The role of rhizosphere phages in soil health

Abstract While the One Health framework has emphasized the importance of soil microbiomes for plant and human health, one of the most diverse and abundant groups—bacterial viruses, i.e. phages—has been mostly neglected. This perspective reviews the significance of phages for plant health in rhizosphere and explores their ecological and evolutionary impacts on soil ecosystems. We first summarize our current understanding of the diversity and ecological roles of phages in soil microbiomes in terms of nutrient cycling, top-down density regulation, and pathogen suppression. We then consider how phages drive bacterial evolution in soils by promoting horizontal gene transfer, encoding auxiliary metabolic genes that increase host bacterial fitness, and selecting for phage-resistant mutants with altered ecology due to trade-offs with pathogen competitiveness and virulence. Finally, we consider challenges and avenues for phage research in soil ecosystems and how to elucidate the significance of phages for microbial ecology and evolution and soil ecosystem functioning in the future. We conclude that similar to bacteria, phages likely play important roles in connecting different One Health compartments, affecting microbiome diversity and functions in soils. From the applied perspective, phages could offer novel approaches to modulate and optimize microbial and microbe–plant interactions to enhance soil health.


Introduction
The One Health concept underscores the interconnectedness of human, animal, plant, and environmental health, emphasizing their interdependence (Banerjee and van der Heijden 2022 ).Originally, the One Health concept centered on the transmission of zoonotic pathogens, vectors of pathogens, and movement and persistence of antibiotic resistance genes (ARGs) across environments (Destoumieux-Garzón et al. 2018 ).Recentl y, soils hav e been proposed to be a critical compartment linking human, animal, and plant health via the sharing of micr oor ganisms (Banerjee and van der Heijden 2022 ).Further, soil microbiomes have been estimated to provide over 40 functions linked with plant growth, nutrient uptake and cycling, provision of essential ecosystem services, and suppression of pathogens, highlighting their importance in One Health fr ame work (Lehmann et al. 2020 , Banerjee andvan der Heijden 2022 ).Especially, plant rhizosphere is considered to play an important role in maintaining high microbial div ersity, whic h has been positiv el y associated with soil health, crop yields, and ecosystem functioning (Fierer 2017, Saleem et al. 2019, Wagg et al. 2019, Banerjee and van der Heijden 2022 ).The benefits of soil biodiversity can be mediated via mutualistic interactions between plants and micr oor ganisms, r esulting in impr ov ed plant gr owth (Triv edi et al. 2020, Li et al. 2021a ).Additionally, high micr obial div ersity can be beneficial in pr e v enting pathogen inv asions and dominance, reducing economic losses on a gricultur al pr oduction (Str ange and Scott 2005, Delgado-Baquerizo et al. 2020, Zheng et al. 2020 ).Suc h pathogen suppr ession can be driven by se v er al differ ent gr oups of micr oor ganisms, including bacteria, fungi, protists, and bacteria-specific viruses-phages (Buee et al. 2009, Wei et al. 2015, Wang et al. 2019, Xiong et al. 2020 ).Of these micr obes, pha ges ar e the most understudied gr oup e v en though they are estimated to be the most abundant entities on Earth (estimate 10 31 ) (Breitbart andRohwer 2005 , Srinivasiah et al. 2013 ), constituting a significant portion of virus-like particles (VLPs) in the environment (Breitbart and Rohwer 2005 ).Previous research has demonstrated that phages play crucial roles in influencing micr obial abundances, div ersity, and functioning of ecosystems across aquatic, human, and terrestrial environments (Wilhelm and Suttle 1999, Brum et al. 2016, Argov et al. 2017 ).Ho w e v er, the role of phages in soil, rhizosphere and plant health remains relativ el y understudied.In this perspective, we review the significance of phages for soil health and microbial diversity within the One Health fr ame w ork b y focusing on both their ecological and evolutionary roles in rhizosphere microbiomes.

