Talk:PLOS/Viral quasispecies

Review by Pilar Domingo-Calap

In the present work, Esteban Domingo and Celia Perales review viral quasispecies. The manuscript is well written and easy to read, highlighting interesting points of the theory and including applications and perspectives. The article sections are well conceived, and provide a useful tool for better understanding viral populations and evolution.

I have some suggestions that, in my opinion, could help the authors improve some sections.

1. As the authors mention, in principle quasispecies applies to any biological entity. However, the manuscript is essentially focused on RNA viruses, which is understandable. Yet, in my opinion, it would be worth discussing to what extent quasispecies apply to some DNA viruses. Although single stranded DNA viruses are included in Table 1, some more discussion would be welcome. In addition, small double stranded DNA viruses code for proteins that block proofreading-repair activity, and they also evolve fast. We showed this in patients recently (Domingo-Calap et al. 2018 Plos Pathogens PMID:30335851), and others have provided similar evidence.

2. The authors cite their own mutation rate estimate of Qbeta, but this has been revisited in more recent work by our own group (Bradwell et al. 2013 Genetics PMID:23852383).

3. On page 6, the genome size range of RNA viruses actually starts at 1.8Kbp (Saccharomyces cerevisiae Killer Virus M1, Campillo-Balderas et al. 2015 Front. Ecol. Evol. doi.org/10.3389/fevo.2015.00143).

4. On page 8, I think convergent evolution should be discussed, since this provides evidence for a deterministic mode of evolution that is well compatible with quasispecies.

5. On page 10, reference 56 is an silico work. The survival of the flattest was subsequently shown experimentally at least twice, by Codoñer et al. 2006 (Plos Pathogens PMID:17196038) and by Sanjuán et al. 2007 (PloS Genetics PMID: 17571922).

6. On page 13, after the sentence “Intra-host evolution is generally more rapid than interhost evolution”, I would add and discuss the reference by Ho et al. 2011 (Molecular Ecology PMID:21740474) about the time-dependence of evolutionary rates.

7. In the “Summary and prospects” section, I suggest introducing a short paragraph with the main ideas and conclusions of the theory from a basic point of view.

Overall, thus, I think that a more balanced overview of the literature would help improve the article, particularly since this is a review paper aimed at a non-specialist readership.

A few typos:

• Page 11, “cooperation” instead of “coopoeration”.
• Page 15, “antiviral” instead of “antviral”.
• Page 5, a parenthesis is missing at the end of “Dynamic heterogeneity” section.

Response to Dra. Domingo-Calap

1) We intended our review to be conceptual and related to mutant spectrum implications. The systems used to develop quasispecies dynamics and its biological implications are mainly RNA viruses. However the reviewer is correct that the concept should be valid for any system that develops mutant spectra. In consequence, we have expanded the first paragraph of the section “Dynamic heterogeneity” to read: “Quasispecies dynamics will operate in any viral or cellular system in which due to high mutation rates (as a result of low fidelity polymerases or environmental alterations) mutant spectra are rapidly generated [22 of the previous version, Shmidt et al. 2017; Takahashi et al. 2016; Domingo-Calap et al. 2018; Sanchez-Campos et al. 2018] (lines 131-133 of the highlighted version).

2) The reason to quote the mutation rate calculated for bacteriophage Qβ in 1976 is because it was the first mutation rate calculated for an RNA virus and its value was integrated in the first description of quasispecies dynamics. We do not want our review to be on mutation rates of viruses. However, to respond to this request, we have expanded the third paragraph of page 3 to read as follows: “The error rate estimated for bacteriophage Qβ has been confirmed, and is comparable to values calculated for other RNA viruses (Drake and Holland, P.N.A.S. 96:13910-13913, 1999; Bradwell et al. Genetics 195: 243-251, 2013)” (lines 73-74 of the highlighted version).

3) The genome size has been expanded to read: “1.8 to 33 Kb”, as requested (line 180 of the highlighted version).

