Controlled prophage induction for the benefit of the host

Bacteriophages have traditionally been regarded as parasites of bacteria. This view is however changing, and in our recent study we found that quorum-sensing-controlled prophage induction strengthened the low-cell-density phenotype of the bacterial host.

Like Comment
Read the paper

Bacteriophages, viruses that infect bacteria, are the most abundant and diverse biological entities in nature and are important driving forces in bacterial diversity, community composition and turnover of nutrients. Some phages are able to integrate into the bacterial chromosome and replicate lysogenically. Once integrated the replicative success of the phage is equivalent with the success of the lysogen and many prophages encode anti-predator traits such as toxins or superinfection exclusion systems in order to protect their host.

In a previous study we observed a wide distribution of a specific group of prophages which shared high similarity to phage φH20 and genome analyses suggested that this group of phages encoded some putative mutualistic genes. In order to analyze whether lysogenic conversion from H20-like phages could justify their global distribution, we tested phenotypic traits of the host, Vibrio anguillarum, compared to an otherwise isogenic prophage-free strain. Simultaneously, another project aimed to investigate whether quorum-sensing (QS) affected regulation of induction of the same group of prophages. We have previously introduced gene deletions in V. anguillarum that successfully lock cells in states that mimic the low- and high-cell-density QS states, respectively.

As we learned that H20 prophage induction was in fact repressed by QS, we realized that merging the two projects would create synergy and raise our understanding of these prophages to a higher level.

In an evolutionary perspective, it is perhaps not a surprise that some phage developmental switches have evolved to factor in bacterial density, given the destructive effect phage propagation can have on dense bacterial populations and conversely the importance host availability has on phage fitness. A few other studies have identified phages that regulate their development by eavesdropping on the level of the bacterial signal molecules that provide molecular information on the density of the bacterial population. In contrast to previous studies, we observed repression of prophage induction in the high-cell-density QS state, which contributes to the ongoing discussion regarding the importance of lysogeny at different cell densities, initiated by the Piggyback-the-Winner model proposed by Knowles et al.

There are conflicting reports on the effect QS has on V. anguillarum biofilms, and thus we thought that it would be relevant to quantify biofilm formation in our QS mutants. We included the prophage-free strains in a comparative analysis and discovered that the H20-like prophage contributed to aggregation at low cell density, which was also the density where we observed the highest amount of biofilm. Since the other phenotypic traits we had tested up to that point, such as motility, were not affected by the prophage, we were thrilled to make this observation. Especially, since prophage-dependent promotion of biofilm was unique for low cell density and hence suggested that this is a specific mechanism evolved to occur at specific environmental cues.

At low cell density, when the response regulator VanO (purple) is phosphorylated and active, it represses the QS pathway, and repression of H20-like prophage (black box) induction is mitigated. V. anguillarum forms biofilms at low cell densities and the elevated level of prophage induction promotes aggregation of cells. As the cell density increases, the master regulator VanT (dark brown), is translated and prophage induction gets repressed, biofilm formation stops and cells disperse. Living cells are represented in green, while light brown cells are undergoing prophage induction. Free phages are released from the red, dead cells. Illustration by Karen Mikkelsen.

In summary, we showed that H20-like prophages are not only parasitic prophages occasionally undergoing spontaneous induction, but regulated entities that are controlled by host communication signals and contribute to biofilm development. In our view, the most exciting perspective of our study is that prophage induction may have evolved to occur also when it can benefit the bacterial host population rather than solely when it benefits the virus at the expense of the host.

If you also find prophage induction, biofilm formation and QS interesting you can read about our study here.

We thank Karen Mikkelsen for her excellent drawing that summarizes our study.

Mads Frederik Hansen & Demeng Tan


Mads Frederik Hansen

PhD Student, Max Planck Institute for Terrestrial Microbiology

I'm a microbiologist with a particular interest in phage-bacteria interactions. I have previously studied and worked at University of Copenhagen, but is currently enrolled as PhD-student in the lab of Prof. Knut Drescher, at the Max Planck Institute for Terrestrial Microbiology.

Comments

Go to the profile of Naresh kumar
4 months ago

Hello, Dear Hansen.


Recently, I read your research paper about the induction of an H20 prophage being repressed by quorum sensing. It was quite interesting. However, the mechanism behind how this happened has not been explained clearly in the paper. For example, how the activation of vanT Protein directly influenced the repression of prophage induction. Is there any hypothesis for this? 

Go to the profile of Naresh kumar
4 months ago

As mentioned in the paper, "There might be a QS-regulated phage receptor on the bacterial surface which might be downregulated at high cell density". If this is true, then how could we say that the Quorum Sensing directly influences the prophage induction?