The di v ersity of phages in the plant rhizosphere
While the significance of phages for microbial ecology and evolution has been extensiv el y studied especially in marine ecosystems (Diaz-Munoz andKoskella 2014 , Breitbart et al. 2018 ), the knowledge of phage diversity and their roles in soil and plant rhizospher e micr obiomes is still r elativ el y limited.In contr ast to aquatic en vironments , soils exhibit higher heterogeneity due to variations in soil particle size, nutrient composition, soil biota, and plant community diversity and composition (Sharma et al. 2011(Sharma et al. , 2016 ) ). Soils exhibit higher heter ogeneity compar ed to aquatic environments, consisting of solid, liquid, and gaseous phases, which can consider abl y v ary acr oss space and time (Moldrup et al. 2001 ).Specifically, soil particle and aggregate sizes, and pore spaces affect hydraulic connectivity (Tecon andOr 2017 , Roux andEmerson 2022 ), which influences the movement of phages and bacteria within the soil matrix along with water (Mckay et al. 1993, Marsh and Wellington 1994, Philippot et al. 2023 ).Well-a ggr egated soils typically possess larger pores and better water infiltration compar ed to poorl y structur ed ones, potentiall y facilitating gr eater dispersal and distribution of pha ges.Similarl y, soils with larger particles (e.g.sandy soils) might allow easier phage movement owing to greater pore space and better water flow.In support of this, pr e vious studies hav e shown that viral communities respond r a pidl y to wetting e v ents, r esulting in a significant increase in vir al ric hness (Santos-Medellin et al. 2023 ).Additionall y, the adsor ption of viruses to soil particles, such as colloidal surfaces, has been found to enhance viral persistence and support higher viral abundances (Lipson andStotzky 1986 , Zhuang andJin 2003 ).It has also been shown that viral communities undergo shifts along thawing permafrost peatland soils, with correlations observed with host community composition, pH, soil moisture content, and soil depth (Emerson et al. 2018 ).As a result, the abundance and diversity of bacteria and phages varies spatially between and within terrestrial habitats and soil types (Williamson et al. 2017, Geisen et al. 2019 ), resulting in relatively higher phage diversity in soils compared to aquatic environments (Guemes et al. 2016 ).The estimated number of distinct viral genotypes (richness) identified in soils varies between 1000 and 1 000 000 depending on the soil type (Ashelford et al. 2003, Williamson et al. 2005, Reavy et al. 2015, Williamson et al. 2017 ).In contr ast, vir al ric hness estimates v ary fr om 532 to 129 000 for marine, and between 400 and 40 000 viral types for freshwater systems (Green et al. 2015 ).Moreover, phage abundances have been estimated to be r elativ el y mor e homogeneous in seaw ater (betw een 10 5 and 10 7 particles per millilitr e) compar ed to soils (between 10 3 and 10 9 particles per gram of soil) (Graham et al. 2019 ).While these diversity and abundance estimates ar e highl y v ariable, likel y due to v ariation in abiotic and biotic factors between and within both habitats, they suggest that terr estrial envir onments harbour higher taxonomic vir al div ersity.Pha ges also exhibit extensiv e mor phological div ersity, including tailed, nontailed, and filamentous forms (Nobrega et al. 2018 ), show high variability in nucleic acids composition (dsDNA, ssDN A, dsRN A, and ssRN A) (Dion et al. 2020 ) and have adopted v arious life cycles, r anging fr om l ytic to l ysogenic and c hr onic (Simmonds et al. 2017 , Ofir andSorek 2018 ).Especially, temperate phages capable of either lysing the host cell or integrating and replicating as part of the host genome are prevalent in soils (Argov et al. 2017, Ho w ar d-Varona et al. 2017, Sharma et al. 2019 ).When integrated into the host genome during lysogeny, temperate phages can facilitate their host survival, thus sustaining phage and bacterial population densities in soils (Williamson et al. 2002, Weinbauer et al. 2003, Chibani-Chennoufi et al. 2004 ).
Soil phage diversity also varies spatially within a given location.One important factor shaping local pha ge comm unity diversity is the presence of plants.For example, it has been shown that vir al-to-bacterial r atios differ between the rhizosphere and bulk soils and that lo w er bacterial abundances in the bulk soil are thought to be the contributing factor explaining r elativ el y lo w er pha ge div ersity (Swanson et al. 2009 ).Similarly, Bi et al .( 2021 ) identified significant differences in phage community composi-tion between rhizosphere and bulk soils, which were mainly attributed to the absence of plant root exudates and plant litter in the bulk soil that acts as nutrient sources for bacteria (Haichar et al. 2008, Zhalnina et al. 2018 ), promoting also higher phage abundances and div ersity.Mor eov er, rhizospher e pha ge comm unities vary between different plant species, due to differences in plant root exudate composition (Zhalnina et al. 2018 ), which can determine the composition and abundances of plant-associated bacterial microbiomes (Wang et al. 2022 ).In addition to spatial v ariation, the div ersity and activity of pha ges can show dynamic tempor al c hanges, driv en by seasonal fluctuations in temper atur e (Coclet et al. 2023), soil moisture (Santos-Medellin et al. 2023 ), plant growth and development (Yang et al. 2023 ), and anthropogenic factors such as crop management (Muscatt et al. 2022 ).These external factors can also influence the intrinsic dynamics between phages and their host, e.g. by triggering population fluctuations and ecological succession in microbial communities, which can feed back to ecosystem functioning (Liao et al. 2023, Santos-Medellin et al. 2023 ).T he emergence of no vel methodologies for purifying, extracting, and quantifying phage assemblages from soils (Williamson et al. 2003 ), coupled with advancements in meta genomic sequencing, has pav ed the way for a more in-depth exploration of the true soil phage diversity (Paez-Espino et al. 2019, Koonin and Yutin 2020, Roux and Emerson 2022, Mabrouk et al. 2023 ), while metatranscriptomic approaches have opened the door to study phage activity and RNA phage diversity (Callanan et al. 2020, Neri et al. 2022 ).With these methods, we are continuousl y discov ering nov el pha ge div ersity that does not exist in the current databases (Roux and Emerson 2022 ), increasing our understanding on the significance and v arious r oles pha ges play for soil and rhizosphere microbiome diversity and functioning in terrestrial ecosystems.

