Birnaviral Hijacking of Endosomal Membranes

Reviewer #1 (Public Review):

Summary:

Zanetti et al. use biophysical and cellular assays to investigate the interaction of the birnavirus VP3 protein with the early endosome lipid PI3P. The major novel finding is that the association of the VP3 protein with an anionic lipid (PI3P) appears to be important for viral replication, as evidenced through a cellular assay on FFUs.

Strengths:

Supports previously published claims that VP3 may associate with early endosomes and bind to PI3P-containing membranes. The claim that mutating a single residue (R200) critically affects early endosome binding and that the same mutation also inhibits viral replication suggests a very important role for this binding in the viral life cycle.

Weaknesses:

The manuscript is relatively narrowly focused: one bimolecular interaction between a host cell lipid and one protein of an unusual avian virus (VP3-PI3P). Aspects of this interaction have been described previously. Additional data would strengthen claims about the specificity and some technical issues should be addressed. Many of the core claims would benefit from additional experimental support to improve consistency.

Reviewer #2 (Public Review):

Summary:

Birnavirus replication factories form alongside early endosomes (EEs) in the host cell cytoplasm. Previous work from the Delgui lab has shown that the VP3 protein of the birnavirus strain infectious bursal disease virus (IBDV) interacts with phosphatidylinositol-3-phosphate (PI3P) within the EE membrane (Gimenez et al., 2018, 2020). Here, Zanetti et al. extend this previous work by biochemically mapping the specific determinants within IBDV VP3 that are required for PI3P binding in vitro, and they employ in silico simulations to propose a biophysical model for VP3-PI3P interactions.

Strengths:

The manuscript is generally well-written, and much of the data is rigorous and solid. The results provide deep knowledge into how birnaviruses might nucleate factories in association with EEs. The combination of approaches (biochemical, imaging, and computational) employed to investigate VP3-PI3P interactions is deemed a strength.

Weaknesses:

(1) Concerns about the sources, sizes, and amounts of recombinant proteins used for co-flotation: Figures 1A, 1B, 1G, and 4A show the results of co-flotation experiments in which recombinant proteins (control His-FYVE v. either full length or mutant His VP3) were either found to be associated with membranes (top) or non-associated (bottom). However, in some experiments, the total amounts of protein in the top + bottom fractions do not appear to be consistent in control v. experimental conditions. For instance, the Figure 4A western blot of His-2xFYVE following co-flotation with PI3P+ membranes shows almost no detectable protein in either top or bottom fractions. Reading the paper, it was difficult to understand which source of protein was used for each experiment (i.e., E. coli or baculovirus-expressed), and this information is contradicted in several places (see lines 358-359 v. 383-384). Also, both the control protein and the His-VP3-FL proteins show up as several bands in the western blots, but they don't appear to be consistent with the sizes of the proteins stated on lines 383-384. For example, line 383 states that His-VP3-FL is ~43 kDa, but the blots show triplet bands that are all below the 35 kDa marker (Figures 1B and 1G). Mass spectrometry information is shown in the supplemental data (describing the different bands for His-VP3-FL) but this is not mentioned in the actual manuscript, causing confusion. Finally, the results appear to differ throughout the paper (see Figures 1B v. 1G and 1A v. 4A).

(2) Possible "other" effects of the R200D mutation on the VP3 protein. The authors performed mutagenesis to identify which residues within patch 2 on VP3 are important for association with PI3P. They found that a VP3 mutant with an engineered R200D change (i) did not associate with PI3P membranes in co-floatation assays, and (ii) did not co-localize with EE markers in transfected cells. Moreover, this mutation resulted in the loss of IBDV viability in reverse genetics studies. The authors interpret these results to indicate that this residue is important for "mediating VP3-PI3P interaction" (line 211) and that this interaction is essential for viral replication. However, it seems possible that this mutation abrogated other aspects of VP3 function (e.g., dimerization or other protein/RNA interactions) aside from or in addition to PI3P binding. Such possibilities are not mentioned by the authors.

(3) Interpretations from computational simulations. The authors performed computational simulations on the VP3 structure to infer how the protein might interact with membranes. Such computational approaches are powerful hypothesis-generating tools. However, additional biochemical evidence beyond what is presented would be required to support the authors' claims that they "unveiled a two-stage modular mechanism" for VP3-PI3P interactions (see lines 55-59). Moreover, given the biochemical data presented for R200D VP3, it was surprising that the authors did not perform computational simulations on this mutant. The inclusion of such an experiment would help tie together the in vitro and in silico data and strengthen the manuscript.