4) We do not see a connection between convergent evolution and quasispecies that justifies that we address this point. Convergent evolution, at least in its classical sense, refers to long term evolution of initially distinct lineages towards similar sequences or structures. Quasispecies deals with a population structure that is relevant to viral adaptability, pathogenesis, and short-term evolution. The only connection we could force between convergent evolution and quasispecies is that a mutant spectrum can converge to give the same consensus sequence. Except for this, we do not see the need to refer to convergent evolution that would take us away from the focus of our article and may even be confusing to the reader. To provide a specific example of possible confusion, in the process of virus extinction by lethal mutagenesis the consensus sequence remains constant, and, in our view, it would be incorrect to attribute such constancy to convergent evolution, at least according to the models of lethal mutagenesis that fit best the experimental observations. Equally complex would be to raise the deterministic versus stochastic quasispecies evolution beyond what we already wrote (lines 94-102 of the highlighted version).

5) Again, we did not want to extend the concept of advantage to the flattest. However, in view of the request of Dr. Domingo-Calap, we have added three additional references, two that preceded the concepts expressed in Wilke et al. 2001, and the reference to Codoñer et al. 2006 suggested by the reviewer (lines 328-329 of the highlighted version). We have not included the reference to Sanjuán et al. 2007 because its claim (last sentence of the Abstract of that paper) that the study revealed a mechanism of resistance to lethal mutagenesis was disproved in subsequent publications of ours (Martin et al. Virology 378:185-92, 2008) and others (O’Dea et al Plos Computational Biology 6: e1000811, 2009). No evidence of the proposal by Sanjuán et al. of mutagenesis-induced robustness as a mechanism of failure to lethal mutagenesis has been obtained, to our knowledge, during more than a decade of work on lethal mutagenesis. Inclusion of the reference to Sanjuán et al. would force us to extend the discussion of a point which is a side issue relative to the quasispecies core, and would increase even more the number of references.

6) Although we intended to restrict the time-dependent effects to the case of viruses, in response to this request, we have added a reference by Ho et al. on the same topic and journal, but from 2015 (rather than 2011) because it includes updated information (Ho SY, Duchene S, Molak M, Shapiro, B. 2015. Time dependent estimates of molecular evolutionary rates. Evidence and causes. Mol. Ecol. 24(2): 6007-6012). The penultimate sentence of the section “Bottlenecks: weakening, liberation, or challenge?” has been extended to read: “…as documented with viruses [Domingo et al. 2012] and other biological systems [Ho et al. 2015] (line 413 of the highlighted version).

7) We have added an introductory statement in the “Summary and prospects” section to read: The main concepts covered in the present article and their domains of applicability are summarized in Table 1. The adequacy of quasispecies theory (versus other formulations of evolutionary dynamics [Page and Nowak, J. Theor. Biol. 219, 93-8, 2002]) as a framework for the error-prone replication of viruses and its consequences stems from its including mutation as an integral part of the replication process [13] (lines 520 to 524 of the highlighted version).

The main criticism of Dr. Domingo-Calap is that our review did not have a balanced overview of the literature. In the modified version we have introduced four of the references that she has suggested and that belong to work carried out by her group or the group of Rafael Sanjuán, despite such papers not mentioning the terms “quasispecies” or “mutant spectra”.

The typos have been corrected.

Review by Fabio Luciani

Overall, this short review on viral quasi species is a high level summary of published work on viral quasi species, nicely report key historical facts that are often neglected. I do think however that the authors should rewrite several sections of this review keeping in mind that the reader may get annoyed of reading each second paragraph without a reference and just a series of sentences about “features or facts “ of quasi species, without any clear context, any clear paper cited and just a remind to a review or two (usually of the same authors)

1) References need a thorough re-asessment. Several key references are missing and there are several paragraphs without actually pointing at a previous work but simply referring to the same review of the authors (E.g Ref 34)

2) I have some issue with them definition of "quasispecies memory” The concept of viral identity that persist despite the specific distribution of mutants , it is true . However as the authors say this is generally a function of the selective force that the host impose. As the authors mention the immune response to viral infection such as in HIV and HCV. I think the authors nee dot make clear that this “memory” is actually a memory of the host as well and not of the quasi species. The Tc ell for instance also have a memory, and the host genomics also affect “memory" Another view of this memory is simply that adaptation shapes almost deterministically the distribution and evolution of quasi species. So I recommend that this paragraph should be more details, with clear examples of what it is meant with “memory of quasispecies” The concept of memory requires that the entity is capable of reproducing a response or an evolutionary trajectory when the same insult or perturbation is imposed upon Please describe and explain better than what it is now.

