1. Microbiology and Infectious Disease
Download icon

A multicellular way of life for a multipartite virus

  1. Anne Sicard
  2. Elodie Pirolles
  3. Romain Gallet
  4. Marie-Stéphanie Vernerey
  5. Michel Yvon
  6. Cica Urbino
  7. Michel Peterschmitt
  8. Serafin Gutierrez
  9. Yannis Michalakis
  10. Stéphane Blanc  Is a corresponding author
  1. BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, France
  2. CIRAD, BGPI, INRA, Montpellier SupAgro, Université de Montpellier, France
  3. MIVEGEC, CNRS, IRD, Université de Montpellier, France
Short Report
  • Cited 1
  • Views 13,818
  • Annotations
Cite this article as: eLife 2019;8:e43599 doi: 10.7554/eLife.43599

Abstract

A founding paradigm in virology is that the spatial unit of the viral replication cycle is an individual cell. Multipartite viruses have a segmented genome where each segment is encapsidated separately. In this situation the viral genome is not recapitulated in a single virus particle but in the viral population. How multipartite viruses manage to efficiently infect individual cells with all segments, thus with the whole genome information, is a long-standing but perhaps deceptive mystery. By localizing and quantifying the genome segments of a nanovirus in host plant tissues we show that they rarely co-occur within individual cells. We further demonstrate that distinct segments accumulate independently in different cells and that the viral system is functional through complementation across cells. Our observation deviates from the classical conceptual framework in virology and opens an alternative possibility (at least for nanoviruses) where the infection can operate at a level above the individual cell level, defining a viral multicellular way of life.

https://doi.org/10.7554/eLife.43599.001

eLife digest

Many viruses are small particles consisting of genetic material surrounded by a coat made of proteins. They are unable to multiply on their own and so they must enter a host cell and trick it into reading their genetic information to produce new virus particles.

It is generally thought that the process of making new virus particles happens independently in each infected cell. This idea assumes that a given particle contains the entire set of genetic material (known as the genome) of that virus, but this is not always the case. Many so-called ‘multipartite’ viruses have genomes that are split into several segments carried in separate particles: in this case, a single particle only contains a portion of the entire viral genome.

Faba bean necrotic stunt virus (or FBNSV for short) is a multipartite virus that infects and causes disease in members of the pea and bean family. There are eight types of FBNSV particle that each carries a distinct genome segment, a small section of the entire viral genome. There is a low probability that a single cell could be infected with all eight different types of particle at the same time and receive the complete FBNSV genome. So how is this virus able to successfully multiply within a plant?

To address this question, Sicard et al. used microscopy to study FBNSV genome segments as they infected the cells of faba bean plants. The experiments confirmed that the eight different segments of the FBNSV genome were not necessarily found together within the same cell, but instead accumulated independently in different cells. This means that a cell infected with FBNSV may be unable to make all of the proteins needed to assemble new virus particles. However, additional experiments demonstrated that infected cells may be exchanging virus proteins, which could enable them to create complete virus particles.

The findings of Sicard et al. demonstrate that FBNSV hijacks groups of host cells to manufacture new virus particles, rather than relying on individual cells as previously thought. It is possible that other multipartite and non-multipartite viruses work a similar manner. Ultimately, this knowledge may reshape what we know about how viruses infect their hosts.

https://doi.org/10.7554/eLife.43599.002

Introduction

Viruses introduce their genome into a susceptible cell and highjack diverse cell functions to complete their replication cycle. Once this ‘cell-autonomous’ cycle is completed, virus particles exit the infected cell and enter a new healthy one where a similar cycle is reiterated. This universal view of the viral way of life has been applied to multipartite viruses since their discovery, over half a century ago (Brakke et al., 1951; Lister, 1966; van Kammen and van Griensven, 1970; Gokhale and Bald, 1987), despite the fact that such a conceptual framework fails to explain the evolution and even the functioning of these viral systems (Iranzo and Manrubia, 2012; Sicard et al., 2016; Lucía-Sanz and Manrubia, 2017).

Depending on the viral species, multipartite viruses have their genome composed of two to eight segments of DNA or RNA (single or double stranded), each encapsidated individually in a separate virus particle. They have been reported to infect frequently plants and fungi where they represent at least 35–40% of the viral genera and families described (Hull, 2014), rarely insects (Hu et al., 2016; Ladner et al., 2016), and only hypothetically vertebrates (Ladner et al., 2016). The benefits of the individual encapsidation of distinct segments are unclear and debated (Sicard et al., 2016; Lucía-Sanz and Manrubia, 2017). In contrast, the reduced probability to infect individual host cells with at least one copy of each segment is unanimously acknowledged as a major cost that increases with the number of segments composing the viral genome (Iranzo and Manrubia, 2012). This cost appears insurmountable, at least for viruses with more than three genome segments, and sets the existence of such highly multipartite viruses as an enigma in general biology (Iranzo and Manrubia, 2012; Sicard et al., 2016; Lucía-Sanz and Manrubia, 2017). Nevertheless, its empirical basis has never been questioned.

Using a highly multipartite virus with eight genome segments, we here propose that the conceptual framework in Virology should be amended to account for such viral systems. Indeed, in our specific experimental model species, we demonstrate that the distinct segments do not need be together in individual cells for the system to be functional. They accumulate independently in and complement across neighboring cells, defining a multicellular way of life.

Results

Distinct viral genome segments do not necessarily co-localize in individual infected cells

Our first aim was to experimentally verify whether the different segments of a multipartite virus have to be together in individual cells for the system to be functional. We used the highly multipartite faba bean necrotic stunt virus (FBNSV, genus Nanovirus, Family Nanoviridae), whose genome is composed of eight circular ssDNA segments (Figure 1A) (Grigoras et al., 2009). We first compared the localization of distinct segments in individual cells of faba bean host plants with specific fluorescent probes and confocal microscopy. In plants submitted to productive viral infection (see Materials and methods), the visualization of pairs of segments respectively labeled with green and red fluorochromes immediately revealed that distinct segments do not necessarily co-occur but are most often found in different cells (Figure 1B–I and Figure 1—figure supplement 1). We tested 7 out of the 28 possible pairs of segments. In all cases we observed a similar situation with one segment most often highly accumulated in the absence of the other in individual cells. Remarkably, this applied even to the pairs from the three segments encoding for the three basic functions of plant viruses: replication (segment R encoding M-Rep), encapsidation (segment S encoding CP) and intra-host movement (segment M encoding MP), suggesting that cells containing all eight segments are extremely rare.

Figure 1 with 2 supplements see all
Localization of distinct FBNSV segments in individual cells.

The FBNSV genome comprises eight single-stranded circular DNA segments (A). The size and name of the segments and encoded proteins are indicated inside the circles. Black stem-loops and blue arrows indicate the replication origins and the coding regions, respectively. The function of each segment is described in Grigoras et al. (2009): DNA-C, cell cycle resetting; DNA-M, within plant viral movement; DNA-N, nuclear shuttle protein; DNA-R, replication; DNA-S, coat protein/encapsidation; DNA-U1, U2 and U4, unknown. B-I show cross sections of infected faba bean petioles, with phloem (ph, the only tissue infected by nanoviruses), and xylem (xy) bundles indicated in B. Pairs of segments are green- and red-FISH probed as indicated in each micrograph. The fluorochromes incorporated into the probes are inverted in B and C for control. Color channels are merged in all images but C where green and red are shown merged and separated to evidence the accumulation of the segment S in cells where the segment R is not detected (exemplified by white arrows). Split channel images corresponding to images B and D-I are shown in Figure 1—figure supplement 1. Nuclei are DAPI-blue stained. Horizontal bars = 10 microns.

https://doi.org/10.7554/eLife.43599.003

Though challenging the current view of the viral cell-autonomous replication cycle, a simple explanation of our observation is that the FBNSV can function while its genome segments occur in distinct neighboring cells. This possibility calls i) for further evidence that the accumulation of a given segment is independent of the accumulation of the others in individual cells and ii) for the proof of concept that the function encoded by a given viral segment can complement the others at a distance, in cells where this segment is absent.

