Community complexity does not weaken pairwise coevolution in a soil bacterial community

Communi t y comp lex i ty does no t weaken p airwise coevol ution in a soil ba c ter ial 1 commun ity 2 Author Affilia tions 3 Zoltan Er dos 1 (z.erdos@ex et e r.ac .uk ) , Daniel P adfie ld 1 (d .pa dfield @exeter . ac.uk) Elze Hes s e 1 4 (e.hesse @ ex eter.a c .u k ), Angus B uck li ng 1 ([email protected] .uk) & Me ag h a n Cas t le di ne 1 5 (m.castledine@exete r. ac.uk) 6 1 Centre for Ecology and C o nservati on, Faculty of Env ir onment, Scienc e and Econom y , 7 University o f Exeter, Penryn, C or nw all, TR10 9F E , U .K. 8 Keywords: C oev olut ion, d ynamics, e v olutio n, so il bact e r ia , compet i t ion , ant agonism 9 Correspo nding aut hor : Zoltan Erdo s ( z.erdos@ex eter. ac. uk) 10 Author con tribu tions (MeR IT): 11 Zoltan Erdos : Inve stigati on; Co nceptualiz ation; writin g - ori gi n a l draft ; for mal analy sis, w r i t in g - 12 review and editing; methodolog y; Vis ualiz ation. D aniel Padfield : formal a n a lysis; revie w and 13 editing. Elze H es se: Co nceptualiz at i o n; w rit i ng - r ev iew and editing ; Superv ision. An g us 14 Buckling : C onceptualization; writin g - review and editing; Supe rvision; Res ources. Meagha n 15 Ca stledine: Investig at i o n; Conceptu aliza t ion; w rit i n g - or i g inal draft; wri ting - review and 16 editing; met hodolo gy . 17 Data availability stateme n t 18 All data and cod e u s ed in the analysi s are availabl e on Git H ub 19 (http s :/ /gi t hub.com/zltnerdos/ antagon istic_ coe voluti on). 20 21 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint 22 Abst ract 23 Exploitativ e i nter ac t ions , such as pred ator -prey and host-parasite interactions, are ubiquit ous 24 in microbial comm unities . These int er a ct i on s shape c om munity density and composition, 25 imposin g str on g selection o n members t o evolve coun termeasures r edu ci n g the negat iv e 26 impacts o f exploitation. Ex ploitat i ve coev oluti on is oft en s tudied betw ee n species pairs in 27 isolati on, w hic h may o ver-estimate the strengt h and relev ance of pairw ise co ev olut io n. Here w e 28 studied how c om mun ity contex t i nfluences coevolut ion between P s eudom o na s fl uorescens 29 (exploited) and Variovorax s p. (ex ploiter ). We evolv ed t he se species in pairw ise coculture and 30 embedd ed within a five-species community to i nvest igate ev olved ch anges in pairwis e 31 interactions. We found ev idence for asy m metrical c oevolutio n: Variovorax ev olved more ra pidly 32 than Pseudomonas , leading to incr eased ex ploitation throug h time, while Ps eudo monas 33 evolv ed increased tolerance to Variov orax with time lag. T h e pairwise c o e volutionary dynamics 34 were not aff e ct ed by the pr e sence of ot her comm unit y members. Understanding how 35 coevoluti onary patter ns change w ith i n c r ea sing communit y complexit y c an hav e important 36 implication s f or c om mun ity persiste nce and fu nc t io n. 37 Int r oduc ti o n 38 Exploitativ e inter a ct i o n s - where one s pecie s benefits at the expense of anot her – c an r e sult in 39 the adaptive evolutio n of defence , c ounter -defence or r ec ip r ocal evolution of these tr ai t s 40 (antagoni stic coevolu tio n) (B rockhurst & Koskella, 2013; G andon et a l., 2 008). Antagon ist ic 41 coevoluti on can hav e far -re aching co ns equences for e cological and evolutionary dyna m i cs of 42 communities; particularly s o in micr obial communit ies where organism s often have la r g e 43 populations and short generatio n tim es (Bar ra cloug h, 2015; G ar butt et a l., 2011; Paterson et 44 al., 2010), meaning ecol ogi cal and evoluti onary processes often happen simultaneousl y 45 (Loreau et al., 2023). Previous work int o an tagon istic coevo l u tio n has pri marily focussed on 46 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint interactions between tr ophic lev els, h osts and paras it e s (Friman & Buc k lin g, 201 3; Gómez & 47 Buckling, 2 011) and predator and prey (Friman et a l., 2011; J ohnke e t al., 2017). H owe v er , it is 48 also important w ith in t roph ic lev els where interactions often occur over exploitation of 49 extracellula r c o mpounds (e.g. evolutio n of r es istance to antibio tics via comp etitio n (K och et al., 50 2014), ev olut ion of i ncreased com petitiv eness and exploitati on in b i of i lm s ( Hansen et al., 200 7; 51 Kim et al., 201 4)). 