3) In the description of the impact that NGS has had on understanding viral quasi species, there is avery dry and incomplete description of what actually has been done in this topic and how critically NGS deep sequencing has changes the way to know viral quasi species. “Indeterminacies", and "Consensus is not enough” are probably the two sections that need to be completely reshaped. These are not clear, don’t really explain concepts and have lack of details One or two examples for all . At the end of the first paragraph of Consensus is not enough the authors sau , “Many events in viral pathogenesis and evolution are due to mutant spectrum modification which cannot be properly interpreted solely on the basis of consensus”" Why, what is the problem? The remaining subsections, starting from “Collective responses” and ending on individual genomes, are again a philosophical description of key concepts but not really easy to follow, especially for people that want to use this review to get a take home message on what is viral qusispecies and why this concept is critical to understand viral evolution and even biomedical research.

4) Another example is “Vaccines should be multi-epitopic” Firstly, I dont think multi-epitopic is a correct definition. Vaccines should be base don the design of antigens that cover multiple epitopes that are recognised by the adaptive immune cells (both T and B cells) is probably a better definition The immune responses again viral quasi species is never clearly explained/defined/presented in this review.This section is vague and doesn’t really have a message. At least I couldn’t see it “with current available vaccines….the order of expected efficacy for RNA viruses and other variable pathogens is: ….” This sentence has no sense to me What is that the authors want to convey?is this a suggested strategy on how vaccines should be designed? Please clarify the concept. Same for Antiviral targets…. 5) In the discussion, I think the authors need to address more critically the take home message, and also maybe acknowlegde where this theory is likely to bring research in virology and medicine. Perhaps acknowledge that the NGS and bioinformatics analysis are the “new way” to look at the sequences that represent these quasi species, and how this could be key to further understand evolution of RNA viruses. Perhaps address the phylogenetic and phylodynamics work that others are doing even if these could be contrasting with the view of the authors of this review. I sense a clear detachment from a branch of viral quasispecies research from other authors that deserve to be mentioned , especially the ones that are applied to specific viruses in humans .

6) When the authors describe the high mutation rate they should at least summarise the key estimates that have been found across theRNA viruses, and maybe cite few NGS papers that have re-estimated few of these rates at a different rate. I also would suggest to cite some work on how the distribution of mutants are actually found in the quasi species. I particularly liked the work of Raul Anding on polio in Nature get years back where he showed with circular RNA templates that the level of detection of deep sequencing usually allows only to study the tip of the iceberg, while the majority of mutants occur at a frequency below the detection

7) Table 1 is not comprehensible , is a sort of piece of the review, no clarity , no clear simple message, lack of structure. This needs to be redone with more specific concepts and bullet points, and also some references that point at the key messages.

8) In the section Current Scope of quasi species, there are a series of research findings simply stated as sentences, not explained, not references. These should all be summarised in a table 2. This can be also in a Supp Table if restrictions are present.

9) Phenotypic reservoir: Can the authors discuss more the relationship between recombination and single point mutations in viral quasi species?Are these independent from each other? or are those correlated? In the same section, the authors should also acknowledge that the unit of infection is not the only major variable changing phenotype. For instance, the transmitted founder virus is also important and a single or a few viruses can actually change the phenotype.

10) Figure 1 : it is hard to understand what is the main message of the representative viral quasispecxies. There is minimal diversity between the four draws I would suggest a more clear draw that highlights the message.