Reviewer #3 (Public Review):

Summary:

infectious bursal disease virus (IBDV) is a birnavirus and an important avian pathogen. Interestingly, IBDV appears to be a unique dsRNA virus that uses early endosomes for RNA replication that is more common for +ssRNA viruses such as for example SARS-CoV-2.

This work builds on previous studies showing that IBDV VP3 interacts with PIP3 during virus replication. The authors provide further biophysical evidence for the interaction and map the interacting domain on VP3.

Strengths:

Detailed characterization of the interaction between VP3 and PIP3 identified R200D mutation as critical for the interaction. Cryo-EM data show that VP3 leads to membrane deformation.

Weaknesses:

The work does not directly show that the identified R200 residues are directly involved in VP3-early endosome recruitment during infection. The majority of work is done with transfected VP3 protein (or in vitro) and not in virus-infected cells.

Additional controls such as the use of PIP3 antagonizing drugs in infected cells together with a colocalization study of VP3 with early endosomes would strengthen the study.

In addition, it would be advisable to include a control for cryo-EM using liposomes that do not contain PIP3 but are incubated with HIS-VP3-FL. This would allow ruling out any unspecific binding that might not be detected on WB.

The authors also do not propose how their findings could be translated into drug development that could be applied to protect poultry during an outbreak. The title of the manuscript is broad and would improve with rewording so that it captures what the authors achieved.

Author response:

eLife assessment

This study presents valuable information on the mechanism of how birnavirus VP3 protein interacts with PI3P in early endosomes. Evidence supporting the proposed two-stage mechanism is incomplete and would benefit from additional supporting experiments, and additional experimentation would also address concerns about data consistency.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Zanetti et al. use biophysical and cellular assays to investigate the interaction of the birnavirus VP3 protein with the early endosome lipid PI3P. The major novel finding is that the association of the VP3 protein with an anionic lipid (PI3P) appears to be important for viral replication, as evidenced through a cellular assay on FFUs.

Strengths:

Supports previously published claims that VP3 may associate with early endosomes and bind to PI3P-containing membranes. The claim that mutating a single residue (R200) critically affects early endosome binding and that the same mutation also inhibits viral replication suggests a very important role for this binding in the viral life cycle.

Weaknesses:

The manuscript is relatively narrowly focused: one bimolecular interaction between a host cell lipid and one protein of an unusual avian virus (VP3-PI3P). Aspects of this interaction have been described previously. Additional data would strengthen claims about the specificity and some technical issues should be addressed. Many of the core claims would benefit from additional experimental support to improve consistency.

We focused our efforts on the characterization of the molecular interaction between the birnaviral protein VP3 and the anionic lipid PI3P, which is found in the host cell. This decision was motivated by our previous research, which made use of cell biology and virology techniques to demonstrate that VP3 facilitates the formation of the viral replication machinery on the cytosolic leaflet of early endosomes due to its inherent endosome-targeting capability (J Virol. 2018 May 14;92(11):e01964-17). Additionally, our previous findings indicated that PI3P, present in early endosomal membranes, is a critical host factor enabling VP3's association with these membranes, thereby promoting viral replication (J Virol. 2021 Feb 24;95(6):e02313-20). Consequently, an in-depth characterization of the VP3/PI3P interaction was necessary and motivated the present work. We plan to incorporate specific recommendations to further substantiate our assertions in the revised version of our manuscript.

Reviewer #2 (Public Review):

Summary:

Birnavirus replication factories form alongside early endosomes (EEs) in the host cell cytoplasm. Previous work from the Delgui lab has shown that the VP3 protein of the birnavirus strain infectious bursal disease virus (IBDV) interacts with phosphatidylinositol-3-phosphate (PI3P) within the EE membrane (Gimenez et al., 2018, 2020). Here, Zanetti et al. extend this previous work by biochemically mapping the specific determinants within IBDV VP3 that are required for PI3P binding in vitro, and they employ in silico simulations to propose a biophysical model for VP3-PI3P interactions.