Ecological impacts of rhizosphere phages in soil ecosystems
The viral shunt: the role of phages in soil nutrient cycle Phages play a crucial role in impacting the mortality of bacteria by lysing and releasing host cell contents into the environment, with significant consequences for the global cycling of nutrients, energy flow, and food web dynamics-a phenomenon known as the "viral shunt" in aquatic systems (Brum andSulliv an 2015 , Br eitbart et al. 2018 ) (Fig. 1 A).For instance, marine pha ges ar e estimated to l yse a ppr oximatel y one-third of ocean micr oor ganisms per day, liber ating carbon at a globally significant scale (Suttle 2005, 2007, Guidi et al. 2016 ).In contr ast, ther e ar e still significant knowledge gaps regarding how soil phages contribute to food webs, decomposition of organic matter, carbon and nutrient cycling, greenhouse gas emissions, and a gricultur al pr oductivity.Virulent pha ges can contr ol host population abundances, thereby influencing rates of microbially mediated processes linked with the cycling of carbon, nitrogen, sulfur, and phosphorus in soils (Williamson et al. 2005, Kim ur a et al. 2008, Allen et al. 2010, Helsley et al. 2014 ).For instance, phages that infect nitrogen-fixing rhizobia bacteria, have been shown to reduce nodulation by pha ge-sensitiv e rhizobia, thus influencing nitrogen fixation in legume-rhizobium symbiosis (Evans et al. 1979, Sharma et al. 2002 ).Recentl y, tempor al c hanges in pha ge comm unities were linked with nutrient cycling during composting, where phages specific to mesophilic and thermophilic bacteria tr ac ked their host densities, triggering bacteria-pha ge comm unity succession via top-down control (Liao et al. 2023 ).Crucially, nutrient turnov er corr elated positiv el y with virus-host r atio, indicativ e of a positiv e r elationship between ecosystem functioning, vir al abundances, and viral activity (Liao et al. 2023 ).Moreover, viruses specific to mesophilic bacteria encoded and expressed several auxiliary metabolic genes (AMGs) linked to carbon cycling, impacting nutrient turnover alongside bacteria (Liao et al. 2023 ).Both temper ate pha ges and virulent pha ges can encode AMGs, whic h can affect the host cell metabolism in ways that promote viral replication and host survival (Ho w ar d-Varona et al. 2020 , Puxty andMillard 2023 ).Phages are hence likely to influence nutrient cycling by metabolicall y r epr ogr amming their host bacteria thr ough the expression of virus-carried AMGs (Breitbart 2012 ) linked with carbon metabolism and degradation of soil organic matter (Trubl et al. 2018, Wu et al. 2021 ).It remains yet unclear how phage-mediated nutrient cycling in the soil cascades through microbial food webs, affecting the plant growth and fertility of soils.Both lab-and fieldscale experiments, are hence required to directly test how viruses might shape the soil microbiome composition and plant growth by turning over bacterial biomass via lysis and by encoding AMGs associated with nutrient cycling.

Virulent phages: suppression of soil-borne bacterial pathogens via top-down density control
Virulent pha ges obligatel y infect and l yse their host cells, exerting significant effects on host cell densities and triggering competiti ve d ynamics within microbial communities (Morella et al. 2018 ) (Fig. 1 B).As a result, virulent phages have been extensively explored for their potential as biological control agents, demonstrating successful applications in the control of plant and zoonotic pathogens in soils (Jones et al. 2007, Buttimer et al. 2017 ).Phages ar e also mor e likel y to surviv e in the soil as long as their host bacteria ar e pr esent compar ed to biocontr ol bacteria that might hav e poor survival in the rhizosphere due to lack of vacant niche space and competition with native microbiota (Jones et al. 2007, Brodeur 2012, Meyer 2013, Buttimer et al. 2017, Kaminsky et al. 2019, AL-Ishaq et al. 2021 ).The earliest evidence of phage therapy applied to plant disease dates back to 1924 when the filtrate from rotten cabbage was used to inhibit the growth of Xanthomonas campestris (Mallmann and Hemstreet 1924 ).Since then, phage therapies have been used in the treatment of various phytopathogens, including Pseudomonas syringae (Frampton et al. 2014, Rombouts et al. 2016 ), Ralstonia solanacearum (Fujiwara et al. 2011, Wang et al. 2019 ), Xanthomonas spp.(Balogh et al. 2008 ), Erwinia amylovora (Schnabel andJones 2001 , Kim et al. 2020 ), as well as zoonotic pathogens present in soils, including Salmonella anatum (Gessel et al. 2004 ) and Rhodococcus equi (Salifu et al. 2013 ).Most of these studies predominantly examined the direct effects of virulent phages on the population densities of target pathogens through cell lysis.As a r esult, ther e is a growing interest in utilizing phage-derived proteins, suc h as endol ysins, for pathogen biocontr ol (O'Flaherty et al. 2009 ).Researc hers hav e identified se v er al tec hniques to enhance the efficacy of phage biocontrol, including optimizing the timing, fr equency, and dosa ge of a pplication (Iriarte et al. 2007, Cui et al. 2019, Li et al. 2020 ) 2016 ) or probiotics (Wang et al. 2017 ).Several products have been de v eloped and alr eady made commerciall y av ailable, including AgriPha ge ™ and Erwipha ge ™ (Gr ace et al. 2021 ).Ho w e v er, most of these pr oducts tar get pathogens in the phyllosphere and no products targeting soil-borne pathogens are yet commercially available to our knowledge.