Response to Dr. Luciani

1) Additional references have been included, as indicated in the response to Dr. Domingo-Calap. If each experimental result on quasispecies were accompanied by one or more specific references rather than reviews (that gather references to many studies), the objective of a concise and conceptual account on “Viral quasispecies” would be impossible. Not only reference 34 (a book by one of us in which quasispecies origins, concepts and implications were comprehensively summarized for the first time) but also reference 13 (a multi-authored book on theoretical and experimental quasispecies), reference 21 (by an independent Russian team), reference 22 (by us with medical emphasis), reference 23 (by an independent Polish Academy group), and reference 78 (a classic text edited by J.J. Holland) are quoted to avoid including hundreds of references (to illustrate this point, it should be considered that in reference 22 a total of 895 literature citations are included). We have added references to many of the statements that did not have one (highlighted in the revised version).

2) Memory is a property of the quasispecies, not of the host, and it is unrelated to virus identity in the sense mentioned by Dr. Luciani. In the original articles from our group in which quasispecies memory was first described (a total of eight publications of which four are quoted in the present review) the molecular basis of memory is described and measured experimentally with different viruses and genetic markers. As explained in page 9, memory refers to a subset of components of the mutant spectrum (present as minorities but at a frequency range significantly above the basal mutation frequency level) that represent a record of the response to selective forces to which a viral quasispecies was subjected in its evolutionary history. This is all explained in page 10 of the revised version, with supporting references.

3) NGS has been an extremely important tool for quasispecies characterization but has not contributed or modified in any substantial manner the main quasispecies features that had been unveiled by molecular or biological cloning and Sanger sequencing. It has expanded the capacity to penetrate into the detailed composition of mutant spectra, but both the existence and the implications of mutant spectra had already been established prior to the advent of NGS. We think that what we state on deep sequencing (based on our own experience with the methodology applied to hepatitis C virus in the laboratory and the clinic) is a fair account of its contribution to quasispecies.

Dr. Luciani indicates that the sections entitled “Indeterminacies” and “Consensus is not enough” require complete reshaping. We have analyzed the arguments. First, we are not clear about the incorrect statements in our text or the type of details that would be needed to improve it. We do not understand what Dr. Luciani means with the statement “one or two examples for all” (does he mean that two examples are needed? or perhaps that two examples would be sufficient? or insufficient?). However, since we sense that indeed there might be some problem of clarity from our side, we have modified the section on “Indeterminacies” as follows: first, since here we deal not only with indeterminacies but also with methodological limitations, we introduce the following new title for this section ”Limitations and indeterminacies” (line 204 of the highlighted version). The first paragraph of the section deals with limitations and the second paragraph with indeterminacies. In the first paragraph we have explicitly mentioned some limitations that are encountered in experimental work that we hope will render our account more precise. The modified paragraph reads as follows: “Each of these procedures implies some limitations: biological cloning can bias the representation in favor of infectious genomes, while molecular cloning can introduce non-infectious (defective) genomes in the analysis (10, 34, 35). Whole genome quasispecies description is still technically challenging due to the artifactual introduction of mutations” (lines 209- 213 of the highlighted version).

The second paragraph explains indeterminacies that are inherent to the nature of dynamic mutant spectra. The problem was announced by J.J. Holland and colleagues 27 years ago (reference 40, and his statement written in page 8) and, to our knowledge, no studies have provided any evidence to the contrary. We think that the paragraph conveys clearly the type of indeterminacies we are talking about. However, we have added at the end of the paragraph: “as evidenced by many experimental studies [6, 12, 16, 34, 35, 45, 47, 52]”, to stress that the indeterminacies are supported by experimental evidence (lines 235-236 of the highlighted version).