52 Communi ty co ntex t may have a s ign i fi cant eff e ct on pairwi se coevoluti on (B arracloug h, 20 15; 53 Blazanin & Turner, 2021; Manriquez et al., 2021). B e ing embedded within a community is l ikely 54 to r e d uc e the occu rrence and spe ed of p airw ise coevolution. Inter a ct i ng wi th mult iple 55 community mem be r s w ill po tentiall y reduce the f r equency of interaction f or a g iven species 56 pair, result in trade-offs b etween ad a ptation to mult iple species (A l seth et al., 2019; Friman & 57 Buckling, 2 013) ) and increas e the magnitude of trade-of fs betwe e n a biotic and b i ot ic 58 adapt ation (Br is coe Runqu i st et al., 2020; Gómez & B uck ling, 20 13; H all et al., 2018; Law r e nce 59 et al., 2012 ; Luján et al . , 2022; Yin et a l ., 2023), all of whic h w i ll reduce the s t rength o f r e cipr oca l 60 selec t ion. Furt hermore, reduced pop ulat i on siz es w it h increas ing commu nity member s wi ll 61 reduce t he supply of mu t ation on wh ich s electi on a ct s ( C astledine et al., 2020; H art et al. , 62 2019 ). Our understanding of h ow c o mmunity complexit y affects coevoluti onary dynamics is 63 howev er primar i ly lim ited t o s t udie s of bacteria-virus ( bac t eriophage ) sys tems. 64 Here, we em ploy a t i m e s hift (Buckl ing & Rai ney, 2 002; G aba & Ebert, 2009) appr oac h to 65 quantify t he impact of commun it y c omplex ity o n t he coev olu tio nary dy n amics o f two f oca l 66 specie s th a t en g age in an ex ploit ati v e interaction. V a r iov orax sp. (AB1) b e nefit s fr om unkn ow n 67 metabolites produced by Pseudomon as fluorescens (AB1) (h e r ea f ter re fer r e d to as Variovorax 68 and Pseudomonas ) , while the growth of Pseudomonas i s red uced by t he p res ence of Va riovorax 69 sugg est in g t he potent i al f or stro ng sel e ct i on for coe volut ion (C as tledine et al., 2024). This so i l 70 microbial community cons ists of five s pecies (including Pseudomo nas and Variovorax ) a nd is 71 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint dominated by competitive interactions (Cast ledine et al. , 2024). B riefly , unde r s imilar 72 expe r imental condition s to our current work, Pseudomonas is neg ativel y im pact e d by all other 73 specie s. While V a r iov orax r educes th e fitness of a ll other sp ec ies, it exp eriences a growth 74 benefit in t he pr es ence of three of the f our other communit y members (w it h hi ghest ex plo itatio n 75 agains t Ps eudo mo nas ). Therefor e , we expec t that communit y con text w ould r e d uc e t he extent 76 of pairwis e coevolution between Vario vorax and P s eudomo na s by dec reasing mu tation supply 77 and the strength of reciprocal sele ct i on. We hypothesiz e that s t ron g sel ect ion w ould r e sult i n 78 Pseudomonas an d Variovorax evolvi ng via arms r ace d y namics (e.g . Pseudomonas ge t t i n g 79 better a t defence ag ainst exploi tati on and Variovorax evolv i ng increas e d ex ploitation ) in 80 coculture, and th a t coev olutio n of t h e f oca l speci es w ould be weakened in a multi species 81 community. 82 Mater ials an d m ethods 83 Experim ental e volution 84 Experimental evolution treat ments w er e s et up to s t udy ad apt ation and pai rwise coev olut i o n 85 between Pseudomonas and V ariovora x in different biotic condit ions (F ig ur e 1). A monoculture 86 evolutio n treatment w as set up to isolate any ef fects of co rrelate d abiotic adaptation, whi ch 87 can alter competitive hierar c hies, mediate exploitativ e dynamic s, and influence nic he 88 partitioni ng wit hin the c ommunit y. A coculture and community e volu tio n treatm e n t was us ed to 89 study coevolu tio n in a pairw ise and community background r es pectivel y, and to ass ess th e 90 impact of communi ty com ple xity o n pairwise coev oluti on between t w o species. 