Evaluation of Potential ARG Packaging by Two Environmental T7-Like Phage during Phage-Host Interaction

Round 1

Reviewer 1 Report

In the manuscript by Liu et al., two podophages are isolated, routinely characterized and tested for antibiotic resistance genes. These genes were found. This work was done to supplement metagenomic testing for these genes, with phages defined via filtration in the metagenomic studies. This phage definition ignores the fact that filtration would also purify bacterial DNA-containing outer membrane vesicles. Thus, the metagenomic studies were not valid studies, as far as I can tell. That is to say, the present study is a significant contribution to the literature for the detection of phage-carried antibiotic resistance genes. However, this study has the following non-productive characteristics, which is the reason that I do not recommend publication as is.

[1] Plaque sizes are given without giving the time of incubation, the medium or the gel used for the upper layer. Thus, these numbers are meaningless. The two plates shown in Figure 1 might have been co-incubated and might not. Thus, the reader cannot tell whether the difference in plaque diameter is significant.

[2] Poor quality electron micrographs of phages are often published (https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/pmc/articles/PMC5029504/). But, the quality of the images in Figure 1 is so low that one cannot be certain that one is seeing phages.

[3] 83-86: The buffer for incubation was not given. Stability of other, similar phages is very dependent on ionic strength and divalent cation concentration.

[4] 180-182: No justification is given for infecting at different bacterial counts and moi. What is the purpose? Also, the titers are very low compared to those of other podophages, such as phi29, P22, T3 and T7.

[5] 191:  A circular sequence does not mean a circular genome. All terminally repetitious, linear phage DNAs produce circular sequences. Podophages typically do not have circular genomes.

[6] 219-228: This text is not comprehensible, with a major sentence structure problem in line 221.

[7] 243: The meaning of G1, etc. is never defined.

Author Response

Thank you very much for the editor’s suggestions for this article. Next, I would try my best to answer the editor’s questions and modify the article.

For the outer membrane vesicle problem: During phage-host interaction, bacteria would indeed cause the formation of outer membrane vesicles under the stress of the lysozyme of phage. Outer membrane vesicles could carry bacterial DNA or plasmids. However, when we observed the morphology of phage under transmission electron microscopy, no vesicles were observed. This may prove that our method of extracting phage DNA could also remove potential bacterial contamination.

Regarding the problem of using metagenomics for research: In this article, since potential packaging was a low probability event, the genome of phage carrying ARGs would be spliced as a foreign source of contamination and discarded. So, the analysis of raw sequence data of ARG was an important way to determine the possibility of potential packaging of phage in phage-host interaction. Finally, qPCR method was used to evaluate the absolute abundance of the potential packaging of ARG fragments in the process of phage-mediated generalized transduction.

For question 1: The amount of upper medium and the time of co-incubation would affect the size of the plaque. That was my negligence, and the details should be added. When we carried out the double-layer plate test, the amount of upper medium added was 5 mL. Bacteria in the logarithmic growth phase and the diluted phage were mixed 1:1, and rested for 15 min. 200 ul mixture was added to the upper medium to incubate at 37°C for 12 hours. The plaques on the picture had been taken after 48h.

For question 2: Compared to other types of phage, the morphology of Podoviridae phage, which “tail” was too short, were relatively disadvantageous. In addition, due to the angle, Podoviridae phage were easily overlooked under transmission electron microscopy. This picture was taken by our team after repeating transmission electron microscope three times. In addition, we also found some T4 phage using this method.

For question 3: I was negligent and should be added in full.

For question 4: The expression was incorrect and had been modified. My original intention was to mix the host bacteria in the logarithmic growth phase with phage according to the ratio of P:B = 10:1, 1:1, 0.1:1, 0.01:1, 0.001:1, and 0.0001:1.

For questions 5 and 6: Thanks for the reminder from the editor. Modifications had been made.

For question 7: Thanks to the editor for your carefulness, I had clarified the meaning of G1 etc. in the experimental steps.

