University of Groningen The ecology and evolution of bacteriophages of mycosphere-inhabiting Paraburkholderia spp. Pratama, Akbar Adjie

The ecology and evolution of bacteriophages of mycosphere-inhabiting Paraburkholderia spp.

Pratama, Akbar Adjie

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2018

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Pratama, A. A. (2018). The ecology and evolution of bacteriophages of mycosphere-inhabiting Paraburkholderia spp. Rijksuniversiteit Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

The significance of mutualistic phages

for bacterial ecology and evolution

Nancy Obeng, Akbar Adjie Pratama and Jan Dirk van Elsas Published in Trends Microbiol. (2016) 24(6): 440-449 doi: 10.1016/j.tim.2015.12.009

Abstract

Bacteria and phages have traditionally been viewed as ‘antagonists’. However, temperate phages can transfer genes, which can broaden their bacterial hosts’ metabolic repertoire, confer or enhance virulence, or eliminate competing organisms, and so enhance bacterial fitness. Recent evidence shows that phages can also promote biofilm formation leading to population-level benefits for their bacterial hosts. Here, we provide a perspective on the ecological and evolutionary consequences for the bacteria interacting with phages, when phage and host interests are aligned. Furthermore, we examine the question whether bacterial hosts can lower immune barriers to phage infection, thereby facilitating infection by beneficial phages. Taking recent evidence together, we suggest that in many cases temperate phages are to be considered as being mutualistic as well as parasitic, at the same time.

5

A history of bacteriophages and bacteria as foes and friends

In natural ecosystems, many bacteria are accompanied – and even outnumbered – by their viruses, the bacteriophages (Breitbart and Rohwer, 2005; Clokie et al., 2011). The interplay between bacterial hosts and phages has been extensively studied since the 1960s, and it has become apparent that phages can act as population-size controllers and gene-transfer agents (Romig and Bodetsky, 1961; Taylor and Thorne, 1963). From more recent work, we have learned that bacterial genomes can be made up of up to 30% of phage inserts or remnants thereof (Casjens, 2003; Comeau et al., 2007; Petrov et al., 2010). These inserts constitute the ‘genomically-fossilized’ legacies of past phage–host encounters. While the role of lytic phages (see Glossary) as modulators of host population sizes is not to be underestimated, we will here focus on temperate phages, which interact with bacterial genomes in intricate ways. As temperate phages can integrate into and excise from bacterial genomes, genes from a previous host can be transferred, by transduction, into the genome of a current host (Figure 5.1, Key Figure). In particular cases, this may trigger the lysogenic conversion of the novel host, thereby potentially increasing its fitness.

Highlights

Temperate phages can provide a plethora of benefits to bacterial hosts, including the ability to respond rapidly to changes in the environment.

Bacterial biofilms can benefit from lysogeny. When phages are induced in a few cells, the lysing cells release nutri-ents and extracellular DNA locally, which may strengthen biofilm structure. Further, lysis breaks biofilm bonds, allowing for enhanced cell dispersal.

In bacterial competition, phages can facilitate bacteriocin release or act as ‘weapons’ against competitors themselves.

Recent evidence indicates that bacter- ial populations may modulate phage infections, as individual cells may lose immunity or selectively allow infections to occur. The capacity of CRISPR-Cas immune systems to discriminate between lytic and temperate phages (based on gene expression) may enable bacterial hosts to selectively permit temperate phages to enter.

Yet, inserted phages often impede host functioning. For example, phages might insert into – and thus disable – functional genes, bacterial metabolism may be slowed down, or induction may lead to host death (Bernhardt et al., 2000; Brüssow