Distinct segments accumulate independently in distinct individual host cells

One may argue, sticking to the paradigm that the viral genetic information is replicated as an integral genome within individual cells, that all FBNSV segments are present in infected cells but that the apparent absence of some in Figure 1 is due to the detection limit of our technique. Because every technique has its limit, whatever the technology implemented, it would not be reasonable to certify that the absence of detection is a proof of the absence of the corresponding segment. We thus imagined an approach where the detection limit becomes irrelevant. For a given pair of segments, we quantified and compared both green and red fluorescence in all individual cells where at least one of the two was observed above background (for detailed quantification procedure see the Materials and methods section). By doing so, we alleviate the problem of the limit of detection and rather question whether the accumulation of one segment of the pair is dependent on that of the other. As a positive control of this approach, we first produced two fluorescent probes, each specifically labeling a different region of the same segment R (probes R1 and R2). Figure 2 illustrates that all cells labeled by R1 are also labeled by R2 (Figure 2A). Moreover, plotting the average intensity of green over red fluorescence for each of these cells resulted in a highly significant linear relationship (correlation coefficient r = 0.90, p=1.98 10−23 in the example of petiole N° 42 in Figure 2B). Four independent repeats of this control similarly showed strong correlations (Table 1). This result validates our approach by demonstrating that when we monitor two viral DNA sequences whose accumulation should be highly correlated, such as two regions of the same segment, we indeed find that they co-localize and accumulate at highly correlated levels.

Figure 2 with 1 supplement see all
Independent accumulation of FBNSV segments in individual cells.

(A) labeling of two distinct regions of the same segment R by a red and green fluorescent probe, respectively. Green and red channels are shown merged (left) and separated (center and right). Nuclei are DAPI-stained in blue. Horizontal bar is 10 microns. The average intensity of the green fluorescence within individual cells is plotted against that of the red in B and C. Green and red probes are targeted against two distinct regions of the same segment R in (B), and against two distinct segments, respectively S and R in (C). The analyzed cells are from petiole N°42 for the positive control in (B) and from petiole N°5 for the pair R/S in (C). The same analysis from three additional petioles for the control (also see Figure 2—figure supplement 1) and nine additional petioles for the segments pair R/S are summarized in Table 1. Panel D plots the ratio of the frequency of two segments of a pair (estimated by qPCR on total DNA extracted from infected tissues) against the ratio of the number of cells they respectively infect within eight petioles from eight different infected plants (data provided in Supplementary file 1: Table S1).

https://doi.org/10.7554/eLife.43599.006
Table 1
Correlation test for the accumulation of two segments in individual cells
https://doi.org/10.7554/eLife.43599.008
Petiole*1234579101112131415161840414243
dpi*25121212271125252520202020202019211927
PairR/SS/MR/MR1/R2
n47186731392820293915194332234616226249
F§3.040.190.151.230.240.460.100.690.711.944.731.742.252.682.431323297254345
r0.250.110.05−0.20−0.08−0.130.08−0.160.140.36−0.470.200.26−0.340.230.990.970.900.94
p††0.090.660.700.280.630.510.750.410.400.190.040.190.140.120.132.89e-151.81e-131.98e-232.81e-23
  1. * Code number of petioles and time of sampling (dpi: days post infection).

    Pair of segments analyzed for each petiole.

  2. n = number of cells detected positive for at least one of the segments of the pair. Fluorescence quantification for each of these cells is given in Supplementary file 2: Table S4.

    §Value of F from regression analysis.

  3. Correlation coefficient.

    ††p-value.

This approach was then extensively used for the pair of segments R/S because they encode the two key viral functions, replication and encapsidation. In total, we tested 10 petioles, each from a different plant infected independently. No significant correlation could be found (Figure 2C, Table 1), whether at early or late stages of infection (Table 1 and also see Figure 1—figure supplement 2), showing that the accumulation of the segment S in a cell is independent of that of the replication-encoding segment R. We extended this analysis to five additional petioles from five independent plants, two for the pair R/M and three for the pair S/M, and again found no significant positive correlation; a result indicative of independent accumulation (Table 1). To further support this conclusion, we used the earlier reported variation in the relative frequency of the segments across infected plants (Sicard et al., 2013). Within an infected petiole, if two segments independently enter and accumulate in individual cells, then their relative frequencies in infected tissues should be proportional to the ratio of the number of cells they respectively infect. We quantified by qPCR the relative frequency of the two segments of the pair in a fragment of eight of the analyzed petioles (Supplementary file 1: Table S1). Figure 2D shows a significant correlation between the relative frequency of the segments in the tissues and the relative number of cells they respectively infect (n = 7, F = 12, R = 0.677, p-value=0.0122). Hence, two different technologies (FISH-confocal microscopy and quantitative real-time PCR) concur to demonstrate that the accumulation of distinct FBNSV segments in individual cells is independent.

The function of a viral gene can act in a cell where the gene itself is absent

To make sense out of the above observation we assumed that a viral function can be present and act in a cell where its encoding segment is not, and thus that the viral system can be functional despite the independent dispersal of its distinct genes in distinct cells. The validation of this assumption was set up with a focus on the replication function. We first confirmed that the segments detected in individual cells are the product of replication. Then we searched for the presence of the protein M-Rep in these cells, to support replication even when its encoding segment R is absent.

Two possible alternatives could lead to the detection of a segment as a fluorescent signal spread out in a large volume of the nucleus. Either one (or very few) copy entered, replicated and invaded the nucleus to detectable level, or a number of copies large enough to be detectable entered and diffused throughout the nucleus with no replication. When looking at images shown in Figure 1, the latter appears unrealistic because it would imply the specific sorting of different segments at their entry in distinct nearby cells. Nevertheless, to experimentally dismiss this possibility, we similarly investigated the distribution of two alleles of the same segment (differing solely by a small inserted marker sequence) for which no specific sorting can be expected. In the absence of sorting, if the segment massively enters cells to detectable level without replication, the two alleles should be co-detected in most cells. In contrast, if one (to very few) copy of this segment enters each individual cell and then accumulates through replication, allele-specific FISH labeling should reveal solely one of the two alleles in most cells. Three independent experiments were carried out for each of the segments S and N where two genetic markers have earlier been inserted (Gallet et al., 2017). They consistently revealed that most individual cells (54% to 100%) contained only one detectable allele (Figure 3—figure supplement 1), confirming that one to very few copies of the segment initially enter individual cells and thus that detection all over the nucleus can only result from replication.

We then confronted the cells where the segment S is detectable (thus where it has replicated) to the detection of either segment R or its protein product M-Rep, by a combination of FISH and immunofluorescence-labeling (Figure 3A and B). While segment R was detectable in only a minority of these cells (approx. 40%), its protein product M-Rep was positively revealed in nearly 85% (Figure 3C), indicating that the expression product of R segment can move to neighboring cells. That the protein M-Rep is undetectable in a few cells replicating segment S (approx. 15%) can be explained by distinct turnovers for the genome segments and their expression products. Indeed, both mRNAs and proteins have a relatively rapid turnover within a cell, whereas the DNA viral sequences can be stored indefinitely into the cell nucleus either encapsidated or as stable minichromosomes (Gronenborn, 2004; Ramesh et al., 2017; Deuschle et al., 2016; Rodríguez-Negrete et al., 2014). Thus, we assume that for all cells where at least one FBNSV segment is detectable the protein M-Rep is or has been present, but that this protein may have disappeared in some cases at the moment of the observation. Consistent with this assumption is the observation of a similar proportion of cells containing the segment R where its own expression product M-Rep is no longer detectable (Figure 3C).

Figure 3 with 1 supplement see all
Accumulation of the protein M-Rep in cells containing either segment S or segment R.

Segment R (A) or S (B) is FISH-labeled with a green fluorescent probe whereas the M-Rep protein is immuno-labeled with a red fluorochrome. Green and red channels are shown separated (top and middle) and merged (bottom). Nuclei are DAPI-stained in blue. Horizontal bar is 10 microns. Yellow arrows point at cells where both the DNA segment and the protein M-Rep are detected, whereas red arrows exemplify cases where only M-Rep is detectable. The boxplots (C) are constructed from Supplementary file 1: Table S1. The proportion of cells with the S segment that also contain the R segment (green left boxplot, estimated from 10 petioles) is compared to the proportion of cells with the S segment that also contain the protein M-Rep (middle orange boxplot, estimated from 14 petioles). The proportion of individual cells with detectable R segment that also contain the protein M-Rep was estimated from 4 infected petioles (blue right boxplot). Difference between left green boxplot and orange middle boxplot is highly significant (GLM model, p-value (>χ12) = 2.94e-12); that between middle orange boxplot and right blue boxplot is not (GLM model, p-value (>χ12) = 0.55).

https://doi.org/10.7554/eLife.43599.009

It is important to note the detection of the M-Rep protein, often with intense signal (Figure 3A & B), in a large number of cells where segment R is absent (Supplementary file 1: Table S1, petioles 36–39). Moreover, Figure 3C (orange middle and bleu-grey right boxplots) indicates that the protein M-Rep is not more associated to cells containing its own encoding segment R than to cells containing other segments. Although they represent indirect evidence, these observations together further support our conclusion that either the mRNA or the protein M-Rep itself can travel from the producing cells (those where segment R accumulates) to other cells of the host, as further discussed below.