Go to the profile of Mads Frederik Hansen
4 months ago

Dear Naresh Kumar,
Thank you for your interest. We did had a working hypothesis; That VanT would bind upstream of the CI-like repressor gene at HCD and upregulate CI expression, and thereby strength lysogeny / prevent induction. This idea was initiated by the identification of a motif upstream of the CI gene, which was similar to a known binding motif of HapR in cholerae (Tsou et al. NAR 2009). So we had some ideas about investigating this molecular mechanism, but never got around to do so unfortunately. But I think Assoc. Prof. Svenningsen is partly looking into this mechanism, so we might know more in near future.
I hope this answered your question.
Mads Frederik Hansen

Go to the profile of Mads Frederik Hansen
4 months ago

Hi again Naresh,
In this case we have to separate things. One phenomenom doesen't nescessarily rule out the other, but can create an artefact in the quantification. It is quite well-investigated that phage defenses are QS regulated (citations in the paper) and among these are downregulation of surface receptors at HCD. S. Abedon has published a hypothesis on "spatial vulnerability", where he talk about the risk of being phage targeting when you are in a group. He refers to spatial arrangements like biofilms, but you can argue that it also applies to the mere situation of being a dense population - So from an evolutionary perspective it makes a lot of sense to reduce phage receptors when you are a large phage target and vulnerable.
All right, we do not know which specific receptor is targeted by phage H20? So we cannot make direct quantifications of the regulation, but we can make an indirect analysis by measuring adsorption of the phage in the different mutants. This is important because it could influence how we quantify prophage induction - if there is a high level of adsorption at one QS state compared to another, it could affect the quantification of free, induced prophages as we measured. This doesn't affect the QS-prophage-induction-mechanism per se, but potentially hinder correct induction quantification. However, when we did the adsorption assay, although we found a slight difference, we did not find a significant difference in adsorption between mutants. And even if we take the adsorption-difference into account, adsorption is actually 'lowest' or 'slowest" at HCD (As the vulnerability hypothesis would suggest), which would mean that even if it creates a bias in the quantification, it only supports our claim; That there are more free induced phages at LCD.

It was quite long, but I hope it make sense to you.

Mads Frederik Hansen

Go to the profile of Naresh kumar
4 months ago

Thank you so much for your time in explaining detailedly about the hypothesis and the adsorption test results. However, I would like to ask you some more doubts regarding this. 


1. I have read the paper published by silpe and bassler (2019) on "A host- produced Quorum-sensing auto-inducer controls a phage lysis-lysogeny decision". In this paper, they proposed that, at high cell density the prophage encoded QS receptor binds (VqmAphage) with the host produced auto-inducer (DPO) to upregulate Qtip which sequesters the cI repressor that allows to launch the lysis pathway. So, this mechanism can't be used to explain your results, right? 


2. The next doubt is regarding the biofilm and aggregation. First, I couldn't differentiate between biofilm and aggregates. From my understanding biofilm consists of EPS matrix, live and dead cell, e-DNA etc. While aggregates are just a group of live bacterial cells. Is my understanding is right? And in your paper, it has been mentioned from the obtained results that at HCD, the QS circuit actually reduced the biofilm formation. If so, how could the bacteria attach to the fish or any host's surface to cause infection? 


And one more thing, can I get a clear view of the exact purpose of your experiments? 


Sorry, if my doubts are so simple. This research field is really new to me and I still have a lot to know about this. I look forward to hearing back from you. Thank you!

Go to the profile of Mads Frederik Hansen
4 months ago

Hi Naresh,


Dont worry. Simple doubts are just as relevant as complex ones - And I just hope I can help clear out some of your questions.


1. No, the elegant work by Justin Silpe and Bonnie Bassler cannot be used to explain our scenario. The overall phenotype is simply opposite. The find highest induction at HCD, while we find it at LCD.

2. In V. cholerae (and some other Vibrios) it well-studied that biofilm formation is most prominent at low-cell-density state. You can for example read the paper by Praveen Singh et al. 2017 in Current Biology. If you then add the work by Jun Zhu & John Mekalanos (Developmental cell 2003) you will find why this make sense from an infectious perspective. Biofilm in the early stage is important for infection and likely also for surviving the journey through the acid environment of many host before reaching the intestines.
Regarding to microcolonies, biofilm, aggregates etc. it is a constant debate. The way I see it, is that they present different stages of a continuum, but where they separate from one another can be difficult to define. The reason why we did it this way, is that Crystal Violet is a fine tool for screening, but not exactly accurate - So we wanted to support the Crystal Violet quantification with another quantification method that is more accurate. As you rightly mention, biofilms is complex mixture of various matrix components including eDNA and dead cells. You can stay these various components of interest (See the work of Thomas Neu for example), but in many cases staining the bacteria will be sufficient to find the size of the "biofilm/aggregate". As mentioned, the definitions are intertwined, and we saw in the microscope that cell aggregates on the bottom where nicely separated and not laying as a blanket. Agreed that the number 28um3 is quite arbitrary, but we decided that this was the size where we think an aggregate becomes "relevant" or "real" aggregation. And in this case you shouldn't put too much into the terms, here aggregate and biofilm cover the same phenotype, but because the latter analysis is made in microscope you can distinguish that biofilms are formed in aggregates rather than blanket, and hence we thought that word would be most suited.


3. I would hope the purpose would be visible in the introduction, and from the behind the paper text. In short we had an interest in phage H20, because Kalatzis et al. 2017 found it distributed worldwide, without any clear and obvious reasons on the genome annotation. So we made a prophage mutant and started searching for interesting phenotypes - and found that the answer was more complex than +/- prophage, but also took QS into account - Which in my humble opinion is super interesting from an evolutionary perspective.


I hope you find these answers useful. Good luck with entering this research field - phage-bacteria-interactions are great and interesting, and I hope you'll make some nice discoveries.


Mads Frederik Hansen

Go to the profile of Naresh kumar
4 months ago

Thank you so much for being so kind to answer all of my questions and clearing my doubts regarding the phage-host interactions and quorum sensing. It was really very nice to had a conversation with you. Once again, thanks for your replies! All the best for your future works as well, Mads Frederik Hansen!