et al., 2004; Coleman et al., 1991). Thus, bacterial hosts face a trade-off between the positive effects phage-carried genes bring and the costs of phage insertion. To curb the negative impact of infections, many hosts have evolved genetic systems that antagonize these: bacterial immune systems. In particular, the recently discovered CRISPR-Cas system interferes with invasion of the host genome (Samson et al., 2013), thereby constituting a central defense strategy. As phages may incidentally ‘break through’ such systems, there is an evolutionary pressure favoring new defense strategies in the infested populations. On the other hand, phage infection strategies also evolve, allowing newly emerging phages to infect the new hosts. This iterative dynamics has been coined an evolutionary arms race or Red Queen dynamics, depending on whether the resulting selection dynamics is characterized by selective sweeps or fluctuating selection (Hall et al., 2011; Råberg et al., 2014; Stern and Sorek, 2011). However, mutualistic interactions between the two partners have also shaped their coevolution (Brüssow et al., 2004; Carrolo et al., 2010; Edlin et al., 1975, 1977; Frias-Lopez et al., 2008; Gödeke et al., 2011; Lopez et al., 2012; Resch et al., 2005; Roossinck, 2011; Rossmann et al., 2015; Schuch and Fischetti, 2009). In any treatise on bacterium–phage interactions, both the positive and negative consequences of phage infection affecting coevolution need to be considered. Here, we examine how phages modulate their hosts’ life, both at the individual and population levels. We provide a perspective on the consequences of such interactions resulting from the tightly intertwined coevolutionary dynamics of bacteria and phages.

Benefits to bacterial hosts gained from phage infections

Given their propensity to ‘tinker’ with bacterial genomes, temperate phages can directly affect host behavior and fitness. For instance, inserted phages (called prophages) can protect their hosts against secondary infections through superinfection exclusion as they prevent similar phage particles from attaching to the host (Lynch et al., 2010; Matos et al., 2013). Moreover, prophages can introduce a plethora of genes that provide functions of different kinds to their hosts (Table 5.1). Such lysogenic conversions can take the form of stable phenotypic changes or increased host plasticity, thereby improving bacterial survival under changing conditions.

Stable changes can be brought about by phage-carried genes that enhance bacterial metabolism (Edlin et al., 1975, 1977; Sekulovic and Fortier, 2015) or extend the host metabolic repertoire, opening up new niches. This phenome-non-has previously been observed in lytic phages, where so-called auxiliary metabolic genes (AMGs) increase metabolism and divert metabolic pathways in diverse ways (as reviewed by Breitbart, (2012)). Here, we extend the scope of AMGs to temperate phages, while

5

changes Phenotypic plasticity of the individual

Traits affecting biofilm formation

phage 1 phage 1

Host cell surface Restriction modification system

BREX

Cas gene CRISPR locus

Innate immunity

Adaptive immunity

Integration

Red queen dynamics Coevolution Arms-race dynamics Fitness 111 Ev olutionar y selec tive pr essur e M icr obial c ommunit y Hor iz on tal gene tr ansf er phage 2 phage 2 Bacteriocin release in bacterial ‘warfare’ S elec tiv e adv an tages gained b y individual bac ter ia

and whole popula

tions Phenotypic plasticity 222 3 32 333 3 3 3 3 3 3 3 3 3 33 3 3 3

Example. Phage SopEΦ can extend S.Typhimurium’s metabolic repertoire by increasing iNOS production

iNOS 333 iNOS 333 333 3 3 iNOS iNOS iNOS NO3 NO3 NO3 NO3 NO3 NO3 NO 3 3 iNOS iNOS - SopEΦ + SopEΦ

. Phage SopEΦ can extend repertoire by increasing iNOS productionTyphimurium repertoire by increasing iNOS production. Phage SopEΦ can extend repertoire by increasing iNOS productionTyphimuriumTyphimuriumTyphimuriumTyphimuriumTyphimurium

NO3 NO NO3 NO NO NO NO3 _{+ SopEΦ} 3 3 iNOS iNOS iNOS iNOS iNOS iNOS 333 333 iNOS iNOS iNOS iNOS 333333 iNOS iNOS iNOS 3 333333 333 iNOS 333 333 iNOS 333 iNOS iNOS iNOS iNOS iNOS 333 333 iNOS iNOS iNOS iNOS iNOS NO NO NO NO NO33 iNOS iNOS iNOS

. Phage SopEΦ can extend . Phage SopEΦ can extend

iNOS

Typhimurium repertoire by increasing iNOS production. Phage SopEΦ can extend S.Typhimurium repertoire by increasing iNOS production. Phage SopEΦ can extend S.Typhimurium. Phage SopEΦ can extend . Phage SopEΦ can extend Typhimurium

iNOS iNOS iNOS Typhimurium’s metabolic Typhimurium iNOS iNOS

X

Example. Phages can mediate bacteriocin release in bacterial ‘warfare’

X

X

X

111

Bacteriocins

Example. Phages can mediate bacteriocin release in bacterial ‘warfare’