Concerning “Consensus is not enough”, Dr. Luciani suggests that it is not a problem that many events in virus evolution cannot be properly interpreted solely on the basis of consensus sequence. We disagree but at the same time, the statement by Dr. Luciani makes us suspect that we have not been sufficiently clear in explaining exactly why ignoring the mutant spectrum is a problem. Therefore, we have added a sentence in the section “Consensus is not enough” that reads as follows: “Experimental results have demonstrated that minority genomes from a mutant spectrum (that cannot be identified by examining the consensus sequence) can include mutations that confer resistance to antiviral inhibitors, neutralizing antibodies or cytotoxic T cells, or that can alter the capacity to induce interferon (IFN) or to respond to IFN, virulence or particle stability, among other phenotypic traits [16, 33, 37, 47, 52-56]. Mutant spectra can also mediate cyclical adaptation to different cell types (Donohue RC, Pfaller CK, Cattaneo R. 2019. Cyclical adaptation of measles virus quasispecies to epithelial and lymphocytic cells: To V, or not to V. PLoS Pathogens 15(2): e1007605)” (lines 243 to 248 of the highlighetd version). We end the sentence by quoting the review articles where specific examples are described plus a recent 2019 paper by Roberto Cattaneo and colleagues on measles virus cyclic adaptation to two relevant cell types, published in PloS Pathogens (a study that was not available when our first version of the present article was submitted). We hope that with the added sentence, the problem of ignoring mutant spectra is clearer.

We disagree that the following sections beginning with “Collective responses: ensembles as units of selection” are philosophical descriptions. They are important effects on the behavior of viral populations based on specific experimental findings that the reader will find in the references quoted in the several sections. All concepts are based on findings in more than one laboratory and that have been contrasted by the scientific community. Contrary to the last statement in point 3 of Dr. Luciani’s review, we consider that we have been very clear on why the quasispecies concept is important to understand viral evolution and the important connections between quasispecies and biomedical research.

4) Although we and others have used “multiepitopic” to refer to vaccines that expose the immune system to multiple B cell and T cell epitopes, we have changed the title and rephrased the subsection on “Vaccines should be multi-epitopic” to read as follows:

Vaccines should expose multiple B cell and T cell epitopes to the immune system. Vaccines should include repertoires of B cell and T cell epitopes to evoke an ample immune response. The broad response should minimize selection of escape mutants that may be present as minority components in mutant spectra, as repeatedly documented experimentally [14, 16, 35, 59]. With the current types of available vaccines, those that best comply with the multiple epitope requirement are, in the order of expected efficacy to confer protection against highly variable viruses (without consideration of safety issues that should be examined for each vaccine): attenuated >… (lines 458-468); the paragraph ends as in the first version.

We assumed that the reader would understand why a complex vaccine should minimize selection of escape mutants and that for this reason the preferred type of vaccine is an attenuated vaccine, and that this is the vaccine design we suggest for highly variable viruses. With the new version we hope that these issues are clear, and that the paragraph makes sense to Dr. Luciani. From the careful reading of our article we believe that the take home message is clear, and the importance of deep sequencing has been duly acknowledged (see also our response to point 3 above). However, to stress that the same concept applies to vaccine and antiviral designs, we end the section on “Antiviral agents should be used in combination” with the following sentence: “Vaccines exposing multiple epitopes and combination therapies follow the same strategy whose aim is to limit possible escape routes to viral quasispecies in the face of the suppressive constraint” (lines 490-492 of the highlighted version).

5) With the modifications and clarifications introduced in the modified version, we think that the description of the connections between quasispecies and biomedical research has been improved over the previous version. In no way quasispecies work is contrasting with the phylodynamics work of others. This has been recently acknowledged by J.L. Geoghegan and E.C. Holmes in an article (Evolutionary virology at 40. Genetics 210: 1151-1162, 2018). Dr. Holmes is one of the maximum experts in virus evolution and phylogenetic and phylodynamic procedures, and in the recent Genetics article the complementary merits of quasispecies and long-term evolution of viruses are elegantly enumerated. Our article does not represent any despise or detachment of the procedures used by our colleagues to approach long-term evolution of viruses, but to try to introduce long term evolution in the present article would destroy its focus. We were requested to address concisely viral quasispecies. However, in the modified Table 1 under the heading “Concept” we have included “Short-term versus long-term virus evolution” with the reference to J.L. Geoghegan and E.C. Holmes, as well as a precedent by V. Agol where this issue was approached for poliovirus.