Discussion/conclusion

Altogether, our results demonstrate that key genome segments of the FBNSV accumulate in individual cells independently of the presence and accumulation of the others. We thus propose that the different parts of the viral genome can complement across distinct neighboring cells and can only sustain the productive infection at a multicellular tissue level. As numerous other plant virus species (Hipper et al., 2013; Folimonova and Tilsner, 2018), nanoviruses are restricted to vascular tissues and replicate in phloem companion and parenchyma cells (Shirasawa-Seo et al., 2005). A model compatible with our observations is that each genome segment entering and expressing within a cell can export its gene products as ‘common goods’ in neighboring cells and/or sieve elements, so that these common goods are redistributed among interconnected cells and complement the system. The demonstration that the protein M-Rep accumulates and functions in cells devoid of its encoding DNA-R is a proof of concept fully supporting this model. Numerous studies have shown that plant mRNAs (Kehr and Kragler, 2018) and proteins can move from cell to cell or long distance (Turnbull and Lopez-Cobollo, 2013; Lopez-Cobollo et al., 2016; Paultre et al., 2016), even from root to shoot. Unfortunately, the mechanisms and specific control of these mRNAs and proteins mobility is vastly elusive. For FBNSV, each segment may export and redistribute its product in a distinct way, some may produce mobile mRNA, others may produce mobile protein, and others may be complemented for movement by the product of a specific segment (or even by a host factor) that may act as a carrier. Specifying these diverse possibilities is beyond the scope of this report. Likewise, whether this non-cell-autonomous model can be extended to viral species invading non-phloem tissues where cell communication is more restricted is unknown. Non phloem-restricted viruses could induce the formation of symplastic domains, either by manipulating the endogenous capacity of the host plant to do so (Faulkner, 2018) or by opening plasmodesmata through the action of their non-cell-autonomous movement protein (Lucas, 2006), but this possibility awaits further investigation. Here we introduce an additional concept in virology, which is compatible with empirical observations and which partially alleviates the insurmountable cost in highly multipartite viral systems such as FBNSV (Iranzo and Manrubia, 2012): because concomitant infection of individual cells by all genomic segments is not necessary, the associated putative cost should be much smaller if not nil at the within-host level. We earlier discussed the fact that the analogous cost upon between-host transmission and the mechanisms of its compensation remain to be understood (Gallet et al., 2018).

Such a multicellular way of life could be adopted in other multicomponent genetic entities, such as other multipartite viruses, segmented viruses which often fail to encapsidate all genome segments together (Luque et al., 2009; Wichgers Schreur and Kortekaas, 2016; Brooke, 2017), satellites, and defective interfering particles, if functional complementation could occur at a supra-cellular level.

Materials and methods

Key resources table
Reagent type
or resource
DesignationSource/referenceIdentifiersAdditional
information
Gene (virus)FBNSV JKI-2000 Segment CGenBankGenBank: GQ150780.1complete sequence
Gene (virus)FBNSV JKI-2000 Segment MGenBankGenBank:
GQ150781.1
complete sequence
Gene (virus)FBNSV JKI-2000 Segment NGenBankGenBank:
GQ150782.1
complete sequence
Gene (virus)FBNSV JKI-2000 Segment RGenBankGenBank: GQ150778.1complete sequence
Gene (virus)FBNSV JKI-2000
Segment S
GenBankGenBank: GQ150779.1complete sequence
Gene (virus)FBNSV JKI-2000 Segment U1GenBankGenBank: GQ150783.1complete sequence
Gene (virus)FBNSV JKI-2000 Segment U2GenBankGenBank: GQ150784.1complete sequence
Gene (virus)FBNSV JKI-2000 Segment U4GenBankGenBank: GQ150785.1complete sequence
Strain, strain background (nanovirus, FBNSV)FBNSV infectious clonereference N° 11Isolate FBNSV-[ET:Hol:97]
AntibodyAnti-M-Rep, rabbit, polyclonalreference N°32FBNYV-M-Rep 8th BleedDilution 1/300
Sequence-based reagentSynthetic short probesthis paper, synthetsized by the company Eurogenetec, Liège, BelgiumNmys2-Red, Nmys7-Green, Smys1-Red, Smys8-Greensee Table S2
Sequence-
based reagent
random primed probe for distinct segmentsthis paperprobes for distinct segments were prepared as described in the Materials and methods section
Sequence-based reagentPrimers to produce the templates for the probesthis paper,
synthesized by the company Eurogenetec, Liège, Belgium
see Table S2
Commercial assay or kitBioprime DNA-Labeling system kitInvitrogen, Carlsbad, CA, USAreference product: 18094011see Materials and methods section
Other: Host plantFaba beanVilmorin, Saint Quentin Fallavier, FranceVicia Faba, cultivar ‘Seville’

Host plant, virus isolate and inoculation

Faba bean (Vicia faba, cv. ‘Sevilla’) was used as the host plant in all experiments, and seeded, maintained and inoculated as described (Sicard et al., 2013; Gallet et al., 2017; Gallet et al., 2018; Sicard et al., 2015). The viral infectious clone is from the species Faba bean necrotic stunt virus, genus Nanovirus, family Nanoviridae. Its construction in agrobacterium plasmids has been earlier described in details (Grigoras et al., 2009), as well as all conditions and procedures for agro-inoculation (Grigoras et al., 2009; Sicard et al., 2013; Gallet et al., 2017; Gallet et al., 2018; Sicard et al., 2015). Plants were sowed as one seed per pot. The plantlets were agro-inoculated nine days later and symptoms appeared from 9 days post inoculation (dpi) till 19 dpi, depending on individual plants. As shown in all the publications cited in this paragraph, the inoculation of plants with this infectious clone routinely yields a productive viral infection, meaning that the FBNSV particles can be purified from these infected plants and very efficiently transmitted to new healthy ones by aphids.

Sampling of infected material

Infected plants were sampled at different days post inoculation: at 11, 12, 13, 17, 20, 25 and 27 dpi (Table 1 and Supplementary file 1: Table S1). Nanoviruses, and thus FBNSV, replicate and accumulate primarily in phloem vascular bundles of the upper part of the plant, in companion cells and phloem parenchyma cells of the stem, petioles and main leaf veins. A 1 to 3 cm-long section of the petiole of the upper leaf level was cut off from each infected plant and immediately fixed and processed for fluorescent in-situ hybridization (FISH) and/or immunofluorescence labeling (IF). In some cases, the petiole was divided into two equal portions, one for FISH and/or IF and the other for DNA extraction and qPCR.

Quantitative real-time PCR

The FBNSV genome is composed of eight circular ssDNA segments of around 1 kb each (Figure 1A). Total DNA extraction from infected petioles and qPCR reactions specific to each of the eight segments were performed as described previously (Gallet et al., 2017), including primer pairs, PCR conditions, and post-PCR data analysis. The relative frequency of two segments of a pair within a given petiole was obtained by dividing the estimated copy number of a segment by that of the other.

Fluorescent in situ hybridization and immunofluorescence labeling

Random priming and incorporation of Alexa Fluor-labeled dUTP (either Alexa Fluor 488 or Alexa Fluor 568 for green and red labeling, respectively) were performed to prepare probes specific to each of the eight segments, with the BioPrime DNA labeling system kit (Invitrogen) and according to the manufacturer instructions except that dUTP-Alexa Fluor were used in place of dCTP-biotine of the kit. For each probe, the template DNA corresponded to a PCR-amplified fragment restricted within the viral gene coding sequence, which is the sequence unique to each segment. The primer pairs used to amplify the coding sequence of the targeted segments are listed in Supplementary file 1: Table S2. In some cases, short probes were synthesized by the company Eurogentec and covalently linked to either ATTO-488 (green) or ATTO-565 (red) fluorochrome (sequences also available in Table S2). The specificity of each probe was tested on membranes, on dotted plasmids each containing one of the eight viral segments, as well as on healthy plants.