Reviewer 2 Report

I have attached a file with my comments to the authors below. I will paste the same comments into this box, just to ensure that they are received:

Final Comments to the authors for Liu, et al. 2020

In this paper, the authors have identified and characterized two novel T7-like lytic bacteriophage of E. coli. They have furthermore demonstrated that a small subpopulation of phage package antibiotic-resistance genes (ARGs) acquired from their host, presumably due to mispackaging (or erroneous packaging) into the phage heads as it intends to package its own genome. The authors suggest that bacteriophage that package host ARGs can act as generalized transducing phage to spread antibiotic resistance genes throughout a population of host bacteria. Identifying mechanisms of spread of ARGs by transduction obviously has important implications for both clinical and agricultural industries and will be of great interest to the readers. Furthermore, erroneous packaging of DNA by phage and the resulting generalized transduction of bacterial genes between different bacterial cells is an important, but not well-understood mechanism, for the movement of not only ARGs, but also bacterial virulence factors and biosynthetic genes. Understanding generalized transduction better may provide valuable insight into bacterial evolution and pathogenesis. The conclusions of this paper seem well-supported. I recommend this paper for publication, but I would urge the authors to consider the points below, particularly the first one under major comments, which would greatly strengthen the major argument of the paper. If that comment cannot be readily addressed, it does not necessarily invalidate the rest of the manuscript, but this reviewer believes that it will make the results of this paper easier to dismiss and potentially of less value to the field overall.

Major comments

• In a 2017 paper published in the ISME Journal, Enault, et al. perform a meta-analysis of the literature which suggest that although many papers are published reporting ARG in bacteriophage, there is very little empirical evidence (not just sequence-based) to suggest that these phage are packaging whole ARG or that they are then using them to confer resistance onto their subsequent hosts. In fact, they conclude (as demonstrated by the title) that there are actually few ARG packaged by bacteriophage. My questions related to this point for this study are (1) was any attempt made to amplify entire ARG from the phage DNA samples and (2) was any attempt made to use purified phage preps to transduce antibiotic resistance genes into one of the other phage-susceptible coli strains that were identified by the authors to demonstrate that antibiotic resistance was actually conferred in a new background? While I recognize that these experiments (particularly the second one) may be technically burdensome, even negative results would help the reader to place the other data in context and would help to determine what level of concern we should have about phage-mediated ARG transduction. A positive result, on the other hand, particularly in the second set of experiments, would tremendously amplify the importance of the findings of this study. They also would help to blunt the major criticisms of these data that were offered in the 2017 paper.

• Throughout the manuscript (and in the title) the phrase ‘erroneous packaging’ is used. While the intention is understood, this reviewer would prefer a little less emphasis on the ‘erroneous’ (or more shading with ‘potentially’ or ‘probably’)—particularly since the packaging mechanisms are not under investigation in this manuscript. The packaging described here is an error only from the limited perspective of what the phage must ‘want’ to do, which is package its own DNA and create viable progeny. However, from the genes’ perspective, and from the perspective of the bacterial host, evolving signals that would increase the chance of an ARG (or virulence factor, or biosynthetic gene) being packaged into a phage head would be favored by selection for both the gene and the bacterial host (indeed, one of the reasons that a bacterial host might be pressured to not selectively evolve away from susceptibility to a lytic phage is the advantage that might be conferred by the ‘rare’ generalized transduction event that would yield genetic diversity). Indeed, the authors’ conclusion that not all ARG are randomly packaged, but that there is some force of selection, minimizes that this is an ‘error’. Particularly since not all ARGs are present and there is no data presented about other bacterial host genes being mispackaged (or numbers relative to presumably correctly packaged phage genes), it does seem that there must be some selection for these particular genes. While I recognize that the authors are technically correct to dub it ‘erroneous’, I am not sure that so much emphasis on that particular aspect is needed. I recommend a clear summary of specialized transduction vs generalized transduction and how generalized transduction can result from mispackaging or erroneous packaging and then sparingly uses the erroneous terminology throughout the rest of the manuscript.

• This is major only because it consumes a large part of the paper, but I am not sure what the fix is (or if there needs to be one). I do not understand the importance of the experiments demonstrating that changing bacterial host density and MOI either do or don’t change the packaging of ARG. While I think that the ultimate conclusion that the environment can influence the erroneous packaging of ARG (but not significantly) is probably correct, I am not sure mechanistically why changing the number of bacterial hosts or MOI would have that effect (for at least one of the phage). I hesitate to say the data should be removed, because I found it to be interesting, but I do NOT think it is critical to the major points of the paper and I think it is a little bit confusing (it is basically left out of the conclusion of the paper, and that is very clear and enticing). Perhaps if the authors could just read through this material again to clarify and to offer more rationale as to their importance or tie it into some mechanism in the discussion (are there other publications where mispackaging is shown to be dependent upon MOI or density?). Also, as mentioned below, it is not clear if any other phage genes or host genes were looked at, might strengthen how this relates to the process of erroneous packaging.