X

X

. Phages can mediate bacteriocin release in bacterial ‘warfare’ Bacteriocins Bacteriocins

X

X

X

X

X

X

Figure 5.1. In nature, a large and diverse pool of phages with host-beneficial genes can be integrated into the genomes of bacteria that are locally present. Bacteria have developed a range of defense mechanisms against phages, namely, restriction–modification (R–M) systems as innate immunity (Dempsey et al., 2005; Jermyn and Boyd, 2002), the bacteriophage exclusion (BREX) system (Goldfarb et al., 2015), and CRISPR-Cas systems (clustered regularly interspaced short palindromic repeats and CRISPR-associated proteins) (Samson et al., 2013) as specific phage defense or adaptive immunity. Phages have coevolved, yielding mechanisms that allow them to evade these bacterial systems. Moreover, the integration of phages carrying host-beneficial genes into host genomes offers selective advantages to the host, thereby shaping the coevolution between bacteria and phage. Depending on the coevolutionary dynamics (i.e., arms race or Red Queen dynamics), genetic variation may be affected. This in turn influences the diversity of phages and bacteria. Phages may affect bacterial fitness at the individual as well as population levels. Phage SopEF, for example, can increase the expression of inducible nitric oxide synthase (iNOS) in Salmonella

enterica serovar Typhimurium, thereby extending its host’s metabolic repertoire (Lopez et al., 2012).

Whole populations may benefit from infection, as phage lysis provides a mechanism for bacteriocin release toward competing strains or species (Nedialkova et al., 2016). Here, the different colors of the phage and host pools reflect variation in host cell-surface receptors corresponding to phage infection.

Glossary

Allelopathic agent: particle or chemical produced by a cell that influences the growth or survival of surrounding cells.

Arms race dynamics: coevolutionary dynamics in which consecutive selective sweeps fix alleles in host and parasite populations. This is expected to result in rapid evolution with transient polymorphisms and low standing genetic variation. Auxiliary metabolic genes (AMGs): genes carried by phages that affect host metabolism, for example by increasing photosynthesis, nucleotide synthesis, or the biosynthesis of alternative electron acceptors.

Bacterial immune systems: bacterial defense systems that target incoming foreign genetic material. Examples are enzymatic restriction– modification (R–M) systems, clustered regularly interspaced short palindromic repeat (CRISPR) loci and CRISPR-associated (CRISPR-Cas) immunity, and bacterial exclusion (BREX) systems.

Bacterial warfare: interference competition between bacteria, which may include the use of bacteriocins or phages that decrease growth or survival of competing strains or species.

Bet-hedging: theoretical framework that assumes individuals optimize their long-term fitness by following strategies that reduce variance in fitness.

Coevolution: evolutionary process in which interacting organisms evolve together as they reciprocally influence each other.

Epistasis: genetic interaction in which the expression of a gene (and the resulting phenotype) depends on the genetic background.

Lysogen: bacterial cell containing one or more prophages within its genome. Lysogenic conversion: phenotypic change in a host bacterium caused by the insertion of a phage.

Lysogeny: state of phage integration into the bacterial genome.

Lytic phage: bacterial virus, which, upon entering its host, produces offspring and lyses the bacterial cell to release its progeny.

Prophage: temperate phage integrated in the bacterial genome.

Red Queen dynamics: coevolutionary dynamics in which allelefrequencies in host and parasite populations fluctuate cyclically. This is expected to result in stable polymorphisms and high standing genetic variation.

Selective immunity: bacterial immunity that tolerates infection by certain phages (e.g., temperate) whilst preventing infection of other (e. g., lytic) phages.

5

Superinfection exclusion/ immunity: prevention of a secondary infection by a phage through (phage- dependent) cell-surface modifications.

Temperate phage: bacterial virus that can integrate into a bacterial genome (or be maintained extrachromosomally), become stabilized in this way, and, upon receiving a cue, can excise and propagate.