Strengths:

The manuscript is generally well-written, and much of the data is rigorous and solid. The results provide deep knowledge into how birnaviruses might nucleate factories in association with EEs. The combination of approaches (biochemical, imaging, and computational) employed to investigate VP3-PI3P interactions is deemed a strength.

Weaknesses:

(1) Concerns about the sources, sizes, and amounts of recombinant proteins used for co-flotation: Figures 1A, 1B, 1G, and 4A show the results of co-flotation experiments in which recombinant proteins (control His-FYVE v. either full length or mutant His VP3) were either found to be associated with membranes (top) or non-associated (bottom). However, in some experiments, the total amounts of protein in the top + bottom fractions do not appear to be consistent in control v. experimental conditions. For instance, the Figure 4A western blot of His-2xFYVE following co-flotation with PI3P+ membranes shows almost no detectable protein in either top or bottom fractions.

Liposome-based methods, such as the co-flotation assay, are well-known and preferred to study protein-phosphoinositide interaction because the phosphoinositides are incorporated in a membrane, the composition of which can mimic cellular membranes. Additionally, by modifying the phosphoinositide incorporated in the liposomes, this technique allows for determining the specificity of the protein binding. However, this approach is rather qualitative, meaning that, after density gradient separation, the protein is found in the top fractions (bound to liposomes) or in the bottom fractions (not bound to liposomes), and our quantifications have the aim of showing the difference in the bound fraction between liposome populations with or without PI3P. Given the setting of the co-flotation assays, each protein-liposome system [2xFYVE-PI3P(-), 2xFYVE-PI3P(+), VP3-PI3P(-), or VP3-PI3P(+)] is assessed separately, and even if the conditions are homogeneous, it’s not surprising to observe differences in the protein level between each one. Indeed, our revised version of the manuscript will include membranes with more similar band intensities.

Reading the paper, it was difficult to understand which source of protein was used for each experiment (i.e., E. coli or baculovirus-expressed), and this information is contradicted in several places (see lines 358-359 v. 383-384). Also, both the control protein and the His-VP3-FL proteins show up as several bands in the western blots, but they don't appear to be consistent with the sizes of the proteins stated on lines 383-384. For example, line 383 states that His-VP3-FL is ~43 kDa, but the blots show triplet bands that are all below the 35 kDa marker (Figures 1B and 1G). Mass spectrometry information is shown in the supplemental data (describing the different bands for His-VP3-FL) but this is not mentioned in the actual manuscript, causing confusion. Finally, the results appear to differ throughout the paper (see Figures 1B v. 1G and 1A v. 4A).

We used two sources of recombinant VP3: baculovirus and Escherichia coli. Initially, we opted for the baculovirus system based on evidence from previous studies that it was suitable for ectopic expression of VP3. Subsequently, we successfully produced VP3 using Escherichia coli and chose to transition to this system due to several technical advantages. Moreover, mass spectrometry analysis did not reveal any post-translational modifications that may have favored retaining the baculoviral system. We confirmed that VP3, produced in either system, exhibited similar behavior in our co-flotation assays. We will clarify all this in the revised version of our manuscript.

(2) Possible "other" effects of the R200D mutation on the VP3 protein. The authors performed mutagenesis to identify which residues within patch 2 on VP3 are important for association with PI3P. They found that a VP3 mutant with an engineered R200D change (i) did not associate with PI3P membranes in co-floatation assays, and (ii) did not co-localize with EE markers in transfected cells. Moreover, this mutation resulted in the loss of IBDV viability in reverse genetics studies. The authors interpret these results to indicate that this residue is important for "mediating VP3-PI3P interaction" (line 211) and that this interaction is essential for viral replication. However, it seems possible that this mutation abrogated other aspects of VP3 function (e.g., dimerization or other protein/RNA interactions) aside from or in addition to PI3P binding. Such possibilities are not mentioned by the authors.