Beyond the ecological impacts of phage therapy in controlling pathogen densities , phages ma y also ha v e indir ect benefits for soil micr obiome div ersity by pr e v enting pathogens fr om monopolizing the niche space, and potentially stabilizing interactions in rhizospher e micr obiomes (Federici et al. 2020 ) (Fig. 1 C).One pr e vious study demonstrated that R. solanacearum -specific phage cocktail buffer ed the r esident rhizospher e micr obiota a gainst c hanges induced by the pathogen invasion (Wang et al. 2019 ).In addition to affecting the bacterial community composition, the application of the phage cocktail also changed the potential functioning of the bacterial community by altering the proportion of taxa that exhibited facilitative or antagonistic pairwise interactions with the pathogen (Wang et al. 2019 ).These experimental findings are in line with other studies where healthy plants were associated with higher abundances of R. solanacearum -specific phages and a higher proportion of bacterial taxa exhibiting antagonism tow ar ds the pathogen (Wei et al. 2019, Yang et al. 2023 ).Inter estingl y, phages specific to bacterial taxa that sho w ed antagonism to w ar ds R. solanacearum , had an indir ect positiv e effect on plant disease by controlling the densities of antagonistic bacteria both in the lab and the greenhouse experiments (Yang et al. 2023 ).While phage effets might not always be so drastic on the surrounding microbiota (Magar et al. 2024 ), these findings suggest that phage efficacy could be context-dependent, shaping and being shaped by the resident microbiota present in the rhizosphere.Such complex ecological feedback and dynamics triggered by virulent pha ges ar e now only starting to be discov er ed and will have implications for microbial soil ecology beyond pathogen density control.

Temperate phages: horizontal transfer of AMGs in soils
Temper ate pha ges ar e widel y ac knowledged for their r ole in mediating horizontal gene transfer between bacteria via transduction and phages could hence drive bacterial evolution by promoting recombination and provision of new genes (Lwoff 1953, Brussow et al. 2004, Bobay et al. 2014, De Paepe et al. 2014, Martin-Galiano and Garcia 2021 ).Earlier microcosm work has demonstrated high efficiency of phage transduction in soils (Zeph et al. 1988 ).Even though the rate of phage transduction in natural soils is still less well-understood, se v er al examples exist on phage-mediated HGT and associated benefits for host bacteria in soils .For instance , pha ges in atr azine-contaminated soils wer e found to be able to acquire the trzN gene, encoding a c hlor ohydr olase r equir ed for atr azine catabolism (Ghosh et al. 2008 ).Additionally, Ross and Topp ( 2015 ) conducted transduction experiments using phages isolated from the soil to infect Escherichia coli K-12 and discov er ed that sublethal antibiotic concentrations could promote phage-mediated HGT of ARGs in a gricultur al soil microbiomes .T hese phage-encoded ARGs include multidrug resistance, pol ymyxin, and β-lactamase r esistance genes (Lekunberri et al. 2017, Moon et al. 2020 ), that can help bacteria resist antimicrobials originating from anthropogenic or environmental sources when produced by plants, bacteria, or fungi.Phages carrying ARGs have been shown to be infective in propagation experiments, indicating their role as vehicles of ARG transmission between bacteria (Larr ana ga et al. 2018 ).While ARGs can be mobilized by phages (Torres-Barceló 2018 ), it is still, ho w e v er, unclear how common this phenomenon is in soil microbiomes (Enault et al. 2017 ).Integration of temperate phages into bacterial genomes as pr opha ges could also c hange bacterial gene expr ession, leading to loss of functioning of certain genes (Chen et al. 2020, Hsu et al. 2020, Zhou et al. 2023 ).For example , Da vies et al . ( 2016 ) found that when transposable phage φ4 integrated randomly into the bacterial c hr omosome, it r esulted in insertional inactiv ation of type IV pilus and Quorum Sensing-associated genes, which was adaptive.Lysogenic phages could further become 'grounded' if mutations in one or mor e pha ge genes r esult in the failure of pr opha ge excision from the host genome, also referred to as 'cryptic' or defectiv e pr opha ges (Ramisetty and Sudhakari 2019 ).These defectiv e pr opha ges can further pr omote genome e volution thr ough propensity for genetic variations including in versions , deletions , and insertions via horizontal gene transfer (Monteiro et al. 2019 , Ramisetty andSudhakari 2019 ).Mor eov er, temper ate pha ges can enhance bacterial adaptability by increasing the mutation supply rate and thus help generating new genetic raw material for selection (Canchaya et al. 2004, Zhang et al. 2022 ).