Transduction: process of horizontal gene transfer, wherein a region of a bacterial genome is packaged into phage particles that, upon release and entrance into a new host, is inserted into the genome of the latter.

highlighting important differences depending on phage life history. A key finding was that, in Salmonella enterica serotype Typhimurium, the temperate phage SopEF can increase the production of inducible nitric oxide synthase (iNOS) (Lopez et al., 2012), facilitating the production of precursors to the electron acceptor nitrate (Barrett and Riggs, 1982; Lopez et al., 2012). This may be fitness-enhancing by extending the pathways used for anaerobic respiration in the inflamed gut the organism often dwells in. Cells endowed with phage SopEF thus take profit from increased metabolic rates under anoxic conditions and outcompete wild-type cells in vivo (Lopez et al., 2012) (Figure 5.1). In lytic phages, AMGs are thought to divert host metabolism toward phage nucleotide synthesis (Breitbart, 2012) and only transiently increase bacterial fitness. While this phenomenon may facilitate the horizontal transmission of phages (Sullivan et al., 2006), AMGs transcribed from integrated phages make lysogenic cells better refuges and so foster vertical transmission. Future studies should explore to what extent the identity of AMGs and phage lifestyles are associated.

Phages can also modulate bacterial fitness using genetic systems that allow the host to respond plastically to changes in the environment. The stress response of a bacterial cell, for example, can be modulated by inserted phages, as observed in Escherichia coli (Wang et al., 2010). Under extreme oxidative, osmotic, or acid-stress conditions, lysogenic E. coli cells containing the prophages CPS-53 and CP4-57 exhibited a more stable metabolism than phage deletion mutants (Wang et al., 2010). Similarly, phage FCD38-2 allowed its host Clostridium difficile to alternate the composition of its cell wall (Sekulovic and Fortier, 2015). Specifically, FCD38-2 was found to ‘switch on’ the expression of cwpV, a gene, which encodes a cell wall protein associated with phase variation (Sekulovic and Fortier, 2015). This may enable the host cells to aggregate and form biofilms. By attachment to neighboring cells, bacterial survival may be enhanced, while vertical phage transmission is facilitated. As phase variation allows cells to respond to temporally variable environments, it may be considered a bet-hedging strategy (De Jong et al., 2011), in which phages modulate

Table 5.1.

Host-beneficial genes carried by temperate phages and their function for the bacterial host

Host beneficial tr ait Phage(s) Host Affect ed gene(s) and/or e xpr ession Refs

Stable phenotypic changes

Superinf ection ex clusion KL3 Bur kholderia cepacia Unkno wn (L ynch et al., 2010) Unkno wn Ent er oc oc cus f aec alis Unkno wn (Mat os et al., 2013) Car bon metabolism λ, P1, P2, and Mu Esc heric hia c oli Unkno wn (E dlin et al., 1975, 1977) Nitr at e metabolism ΦCD38-2 Clostridium difficile Inducible nitric o xide s ynthase (Lopez et al., 2012) Virulence f act ors CTX Φ Vibrio c holer ae VP -2 pathogenicity island (Dempse y et al., 2005) SM1 Str ept oc oc cus mitis pblA and pblB (plat elet-binding pr ot eins)

(Gomez and Buckling, 2013)

sopEΦ Salmone lla ent eric a ser ov ar Typhimurium sopE (r esponsible f or int estinal inflammation) (Br ockhurst et al., 2007) Lambdoid phages Esc heric hia c oli Shig a t oxins st x-1 and st x-2 (Brüsso w et al., 2004) ΦCD38-2 Clostridium difficile Regulation of t oxin e xpr ession (Sek ulo vic and F ortier , 2015) ΦCTX Pseudomonas aeruginosa Cyt ot oxin (F

ellous and Sal

vaudon, 2009)

Phenotypic plasticity of the indi

vidual Str ess t oler ance CPS-53 and CP4- 57 Esc heric hia c oli Upr egulation of gener al str ess r esponse regulat ors under o xidati ve, osmotic, and acid str ess (W ang et al., 2010) Phase v ariation ΦCD38-2 Clostridium f acile cwpV (cell w all pr ot ein) (Sek ulo vic and F ortier , 2015) Quorum sensing phiCDM1 Clostridium difficile Agr QS cassett e without agrA (r esponse regulat or) (Bensing et al., 2001) λ Esc heric hia c oli Unkno wn (Wint er et al., 2010)

5

Table 5.1.