• Line 72—For the supplementary table with the MIC and sensitivities, is there a way the authors could visually indicate those antibiotics for which the organisms are resistant (add a parenthetical R, or bold, or add an extra column for interpretation?); while the data is there, and I like the fact that the values are shown, it is a bit unwieldy to have to keep comparing the values to the control to determine if the clone is resistant or sensitive to a particular drug; also, the gene designations for resistance are used here, but it does not indicate which gene confers resistance to which antibiotic; that is done later in the results, but is far removed from this point. Also, were multiple replicates done to determine MICs?

• Would it be possible to make a table that shows which phage came from which strain, what the ARGs found in the strain are and what the ARGs found in the phage are? All the information is in the text, but I find that I had to go back and forth a LOT to make sure I was understanding the flow here. It could even be a supplemental table (but would probably have more value in the main body of the text).

Minor comments

line 14—not sure if this is true; a search of antibiotic-resistance genes in E. coli phage returned more than 500 articles (not all probably relevant, but a quick scan through the first fifty suggests many of them were)

line 39—while I think this sentence is discussing the overall development of ARG in bacteria, the way the sentence ends with the transduction of phage suggests that the 2007 reference (#9) is not the only reference that should be used here; I recommend citing the Colavecchio 2017 Frontiers in Microbiology Review (specifically looks at ARG transduction by phage of the Enterobacteriaceae). If not here, then somewhere because it is very relevant to the data presented in this manuscript. (Although this might run counter to the narrative that phage ARG transduction is not well-studied, which I don’t think is true, but it IS certainly less well-studied than the other mechanisms of HGT of ARG.)

lines 43 thru 46—I would be explicit about the difference between a specialized transducing phage (packages specific DNA, usually the phage genome) and generalized transducing phage (packages fragments of the host genome, usually due to mispackaging or erroneous packaging as it is dubbed here). If you are not explicit, then it is not clear to what the parenthetical (erroneous packaging) is actually referring.

Line 48-49—'However, as the most abundant biological entities on earth, phage can exist

49 in a variety of ecological environments and therefore can also carry ARGs derived from their hosts’ This is not clear and does not follow previous sentence about rarity of generalized transduction. I believe the authors are trying to make the point that even though generalized transduction is rare in any one replication cycle, at the numbers that phage are found, levels of generalized transduction could be significant. The fact that they are abundant does not necessarily mean that they can carry ARGs derived from the host. Should probably split this into two sentences/thoughts.

Line 58— ‘potential contribution’; just because phage metagenomes have ARGs doesn’t mean they contribute meaningfully to the spread of antibiotic resistance.

Line 61—the authors are not really exploring erroneous packaging (looking at sequences for uptake, mechanisms like misfolding for mispackaging, etc.)—suggest rewording

Line 145—for table S3, would it be possible to give the approximate size of the entire marker? If the phage is erroneously packaging ARGs, there is no guarantee that it would package an entire gene; as indicated above, I was wondering what percentage of the gene you were actually targeting; was any attempt made to amplify any of the genes in their entirety (which would be necessary to confer antibiotic resistance)?

Line 158—is this respectively (HZP2 infected HZA135 and HZ2R8 infected HZA50)? Or do both infect both?

Line 195-199—Were the unknown genes similar between the two (i.e., shared between them?)

Line 212-218—any thoughts on why there were so many more ARG hits in HZA135 relative to HZP2? Relative abundance in the respective backgrounds?

Line 218—no mention of cmlA in strains described in methods (although Table S1 does show high levels of resistance).

Line 219—E. coli (not coil)

Line 221-222—the structural similarity sentence does not make sense (I think the authors are trying to say that based on structural similarities compared to the other T7 phages, the two they have identified here are most closely related to each other, but I think words are missing).