In addition to facilitating horizontal gene tr ansfer, temper ate pha ges ar e also ca pable of altering host metabolism thr ough the expression of AMGs (Yu et al. 2015, Sun et al. 2023a ) (Fig. 1 D).AMGs originate from bacterial cells but are carried by phages to enhance their own and their host's fitness and can also contribute to the breadth of the phage host r ange (Shar on et al. 2011 ).In c y anophages , AMGs ha ve been associated with functions such as photosynthesis, nucleic acid synthesis, metabolism, and stress tolerance (Thompson et al. 2011, Kelly et al. 2013, Enav et al. 2014 ).Soil environments are also important reservoirs for viruses that encode AMGs (Sun et al. 2023a ).Phages can also provide their host bacteria beneficial tr aits thr ough l ysogenic conv ersion wher e they integrate into bacterial chromosome as prophages .T hese traits can, e.g.include virulence traits (Fortier and Sekulovic 2013, Matos et al. 2013, Taylor et al. 2019 ) or AMGs and different soil en vironments ma y enric h specific AMGs with div erse ecological functions .For example , phage AMGs ha ve been associated with the breakdown of harmful pollutants, including atrazine degradation gene trzN in atrazine-contaminated soil (Ghosh et al. 2008 ), the arsenic resistance gene arsC in lysogenic soil viruses (Tang et al. 2019 ), and the virus-encoded L-2-haloacid dehalogenase gene (L-DEX) in or ganoc hlorine-contaminated soil (Zheng et al. 2022 ).Mor eov er, the pr esence of pha ge AMGs in v ermicompost has been linked to both metabolism and pesticide biodegradation (Chao et al. 2023 ), while c hr omium-induced str ess can enric h AMGs that contribute to microbial heavy metal detoxification and survival in stressful soil environments (Huang et al. 2021 ).Beyond helping host bacteria to survive in contaminated en vironments , phage AMGs can also be involved in the carbon and nitrogen cycling in a gricultur al soils (Roux and Emerson 2022 ), and participate in carbon and sulfur transformation in agricultural slurry (Cook et al. 2021 ).Some AMGs have also been associated with energy acquisition via o xidati v e r espir ation, degr adation of organic matter, and plant-beneficial functions in the rhizosphere (Braga et al. 2021, Wu et al. 2021 ).Selection could, hence act on both host bacteria and pr opha ges, enric hing their frequencies if prophage improve host bacterial fitness r elativ e to pr opha ge-fr ee cells.
Also, some plant-beneficial AMGs have been detected in phage genomes that possibly contributed to plant-microbe interactions (Br a ga et al. 2021 ).For example, succinoglycan and acetolactate biosynthesis play important roles in nodule formation and plantgr owth pr omotion and hav e been found to be encoded by phages (Ryu et al. 2003, Mendis et al. 2016 Phage-encoded AMGs can also carry pathogenicity factors, such as effector proteins, that could help pathogenic bacteria to e v ade plant imm unity (Gr eenr od et al. 2022 ).For example, it was r ecentl y shown that hopAR1 effector protein is encoded by a prophage that can transmit these virulence factors between different P. syringae bacterial genotypes (Hulin et al. 2023 ).From the evolutionary perspective, selection could, thus act on both host bacteria and pr opha ges, enric hing their frequencies if prophages improve host bacterial fitness relativ e to pr opha ge-fr ee cells .A further in v estigation into the r oles of phage AMGs in soil environments is needed to better understand their associated functions and evolutionary advantages for bacterial hosts and surrounding microbiota and plants.