Each petiole harvested from an infected plant as described above was fixed by gentle stirring overnight at 4°C in PBS buffer containing 4% paraformaldehyde and 0.2% Tween-20. After one rinse with 2 mL of PBS buffer containing 0.1M glycine, the petiole was transferred into ethanol 70% and stored at 4°C until use (max. storage time: 1 month).

Fixed petioles were then placed individually into Eppendorf tubes containing 8% low-melting agarose in water at 40°C. Petiole trunks were maintained into an upright position until the agarose cooled down and polymerized, and placed at 4°C for 4 hr. The jellified agarose blocks were extorted from the tubes and cross-sections of 80–100 microns were produced with a Vibratome HM650V (Microm), set up in mode CPC, program 50, speed 23, frequency 100, and amplitude 0.6. Cross-sections were treated in 1 mL of Carnoy solution (six volumes chloroform, three volumes ethanol, and 1 vol acetic acid) during 1 hr under gentle stirring and then rinsed once in PBS buffer. An RNAse treatment (100 μg/mL of RNAse1 in PBS) was applied to all samples for 45 min. at 37°C, in order to eliminate viral mRNAs, followed by one rinse in PBS and three additional rinses in hybridization buffer (20 mM Tris-HCl pH8, 0.9M NaCl, 0.01% SDS, 30% Formamide).

Fluorescent probes were diluted thirty times in hybridization buffer (10 μL in 300 μL total), denatured 10 min. at 100°C and then rapidly cooled on ice for 15 min. Petiole cross sections were then incubated overnight at 37°C in the diluted and heat-denatured probes solutions into embryo dishes sealed with parafilm membranes. After three rinses of 5 min with hybridization buffer and one with PBS, petiole sections were mounted on microscopy slides in Vectashield antifade mounting medium containing 1.5μg/mL DAPI for staining nuclei.

Some samples were further treated for immuno-fluorescent labeling of M-Rep protein. In these cases the FISH-treated petiole cross sections were collected after the last PBS rinse and blocked for 1 hr and 30 min. in PBS + 5% BSA. The incubation with the primary antibody (M-Rep specific antibody ‘FBNYV-M-Rep 8th bleed’ (Vega-Arreguín et al., 2005), diluted 1/300) was in PBS + 5% BSA overnight at 4°C, and that with the secondary antibody (goat anti-rabbit Alexa Fluor 594 IgG conjugate, diluted 1/250, Life Technologies) was in PBS + 5% BSA for 1 hr at 37°C. Samples were submitted to three rinses of 10 min each in PBS + 0.05%Tween-20 at room temperature after incubations with primary and secondary antibodies. Samples were finally transferred into PBS and then mounted on microscopy slides as above.

Confocal microscopy observations and fluorescence quantification

Observations were all performed in sequential mode using a Zeiss LSM700 confocal microscope. Alexa Fluor 488 and ATTO 488 were excited with a 488 nm laser and the variable secondary dichroic (VSD) beam splitter was set to recover fluorescence up to 535 nm. Alexa Fluor 568/594 and ATTO 565 were excited with a 555 nm laser and the VSD was set to recover fluorescence up to 615 nm. DAPI was excited with a 405 nm laser and the VSD was set to recover fluorescence up to 626 nm, with the additional use of a short pass SP490 eliminating wave lengths > 490 nm. Images were acquired with 40x or 63x objectives at variable resolutions, depending on their intended use, with a pinhole aperture of 1 airy unit. Acquisitions were either in plane or stack mode, and the images used from stacks correspond to maximum intensity projections. The settings for the acquisition of images shown in Figures 1, 2 and 3, Figure 1—figure supplement 1 and Figure 3—figure supplement 1 are detailed in Supplementary file 1: Tables S3. All images used for fluorescence quantification were acquired with the 40x objective, and the resolution chosen (512 × 512) corresponded to a compromise between image quality and acquisition time. All were acquired in stack mode and quantification performed on maximum intensity projections. The absence of potential biases induced by image acquisition at distinct resolution for fluorescence quantification has been verified as described in the next section.

For constructing Figures 1, 2 and 3, Figure 1—figure supplement 1 and Figure 3—figure supplement 1, as well as for selecting all cells where the fluorescence has been quantified, raw images were processed using ImageJ software version 1.50c4. The overall intensity of green and red signals was adjusted for each individual image up to the limit where the green and red auto-fluorescence of the cell wall, cytoplasm and nuclei of the xylem and mesophyll cells visually disappeared (the FBNSV being phloem-restricted, it does not invade xylem and mesophyll cells). Individual cells in phloem bundles were considered containing a segment when either green or red fluorescence (or both) could be visualized above background. Above-background fluorescence was never observed in non-phloem cells, consistent with the phloem restriction of this virus.

Infection by FBNSV distorts the anatomy of the phloem tissues and so distinguishing between phloem parenchyma cells, companion cells and sieve elements is difficult and sometimes impossible in our petiole cross-sections. For this reason, and because we were primarily interested by cells where the viral DNA replicates (ssDNA viruses replicate and accumulate in the nucleus of their host cells), we quantified the fluorescence only in those cells where a nucleus was clearly revealed by the DAPI staining, so in cells other than the sieve tubes. The DAPI staining allowed to precisely encircle the nucleus, and thus to define an ovoid corresponding area in each selected individual cell. Both green and red fluorescence were then quantified for each pixel within this area and the average pixel fluorescence intensity was estimated. This gave one average intensity value (in arbitrary fluorescence units) for green and one for red fluorescence in each selected individual cell within an infected petiole and all quantitative results are provided in Supplementary file 2: Table S4.

Reliability of the quantification of fluorescence within individual cells of infected petioles

As indicated above, the setting of the confocal microscope was adjusted depending on the planned ulterior use of images. For example, to prepare Figure 1, the plane mode, higher resolution (>512×512) and higher magnification (63x objective) were generally preferred. In contrast, for fluorescence quantification, the stack mode and maximum intensity projection images appeared best appropriate in order to capture the whole fluorescence of the nuclei. For obvious practical reasons, the numerous images needed for quantification were acquired more rapidly at the feasible resolution of 512 × 512, and with a lower magnification (40x objective) in order to screen microscopy fields potentially allowing the scoring of more segment-containing cells (examples of such images are shown in Figure 1B, and Figure 3A). We wished to confirm that different resolution settings do not differentially affect green and red fluorescence. Another potential concern was the time spanning between microscopy slide preparation and the end of the observation/quantification. Due to the experimental design, it was important to confirm that the stability of red and green probes was similar and that a possible differential degradation could not bias our localization and correlation/regression analysis.

We first quantified the fluorescence of R1-Green and R2-Red probes in petiole N°43 with fast acquired images (resolution 512 × 512). We then repeated the observation 2 weeks later on the same microscopy slides with a resolution 1024 × 1024. The cells quantified at the two dates were not necessarily the same as we could not be sure to retrieve all of them. The regression analysis from these two sets of images gave remarkably similar results, demonstrating that the resolution setting and the aging of the preparation could only have negligible effect (if at all) on our conclusions. Results from these controls are shown as Figure 2—figure supplement 1.

Statistical analysis

Linear regression analyses were performed with the JMP 13.2.0 software. Each petiole was analyzed separately. We did not analyze the fluorescence quantification data set as a whole (pooling cells from distinct petioles) because the virus-associated fluorescence intensity is not directly comparable. The total viral load varies in between distinct petioles because of a ‘natural’ variance across infected plants and because some were harvested at different dpi. The auto-fluorescence of the infected tissues also varied across petioles because of the time of infection, the total viral load and presumably other unknown factors.

The proportion of cells containing both S and R segments, both S segment and M-Rep protein, or both R segment and M-Rep protein (Figure 3C, Supplementary file 1: Table S1) was analyzed with Generalized linear models (GLM) with ‘treatment’ (S segment – R segment/S segment – M-Rep/R segment – M-Rep) as a categorical explanatory factor, and a quasi-binomial error type. The GLM model was computed using R software 3.1.3. The name of each statistical test and its outcome is indicated in the text and figure legends.

Data availability

All data are available in the manuscript and in Supplementary files.

Raw data of all quantified green and red fluorescence within individual cells of infected plants are provided as a separate Excel file, Supplementary file 2: Table S4.

