B - D ), the dashed circles mark the nucleolus. See also S6 Fig .

Although the proteins encoded by TYLCV display specific localizations in the plant cell, all of them, with the exception of C4 (at the plasma membrane and weakly in chloroplasts), can be clear and consistently found in the nucleus (nucleoplasm and/or subnuclear compartments) in basal conditions when expressed alone fused to GFP (Rep: nucleoplasm; C2: nucleoplasm; C3: nucleoplasm, nucleolus, and nuclear speckles; CP: nucleolus and weakly in the nucleoplasm; V2: Cajal body and weakly in the nucleoplasm–in addition to endoplasmic reticulum) ( Fig 3A ). Interestingly, in the presence of the virus, several viral proteins fused to GFP, namely C2, C3, C4, and CP, experienced noticeable changes in their subcellular distribution ( Fig 3A ): C2-GFP, which is excluded from the nucleolus in the absence of the virus, accumulates in this compartment in infected cells; C3-GFP, on the contrary, is excluded from the nucleolus in the presence of the virus; C4-GFP is depleted from the plasma membrane and accumulates in chloroplasts; and CP-GFP re-localizes from the nucleolus to the nucleoplasm, where it accumulates in unidentified structures. These changes had been previously reported for C4-GFP and CP-GFP; while in the case of C4-GFP, Rep alone can trigger its re-localization from the plasma membrane to chloroplasts [ 41 ], no individual protein was sufficient to modify the subnuclear pattern of CP [ 31 ].

The C2/CP module specifically reshapes the host transcriptome

With the purpose of assessing if the functional landscape of C2 might be expanded when in the presence of CP, and considering that the C2 protein from geminiviruses has been previously described to impact host gene expression [46–51], we decided to investigate the transcriptional changes triggered by C2 in the presence or absence of CP as a readout for the activity of the former. To this aim, we transiently expressed C2, CP, or C2+CP in N. benthamiana leaves and determined the resulting changes in the plant transcriptome by RNA sequencing (RNA-seq). As shown in Fig 4A, C2 alone caused the differential expression of 211 genes (139 up-regulated, 71 down-regulated), while expression of CP did not significantly affect the plant transcriptional landscape; simultaneous expression of C2 and CP resulted in a moderate increase in the number of differentially expressed genes (DEGs) to 263 (72 up-regulated, 191 down-regulated) (Figs 4A, S7A and S7B; S2 Table; validation of the RNA-seq results by RT-qPCR is presented in S7C Fig.). Strikingly, however, the identity and behavior of DEGs was dramatically changed by the presence of CP (Fig 4B and 4C), indicating that C2 and CP have a synergistic effect on the host transcriptome. Functional enrichment analysis unveiled that addition of CP indeed shifted the functional gene ontology (GO) categories transcriptionally reprogrammed by C2, and that certain categories appear as statistically over-represented in the subset of down-regulated genes only when both viral proteins are simultaneously expressed (Fig 4D and 4E; S3 Table).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. C2 and CP functionally interact in planta and modify the transcriptome of N. benthamiana in an interdependent manner. (A) Number of differentially expressed genes (DEGs) upon expression of C2, CP, or C2+CP in N. benthamiana leaves. UR: up-regulated; DR: down-regulated; ND: not detected; EV: empty vector. Full lists can be found in S2 Table. (B) Venn diagram of DEGs upon expression of C2 or C2+CP in N. benthamiana. UR: up-regulated; DR: down-regulated; EV: empty vector. (C) Heatmap with hierarchical clustering from samples in (A). The colour scale indicates the Z-score. EV: empty vector. (D) Functional enrichment analysis of up-regulated (UR) or down-regulated (DR) genes in the indicated samples. Gene Ontology (GO) categories from the Biological Process ontology enriched with a p-value<0.01 (up to top 10) are shown; functional enrichment was performed using the orthologues in Arabidopsis thaliana. “C2+CP vs. EV (only)” denotes the subset of genes that are down-regulated in this sample only, and not in the samples expressing the viral proteins separately. The colour scale indicates the -log10 (p-value), showing the significance of GO term enrichment. EV: empty vector. For a full list, see S3 Table. (E) Venn diagram of the GO terms (Biological Process ontology) over-represented in the subsets of down-regulated genes (p-value<0.01) in the different samples. DR: down-regulated; EV: empty vector. For a full list, see S3 Table. See also S7 Fig; S2 and S3 Tables. https://doi.org/10.1371/journal.ppat.1010909.g004