Phage-bacteria coevolution in the rhizosphere
Ra pid pha ge-bacteria coe v olution plays a piv otal r ole in sha ping the dynamics of microbial communities and ecosystem functioning in rhizosphere microbiomes (Koskella and Taylor 2018 , Fields and Friman 2022 ).To r esist pha ge infection and l ysis, soil bacteria have developed a plethora of resistance and defence mechanisms , while phages ha ve evolved numerous ways to overcome them, resulting in a long-term, coevolutionary arms race (Bernheim and Sorek 2020 ) (Fig. 1 E).To initiate the infections, phages first need to adsorb to bacterial surfaces to inject their genetic material inside the bacterial cells.To escape this, bacteria have e volv ed numer ous ways to pr e v ent pha ge adsor ption, including losing the phage receptor, reducing the expression of the receptor, modification of the r eceptor thr ough m utations, or pr oducing proteins that block phage adsorption by masking the receptor (Rostol and Marraffini 2019 ).Once the phage has adsorbed to the surface of the bacteria, bacteria can further alter the permeability of the surface, pr e v enting the injection of phage DNA (Labrie et al. 2010 ).If phages are successful in infecting their DNA or RNA within bacterial cells, bacteria can activate a second line of defences to recognize and degrade the phage nucleic acids.In summary, these defence strategies can be divided into two steps: first detecting the infection of incoming phage with sensors, and second, the initiation of interference on phage reproduction cycle (Georjon and Bernheim 2023 ).These mechanisms include, e.g.restriction-modification (R-M) (Vasu and Nagaraja 2013 ) and the CRISPR-Cas systems (Barrangou et al. 2007 ) that recognize and destr oy incoming pha ge nucleic acids.If pha ges can avoid the r ecognition by these systems and pr ogr am the host to start synthesizing phage proteins, other bacterial defence systems, such as ToxIN, DarTG, and CBASS (Georjon and Bernheim 2023 ), can detect the viral proteins and trigger abortive cell death, which will eliminate infecting pha ges, pr omoting the surviv al of surr ounding uninfected bacterial cells (Gao et al. 2022 ).For example, phage infections can acti vate to xin functions, leading to interference with bacterial DNA replication and protein translation, and prevention of pha ge r epr oduction.Alternativ el y, pha ge infection can trigger the production of cyclic oligonucleotide-based antiphage signaling system (CBASS), which activates effector protein that destroys the host cell and block phage propagation (Georjon and Bernheim 2023 ).Ho w e v er, this hier arc hical activ ation is onl y as one potential explanation for the presence of multiple different types of defence systems, and alternativ el y or additionall y, complementary and synergistic effect between defence systems, might enable bacteria to adopt flexible defence str ategies a gainst pha ges (Zhuang and Jin 2003 ).
Inter estingl y, antipha ge defence systems can also be carried by pr opha ges, whic h include r epr essor-mediated imm unity, exclusion-like systems, and r estriction mec hanisms, that can offer protection against a broad range of related and unrelated viruses (Kita et al. 2003, Dedrick et al. 2017, Patel and Maxw ell 2023 ).Ow en et al. ( 2021) identified BstA pr otein-a famil y of pr opha ge-encoded pha ge-defence pr oteins-in v arious Gr amnegativ e bacteria, whic h enable pr opha ges to defend host cells a gainst exogenous pha ge attac ks without sacrificing their ability to replicate lytically .Additionally , Patel et al. ( 2024 ) discovered that a pr opha ge can encode Tab pr otein, whic h mediates the antipha ge defence by blocking virion assembly of invading phages.Furthermor e, pr opha ges hav e been shown to serve as primary reservoirs and distributors of defence systems in E. coli , which can be located in specific genomic regions, i.e .defence system islands (Rousset et al. 2022, Vassallo et al. 2022 ).
Thr ough coe volution, pha ges hav e e volv ed m ultiple counter adaptations to avoid bacterial defence systems or infect bacterial cells with mutated or alternative receptors.For example, phage λ that primarily uses the LamB as its receptor, has been shown to r a pidl y coe volv e to bind mor e efficientl y to this r eceptor and e v en e volv e to recognize alternative E. coli receptor (OmpF) in response to E. coli resistance ev olution (Mey er et al. 2012 ).Such phage coe volutionary c hanges ar e often ac hie v ed via c hanges in pha ge tail fibers and other host-r ecognition pr oteins (Nobr ega et al. 2018, Altamirano and Barr 2021, Borin et al. 2023 ).When phage nucleic acids enter the host bacteria, phages can employ various antidefence system strategies, including antirestriction modification and anti-CRISPR proteins .For instance , in coliphage P1 (Myo viridae), the proteins DarA and DarB are coinjected into the host cell with the phage genome.Both proteins bind to phage DNA, masking type I R-M recognition sites, preventing the degradation of phage DNA (Atanasiu et al. 2002 ).Moreo ver, phages ha ve evolved anti-CRISPR proteins (Acrs) that inhibit the cleavage of Cas proteins .For example , AcrIIC4 is a broad-spectrum Acr that binds between the two recognition domains of Cas9, REC1 and REC2, r estricting the mov ement of the REC2 domain and thereby maintaining the integrity of its genome (Sun et al. 2023b ).Pr e vious studies hav e demonstr ated that pha ge-bacteria coe volution follows fluctuating selection in the soil micr ocosms, wher e pha ges and bacteria adapt to their contemporary counterparts in time (Gomez and Buckling 2011 ).While a laboratory study sho w ed that the rate of evolution increases with soil-inhabiting Pseudomonas fluorescens SBW25 in the presence of a phage (Pal et al. 2007 ), a follow-up microcosm study contradicted this finding, where the presence of virulent phages and a natural soil virome negativ el y affected the evolution of SBW25 (Gomez and Buckling 2013 ).While not specifically examining the coev olution betw een bacteria and phages, another study revealed that phages can rapidly select for resistant bacteria in the tomato rhizospher e, whic h leads to trade-off with bacterial growth and competitive ability (Wang et al. 2019 ).With the advancement of bioinformatic appr oac hes, numer ous nov el defence systems and antidefence systems ar e continuousl y discov er ed (Gao et al. 2020, Nussenzweig and Marraffini 2020, Millman et al. 2022 ), and it is now becoming important to try to understand their significance, synergies and r elativ e importance for pha ge-bacteria coe volution in natur al envir onments, including rhizospher e.For example, as r eceptor and defence system-based resistances are not mutually exclusive (Alseth et al. 2019, Wang et al. 2023 ), it is important to quantify their costs and benefits in ecologicall y r ealistic contexts to assess their importance for pha ge-bacteria coe volution in terrestrial ecosystems.Understanding the underlying variation and r ate of pha ge r esistance e volution in a gricultur al envir onments is especially important for the long-term success of phage biocontr ol a pplications.