To allow repeat/reproduce all correlation tests, the 508 raw/unprocessed images (.lsm format) used for preparing all figures and for fluorescence quantification in individual cells have been deposited in the public repository figshare. They can be accessed at the DOI: 10.6084/m9.figshare.5981968.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
    Plant Virology, 1104
    1. R Hull
    (2014)
    Academic press.
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32

Decision letter

  1. Fernando García-Arenal
    Reviewing Editor; Technical University of Madrid, Spain
  2. Detlef Weigel
    Senior Editor; Max Planck Institute for Developmental Biology, Germany
  3. Fernando García-Arenal
    Reviewer; Technical University of Madrid, Spain

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "A multicellular way of life for a multipartite virus" for consideration by eLife. Your article has been reviewed by three peer reviewers, including Fernando García-Arenal as the guest Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Detlef Weigel as the Senior Editor.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

The paper has been reviewed now by three experts. All of them consider that the results are original, novel, relevant and important. All three reviewers, though, also consider that major claims, particularly proposing a new paradigm in (plant) virology based on a multicellular way of life, are not fully justified by the observations, and that the presentation and interpretation of the results should be toned down. This would not diminish the importance of the results, but it should help to highlight the observations that are new here. Below we summarise the major comments, as well as new experiments suggested that are strongly encouraged to sustain the present major claims.

1) The visualization of pairs of segments respectively labelled with green and red fluorochromes revealed that distinct segments do not necessarily co-occur but are most often found in different cells. Authors tested 7 out of the 28 possible pairs of segments. As shown in Table S1 in Supplementary file 1 the number of cells harbouring both segments is almost the same as the one describing the number of cells that more accumulates one of them. This cannot be an indication that a cell is mainly affected by one, and only one, segment. Also, the presented results, do not discard that in addition to the detected segment the rest of segments were at undetectable amounts in the same cell, a problem of which the authors are aware. Or it could happen that the mRNA was still active, which seems not to have been considered. In fact, the procedure for the FISH technique has a RNase treatment that eliminates RNAs present in the cell, thus, this possibility cannot be ruled out.

2) It is demonstrated that the function of a viral gene can act in a cell where the gene itself is not detected and these observations suggest that either the mRNA or the protein M-Rep itself can travel from the producing cells (those where segment R accumulates) to other cells of the host. Unfortunately, no evidence for such trafficking is given and thus, by the moment, is speculative. Authors must demonstrate that M-Rep behaves as a non-cellular autonomous protein or than its mRNA is able to traffic from cell to cell. Would the authors' result indicate that all the proteins encoded by FBNSV be non-cell-autonomous in their function or only some?

3) In the Discussion it should be considered the tissue tropism of FBNSV, and the particularities of plasmodesmata between the different phloem cell types, which may limit the generalisation of the presented observations to other viral systems.

These three points can be easily addressed in a revised Discussion section. A modification of the title would also be advisable.

4) The experiments demonstrate that segments of a unique type accumulate preferentially in individual cells, where they are found jointly with the replicase protein but not necessarily with the replicase gene. This observation, however, does not justify that a sufficiently high MOI is not needed and that the viral system is functional through complementation across cells. The experiments reported indicate a sort of specialization of different genes, such that one type is preferentially expressed in each cell. This might be a winner-takes-all strategy that does not necessarily entail isolation of the individual genes upon infection. Task allocation is an often used strategy (e.g. in social insects), and it would be great if it would have been discovered by multipartite viruses in order to optimize segment production. But, again, it would not solve the high MOI requirement for between-host propagation.

5) An experiment was proposed which would clarify these issues: Prepare two bacteria with only one type of segment (say R and S, as in most experiments performed). Inoculate a leaf at increasing distance with ONE copy of each segment. Demonstrate that the S segment does replicate. This experiment would answer two main concerns of the reviewers regarding the claims made: (i) Even if individual segments propagate independently and are affected by severe bottlenecks (such that they reach the plant in very low numbers), replication of any segment is possible; (ii) The amount of replicase protein is high enough so as to reach cells far apart and cause the replication of other "waiting" segments before either of them degrades.

6) Regarding the relationship between the experiments reported and the MOI required upon between-host transmission, two processes are apparently mixed here. The first one is between-plant transmission: as discussed above, this process requires an MOI so high for multipartite viruses with several segments that it has not been observed in nature: It has to be pointed out that the MOI as described e.g. in 5 has never been empirically quantified. Actually, what studies show is that the number of founding particles for the population infecting a given individual plant is low. But this is a number related to genealogy, not to transmission. (An analogous situation is population bottlenecks: the ancestry of a current population can be traced back to a few, even a single, individual, this fact not implying that the ancestral population was that small. Iranzo and Manrubia, 2012, concerns the size of the ancestral population, not the founding population.) The second one is the (theoretical and empirical) search of possible mechanisms that can alleviate that hypothetical high MOI. It has been put forward, with good empirical support, that viral particles might form groups or clusters upon propagation (and this might be relatively common).

7) Finally, there should be a discussion of the cost of producing very high amounts of mRNA and/or replicase protein that will be lost (as demonstrated by the experiments), if the interpretation of the authors proves to be correct.

https://doi.org/10.7554/eLife.43599.017

Author response

The paper has been reviewed now by three experts. All of them consider that the results are original, novel, relevant and important. All three reviewers, though, also consider that major claims, particularly proposing a new paradigm in (plant) virology based on a multicellular way of life, are not fully justified by the observations, and that the presentation and interpretation of the results should be toned down. This would not diminish the importance of the results, but it should help to highlight the observations that are new here. Below we summarise the major comments, as well as new experiments suggested that are strongly encouraged to sustain the present major claims.

We agree that in various instances in the text of the original version of our manuscript, the writing was inappropriate and might have suggested that we claimed that the reported observation applied to all multipartite viruses and established a new paradigm in Virology; this was not our intention. To alleviate this problem, and to comply with the request from the reviewer to tone down this claim, we have modified those sentences to explicitly state that our discovery, at this point concerns the FBNSV infecting faba bean host plants.

More specifically:

In the Abstract the sentence “Our observation opens a new conceptual framework in virology where the infection operates at a level above the individual cell level, defining a viral multicellular way of life” is now replaced by the sentence “Our observation deviates from the classical conceptual framework in virology and opens an alternative possibility (at least for nanoviruses) where the infection can operate at a level above the individual cell level, defining a viral multicellular way of life”

At the end of the Introduction, the sentence “Using a highly multipartite virus with eight genome segments, we here suggest that the dominant conceptual framework in virology must be changed to comprehend these viral systems” is amended to “Using a highly multipartite virus with eight genome segments, we here propose that the conceptual framework in virology should be amended to account for such viral systems. Indeed, in our specific experimental model species, we demonstrate….way of life”.

In the Discussion, the sentence “Here we simply coin a new concept in virology, which is compatible with empirical observations and alleviates the question of the insurmountable cost in highly multipartite viral systems such as FBNSV…” is changed for “Here we introduce an additional concept in virology, which is compatible with empirical observations and which partially alleviates the insurmountable cost in highly multipartite viral systems such as FBNSV.”

1) The visualization of pairs of segments respectively labelled with green and red fluorochromes revealed that distinct segments do not necessarily co-occur but are most often found in different cells. Authors tested 7 out of the 28 possible pairs of segments. As shown in Table S1 in Supplementary file 1 the number of cells harbouring both segments is almost the same as the one describing the number of cells that more accumulates one of them. This cannot be an indication that a cell is mainly affected by one, and only one, segment. Also, the presented results, do not discard that in addition to the detected segment the rest of segments were at undetectable amounts in the same cell, a problem of which the authors are aware. Or it could happen that the mRNA was still active, which seems not to have been considered. In fact, the procedure for the FISH technique has a RNase treatment that eliminates RNAs present in the cell, thus, this possibility cannot be ruled out.

First, to be factual, it is not totally accurate to state “the number of cells harbouring both segments is almost the same as the one describing the number of cells that more accumulates one of them”. Figure 3C shows that, overall replicates, 42% of the cells contain the two segments of the pair R/S whereas 58% contain only one segment of the pair. From Table S1 one can calculate that these proportions are respectively 32% and 68% for the pair S/M, and are indeed more even for R/M with 46% and 54%.