91 The community evoluti on treatment had been ca rried out previously and results h a ve b e en 92 published in Castledine et al. ( 2020). W e r a n domly selected eig h t replica t e com munities in our 93 current study. Mono- and coculture e voluti on t reatments were init iated wit h eig ht replica t es 94 (eight replicates for each species per monoculture) us ing the s ame P s eudomo na s and 95 Variovorax ancestor s and culturing proto c ol. B riefly , ba cterial isolates ob taine d fr om s o i l 96 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint samples, including A c h romobact er sp. , Ochrob actr um sp . , Pseudomonas sp ., 97 Stenotro phomonas sp. , and Va riovorax sp. , w e r e identified based on their distinc t colo ny 98 morphologies o n Kin g' s medium B (K B) aga r. Ea ch s pecies wa s c ultivated fr om a single colony 99 i n i s o l a t i o n f o r t w o d a y s i n 6 m l 1 / 6 4 T r y p t i c S o y B r o t h ( T S B ) m e d i u m a t 2 8 ° C i n g l a s s 100 microcosms. Inoculated abund a nce s ( c o l o ny f orming u nit, CFU) of e ach species w ere 101 estimated app roximately from optical densities (O D 600 ; w aveleng t h 600 nm) after t w o day s of 102 growth (equations for conv er ting OD 600 to C FU /μL des cr i b e d prev iously (C as tl edine et a l., 2024) 103 and adjusted to: 2x 10 6 CFU/µl . R e plica t e lines o f commun ities (all s peci es combined), 104 cocultures ( Pseudomonas and Variov or ax ) and monocultur es ( Pseudomonas , Variovorax ) w e re 105 establishe d using a 20 μL inoculum from e ach spec ies int o fresh 6 mL 1 / 64 TSB . C ultur es 106 underwent weekly serial 10 0- fold dilu tion s (transfer of 1% i nocula in to fres h media) over ten 107 wee ks, w it h samples frozen every s eco nd transfer (−70 ° C in gl ycerol, fina l concentrat ion: 2 5%). 108 Frozen samples from the ancestor (us ed a s inoculum above ) , 6- and 10-wee k-old cultures w ere 109 plated onto KB agar and i ncubat e d f or 2 days at 2 8 ° C . Six clones of Pseudomonas and 110 Variovorax each pe r commun i t y , coculture and monocult ure evolutio n lin e (replicate) w er e 111 isolated from each t imepoint and gro wn for 48 h our s in 1/64 TSB before be ing combi ned and 112 frozen at −70 °C in gl ycerol t o be u sed i n t ime-sh ift as says. 113 Time -shift a ssays 114 To test w he t her the two focal spec ies, Pseudomonas and Variovorax , hav e c oevolv ed o ver the 115 10-week period we cond uc t ed t ime- shift assays , where o ne species was c u l t ured w ith the ot her 116 from a population of th e pas t , contemporary or futu re t i m e p oi n ts. T hese a ssay s were car ried 117 out in t he abs ence of the rest of the c o m munity. This approach allows for s ig natur e s of diffe rent 118 types of coevolutio n to be det e ct e d. Arms rac e dynamics (AR D , i.e. selection of defence and 119 counter defence) is directional, with focal species h a ving a greater ab undanc e when com peting 120 with t he ot her species isolated from a past time point, a n d low er densitie s w hen com petin g 121 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint agains t com peti tors from a future t i m e point . I f i nstead selectio n o n specific genotypes 122 fluctuates throu g h t i m e ( fluctuatin g s e lection dy namic, FS D), p erfo r mance may be pa r ticula r ly 123 high o r low for contemporaneous inter actions ( H all et al., 2011). 124 Populations of P s eudomo nas and Va ri ov o r a x from one of the thr e e diff e r ent timepoint s of t he 125 evolutio n e xperiment (ancest or, 6 week and 10 w eek) we r e r ea ssembled, res ulting in 9 pairwis e 126 combination s per tr e at ment (Figure 1B). Cultures of Pseudomo nas a n d Variovorax w e r e 127 ass em ble d us ing all 6 clones fr om th e same t r e at ment r eplicate. Eight r epli cate mic r ocos m s 128 were set up for e ach com bi n a t ion wit hin t he t hree different ev olution ba ckgrounds except for 129 monocultur e s, see below. T he culture conditions of all treatments were established as 130 described for the initial ex periment a l e voluti on experiment, w it h a pp rox imat e l y 2x10 6 C FUs per 131 specie s in oculated into fresh microco sms. After one week , cultur e samples we r e cr y ogenically 132 frozen and then pla t ed onto KB aga r . Population densities were estimate d by countin g the 133 number of CFUs (10 5 diluted) aft er two day s of growt h at 28 °C. We use proportion of 134 Pseudomonas to in terpret coevolutionary dynamics . Pr oportion of spec ies is often mor e 135 insi g ht ful t han d e n sit ies in describi ng coev olut ion as it hi g hl igh ts sh i f ts in rel ative a b undance of 136 each spec ies. These s hift s ar e i ndi cative of selecti v e pr e ssure imposed by interspecies 137 interaction (e.g. parasi t is m ) and is le s s affected by f luc t u ations in tot al popu lat i on s ize t ha t c an 138 be influenced by v ar i o us factor s such a s stoc hastic variation i n density betw ee n microcosms. 