The consequences of phage resistance evolution for bacterial fitness
Research thus far has demonstrated that bacteria can r a pidl y e volv e r esistance to pha ges in the soil (Gomez and Buc kling 2011 ) and the rhizosphere (Wang et al. 2019 ).Ho w ever, ev olving resistance to phages often comes with costs due to loss or reduced functioning of associated r eceptor genes, whic h could c hange how bacteria interact with other microbes (Fig. 1 F).For example, evolution of broad phage resistance of phytopathogenic R. solanacearum w as sho wn to lead to r elativ el y higher costs of r esistance in terms of reduced pathogen growth and competitiveness with nonresistant, ancestral R. solanacearum genotype (Wang et al. 2019 ).Costs of resistance could also affect how bacteria interact with plants or behave in their environment because receptor mutations often occur in surface proteins that may also use to attach on plant surfaces or move and navigate in the rhizosphere or soil matrix (Ad d y et al. 2012, Ahmad et al. 2014, Narulita et al. 2016 ).For example, pha ge r esistance m utations in genes encoding type IV pilus and type II secretion systems important for bacterial movement and secretion of exoenzymes , respectively, ha ve been linked to both pha ge r esistance and loss of virulence in R. solanacearum (Narulita et al. 2016, Xavier et al. 2022, Wang et al. 2023 ).Mor eov er, m utations in the quorum-sensing (QS) signalling receptor gene, phcS , have been shown to lead to phage resistance and loss of virulence in R. solanacearum e v en though the underlying molecular mechanisms remain unclear (Wang et al. 2023 ).Finally, also the upregulation of CBASS and type I restriction-modification phage defence systems in r esponse to pha ge exposur wer e found to correlate with reduced expression of motility and virulence-associated genes, including pilus biosynthesis and type II and III secretion systems, in R. solanacearum (Wang et al. 2023 ).Together these findings suggest that both phage resistance mutations and upregulation of phage defence systems could result in trade-offs with pathogen virulence and fitness.
The evolution of phage resistance could also make bacteria more susceptible to other stresses.For example, pha ge-r esistance m utations hav e been found to sensitize bacterial pathogens to antimicrobial compounds (Torres-Barcelo and Hochberg 2016 ), and in the clinical context, phage-antibiotic combinations have been found to have superior efficacy to mono-treatments because pha ge r esistant bacteria became mor e sensitiv e to antibiotics due to mutations in genes that increase antibiotic efficacy (Chan et al. 2016, Altamirano et al. 2022 ).While the evolution of generalist resistance to both antibiotics and phages is also possible (Moulton-Bro wn and F riman 2018 , Burmeister and Turner 2020 ), similar sensitization of R. solanacearum to antibiotics produced by Bacillus am yloliquef aciens soil bacterium has been r eported because of pha ge r esistance e volution (Wang et al. 2017 ).If suc h tr ade-offs ar e mor e common among soil bacteria, pha ges could also indir ectl y affect antibiotics-mediated competition between different microbes in soil microbiomes.Finally, the magnitude of the cost of pha ge r esistance is known to v ary depending on the envir onmental conditions, such as nutrient availability, spatial structure, and the strength of resource competition (Brockhurst et al. 2003, Lopez-Pascua et al. 2010, Gomez and Buckling 2011, Gomez et al. 2015, Alseth et al. 2019, Che v aller eau et al. 2022 ).As a result, the e volution of pha ge r esistance could be constrained by its associated costs in complex rhizosphere microbiomes compared to more benign lab en vironments .More realistic experiments using soil and plant systems ar e, hence r equir ed to better understand the fitness costs of phage-bacteria coevolution for both partners in plant rhizosphere microbiomes.