Line 222-223—similar to previous comment, this sentence does not make sense

Lines 219-228—this section is a little confusing, because it starts with a comparison to three T7-like phage, but it is not clear from the language whether the comparison with T7-like phage was the authors’ choice or based on similarities in the entirety of the phage genome (it just says that the top 3 T7 were chosen, but not that the top three hits were T7-like). This confusion is amplified because in the same paragraph other comparisons are made, but not limited to the three T7-like phage selected. For the other comparisons, the authors seem to be reporting back on blastn analyses for each gene—which is fine, but confusing with the set up of the paragraph. Perhaps the paragraph could be set up swapping figures 4 and 5 and then do blastn comparisons for large subunit (currently figure 5A) and DNA polymerase (currently figure 5B). Then indicate that overall comparison suggests similarities to T7-like phage (consistent with polymerase) and then do the overall sequence comparison to show the structural relatedness of the two phage identified in this study with three T7.

Line 236—As per the major comment above, I think this could just be packaging of ARG (and in the figure legend); there is no data about the ‘non-erroneous’ packaging, and the authors are not really elucidating the mechanisms of erroneous packaging. I think that packaging could be used here and then the discussion of why it is likely mispackaging (or erroneous packaging) could be made in the discussion.

Line 238—Am I mistaken or is cmlA not mentioned as an ARG in the first section of the methods?

Line 242—I’m not sure I understand how this section relates to ‘erroneous’ packaging? Did the authors quantify any phage genes? Did the authors check for any other E. coli genes? The authors mention the 16s rRNA DNA in the primer section of the methods, did they screen for those genes in these samples? What did they do with those primers (perhaps I missed it).

Line 259-260—there is something wrong with this sentence

Line 261-263—this paper didn’t really look at the packaging process either; that would require some sort of comparison with phage genes, other host genes, and probably a search for signature motifs

Line 266—period needed, not comma

Line 270-271—does the burst size usually apply to the bacterial strain? Or the phage? Bacterial strains are listed here

Line 289-290—reference for this? As in major comment above, I don’t know that ‘lytic phage rely’ is the proper phrasing—erroneous packaging is probably disadvantageous to the individual phage (the DNA gets into the host, but doesn’t get back out because it doesn’t have all the necessary machinery), but it does explain why a bacterial population might remain susceptible to a phage—some risk death, but others may acquire new genes (like ARGs) through generalized transduction.

Line 290-291—two ‘although’s and something is wrong with this sentence

Line 300—As suggested in major comment, here and elsewhere, since the actual packaging mechanism is not under investigation, I would occasionally shade this with ‘potentially erroneous packaging’.

Line 322-324—not sure why plasmid maintenance is further evidence for phage mediated ARG transduction?

Line 324-325—did you test for other host genes? Do you mean preference for specific ARG or ARG over all other host genes? Not sure you have demonstrated that last point with the data presented.

Author Response

Thank you very much for editor’s suggestions for this article. Next, I will try my best to answer the questions and modify the article.

Answer the maior comments:

(1) I did not try to amplify the entire ARG from the phage DNA. Because shortcomings of qPCR method, the size of qPCR products should not be too long. What the editor said was correct. Only when the intact ARGs was detected in phage DNA, could it prove that the phage has the potential for further transduction. Therefore, I plan to design primers (covering >87% of the full length of target ARGs) for PCR.

(2) As the editor said about the test of transductant, our team had already explored it under laboratory conditions or in a soil model. Due to false positive results (this may be because the antibiotic resistance gene I tested has not been packaged completely by phage), the method was still being modified and optimized.

(3) Regarding the word "erroneous", I agreed with what the editor said. Related articles had shown that generalized transduction, one type of bacterial DNA transfer by phages, can create conditions where not only the recipient host but also the transducing phage benefit. I would make changes in accordance with the editor’s suggestions.

(4) The reason for using different bacterial host densities and MOI was that our team had detected the ARGs of bacteria and phage DNA in environmental samples using qPCR methods and metagenomic sequencing, and found that bacterial density has influence on phage carrying ARGs. Meanwhile, A report (Identification of a bacteriophage from an environmental multidrug-resistant E. coli isolate and its function in horizontal transfer of ARGs) suggested that transduction might be affected by many environmental factors, such as the concentration of host bacteria and phages. For this purpose, the bacterial densities and MOI were designed to explore whether it would affect the potential packaging during phage-host interaction.

(5) The supplementary table with the MIC and sensitivities had been modified. The table had been added to the article as suggested by the editor.

Answer the minor comments:

I had revised the article based on editor’s questions and suggestions. I was very grateful for editor's care and patience.