The arginine amino acid at position 200 of VP3 is not located in any of the protein regions associated with its other known functions. VP3 has a dimerization domain located in the second helical domain, where different amino acids across the three helices form a total of 81 interprotomeric close contacts; however, R200 is not involved in these contacts (Structure. 2008 Jan;16(1):29-37). VP3 also has an oligomerization domain mapped within the 42 C-terminal residues of the polypeptide, i.e., the segment of the protein composed by the residues at positions 216-257 (J Virol. 2003 Jun;77(11):6438–6449). Regarding VP3’s ability to bind RNA, it is facilitated by a region of positively charged amino acids, identified as P1, which includes K99, R102, K105, and K106 (PLoS One. 2012;7(9):e45957). Furthermore, our findings indicate that the R200D mutant retains a folding pattern similar to the wild-type protein, as shown in Figure 4B. All these lead us to conclude that the loss of replication capacity of R200D viruses results from impaired, or even lost, VP3-PI3P interaction.

(3) Interpretations from computational simulations. The authors performed computational simulations on the VP3 structure to infer how the protein might interact with membranes. Such computational approaches are powerful hypothesis-generating tools. However, additional biochemical evidence beyond what is presented would be required to support the authors' claims that they "unveiled a two-stage modular mechanism" for VP3-PI3P interactions (see lines 55-59). Moreover, given the biochemical data presented for R200D VP3, it was surprising that the authors did not perform computational simulations on this mutant. The inclusion of such an experiment would help tie together the in vitro and in silico data and strengthen the manuscript.

We acknowledge that the language used may have overstated the "unveiling" of the two-stage binding mechanism for VP3 on membranes containing PI3P. We intended to propose, rather than confirm, this mechanism, largely based on our coarse-grained simulations. Accordingly, we will revise the manuscript to temper our claims and frame them more appropriately. Regarding the absence of computer simulations for the R200D VP3 mutant, these were indeed conducted, and the results are detailed in Figure 14 of the supplementary material. We realize this was not adequately emphasized in the main manuscript, an oversight we will correct in the revised version.

Reviewer #3 (Public Review):

Summary:

Infectious bursal disease virus (IBDV) is a birnavirus and an important avian pathogen. Interestingly, IBDV appears to be a unique dsRNA virus that uses early endosomes for RNA replication that is more common for +ssRNA viruses such as for example SARS-CoV-2.

This work builds on previous studies showing that IBDV VP3 interacts with PIP3 during virus replication. The authors provide further biophysical evidence for the interaction and map the interacting domain on VP3.

Strengths: Detailed characterization of the interaction between VP3 and PIP3 identified R200D mutation as critical for the interaction. Cryo-EM data show that VP3 leads to membrane deformation.

Weaknesses:

The work does not directly show that the identified R200 residues are directly involved in VP3-early endosome recruitment during infection. The majority of work is done with transfected VP3 protein (or in vitro) and not in virus-infected cells. Additional controls such as the use of PIP3 antagonizing drugs in infected cells together with a colocalization study of VP3 with early endosomes would strengthen the study. In addition, it would be advisable to include a control for cryo-EM using liposomes that do not contain PIP3 but are incubated with HIS-VP3-FL. This would allow ruling out any unspecific binding that might not be detected on WB.

The authors also do not propose how their findings could be translated into drug development that could be applied to protect poultry during an outbreak. The title of the manuscript is broad and would improve with rewording so that it captures what the authors achieved.

In previous works from our group, we demonstrated the crucial role of the VP3 P2 region in targeting the early endosomal membranes and for viral replication, including the use of PI3K inhibitors to deplete PI3P, showing that both the control RFP-2xFYVE and VP3 lost their ability to associate with the early endosomal membranes (J Virol. 2018 May 14;92(11):e01964-17; J Virol. 2021 Feb 24;95(6):e02313-20). In the present work, to further characterize the role of R200 in binding to early endosomes and for viral replication, we show that: i) the transfected VP3 R200D protein loses the ability to bind to early endosomes in immunofluorescence assays (Figure 2E and Figure 3); ii) the recombinant VP3 R200D protein loses the ability to bind to liposomes PI3P(+) in co-flotation assays (Figure 4A); and, iii) the mutant virus R200D loses replication capacity (Figure 4C).

Regarding the cryo-EM comment: we will include images where we used liposomes PIP3(-) in the revised version of our manuscript.

We will also modify the title of the manuscript.

Regarding the question of how our findings could be translated into drug development, indeed, VP3-PI3P binding constitutes a good target for drugs that counteract infectious bursal disease. However, we did not mention this idea in the manuscript, first because it is somewhat speculative and second because infected farms do not implement any specific treatment. The control is based on vaccination. We will mention these aspects of the infection in the revised version of our manuscript.