6) Comments on mutation rates and additional references have been added in response to the evaluation of Dr. Domingo-Calap, and the Nature paper by Andino’s group is quoted in our review. From the reading of the literature and from the presentations of Dr. Andino (and discussions with him), no significant discrepancies between mutation rates (that most often are mutation frequencies) estimated by NGS and other procedures have been found. We will be happy to evaluate and comment on discrepancies if we are given the information.

7) We agree that Table 1 was not well organized. We have modified it to improve its structure, introducing ten points that in our view underline important aspects. We quote the Table under “Summary and prospects” to serve as an overview of main points, with literature references.

8) With the new format and contents of Table 1, do not see the need to introduce an additional table.

9) Both mutation and recombination are mechanisms of genetic variation that affect the composition of mutant spectra but the two mechanisms of variation are not necessarily linked. Some RNA viruses display very limited (or absent) recombination while exhibiting high mutation rates (reference 33). We agree that a single founder virus can change the phenotype, thus enhancing the relevance of bottleneck events. This has been added to page 13: “A founder virus can introduce a different phenotype for the ensuing evolution” (lines 405-406 of the highlighted version).

10) Figure 1 is a schematic representation of an evolving mutant spectrum since the four distributions drawn have a related but different composition with the same consensus sequence. This representation and the accompanying equations recapitulate the core quasispecies concepts that have greatly influenced the field of experimental evolution with viruses. Again, we suspect that we have not been clear enough in explaining the figure and, therefore, we have expanded its legend to explain its meaning (lines 853-854 of the highlighted version).

Review by Josep Sardanyés

The authors are reviewing the paramount concept of virus quasispecies, emphasising in the importance of Quasipecies Theory for the understanding of virus population dynamics and both genotypic and phenotypic heterogeneity.

I would suggest to clearly define the concept of error threshold and lethal mutagenesis, which are related but are different. The error threshold is a shift in sequence space that involves the reconversion of the so-called master (high-fitness) sequence into mutants. Lethal mutagenesis is a demographic extinction phenomenon due to a large amount of mutations. These two concepts have been widely confused in some literature and would be really interesting to define them properly in this Topic Page. Some research has connected both phenomenon, providing evidences from the transition from one to another in time-continuous models, and thus not restricting lethal mutagenesis as a purely demographic phenomenon tied to intrinsic or demographic noise. (Viral replication modes in single-peak fitness landscapes: a dynamical systems analysis. J Fornés et al J. theor. Biol. 460, 170-183 (2019))

In the first paragraph of the Topic Page, Section Viral Quasispecies, I would add: Fueled by high mutation rates and extremely large population sizes, mutants arise continually ...

When the authors mention in Page 3 that the quasispecies nature of RNA viruses has been identified for VSV and QB, they should also add other examples, such as hepatitis C virus or HIV-1.

Missing parenthesis in Page 5: QUENTIN [28]).

Page 10. The term "advantage of the flattest" it is usually dubbed as "survival of the flattest". Please note that the error threshold can be also interpreted as the survival of the flattest (Claus O. Wilke, Personal Communication - Workshop, Kavli Institute of Theoretical Physics, Santa Barbara, January - March 2011), where the pool of mutants (flat species) outcompetes the master sequence (fit) in sequence space.

The authors should emphasise that both error threshold and lethal mutagenesis are highly fitness-landscape dependent. Several works have explored other, let's say, more virologically realistic fitness landscapes. Including the so-called mutational fitness effects on quasipecies population dynamics. Some notes should be add on this.

In the list of Section: Quasispecies and viral disease, the authors should clarify why they differentiate in the four parameters between: replication rate, viral load, and replicative fitness. They seem to be interdependent. For example, what is the difference between replication rate and replicative fitness?

Caption of Fig. 1. consenssu should be consensus

Figure 1. I am not sure about what is the expression of the dilution factor Phi_i that the authors are using in the equation dx_i / dt, but this term is usually multiplied by x_i, since it is proportional to the amount of the variable x_i.