Host-beneficial genes carried by temperate phages and their function for the bacterial host (Continued)

Host beneficial tr ait Phage(s) Host Affect ed gene(s) and/or e xpr ession Refs Tr aits aff ecting

biofilm formation or bact

erial ‘w arf ar e’ Motility CP4-57 Esc heric hia c oli Regulat ors of lif esty le s wit ch fr om hig h to lo w motility (Nanda et al., 2015) Metabolism Unkno wn Esc heric hia c oli Upr

egulation of genes mediating

ex

opol

ysaccharide e

xcr

etion

(Schuch and Fischetti, 2009)

Ly sis within in biofilms SV 1 Str ept oc oc cus pneumoniae sv1 (aut ol ysin) (Carr olo et al., 2010) Unkno wn Shew ane lla oneidensis MR1 Unkno wn (Gödek e et al., 2011) Unkno wn Str ept oc oc cus pneunomiae Aut ol ysin (F rias et al., 2009) Ly sis leading t o bact eriocin r elease ST64B Salmone lla ent eric a ser ov ar Typhimurium ST64B l ysis genes (SL1344_1955-SL1344_1957) (Cascales et al., 2007)

the bacterial strategy. However, it is questionable to what extent environmental cues affect phase variation, and whether it may still be considered bet-hedging when phages render the bacterial population genetically heterogeneous (Shimizu et al., 2009). Clearly, this phage–host interaction should be regarded as mutualistic, as it enhances the survival of host and phage alike.

The lines between phage–host antagonism and

mutualism blur at the population scale

Phage-borne traits can modify bacterial populations and the way they operate. Depending on the consequences of lysogenic conversion, the selection pressures shaping the bacterium - phage coevolution may differ. For instance, direct benefits of increased metabolism confer a selective advantage to individual cells, resulting in successful subpopulations. Concerning traits that affect neighboring bacteria, phages also influence population-level selective pressures. Here, the negative consequences of infection that an individual may experience may actually lead to population- level benefits or vice versa, making antagonism as well as mutualism highly context-dependent.

When phages encode virulence factors (VFs) that enable hosts to become pathogenic (i.e., lambdoid coliphages providing the Shiga-toxin to E. coli (Brüssow et al., 2004; Shimizu et al., 2009)), these phage-encoded secreted toxins may be considered ‘public goods’ to the local bacterial population (West et al., 2007). In other words, individual lysogenized cells ‘pay’ the cost of VF production and secretion, while all neighboring cells, including non-lysogen and non-producer ones, benefit from these (West et al., 2007). The latter neither suffers from the general cost of phage infection nor from the burden of VF production, and are thus favored within a population. Yet, to invade a eukaryotic host, sufficient cooperation between the bacterial cells is required to sustain virulence. Here, high rates of phage infection might support bacterial cooperation. Alternatively, bacteria can circumvent the cost of a potentially destructive prophage by degrading the harmful genes whilst maintaining VFs (Bobay et al., 2014).

Phages further affect bacterial populations that occur in biofilms (Nanda et al., 2015). In the first steps of biofilm formation, some phages may reduce bacterial motility, which spurs the settlement of cells in aggregates (Wang et al., 2009). Within the early biofilm, phages can further modulate bacterial metabolism, providing extracellular polysaccharides that are required for biofilm structure, or act as structural building blocks themselves (Costerton et al., 1999; Schuch and Fischetti, 2009; Secor et al., 2015). Phages of Bacillus anthracis, for example, may thus contribute to biofilm formation in (semi-artificial) soil and host settings (Schuch and Fischetti,

5

2009). Furthermore, lysis of a subset of cells within a biofilm can enhance biofilm formation and maintenance. As recently reviewed (Nanda et al., 2015), bacterial and phage gene expression can be differentially affected by the (nutrient-and oxygen-limited) biofilm environment. Here, high cell densities and abundant reactive oxygen species may incite phage induction in a context-dependent manner (Resch et al., 2005; Webb et al., 2003). The ensuing local lysis may improve biofilm formation by (i) providing additional nutrients for the growth of neighboring cells, (ii) providing extracellular DNA (eDNA), which is an integral part of the biofilm matrix (Carrolo et al., 2010; Frias et al., 2009; Gödeke et al., 2011), and (iii) creating hollow centers in the biofilm, which facilitates bacterial dispersal (Rice et al., 2009) (Figure 5.2). Sometimes, such seemingly stochastic lysis can be prompted by the coordinated action of phage and bacterial lysin (Frias et al., 2009). Here, one might consider this common phage/host gene expression itself to be mutualistic, as host persistence is fortified, thus serving host and phage alike. External forces, such as desiccation, mechanical or antibiotic stresses (Secor et al., 2015) acting on the biofilm, may then indirectly favor the lysogeny of some individuals in the population.