139 Three r eplica t es in the monoculture ev olutio n line (f or both Pseudomonas and Variovorax ) w ere 140 contaminated prior t o w eek 6, therefo r e monocult ure g r ow th assays were ca rried out with only 141 5 replicates. Three repli cates of the commu ni t y evoluti on line w ere c on taminated during t he 142 time-sh ift ass ay, and t hese w er e removed from the analys is . The t o tal density of pairwis e 143 combination s w er e assess ed usin g t h e total abundance (combined C FU ) of Pse u domonas and 144 Variovorax in each microcosm. 145 146 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint Statisti cal analyses 147 All data analys es we r e c arried out in R v 4.3.1 (R Core Te am, 2025). First, w e analyse d whether 148 the dens i ty o f each species varied i n t hei r i nteraction w ith con temporary c ou nterparts o v er 149 time. For this m odel, th e time-s hift a ssay s were subset t o retain onl y th e bacterial densities 150 from their cont emporary assay s. The d e nsit y of each species w as analys ed in l inear mix ed-151 effects models (LMM) s eparat el y, w it h lo g 10 transformed density as the r e sponse v ar i able, 152 treatm e nt , time and their interact i on s as ex planat ory variable and random int ercepts fitted for 153 each replic at e t o account for n on- independence of observa t ion s . 154 One r eplic at e of t he t ime- s h ift assays i n t he coculture treatm ent ( Pseu domonas t ime 6, 155 Variovorax t i m e 10) showe d an unu sually hig h Variov or ax count and w as found to be an 156 influential outl ier (G rub bs - test) and was removed from the analys is. To test how com munit y 157 complexi t y affects coev olutionar y dy namic s we used a binomial generaliz ed linear m i xed-158 effects models (GLMMs ) with a log it lin k function, the proportion of Pseudomonas as combined 159 binary response variabl e and tre atm e nt (complexity) x evoluti onary time of P s eudomo na s x 160 evolutio nary time of Variovorax fitted as fixed effects, as well as t he ir 3-wa y i n t eraction. We 161 included random inter c ept s f or each replicate line t o account for non -independence of 162 observations. To study coevolut io n w i th i n treatments, we tested the effect o f coev olut io nary 163 time of bo t h Pseudomo nas and Va riovorax , plus t heir in teraction, on the p roportion of 164 Pseudomonas in separate models p er tr e at ment us ing binom ial G LMMs w ith a log it l ink 165 function for ea ch e voluti onary backg r ound. Total densit y of pa ir w ise combination s w er e 166 analy sed w ith a line ar mixe d- effects model (LMM) wit h l og 10 tr a n s for med total dens it y as the 167 response variable, treatment, evolutionary t i m e of P seudomonas a n d Va riovorax and their 168 interactions as explanatory v ariables and random intercepts fitted for ea ch r e plic at e to account 169 for non -i ndependence of observa t ion s. T ime was included as dis cr e t e v ariable in all analy sis . 170 Thes e analyses employed LMMs and GLMMs usi ng t he ’ lme4’ pack a ge (Bat e s et al., 2015). For 171 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint these LMMs and G LMMs, we used t he ‘ DH A R M a ’ p ackage (Hartig, 2018) to c heck resi du a l 172 behavi o ur and mod el s imp l ification was c arried out using likeliho od rat io test. Post- h oc 173 multiple comparison test s o n the m ost parsimoni ous models we r e carri ed out us ing the R 174 packag e ‘ emm eans’ (Lenth, 2 0 23) , usi ng T ukey adjustm e nt . 