We totally agree that this cannot indicate that the cell is infected by one and only one segment, and we do not think we have written it anywhere in the paper. This is an indication that the majority of the cells contain solely one of the segments of the tested pair, and that the cells that contain all 8 segments are most likely utterly rare. In fact an accurate estimate of the proportion of cells affected by all 8 segments is not feasible with our current data as it would require the investigation of numerous additional pairs of segments. However, it is clear from the three pairs quantified that the proportion of susceptible cells having the three segments S/R/M together is close to ~ 0.1, and thus that the proportion of cells containing all 8 segments should be much smaller. We conclude that the immense majority of the cells containing viral material do not contain the integral viral genome (all 8 segments). We have accordingly added a few words (in the Results) at the end of the last sentence of the paragraph “Remarkably, this applied even to the pairs from the three segments encoding for the three basic functions of plant viruses: replication (segment R encoding M-Rep), encapsidation (segment S encoding CP) and intra-host movement (segment M encoding MP), suggesting that cells containing all eight segments are extremely rare”.

That a segment could be present at undetectable levels is addressed and discussed in the paper, because all techniques obviously have their limitations. This is why we developed the fluorescence quantification approach, that informs on the interdependency of the distinct segments for their respective accumulation in individual cells and that is not affected by the detection limit; we assume this is the reason why this comment states “a problem of which the authors are aware”. So we are not sure whether this comment calls for a specific response on our side.

Because we can interpret the meaning of “mRNA was still active” in different ways, our response to the corresponding comment is plural. If it means that a mRNA could be present in a cell when the segment encoding it is not (and has never been present), then this is indeed a possibility related to mobile mRNAs [1-4] that is amply discussed in the manuscript. If it means that a DNA segment has been present in the cell but has disappeared while its mRNA is still there (and active), then it goes against our current understanding of how viral infections function. For all plant viruses investigated thus far, the genomic material is stored indefinitely within infected cells into mature viral particles (or in rarer cases as nucleo-protein complexes); the viral genome remains when the viral replication cycle is terminated and when viral mRNA (and translation) have long disappeared (for example see [5]). ssDNA viruses are believed to stably and indefinitely store their DNA as minichromosome (dsDNA) or as encapsidated ssDNA in the nucleus, and so are no exception to this rule (this is discussed with due references when explaining why some cells have the DNA segment accumulated but no detectable M-Rep protein)

Finally, we were aware that some mRNAs could be mobile [1-4] and so could be present in cells where their encoding DNA is not. This is why we used RNAse treatment when detecting and localizing DNA segments, because we did not want the possible presence of mobile mRNA to bias our interpretation on the localization of the DNA segments. To conclude, we do not rule out the possibility of the presence of mRNA in the cells that do not contain the cognate segment, this is even how we think the system could work (discussed in the paper when proposing the conceptual model)

2) It is demonstrated that the function of a viral gene can act in a cell where the gene itself is not detected and these observations suggest that either the mRNA or the protein M-Rep itself can travel from the producing cells (those where segment R accumulates) to other cells of the host. Unfortunately, no evidence for such trafficking is given and thus, by the moment, is speculative. Authors must demonstrate that M-Rep behaves as a non-cellular autonomous protein or than its mRNA is able to traffic from cell to cell. Would the authors' result indicate that all the proteins encoded by FBNSV be non-cell-autonomous in their function or only some?

Indeed, we do not know whether it is the mRNA or the M-Rep protein itself that travels to non-producing cells. The direct demonstration of the movement of the mRNA or the protein M-Rep is clearly part of our future projects. Unfortunately, to develop approaches other than the detection of DNA or proteins proposed here, we need to rely on tools that are not available in faba bean. We can e.g. imagine transient expression of a fusion M-Rep-GFP in identified cells and monitor its traffic to neighboring cells, or express the corresponding mRNA with a root- or companion cell-specific promotor in transgenic plants and detect it in distant green tissues. In both cases the corresponding tools are available solely for model plant species (i.e. A. thaliana or other model plants) that are non-host to FBNSV and where nanovirus gene expression products may thus not be fully functional. All relevant approaches for such direct demonstration will engender months/years of additional work and are in the end extremely “high risk” because currently implementable only in host species or tissues that are not naturally invaded by nanoviruses. We are clearly eager to do this research effort, but not reasonably for the present submitted short report (also see additional experiments in response to comment point 5 below).

Most importantly, we cannot fully agree with the statement that the movement of either the mRNA or the M-Rep protein is speculative at this point and that we have no evidence for it. Though indirect, we do provide two strong pieces of evidence: i) the presence of the protein M-Rep in numerous cells where the R segment is not detected and, perhaps the more convincing, ii) the demonstration that the protein M-Rep is not more associated with cells accumulating its encoding segment R than with cells which do not accumulate R but other segments. To stress that this is indirect evidence, we have accordingly amended the Result section by adding a few words in the last sentence of the section: “Although they represent indirect evidence, these observations together further support our conclusion that either the mRNA or the protein M-Rep itself can travel from the producing cells (those where segment R accumulates) to other cells of the host, as discussed below.”

It is not clear to us whether the last question, “Would the authors' result indicate that all the proteins encoded by FBNSV be non-cell-autonomous in their function or only some”, refers to our argumentation that a function may be present in cells where the genetic information is absent, or whether the question is more specific and asks whether all proteins (or their corresponding mRNA) need to be self-mobile or not. Whatever the question is, we do not know the answer; we can reply however that it is not mandatory that all proteins (or mRNAs) are non-cell-autonomous, though obviously the alleviation of the within-host putative cost of multipartitism through the non-cell-autonomous functioning is positively correlated to the number of non-cell-autonomous segment’s functions. Concerning the more specific question of the potential self-mobility, without a precise description of the mode of action of all viral genes, it is impossible to predict whether each of them should encode a self-mobile function. If the effector of a viral function needs to co-act or physically interact with the effector of another viral function in individual cells, then at least one of them should be mobile (self-mobile or transported by a carrier). In contrast, if a viral gene function is the triggering of an endogenous systemic plant response (e.g. hormone signal) that can propagate in plant tissues, then the self-mobility of this viral gene effector may not be required. At this point, we have no elements to soundly speculate on these numerous possibilities.

3) In the Discussion it should be considered the tissue tropism of FBNSV, and the particularities of plasmodesmata between the different phloem cell types, which may limit the generalisation of the presented observations to other viral systems.

These three points can be easily addressed in a revised Discussion section. A modification of the title would also be advisable.

We are aware that the connections between companion cells (CC) and sieve elements (SE) are different than those between other tissues of the plant. The size exclusion limit (SEL) of the plasmodesmata (PD) is close to 60 KDa between CC and SE and potentially much smaller in other tissues [6,7]. All FBNSV protein monomers are small enough to freely traffic between CC and SE but this may not be true in other tissues. We thus understand the concern of the reviewers regarding the generalization of the presented observation, particularly to other multipartite viruses that are not phloem-limited and invade mesophyll and epidermis. Plants are capable of creating symplastic domains within which the cells are connected by widely opened plasmodesmata where large so-called non-cell-autonomous molecules and macromolecular complexes can freely traffic 7. These domains can appear and disappear during development and more specifically during organ formation and differentiation [7-9]. It is not impossible to imagine that some viruses could manipulate this plant inherent capacity and induce the appearance of symplastic domains where they could freely exchange viral products across distinct cells. Perhaps more directly relevant is the fact that all plant viruses produce movement proteins (MP) that could create symplastic domains. It is well and long established that viral MPs are non-cell-autonomous proteins that move away from the producing cells and enlarge the SEL of PD in mesophyll or epidermal cells [10]. This phenomenon has been frequently reviewed in the last decades and the movement profiles of MPs can even be quantified [11]. The point we want to make here is that, although we do not know how far our observations will be generalizable, there is no theoretical impediment to intercellular exchanges of viral material (proteins or mRNAs) in non-phloem tissues. At this stage, this possibility is pure speculation and this is why we initially chose not to discuss it. However, to comply with this comment we have inserted the following sentence in the discussion of the revised version: “Likewise, whether this non-cell-autonomous model can be extended to viral species invading non-phloem tissues where cell communication is more restricted is unknown. Non phloem-restricted viruses could induce the formation of symplastic domains, either by manipulating the endogenous capacity of the host plant to do so [7] or by opening plasmodesmata through the action of their non-cell-autonomous movement protein [10], but this possibility awaits further investigation.”

The last part of this reviewers’ comment is suggesting to also modify the title. While we agree to tone down our claims in several instances along the text (see our related modifications in response to the General comments above), we believe the title is adequately toned and we would like to maintain it as is. Indeed, it illustrates our observation and explicitly states that this observation is reported in one case, “for a multipartite virus”.