But why would an individual host cell ultimately lyse? In the case of E. coli cells infected with (lytic) phage T4rII, neighboring cells gain great benefits when infected cells die, while – to the population – individual cell death only presents a marginal cost (Refardt et al., 2013; Resch et al., 2005). This appears altruistic, as it benefits the surviving cells (recipients) at the detriment of the lysing cell (actor). How and to what extent environmental conditions in the biofilm trigger the heterogeneous lysis due to temperate phages, or whether this process is purely stochastic, still need to be determined. Overall, when a sufficient number of cells in a biofilm is lysogenized and a subset of these produces eDNA by lysis, the biofilm becomes more stable. In line with this reasoning, recent experiments (Carrolo et al., 2010; Frias et al., 2009; Gödeke et al., 2011; Resch et al., 2005; Schuch and Fischetti, 2009) have shown that biofilms containing lysogens are indeed more proliferative than their nonlysogen counterparts.

Phage-induced lysis can furthermore play a role in bacterial warfare. When bacteria compete for resources, they may produce toxins that are harmful to the competitor. In the case of one class of toxins, bacteriocins, their export is often mediated by specific proteins encoded in bacteriocin operons (Cascales et al., 2007). However, several operons encoding group B colicins (produced by E. coli and S. enterica serovar Typhimurium (Cascales et al., 2007)) do not encode export proteins and the respective bacterial cells are therefore theoretically hampered in colicin release. Interestingly, this may be compensated for by release due to phage-induced lysis (Figure 5.1). Thus, in such cases, phages mediate the release of bacteriocins,

and so the fitness of phage-harboring strains is enhanced (Nedialkova et al., 2016). Bacteria thus produce bacteriocins, and phages aid in the transport of the public good, to the benefit of the host population.

Finally, phages released from lysogenized cells in biofilms can act as allelopathic agents (Brown et al., 2006; Stewart and Levin, 1984). Surviving lysogens may take profit from the decreased population density (Rossmann et al., 2015), while newly lysogenized cells avert immediate death and acquire potential future weapons against competitors. At the same time, the phage population benefits from increased transmission. Thus, we posit that selection should favor lysogeny in bacterial populations, particularly when the ensuing phage-induced benefits have a significant ecological impact. Biofilm Hollow area Phage eDNA Polysaccharide Protein Dispersal cell Dead cell Nutrient limitation Unfavorable oxygen concentration High cell density

ROS accumulation eDNA Surface attachment Aggregate cell Aggregate cell Aggregate cell

Figure 5.2. Prophages can act as drivers of biofilm formation. In all ecosystems, from natural

ones to clinical settings, bacteria commonly aggregate in biofilms. To form a biofilm, cells first disperse in order to colonize the surface at which the biofilm forms (Rossmann et al., 2015). Within biofilms, prophage release is a frequent event. It is related to nutrient and oxygen limitation, high cell densities, and reactive oxygen species (ROS) accumulation (Resch et al., 2005; Webb et al., 2003). While lysis may be undesirable for single cells, it seems to be advantageous for the whole population. During biofilm formation, phages help to degrade the biofilm matrix, break biofilm bonds, and eliminate cells, thereby forming hollow areas in the center of microcolonies (Rice et al., 2009). Thus, nutrients are provided for the surviving cells.

5

Challenges and loopholes of bacterial immunity to exogenous DNA

In order for phages to insert and bacterial hosts to reap beneficial effects from the novel inserts, phages need to overcome host cell defenses (Figure 5.1). The adsorption, invasion, integration, and gene expression of the infecting phages are thus critical.