Challenges and avenues for future research
While r esearc h on pha ge genetics and molecular biology has been adv ancing in lea ps and bounds along with the discovery of myriad of new defence systems (Georjon and Bernheim 2023 ), researc h on pha ge ecology and e volution in terr estrial envir onments is trailing back.While tracking interactions between focal species pairs, such as plant pathogenic bacteria and their specific phages (Gomez andBuckling 2011 , Wang et al. 2019 ), has helped to understand pairwise coevolution in soils and the rhizosphere, this view is simplistic as most bacteria in the soils are likely to have their own specific phages, and hence potential to coevolve and interact with them.While metagenomic sequencing and separation of phage and bacteria fractions before sequencing will has helped to unr av el the true pha ge div ersity in terr estrial systems (Roux and Emerson 2022 ), it is still challenging to infer interactions based on sequence data (Wu et al. 2023 ).This will undoubtedly change when we discover more about the genetics and molecular biology underl ying pha ge-bacteria inter actions, the adv ancement of bioinformatics and computational techniques (Gaborieau et al. 2023 ) and the use of more realistic model ecosystems, such as rhizoboxes (Wei et al. 2019 ).The structure of soils also creates limits for understanding phage-bacteria interactions at different spatial scales .For example , it is difficult to determine phage-bacteria population and metapopulation bor ders.Ho w far can phages migr ate passiv el y or by hitc hhiking with their host bacteria?How long ar e pha ges able to persist in environments in the absence of their hosts?All these questions remain yet to be answered.Going forw ar d, it is also important to employ omics techniques to understand the role of temperate phages for the horizontal transfer of AMGs and how temper ate pha ges inter act with other mobile genetic elements such as plasmids , conjugative elements , and phage satellites (Rocha and Bikard 2022 ).For example, what is the relative contribution of phages to the accessory genome of bacteria in terrestrial ecosystems?What are the k e y roles of phages for their host metabolism and how is pha ge div ersity linked with microbial and terrestrial ecosystem diversity?
More work on how to potentially harness phage ecology and evolution for the benefit of phage applications and ecosystem functioning in terrestrial ecosystems is also needed.As phages are often specific to their target bacteria, they could potentially be used to precision-edit bacterial communities by removing specific bacterial taxa or functions (Wang et al. 2019 ).For example, in addition to targeting plant pathogenic bacteria, one could use pha ges to tar get bacteria that inter act with the pathogen in the rhizospher e micr obiomes (Li et al. 2021b, Yang et al. 2023 ) or target bacteria that carry ARGs that are located in conjugative plasmids, r equiring pilus expr ession for their tr ansmission, whic h makes them also susceptible to phage infections (Jalasvuori et al. 2011 ).Alternativ el y, it might be possible just to target specific key taxa known for ARG carriage or just reduce overall bacterial abundances using nonspecific phage communities as has been done with the treatment of se wa ge systems (Yu et al. 2017 ).In contrast to removing bacterial taxa or functions, temperate phages could also be used to intr oduce ne w beneficial functions in the soil microbiomes .For example , if plant gr owth-pr omoting AMGs ar e common among temperate phages, such phages might be used to deliver plant-beneficial functions into rhizosphere microbiomes to potentially improve plant health and crop yields.Such beneficial functions could be identified, e.g. from so-called suppressive soils that can constrain plant pathogen infections and promote plant growth (Garbeva et al. 2004, Peralta et al. 2018 ).Instead of identifying specific phage species, employing phage communities as part of rhizosphere soil transplants could be used as an initial screen to identify beneficial microbial communities.For example, soil transplant from healthy tomato plants was shown to constrain R. solanacearum pathogen invasion in the next tomato generation and this effect was likely driven by both pathogensuppressing bacteria and R. solanacearum -specific phages present in the transplanted soil (Wei et al. 2019, Yang et al. 2023 ).Finally, ecological theory and experiments suggest that biodiversity corr elates positiv el y with ecosystem functioning and this pattern has been also shown to hold in terrestrial ecosystems (Pennekamp et al. 2018, Jochum et al. 2020 ).Pha ge div ersity could, hence be important by promoting bacterial community stability and providing a mor e div erse suite of AMGs and accessory genome functions for the bacterial and plant community.To address all these questions, mor e r esearc h on pha ge ecology and evolution in terrestrial ecosystems is required where phages are recognized as a vital component of soil microbiomes with clear links with human and plant compartments within the One Health fr ame work.

Figure 1 .
Figure 1.Summary of ecological and evolutionary roles of rhizosphere phages in soil health.The ecological impacts of rhizosphere phages on soil ecosystems include phage effects on nutrient cycling via lysis of bacterial cells (A), top-down density regulation of bacteria (B) and positive phage effects on microbial community diversity , stability , and composition via Kill-the-Winner dynamics (C).The evolutionary impacts of rhizosphere phages on soil ecosystems include horizontal gene transfer, including phage-encoded AMG that increase host bacterial fitness (D), phage-bacteria coevolution, and (E) selection for pha ge-r esistant m utants with alter ed ecology due to genetic corr elations suc h as tr ade-offs with pathogen virulence in planta (F).