Response to Dr. Sardanyés

We have defined error threshold and lethal mutagenesis, as suggested. Although no unanimous definitions are found in the literature of quasispecies, theoretical biology and experimental virology, we have included the definitions that we believe most reflect the majority of studies. The error threshold relationship was mentioned at the end of the first paragraph under “Historical origins” (page 3). Here we have defined it as: “the maximum mutation rate at which the master (or dominant) sequence can stabilize the mutant ensemble” (lines 50-51 of the highlighted version). We have expanded this part to read: “Violation of the error threshold results in loss of dominance of the master sequence and drift of the population in sequence space” (lines 52-54 of the highlighted version) with three relevant references [2-5], one of them being the one by Dr. Sardanyés and colleagues that he suggests we should quote. A definition of lethal mutagenesis has been introduced at the beginning of the subsection entitled “Lethal mutagenesis” (lines 493-495 of the highlighted version). The definition we have written because in our view harmonizes theoretical and experimental findings is the following:” Lethal mutagenesis is defined as the error rate at which a virus can no longer maintain its inheritable genetic information due to an excess of mutations, resulting in virus extinction”.

In the first paragraph of the section “viral quasispecies” the terms “and extremely large population sizes” have been added, as requested. This is a good point.

In page 3, we emphasized Qβ and VSV for historical reasons. However, in agreement with the suggestion of Dr. Sardanyés, at the end of this section we have added the following sentence: “Particularly detailed studies on quasispecies dynamics have been carried out with HIV-1 and HCV, with a reference to one of our reviews [16], and two additional ones that we consider should be quoted if we mention explicitly these two viruses: Meyerhans et al. Cell 1989 as key to HIV-1 and Farzi, Seminars in Virology, 2011 as a summary for HCV (references 17, 18).

The parenthesis in page 5 has been added.

We have added “survival of the flattest” to “advantage of the flattest” since we have seen both expressions used in the literature and discussions. Dr. Sardanyés is right that one of the models of error threshold is based on survival of the flattest. Although it is not our preferred model for the error threshold, to cover all published models we have added the following sentence in this paragraph: “Survival of the flattest has been proposed as an ingredient of some models of the error threshold [74]”. The reference (Tejero et al.) summarizes several models of error threshold and lethal mutagenesis (lines 328 to 329 of the highlighted version).

Following the suggestion of Dr. Sardanyés, we have added a comment on the dependence of the error threshold and lethal mutagenesis on the fitness landscape, with an indication that both can occur in complex fitness landscapes, a point that has been studied and presented in great detail by Peter Schuster in a recent publication. The sentence added to page 16 is the following: “Both the error threshold and lethal mutagenesis are highly fitness landscape-dependent, but both can occur in complex fitness landscapes as those pertinent to viral populations [5]” (lines 497 to 499 of the highlighted version).

In the last bullet of the section on “Quasispecies and viral disease” we have clarified the differences between replication rate, viral load and replicative fitness by writing what we mean in parenthesis. The definitions given in parenthesis (page 14) are the following: for replication rate “the rate at which viral RNA or DNA is synthesized intracellularly for viral progeny production”; for viral load: “the total amount of virus quantified in an infected host or host compartment”; for replicative fitness: “the yield of infectious particles that can contribute to the next generation” (lines 438-442 of the highlighted version).

The typo in the legend for Fig. 1 has been corrected.

The term Φi in equation 1 is a flux factor, as indicated in the right panel. It is included in the quasispecies equations, and in our case the equation was reviewed by Peter Schuster for a previous publication of our group.

Second Review by Pilar Domingo-Calap

In this revised version of the manuscript entitled “Viral quasispecies”, some of my concerns have not been addressed. In addition, I consider that some answers are out of context.