Adsorption depends on specific interactions between bacterial cell-surface receptors (CSRs) and phage receptor-binding proteins (RBPs). Both are prerequisites for successful adsorption (Samson et al., 2013). The reciprocal evolution between bacteria and phages has created diversification in CSRs, often driven by the need to avert phage adsorption. At the same time, phages have evolved strategies to overcome such barriers. Interestingly, mutations in RBP genes have been found to enable lambdoid phages to not only bind to the CSR LamB, but also to ompF (Meyer et al., 2012). Moreover, in certain cases, phages have partially acquired bacterial immune systems, such as restriction– modification systems (Dempsey et al., 2005; Jermyn and Boyd, 2002) or recognition sequences of the CRISPR-Cas system (Samson et al., 2013), allowing them to remain undetected and be maintained by bacterial hosts. Here, we posit that, when there are benefits associated with lysogeny, the net selection pressure against phages should decrease and attenuate the ‘arms race’. So, would bacterial immunity be lowered in such cases, facilitating a controlled level of infection? Based on empirical evidence and theoretical considerations (Gophna et al., 2015; Jiang et al., 2013), one mechanism, which enables bacterial cells to take up phages, might be akin to a loss or modulation of immunity. In Staphylococcus epidermidis, individual cells within populations were found to stochastically lose their CRISPR-Cas immunity, as key recognition sequences acquired mutations (Jiang et al., 2013). The resulting heterogeneous host cell populations could thus, to a certain extent, acquire novel genetic material without being completely vulnerable to phage attack. Individual cells that allow phages to enter can be seen to ‘gamble with their life’, as they will often be also exposed to lytic phages. To differentiate whether this may be a response to temporal variation, that is, bet-hedging, or a response to environmental changes, we will need to elucidate the respective selection pressures, often subtle in nature, that operate in the system (De Jong et al., 2011).

Instead of losing their immunity, host cells may develop ‘selective immunity’, which allows to tolerate potentially beneficial infections by temperate phages, yet to immediately degrade lytic phages (Goldberg et al., 2014). For instance, the type III CRISPR-Cas system in S. epidermidis apparently targets phages that actively transcribe recognition sequences whilst tolerating temperate phages that do not express lysis genes until induction (Goldberg et al., 2014). As a result, the latter may insert into host genomes and potentially enhance bacterial fitness. Correspondingly,

CRISPR-Cas immunity does not seem to prevent horizontal gene transfer in the long term (Gophna et al., 2015). Whether this ‘selective immunity’ is limited to type III CRISPR-Cas systems (Goldberg et al., 2014) or extends to other systems has yet to be explored. More studies on the selection pressures mediating CRISPR-Cas immunity to phages that carry host-beneficial genes are clearly required. Future analyses should also consider other bacterial defense systems, such as the recently described bacteriophage exclusion (BREX) system. This system blocks the replication of phages that contain the nonpalindromic TAGGAG repeats (Goldfarb et al., 2015). As some phages can overcome this barrier, the BREX system might be subjected to similar dynamics as the CRISPR-Cas system (Goldfarb et al., 2015).

In the long term, the fate of prophages after integration will be determined by the conditions within the host. Here, epistasis can attenuate the benefits of the genomic changes (Scanlan et al., 2015), and deletion bias may partially or completely remove costly or deleterious phage genes or even whole prophages. Alternatively, these can mutate to less-harmful sequences and thus be tolerated in the host population (Bobay et al., 2014). Undoubtedly, phage hosts face the trade-offs between the costs of expanded genomes (Bernhardt et al., 2000; Lawrence et al., 2001) and the putative benefits, which they strive to optimize at both the individual and population levels.

Concluding remarks

Several recent studies have indicated that phages drive bacterial adaptation and fitness at different levels, often by introducing beneficial genes (Brüssow et al., 2004; Carrolo et al., 2010; Edlin et al., 1975, 1977; Frias-Lopez et al., 2008; Gödeke et al., 2011; Lopez et al., 2012; Resch et al., 2005; Roossinck, 2011; Rossmann et al., 2015; Schuch and Fischetti, 2009). The genome–genome interactions leading to successful combinations are often mutualistic rather than purely parasitic. The coevolutionary consequences of these interactions are complex, as they involve processes from the individual to the population level. A presumably stochastic lowering of the barriers in bacterial hosts to infection by phages appears to be involved. As phages can aid in biofilm formation, in the release of bacteriocins, and in the fight against competitors, they can increase the adaptive and competitive ability of bacteria within populations. In the light of such phage-provided benefits, lysogeny is expected to be often favored in bacterial populations. While limited by the possible costs, it is therefore likely that lysogeny is common among bacteria as it is driven by the large potential for beneficial effects (Campbell, 1961). In spite of our increased understanding of the trade-offs that govern phage–host relationships, there is still much unknown about how these are regulated, in particular when temperate phages carry genes for host fitness-enhancing traits (see Outstanding Questions). First, do we understand all immune systems that