4) The experiments demonstrate that segments of a unique type accumulate preferentially in individual cells, where they are found jointly with the replicase protein but not necessarily with the replicase gene. This observation, however, does not justify that a sufficiently high MOI is not needed and that the viral system is functional through complementation across cells. The experiments reported indicate a sort of specialization of different genes, such that one type is preferentially expressed in each cell. This might be a winner-takes-all strategy that does not necessarily entail isolation of the individual genes upon infection. Task allocation is an often used strategy (e.g. in social insects), and it would be great if it would have been discovered by multipartite viruses in order to optimize segment production. But, again, it would not solve the high MOI requirement for between-host propagation.

The winner-takes-all strategy is proposed as an alternative view that we did not consider before, and we thank the reviewers for this comment. If we understand the comment properly the reviewer hypothesis is that the MOI could be high, with potentially several copies of each segment type entering within individual cells, but a ‘winner’ would replicate predominantly within each cell; the reviewer comment is not explicit on what the ‘winner’ would be and we see two possibilities that we discuss in turn: (i) the winner is the type of segment, e.g. R or S etc., and thus all molecules of this segment type initially present within a cell can replicate or (ii) the winner is a single molecule, of some segment type, that will engender all offspring within a cell through replication, other molecules of the same or different segment type initially present within this cell being left aside.

For the case (i) we ought to remind that all FBNSV segment types have a very conserved replication origin and that selectively replicating one segment type only in a cell, and another segment type only in another cell, is hardly conceivable (unless solely this type is present, as we propose). Moreover, our experiment with the genetic markers inserted in segment S or N pleads against this possibility: if the MOI was high, following the reviewers’ hypothesis, the observation that most cells accumulate only one of the alleles of each segment would not only imply the existence of a specific replication mechanism, but the specificity would need to operate at the allele and not the segment level; we fail to see any basis for such specificity and think that it is much more parsimonious to conclude, on the basis of our observations, that the MOI is low.

For the case (ii), assuming a high cellular MOI implies that many segments, even many copies of each segment type, are present and available in each cell but that a single one will efficiently replicate (the winner in the reviewers’ view). Again, all segments having the same replication origin we do not see how a single copy could be the only replication target, unless the M-Rep and replication complexes are in extremely limiting number, down to one. When looking at the distribution of M-Rep protein all over the nucleus (Figure 3), it is evident that many more than one copy is present. Nevertheless if we are to play an intellectual game, we could argue from our observation that the winner appears to be “designated” randomly: the segment type accumulating differs from cell to cell (Figure 1 and Figure 1—figure supplement 1) and the frequency of a segment in a tissue correlates with the number of cells where it is the winner (where it accumulates; Figure 2D). Would this be true, stochasticity in the initial stages of a plant infection would greatly determine the relative amount of the distinct segments accumulating later and so the genome formula. We have previously published that the genome formula is highly constrained to a reproducible pattern that is not dependent on the initial segment frequency, a fact that is not compatible with this speculation. This argument also applies to case (i) above.

In conclusion, whatever scenario we can think of for the winner-takes-all proposition, it is hard to reconcile it with what is observed on the biology of FBNSV both in this and in earlier published work.

The last sentence of this comment questions the transmission between hosts and is addressed in our response to the comment point 6 below.

5) An experiment was proposed which would clarify these issues: Prepare two bacteria with only one type of segment (say R and S, as in most experiments performed). Inoculate a leaf at increasing distance with ONE copy of each segment. Demonstrate that the S segment does replicate. This experiment would answer two main concerns of the reviewers regarding the claims made: (i) Even if individual segments propagate independently and are affected by severe bottlenecks (such that they reach the plant in very low numbers), replication of any segment is possible; (ii) The amount of replicase protein is high enough so as to reach cells far apart and cause the replication of other "waiting" segments before either of them degrades.

Despite our demonstration that segments can massively accumulate in cells where the segment R encoding the replication function is not detectable, that the intensity of replication of segments S and/or M does not correlate with the amount of segment R accumulating in the same cell, that the protein M-Rep accumulates in cells where its encoding segment is not detectable, and that the protein M-Rep does not colocalize more with its encoding segment than with another segment (S), we understand from this comment that the reviewers would appreciate an additional proof of the absence of R in virus replication competent cells.

We considered this comment/suggestion extremely carefully. We have accordingly performed an additional experiment (see Author response image 1) and, unfortunately, it confirms that the approach proposed by the reviewers is not feasible. There were several a priori drawbacks for the suggested experiment:

– It is totally out of the virus infection context and so interpretation could be questionable

– The products of segments other than S and R could be mandatory for the trafficking process

– The distance at which across-cell complementation is efficient could be too small for this experiment

– For unknown reasons, FBNSV clone is infectious solely when agro-infiltrated in the stem, not in leaves

– Wounding in the infiltrated zone can induce auto-fluorescence

– The agrobacterium plasmid with inserted viral sequences could be present in the observed cell

– We have no idea of the plasmid copy number and related artefactual fluorescence in observed cells

– We have no idea of the number of cells initially agro-inoculated and/or infected, and whether they would be too rare to be efficiently retrieved.

To collectively evaluate all these potential drawbacks, we decided to simply agro-inoculate faba bean stems as described for plant infection (with agro-bacteria carrying all 8 segments). This way, we are sure the viral infection is initiated in the inoculated region, which we sectioned, fixed, labeled and observed in search for fluorescent S (green) and/or R (red) signal. Our reasoning was that if no fluorescence could be detected in the nucleus of individual cells of these agro-infiltrated stem-tissues, it would be evidence that the approach proposed cannot work. Author response image 1 shows wound-related auto-fluorescence (Panel b), attesting that we are in the relevant inoculated area. Panel (c), (d) and (e) show typical microscopy fields including phloem bundles where no fluorescent cells could be observed, whether at 3, 5 or 10 days post-inoculation, and despite observing 49, 25 and 37 sections for each date, respectively. Panel (a) shows a similar experiment but in systemically infected tissues (just as in Figure 1 of the manuscript), controlling for the quality of our fluorescent probes. In this control case (in 15 sections), 41 fluorescent cells could be detected: 16 green only, 20 red only, and 5 green/red. We conclude that the cells where the infection initially starts are much too rare to be retrieved in panels (c)(d)(e), like looking for a needle in a haystack. Should we observe hundreds of sections from stems inoculated as suggested by the reviewers, we may eventually find a couple of fluorescent cells (or not), but would such rare occurrence be more convincing than all results shown in the paper in systemically infected tissues?

Author response image 1
Attempt at detecting viral segments S and R in agro-infiltrated region of the stem.

Trunks of faba bean stems agro-infiltrated with a mixture of agro-bacteria containing the eight FBNSV segments were sectioned and FISH-processed exactly as described in the Materials and methods section. Green-labeled S segments and Red-labeled R segments are visible solely in the petiole sections from systemically infected upper part of the plant (a) and not in the agro-infiltrated section of the stem (b-e). The wound (w) inflicted by the agro-infiltration process induces autofluorescence of the plant tissue that appears as a line of green and red fluorescence at the boundary of the wound (b). No fluorescent cells could be observed at three distinct dates, 3, 5 or 10 dpi, for which representative images are respectively shown in (c), (d) and (e). Horizontal bar is 50 micrometers.

6) Regarding the relationship between the experiments reported and the MOI required upon between-host transmission, two processes are apparently mixed here. The first one is between-plant transmission: as discussed above, this process requires an MOI so high for multipartite viruses with several segments that it has not been observed in nature: It has to be pointed out that the MOI as described e.g. in 5 has never been empirically quantified. Actually, what studies show is that the number of founding particles for the population infecting a given individual plant is low. But this is a number related to genealogy, not to transmission. (An analogous situation is population bottlenecks: the ancestry of a current population can be traced back to a few, even a single, individual, this fact not implying that the ancestral population was that small. Iranzo and Manrubia, 2012, concerns the size of the ancestral population, not the founding population.) The second one is the (theoretical and empirical) search of possible mechanisms that can alleviate that hypothetical high MOI. It has been put forward, with good empirical support, that viral particles might form groups or clusters upon propagation (and this might be relatively common).

We are not sure we correctly understand this comment, which discusses the MOI upon between host transmission, a question that we do not address in this manuscript. Perhaps the wording in the Discussion section was not clear enough on this point and we clarified with an explicit sentence in the corresponding lines: “Here we introduce an additional concept in virology, which is compatible with empirical observations and which partially alleviates the insurmountable cost in highly multipartite viral systems such as FBNSV: because concomitant infection of individual cells by all genomic segments is not necessary, the associated putative cost should be much smaller if not nil at the within host level. We earlier discussed the fact that the analogous cost upon between-host transmission and the mechanisms of its compensation remain to be understood [12].”