One of my suggestions was to include a more balanced overview of the literature, a point also suggested by reviewer Dr. Luciani. The core of this manuscript is mainly based in FMDV, work done by Dr. Esteban Domingo and his group. I think that quasispecies have been shown in a large variety of RNA viruses, and in some small DNA viruses. If the aim of this manuscript is to highlight the principles of viral quasispecies, I agree with Dr. Luciani that including phylogenetic and phylodynamics work from others will give a better idea about the general concerns of viral quasispecies. In this version, more than 30% of the reference list are works from the authors of this review. In my opinion, literature should be revised prior to its publication.

Regarding the new version of the manuscript:

1.“Variants that drive responses” section should be renamed, it is not clear which responses.

2.Poliovirus mutants of polymerase fidelity: mutation is G46S instead of G46G.

3.“New antiviral strategies: lethal mutagenesis” section should be renamed. Firstly, lethal mutagenesis is not new. In addition, more strategies are included in the section (not only lethal mutagenesis). There is also an important error of concept, it is written: “lethal mutagenesis is the error rate at which a virus can no longer maintain its genetic information”. Lethal mutagenesis is NOT an error rate. It is a demographic process in which the population becomes extinct. Lastly, the section starts and ends with the same sentence and same reference, please, remove one of them.

4.“Summary and prospects” section should be revised. As previously mentioned in my first revision, main ideas and conclusions of the theory should be mentioned, and not just a short sentence as included by the authors.

5.There are many typos in the revised version (g.e. coopoeration, host-to host, bottelneck), please, revise carefully.

Second Response to Dra. Domingo-Calap

1. We have added 4 additional references covering RNA and DNA viruses and environmental microbiology (van Boheemen et al., 2017, Vlok et al., 2019, Hirose et al., 2018, Gisder et al., 2018, references 38, 39, 40 and 41). The vast majority of studies on viral quasispecies have concerned RNA viruses, irrespective of the concept applying also to several DNA viruses. With these additional references (insert in page 5), the total number of references is 111, and 75 of them are not ours. We would appreciate that if a bias in favor of our work persists, it should be considered that one of the authors (ED) has been working on viral quasispecies for 40 years. Also, we would appreciate that it be also considered that out of the 37 references to our previous work (not only FMDV), 10 are review articles that include many studies of other authors. Dr. Domingo-Calap invokes Dr. Luciani in the request to give a more balanced view of the literature, but Dr. Luciani has approved the revised version of our manucript. We surveyed the literature that includes “quasispecies” as a keyword prior to the writing of the manuscript, and now again to prepare this resubmission.

2. We have modified the subheading in line 251 to read “Variants that drive responses to selective constraints”.

3. The typo G64G has been corrected (line 328)

4. We have modified the subheading in line 444 to read “Antiviral strategies in response to the quasispecies challenge”. The terms “error threshold” and “lethal mutagenesis” were re-written in response to a query of Dr. Sardanyés who has approved the revised version of our manuscript. However, we have modified slightly the definition of lethal mutagenesis as follows: “Lethal mutagenesis is the process of virus extinction at the error rate at which a virus can no longer maintain its genetic information”. The demographic process mentioned by Dr. Domingo-Calap is a consequence of exceeding the critical error rate, not the primary reason of the extinction.

We have removed the reference at the end of the section “Antiviral strategies in response to the quasispecies challenge”.

Upon reconsideration of the section “Summary and prospects” we think it conveys the main points we want to emphasize, considering that the article is intended for a broad readership.

The typos have been corrected.

Second Review by Fabio Luciani

The authors have addressed my comments and I am satisfied with the manuscript.

Second Review by Josep Sardanyés

The authors have addressed my comments thoroughly and I am happy to communicate that this can be accepted for publication.

Wikification

I've read through article and peer reviews again and I think that even though the self-citation rate is high, it does not misrepresent the overall state of the field. I think it is therefore reasonable to move forwards with the publication process.

As a last step, please look though and see if there are any hyperlinks to wikipedia pages that could usefully be added. T Shafee (talk) 22:57, 27 April 2019 (PDT)

References now inline formatted with {{citation}} template. T Shafee (talk) 00:06, 19 July 2019 (PDT)