175 Resul t s 176 Popu latio n den si ties chan ge acr o ss contempo ra ry species pa i r s evolved with in 177 commun ity 178 We first determined how t he r el at iv e succe ss of Variovorax and Pseudomonas differ ed betw een 179 treatm e nt s and through t ime. Specifi cally , w e dete r mined the density of each species w hen 180 cultured with their c o ntemporary co unterparts. The effec t of treatment ( monocultur e , co-181 culture, 5 species communit y ) on P s eudom onas densi ty did n ot differ thr oug h t ime (LM M: 182 community complexit y x evolut ionary time: χ2 2 = 3.06, p = 0.22, Figure 2), n or was there an 183 overall ef fect of time (LMM: evolution ary time: χ2 1 = 3.14, p = 0.077). Tre at ment did howev er 184 have a significant effect on Pseudomo nas de n s ity (LMM: community complex ity: χ2 3 = 21.20, p 185 < 0.00 1, Figure 2). Pseudomo nas po pulations t hat had ev ol ved in c omm uni t y (4.52 log 1 0 186 CF U /ml [4.43, 4.60]; Tukey H SD : es t imate = 0.21, t-ratio = 3.1, p = 0.01 8) r e ached significantl y 187 lower dens ities compared to the an cestor and tho se ev ol ved as cocult ures ( Tukey HSD: 188 estimate = 0.14, t-r atio = 2.77, p = 0.051). Mean Pseudomonas population densi ty in 189 contemporar y combinat i on s of mo no culture evolv ed P s eudomo na s and Va ri ov o r a x (4.79 lo g₁₀ 190 CF U /ml; 95% C I: [4.69, 4.88]) did not d iffer s ignificantly from that of the ancesto r P s eudom onas 191 and ancestor Variovorax combination (4.73 log₁₀ CFU/ml; 95% CI: [4.62, 4.84]; T ukey HSD: 192 estimate = - 0.0 57, t-ratio = -0.8 0, p = 0.85). Si m il ar ly , m ea n Pseudo monas densi t y in 193 contemporar y com bi n a t ions of cocult ure evolved lineag es (4.65 log₁₀ C FU / ml; 95% C I: [4.57, 194 4.73]; Tukey HSD : estimat e = 0.078, t -ratio = 1.19, p = 0.64) wer e not diffe r e nt to the ancestor 195 combination. Evol v in g in a communi ty, and cocult ure to a lesser ex tent, results in low er 196 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint Pseudomonas dens it i es when cultur ed with Variovorax , suggesting a change in their interaction 197 compare d to t he ancest ors o r m onocu lture ev olved l ine ages. 198 Similarly to Pseudomonas , the dens ity of Variovorax w as unaf fected by an i nt eraction betw een 199 treatm e nt a nd evolutionary t ime (LMM: commun i t y complexit y x evolut ionar y time: χ2 2 = 4. 2 9, p 200 = 0.12, Figure 2) or ev olu t i o nar y t i m e itse lf (LM M: e volutio nary time: χ 2 1 = 1 . 43 , p = 0 . 2 3 ) , b u t 201 treatm e nt had a s ignificant eff e ct on Var i ovorax density (LMM: communit y c o mplex ity: χ2 3 = 202 21.14, p < 0.001, Fi gure 2). Me an c ont emporary pop ula t ion densities o f Var i ovorax in t h e 203 ancestor c om bina t ion ( 4.67 lo g ₁ ₀ CFU/m l ; 95% C I: [4.55, 4.80]) w er e not differ ent to t ho s e 204 evolv ed in monoculture (4.70 log₁₀ CFU/ml ; 95% CI: [4.60, 4.80]; Tukey HS D: e stimat e = -0.024, 205 t-r atio = -0.30, p = 0.99) or coculture (4.83 l o g ₁ ₀ CFU/m l ; 95% CI: [4.74, 4.92]; T ukey HSD : 206 estimate = -0.15, t- ratio = -2.00, p = 0.21). How ever, mean Var i ovorax densit ies in th e 207 community evolut io n t reatment reached si gnifican tly higher densities than t he anc estor in t heir 208 respective contempo rary cocultur es (4.95 lo g₁₀ C FU /ml ; 95% CI: [4.8 6, 5.05]; Tukey HSD: 209 estimate = - 0.28, t-ratio = -3.49, p = 0.007). These r es ult s confirm that community co ntext has a 210 sig n ificant impact on t he adaptation of spe cies, and th a t coev oluti on l e ad to a change in 211 interaction between species ( w hen co mpare d to ancestor or mo noculture evo l ved li nes ). 212 Antagonistic coevolution evident between species pairs 213 Time shift assay s were used to c har ac t eris e c o e voluti on i n c o c ult ure and in a com munity by 214 growing Pseudomonas w i t h Variovor ax isolated from different time-poin t s and v ice-ve r s a 215 (Figure 1). To co ntr ol for abiot ic adaptation, pairs of mon oculture lines ( Pseudomonas and 216 Variovorax cultured al o ne ) ev o l ved alongs ide the coc u l t ure and com munit y treatment s w ere 217 also subjected t o time shift a ssays . 