Despite this not being the object of the current submitted manuscript, concerning the specific comment that “this is a number related to genealogy, not to transmission”, we wish to mention that one of our estimation methods in a recent paper discussing the plant-to-plant transmission of FBNSV by aphid vectors 12 relies entirely on transmission, and not at all on genealogy, and nevertheless indicates a relatively low founding population when a plant is inoculated by a single or by ten aphids. So it is possible that there is a small MOI in between hosts and, as discussed in [12], we do not know yet how the virus efficiently overcomes this problem.

Regarding the two last sentences of this comment, our results indicate that within host high MOI is not required for FBNSV because the segments do not need be together in every infected individual cell for the viral system to be functional. We demonstrate that the accumulation of one segment is independent of the accumulation of the others. Thus, observations in our viral system do not provide any support for (even go against) the hypothesis that groups/clusters of viral particles propagate ensuring that all segments always co-infect individual cells.

7) Finally, there should be a discussion of the cost of producing very high amounts of mRNA and/or replicase protein that will be lost (as demonstrated by the experiments), if the interpretation of the authors proves to be correct.

This comment is based on assumptions from the reviewers that are questionable:

First, our interpretation is that most susceptible cells where the mRNA/protein M-Rep will move can potentially contain a FBNSV segment and so does not necessarily induce a waste.

Second, producing viral material that will be lost is not an issue that is specifically relevant to our study, and it may even not be relevant in virology. Indeed, all viruses “waste” a huge proportion of whatever viral material is produced in the infected host. If we adopt the view expressed in this comment, wasting appears as the general lifestyle of viruses. For example, viruses build up populations of billions of particles within hosts, but only a few units are transmitted to next hosts in most cases [13]. Infected cells (whatever bacterial, animal or plant cells), in addition to numerous virus particles, also accumulate inclusion bodies that are often considered as storage sites for excess material, and this is left behind just as most of the virions produced which will not infect other cells.

Third, whether and how all these “wastes” translate as a cost for the virus is unclear because they are produced by the infected host. But this is a debate that is very far from the message of our paper and we would like to avoid entering it.

In summary, if agreed with the editors and reviewers, we would prefer not to discuss this point, because i) we are not sure our interpretation generates more waste than any other interpretation, ii) it is not a question addressed in the paper, iii) this question is not more acute with the here-reported peculiar FBNSV way of life than with any other described viral cycle.

List of cited references:

1 Winter, N. & Kragler, F. Conceptual and Methodological Considerations on mRNA And Proteins as Intercellular and Long-distance Signals. Plant & cell physiology, doi:10.1093/pcp/pcy140 (2018).

2 Reagan, B. C., Ganusova, E. E., Fernandez, J. C., McCray, T. N. & Burch-Smith, T. M. RNA on the move: The plasmodesmata perspective. Plant Sci275, 1-10, doi:10.1016/j.plantsci.2018.07.001 (2018).

3 Liu, L. & Chen, X. Intercellular and systemic trafficking of RNAs in plants. Nat Plants4, 869-878, doi:10.1038/s41477-018-0288-5 (2018).

4 Kehr, J. & Kragler, F. Long distance RNA movement. The New phytologist218, 29-40, doi:10.1111/nph.15025 (2018).

5 Gutierrez, S. et al. The Multiplicity of Cellular Infection Changes Depending on the Route of Cell Infection in a Plant Virus. J Virol89, 9665-9675, doi:10.1128/JVI.00537-15 (2015).

6 Knox, K. & Oparka, K. Illuminating the translocation stream. Current opinion in plant biology43, 113-118, doi:10.1016/j.pbi.2018.03.009 (2018).

7 Faulkner, C. Plasmodesmata and the symplast. Curr Biol28, R1374-R1378, doi:10.1016/j.cub.2018.11.004 (2018).

8 Otero, S., Helariutta, Y. & Benitez-Alfonso, Y. Symplastic communication in organ formation and tissue patterning. Current opinion in plant biology29, 21-28, doi:10.1016/j.pbi.2015.10.007 (2016).

9 Tilsner, J., Nicolas, W., Rosado, A. & Bayer, E. M. Staying Tight: Plasmodesmal Membrane Contact Sites and the Control of Cell-to-Cell Connectivity in Plants. Annual review of plant biology67, 337-364, doi:10.1146/annurev-arplant-043015-111840 (2016).

10 Lucas, W. J. Plant viral movement proteins: agents for cell-to-cell trafficking of viral genomes. Virology344, 169-184, doi:10.1016/j.virol.2005.09.026 (2006).

11 Trutnyeva, K., Ruggenthaler, P. & Waigmann, E. Movement profiles: a tool for quantitative analysis of cell-to-cell movement of plant viral movement proteins. Methods in molecular biology (Clifton, N.J451, 317-329, doi:10.1007/978-1-59745-102-4_22 (2008).

12 Gallet, R. et al. Small Bottleneck Size in a Highly Multipartite Virus during a Complete Infection Cycle. J Virol92, doi:10.1128/JVI.00139-18 (2018).

13 Gutierrez, S., Michalakis, Y. & Blanc, S. Virus population bottlenecks during within-host progression and host-to-host transmission. Current opinion in virology2, 546-555 (2012).

https://doi.org/10.7554/eLife.43599.018

Article and author information

Author details

  1. Anne Sicard

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Conceptualization, Data curation, Investigation, Visualization, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  2. Elodie Pirolles

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  3. Romain Gallet

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Formal analysis, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Marie-Stéphanie Vernerey

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Data curation, Investigation, Methodology
    Competing interests
    No competing interests declared
  5. Michel Yvon

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
  6. Cica Urbino

    1. BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    2. CIRAD, BGPI, INRA, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Supervision, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
  7. Michel Peterschmitt

    1. BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    2. CIRAD, BGPI, INRA, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Investigation, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  8. Serafin Gutierrez

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Conceptualization, Supervision, Investigation, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  9. Yannis Michalakis

    MIVEGEC, CNRS, IRD, Université de Montpellier, Montpellier, France
    Contribution
    Conceptualization, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing—original draft, Project administration
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1929-0848
  10. Stéphane Blanc

    BGPI, INRA, CIRAD, Montpellier SupAgro, Université de Montpellier, Montpellier, France
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    stephane.blanc@inra.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3412-0989

Funding

Institut National de la Recherche Agronomique

  • Anne Sicard
  • Elodie Pirolles
  • Romain Gallet
  • Marie-Stéphanie Vernerey
  • Michel Yvon
  • Serafin Gutierrez
  • Stéphane Blanc

Centre National de la Recherche Scientifique

  • Elodie Pirolles
  • Yannis Michalakis

Institut de Recherche pour le Développement

  • Yannis Michalakis

Agence Nationale de la Recherche (ANR-14-CE02-0014)

  • Anne Sicard
  • Elodie Pirolles
  • Romain Gallet
  • Marie-Stéphanie Vernerey
  • Michel Yvon
  • Yannis Michalakis

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We are grateful to Sophie Leblaye for greenhouse work, plant production and virus inoculation. We also thank T Timchenko, B Gronenborn and J Vetten for kindly providing the antibody directed against the viral protein M-Rep. This work was supported by INRA dpt. SPE and by the grant ‘Nano’ N° ANR-14-CE02-0014 from the French national funding agency (ANR). MP, CU and SG acknowledge support from CIRAD. YM acknowledges support from IRD and CNRS.

Senior Editor

  1. Detlef Weigel, Max Planck Institute for Developmental Biology, Germany

Reviewing Editor

  1. Fernando García-Arenal, Technical University of Madrid, Spain

Reviewer

  1. Fernando García-Arenal, Technical University of Madrid, Spain

Publication history

  1. Received: November 13, 2018
  2. Accepted: February 26, 2019
  3. Accepted Manuscript published: March 12, 2019 (version 1)
  4. Version of Record published: March 12, 2019 (version 2)

Copyright

© 2019, Sicard et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 13,818
    Page views
  • 1,058
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Microbiology and Infectious Disease
    Emma J Gehrke et al.
    Research Article
    1. Microbiology and Infectious Disease
    Pratibha C Koneru et al.
    Research Article Updated