5

function as micro lifeguards in bacterial hosts? Further, what molecular mechanisms govern the population-scale regulation of phage maintenance, gene expression, and lysis, and how is this triggered by the environment? Have we seen the whole plethora of phage-provided benefits, or have we just opened a true Pandora’s Box? Future studies should place a focus on the intricacies of temperate phage–host interactions, quantifying the fitness effects phage-carried beneficial genes have on bacterial cells and their populations. This extends to a need to understand the trade-offs involved for the phages themselves, as they maintain host-beneficial genes (Hatfull and Hendrix, 2011) and determine life history (Maslov and Sneppen, 2015). Further-more, mathematical models will need to more specifically account for the different types and levels of bacterium–phage interactions (Koonin and Wolf, 2015) and integrate the complex interplay of the benefits and costs of lysogeny.

Moreover, the aforementioned mutualistic interactions may affect the coevolutionary dynamics, which was originally characterized solely in terms of selection pressures on host defense and phage infectivity. So far, it has been found that the costs of resistance and infectivity as well as population size, host specificity, and habitat structure may affect whether coevolution exhibits an arms race or rather a Red Queen dynamics and how this affects the standing genetic variation (Gomez and Buckling, 2013; Hall et al., 2011; Tellier et al., 2014). Due to mutualistic effects at the population level, we expect the dynamics of coevolution with temperate phages to differ from that observed with lytic phages (Brockhurst et al., 2007; Hall et al., 2011). Likely, such a lowered cost to infection (or, inversely, a higher cost to host resistance (Hall et al., 2011)) decreases the pressure on resistance development. It would thus lead to Red Queen dynamics with fluctuating selection; likely yielding increased diversity of phages and the bacterial hosts.

During phage–host coevolution, the (stochastic) transfer of beneficial genes from bacterial host genomes to phages may, indirectly, foster the fitness of phages, as some phages may encounter a new host, endow it with novel beneficial traits, and be perpetuated. Moreover, it is eminently possible that particular hosts have started to ‘rely on’ phages, as these provide beneficial traits which are not commonly encoded by the bacterial genome. Temperate phages carrying host-beneficial genes may in such cases be regarded as mutualists. In particular when phage-conferred traits cannot readily evolve de novo and are not fixed in bacterial genomes, bacteria may become dependent on these (at least temporarily), making phages truly resemble ‘genetic symbionts’ (Fellous and Salvaudon, 2009). Finally, understanding the dynamics of the interactions between temperate phages and their hosts will foster our theories on the intricacies of these microbial interactions, and play a pivotal role in, for instance, the development of phage therapy as an alternative to antibiotics. We will need to

investigate, in particular, the events at the single-cell level and explore how these affect population-level outcomes. Calling the antagonistic account of coevolution between bacteria and temperate phages into question, we now have to integrate the interplay of both mutualistic and antagonistic interactions into a comprehensive framework of coevolutionary dynamics, which ranges from single cells to whole populations.

Outstanding Questions

How do bacteria selectively integrate, maintain, modulate, and transcribe potentially beneficial genes within their prophages?

And how do such genes overcome the selective pressures within the bacterial genome (i.e., dele-tion bias)?

How do phage abundance and infec-tion rates affect bacterial social inter-actions, such as the cooperative production of ‘public goods’ or com-munication through quorum sensing?

How are host defense mechanisms against phages affected by abiotic and biotic conditions, and how do phages overcome these?

To what extent can bacteria selectively admit temperate and reject lytic phages using immune systems, includ-ing and even beyond the type III CRISPR-Cas system?

How does the coevolutionary dynam- ics between bacteria and temperate phages differ from that with lytic phages?

What is the influence of beneficial phage genes on this dynamics?

How does the coevolutionary dynamics change in a system in which both temperate and lytic phages (which can diminish the host cell pool rapidly) are present?

Acknowledgments

This work was supported through a scholarship of the German National Academic Foundation to N.O. and a scholarship of the Indonesia Endowment Fund for Education (LPDP - Lembaga Pengelolaan Dana Pendidikan, Departemen Keuangan, Republik Indonesia) to A.A.P. Further, we would like to thank Dr Gail Teitzel for structural comments on an early draft and, importantly, two anonymous reviewers for their critical and helpful comments.