218 The results of the t ime sh ift a ssay s ar e indi cat i ve of ant a g oni stic c oevoluti on betwe en 219 Pseudomonas and Variovorax in bot h c ocult ure (G LMM: Pseudomonas time x Variov or ax time : 220 χ2 3 = 7.15, p = 0.067, F igure 3B) and in t he comm uni ty ( G LMM: P s eudom onas tim e x Variovorax 221 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint time: χ2 3 = 21.85, p < 0.0 01, Fi gur e 3C). Variovorax s howed a tendency to i ncrease in 222 competitivene ss aga ins t Pseudomon as t hrough time i n bot h th e cocultur e and comm uni t y 223 treatm e nt s . Compared t o the ancest r al Variovorax, popul ations evolved f or either 6- o r 1 0-224 wee ks in both coculture (Figure 3B) and community (Fig ur e 3C) treatment s reduced the 225 propor tion of ancestral Pseudomonas . There w a s no s ign ificant difference b etween the effect 226 of Variovorax coevolved for 6- or 10-week s on the proport ion of ancestral Pseudomonas for the 227 coculture treatment (T ukey HSD coculture : es t imate = - 0.027, z - ratio = - 0.1 5, p = 0.88), h owev er 228 ancestral P s eudom onas propor tion was significant ly hi gher in t he community-e volve d 229 treatm e nt , when culture d w it h 10-w eek Va ri ov o r a x comp are d to 6-w eek Variovorax (Tukey 230 HSD community : es t imate = -0.50, z - ratio = -2.69, p = 0.007, Figure 3BC ) . T his latt er result could be 231 attribut ed to eff ects ass ociated with o ther s pecies in the communi ty evolut i on treatment ( e . g . 232 adapt ation to other species ). 233 Pseudomonas underwent recip rocal a daptation to Variovorax . The propor tion of P s eudom onas 234 cultured with 6-week Variovorax i n c r e a s e d w i t h Pseudomonas e volut ion ary time for both 235 coculture (Tukey HS D ancestor – 6 week : e s t i m a t e = - 0 . 6 6 , z - r a t i o = - 3 . 45 , p = 0 . 0 01 6 ; T u k e y H S D ancestor – 236 10 week : estimate = -0.85, z-ratio = -4.37, p < 0.001 , F igure 4B) and commun ity ( T ukey H SD ancestor – 10 237 week : estimate = -0.79, z-rat i o = -3.07, p = 0.0061 , Figure 4C) treatments. W hil e P s eudom onas 238 show ed an ove r all increas e i n r esis t a nce through time against 6 -week Vari ovorax , r e s i st an ce 239 declined ag ainst ancestral Variovorax for both tr ea t ments (Fig ur e 4B C ). Resi stance to 1 0-week 240 Variovorax was not diffe rent betwee n any of the Pseudomonas evoluti on ary t i m e p oi n ts for 241 either th e coc u l t ure or comm unit y t reatment. These dynamic s ar e cons i ste nt wit h fluct ua t in g 242 selec t ion ac t i ng on P s eudomo nas , such that i t became specificall y ad apted to evolvin g 243 Variovorax , while becoming maladapte d to ances tral Variov o rax . 244 To contr ol for adaptation t o abiot ic conditi ons potent ially bein g i nt e r preted as coev oluti on, 245 monocultur e lines w ere e volved alon gsi d e the coculture and community t reat ments. We did 246 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint not det e ct i n teractio ns that were s ug gesti ve of c oev olut ion betw een P s eudomo na s and 247 Variovorax e v o l v e d i n m o n o c u l t u r e ( G L M M: Pseudomon a s time x Var i ovorax ti me: χ 2 3 = 4.02, p = 248 0.26). There was no indica t i on of Variovorax adapting to abiotic condit ion s or th a t any a biot ic 249 adapt ation affect e d its i n teraction wi t h Pseudomonas ( G LMM: Variovorax t i m e χ 2 2 = 0 . 8 8 , p = 250 0.64, Fi gure 3A). We found some ev idence for abiotic adaptation in P s e udomo na s, whi ch 251 affe cted interactions with Variovorax ev olved in mo nocult ure, w ith an increase in t he proportion 252 of Pseudomonas a t w e e k 6 ( G L M M : P seudomonas t i m e χ 2 2 = 10.06, p = 0.0 07, Figure 4A), but 253 not week 1 0. 254 Community complexity did not affect coevolution of species 255 Con trary to our ex pectation t hat e v olut io n i n a multispeci es commu n i t y would weak en 256 coevoluti on, we found n o differ ence i n coev olutio nary d y namics between t h e coculture and 257 community tr e at ments ( G LMM: Pseudomonas time x Variovorax time x com plex ity interactio n 258 χ 2 3 = 3 . 54, p = 0.32). Pseudomonas p r oportion was not differentially a ffect ed by commu ni t y 259 complexi t y o v er ev olut i o nar y t ime for each s pecies (G LM M: Pseudomonas time x complexit y 260 interaction and Variovorax time x com plex ity i nteraction: χ 2 2 pseudomonas = 4. 0 1 , p pseudomonas = 0 . 