To investigate the relevance of the re-localization of C2 to the nucleolus (Fig 3B and 3C) for this effect, we selected DEGs specifically affected by the co-expression of C2 and CP, and tested the ability of C2-GFP (which re-localizes to the nucleolus in the presence of CP) or GFP-C2 (which does not re-localize to the nucleolus in the presence of CP) to affect their transcript accumulation when transiently co-expressed with CP in N. benthamiana leaves, as measured by RT-qPCR. As shown in Fig 5, only C2+CP and C2-GFP+CP, but not GFP-C2+CP, affect the expression of the selected genes compared with C2, C2-GFP, or GFP-C2, respectively. This result suggests that the modification in subnuclear localization of C2 mediated by CP is likely required for the impact of the combination of these proteins on host gene expression.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Expression of selected DEGs upon transient expression of C2, C2-GFP, or GFP-C2 in the presence and absence of CP in N. benthamiana leaves. Gene expression was measured by RT-qPCR. The samples expressing CP or empty vector (EV) are used as control. Expression values are the mean of at least three biological replicates. Error bars represent SD. Asterisks indicate a statistically significant difference (*: p<0.05, **: p<0.01) according to a two-tailed comparison t-test. NbACT2 was used as the normalizer. https://doi.org/10.1371/journal.ppat.1010909.g005

Next, we investigated the contribution of C2 and CP to the virus-induced transcriptional reprogramming in the context of the viral infection. We reasoned that, if C2 and CP together affect the transcriptional landscape of the host in a different manner than C2 or CP alone, then the transcriptional changes triggered by mutated versions of the virus unable to produce either C2 or CP should present overlapping differences compared to the changes triggered by the wild-type (WT) virus. Following this rationale, we compared the transcriptome of N. benthamiana leaves infected with the WT virus or mutated versions unable to produce C2 (TYLCV-C2mut) or CP (TYLCV-CPmut1), with respect to the empty vector (EV) control (Fig 6A) or to the WT virus (Fig 6B). As expected, both point mutants were unable to establish a full systemic infection, indicating that the corresponding viral proteins are most likely not produced from the mutated genes (S8A and S8B Fig). Of note, although the CP null mutant (TYLCV-CPmut1) accumulated to lower levels in these assays, no significant changes in the accumulation of viral transcripts were detected among these viral variants in local infection assays (S6B and S8C–S8F Fig). Importantly, and despite the fact that expression of CP alone did not result in detectable transcriptional changes (Fig 4A), mutation of CP in the viral genome led to the differential expression of 3,256 genes when compared to the WT infection, supporting the notion that CP modulates host gene expression in combination, physical or functional, with other viral proteins; remarkably, 2,591 of these DEGs (79.5%) overlapped with those caused by the loss of C2 (Figs 6C, 6D and S8G; S2 Table; validation of the RNA-seq results is presented in S8H Fig), indicating that C2 and CP cooperatively mediate changes in host gene expression during the infection. Functional categories over-represented among the up-regulated genes in the presence of the WT virus appear as down-regulated in the subset of DEGs commonly triggered by the C2- and CP-deficient viruses compared to the WT version (Figs 6E and S9; S4 and S5 Tables), suggesting that the C2/CP module is responsible for the transcriptional changes of genes associated to these GO terms. A complete overview of the functional enrichment in the different subsets of DEGs can be found in S9 Fig and S5 Table.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. C2 and CP functionally interact in planta in the context of the viral infection. (A, B) Number of differentially expressed genes (DEGs) upon infection by TYLCV WT or C2-null or CP-null mutant variants (TYLCV-C2mut and TYLCV-CPmut1, respectively) in N. benthamiana leaves compared to the empty vector control (A), or to TYLCV WT (B). UR: up-regulated; DR: down-regulated; EV: empty vector. Full lists can be found in S2 Table. (C) Venn diagrams of DEGs upon infection by TYLCV C2-null and TYLCV CP-null mutants (TYLCV-C2mut and TYLCV-CPmut1, respectively) compared to TYLCV WT. UR: up-regulated; DR: down-regulated. (D) Heatmap with hierarchical clustering from (A). The colour scale indicates the Z-score. (E) Functional enrichment analysis of the subsets of up-regulated (UR) or down-regulated (DR) genes in the indicated samples. Gene Ontology (GO) categories from the Biological Process ontology enriched with a p-value<0.01 (up to top 10) are shown; functional enrichment was performed using the orthologues in A. thaliana. The colour scale indicates the -log10 (p-value), showing the significance of GO term enrichment. For a full list, see S4 Table. See also S8 and S9 Figs; S2 and S4 Tables. https://doi.org/10.1371/journal.ppat.1010909.g006

Taken together, our results demonstrate that TYLCV proteins form an intricate network of interactions that potentially vastly increase the complexity of the virus-host interface, and that viral proteins can have additional effects on the host cell when in combination. Given that intra-viral protein-protein interactions have been reported for viruses belonging to independently evolved families and infecting hosts belonging to different kingdoms of life, we propose that this might be an evolutionary strategy of viruses to expand their functional repertoire while maintaining small genomes, which would call for a reconsideration of our approaches to the study of viral protein function and virus-host interactions.

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010909

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/