1 4; 261 χ 2 2 variovorax = 4.21, p variovorax = 0.1 2). Howev er , selection pressures are clearly di f f erent betwe en the 262 two t reatments. Pseudomonas evolv ing in comm unit y context di splayed lower p roport ions 263 compare d to cocultur es (Tukey HSD: estimate = 0.33, z -ratio = 5.07, p < 0.00 1), suggest i n g that 264 Pseudomonas adaptation to V a r iov or ax is w eakened in a community conte xt. T h i s is driven by 265 both an increase in Variovorax density (Tukey HSD: estimate = - 0.08, t-ratio = -3.1 0, p = 0.003) 266 and a de cre ase in Pseudomonas dens ity ( T ukey H SD : es t i m a t e = 0.06, t- ratio = 2.12, p = 0.037, 267 Figure 5C). 268 Changes in total density driven by Variovorax evolution 269 Total dens it y (of Pseudomo nas and Va riov or ax cocultures) could change o ve r ev o l u tio nary time 270 due to a change in s pecies int eraction s (e.g. increased exploitation by V a r iovorax could res u l t in 271 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint a decli n e in total density). We observ ed an increase in tot al density as a resu lt of an increase in 272 exploi t ation of evol ved Variovorax , an d this effect w as indep endent of evolutionary time and 273 backg r ound (community complex ity) (LMM: Pse u domonas time x Variovora x time x comm unity 274 complexi t y i nteraction: χ 2 6 = 5 . 1 1 , p = 0 . 5 3 , F i g u r e 5 ) . T o t a l d e n s i t y w a s n o t d i f f e r e n t i a l l y 275 affe cted by community complex it y ove r ev oluti onary time for each species (LMM: 2-way 276 interaction for Pse u domonas time x complexity and Variovorax time x comple xity: χ 2 4 pseudomonas = 277 6.58, p pseudomonas = 0 . 1 6 ; χ 2 4 variovorax = 2 . 7 2 , p vario vora x = 0.61). Total density w as als o no t affected by 278 an inter a ct i on betwee n the t w o s peci es evolutio nary time ( L MM: P s eudomonas x Variovorax 279 time interaction: χ2 3 = 5.90, p = 0.12). Only the ev olutio nary time of Variov o rax had a sig n i ficant 280 effect on the d ensity of th e assembled co-cultures (LMM e ff e ct o f V a riov or a x time: χ2 2 = 7 . 5 3 , p 281 = 0.024). This i s driven by an increase in Variov o rax densit y at w eek 6 (Tukey HSD: estimate = -282 0.054, t-ratio = -2.57, p = 0.0 47) leading t o an i ncrease in t ot al densi ty com pa r ed t o t he ancest or 283 (Figure 5BC). T her e f ore , incre ase d e xploitation of P s eudom onas by Variovorax in cr e as ed 284 community densit y despite dec r eas es in P s eudom onas den sit y. 285 286 Dis cu s sion 287 In t hi s s t udy , w e soug ht to understand the effect of community complexity on e xploitat ion-288 mediated pairw ise coev olution in a soil m i cr obia l communi ty. In o ur model sy stem 289 Pseudomonas (and s om e ot her members of the community) are ex ploi ted by Variovorax 290 mediated by intera ctions over metabolites (Cast ledi n e e t al., 2024). While w e found evi d e nce 291 for ant a gonis tic coe voluti on, commun ity c omplex it y d id no t sign ificantly aff ect coev o l u t ionary 292 dynamics . This i s despite the comm un i t y co ntext pr oviding m ore species for Variovorax t o 293 exploi t (ther e f ore reducing selectio n on Pseudomo nas specifically ) ; and P s eudom onas 294 expe r ie n c ing competit io n fr om more spec ies, therefor e inh ibiti ng population si z es and/or 295 offering conflictin g selec t i on press u res . T his sugge sts t ha t reciprocal s elec t ion betwe en 296 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint Pseudomonas and Variovorax is suffic iently stro ng to buff e r the ef fect of th e alt e r e d se lection 297 pressures cr ea t ed by the communi ty cont ex t. 298 While there are ex amples of wi t hin-tro phic level s t udies of b a ct e r ia l int eractions showi ng h ow 299 specie s inter a ct ions can lead to di f fer ent evolutionary outcomes i n both coculture and 300 multispecies com munit ies ( Chang et al., 2020; C hen & Zhang, 202 4; Pearl Miz r ahi et a l., 2023; 301 Piccardi et al. , 2024), to our know led ge this is t he first study ex plicitly demonstr a t in g w ith in-302 trophic level antagoni stic c oevolut io n betw een bacteria. The mech anism s under l ying any 303 observed ada ptations in thi s s tudy are unknown. How e ver, fr om previous w or k w e know that 304 Variovorax benefits from the pr es enc e a nd metabolic activity of P s eudom onas i n our mod e l 305 (Cas t le dine et al., 2024), t herefore me chanisms of Pseudomonas r e s i s t a n c e t o Variov or ax a r e 306 likely related to alteration of metabolites, while Variovorax becomes mor e efficient at using 307 metabolites produced by Pseudomonas or a dapts to utilise altered meta bolites. Analysi n g the 308 exa ct na t ure of this interaction (and evoluti onary m e chanisms thereof), is m ade difficult owing 309 to t he complex nutr ie n t m e d i um . 310 In both pai r w ise cocultures and com mu ni t ie s, V ariov or a x became mor e ef ficie nt at exploit ing 311 Pseudomonas , w ith Variovorax reach ing hi g her densit ies r e lat i ve to Pse u do monas thr ough 312 time. Exploitatio n of P seudomonas b y Variovorax s howed an ove r all i ncrease through time that 313 is c on sis tent wi th arms race dynamics (G andon et al., 200 8). In turn, Pseudo monas adapted by 314 becoming more resistant to e xploitati on, albeit with a time la g, t o 6-we ek e volved Variovorax . 315 How ever, there was no observed increas e over all in res istance of Pseudomonas aga inst 316 Variovorax ov er t ime, w ith t he i ncr ea se in resistance to wee k 6 Variovorax being ac com pani ed 317 by a decre ase in r esi stance to ancestral V a riov orax. This sugges t s that fluct ua t ing se lectio n 318 dynamics (FSD ) acting on Pseudomon as playe d an impor tant role in the c oevoluti on with in t his 319 sy stem ( Hall et al., 2011). O ur knowledge o f ant ag onist ic coevolut ionary dyn amics in mic r obes 320 is mostl y based on bacte ria-phage m odels, w here s election press ur es an d th e ev olut ionary 321 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint potential ar e diff erent t o wi t hin- trophi c coevolutio n (B uck lin g & Rainey, 20 02; Gandon et al., 322 2008; Góm e z & Buckling, 2011). Infect ion w it h lytic phage l eading to cell dea th imposes str on g 323 selec t ion o n bacteria to a dapt defenc es, w hereas comp et i t iv e i nter a ct i on b etw een bact e r ia – 324 especi ally t ho se in v o l v in g i nteract i on s ov er metabolites - mi ght be less s pecific, leading to 325 wea ker s election. FS D in hos t -pathog en sys t e m s bec om es more prevalent due to increasing 326 costs of higher infecti vit y/resis t ance for the phag e and bacteria, resp ectively ( Hall et al., 2011), 327 resulting in sel ect ion o n st a nd i ng gen etic variation. Thi s might explain w hy most work to date, 328 using a bacteria-phage model sy stem r eported coevolv ing part ne r s both exhi bit i ng ei ther ARD 329 or FSD. O ur res ult s suggest t hat th is i s not the cas e for with in trop hi c- lev el coev o l u ti on, w here 330 coevolving partners show differ e nt coevolutio nary dynamics . Furt her w or k is needed to ex plore 331 the wider implic at i o n s of such h igh l y a sy m metric coev olu tio nary dynamics a nd their impact o n 332 ecology and ev o lut ion . 333 A common crit icism o f labor a t ory studi es of coev olut ion is that t hey are overly simplis tic in t heir 334 conditi ons and far removed f rom nature to of fer ins ig ht into mor e co mplex ev olut io nary 335 dynamics . Previous s t udies have f ound t hat ev en small i ncr e a ses in com munity complexit y can 336 sig n ificantly affect coevoluti on (B a r raclough, 2015; Blaz ani n & Turner, 2021; Cas t le d i ne, 337 Sierocinski , et a l., 2022; Manriquez et al. , 2021). Our results instead find that pairwis e 338 coevoluti on can pr edic t coevoluti on i n community c ontext s whic h may be due to se lection 339 pressures being sufficientl y s tro ng . Similar cases of parallel evolutionary dynamics may be 340 observed i n w ider contexts suc h as ph age therapy, w here bacter i a ex perienc e strong sel ect ion 341 to phage in patients and i n c losed laborator y conditio ns ( Cas t ledi n e , P a df i el d, et a l ., 2022). A s 342 Pseudomonas is the specie s Variovorax derives the stron ges t fit ness be n e f it from, interaction 343 intens itie s and reciprocal selec t ion may hav e been suf ficiently stro ng for coev olut ion des pit e 344 co-occurr i ng communit y members. M utation supply rates may h a ve als o be en non-s ig nificantl y 345 affe cted by oth e r community membe rs as Var i ovorax can generally ex ploit at least two ot her 346 community members (o ther t han Ps eudomonas ) w h i ch may allow it to maintain s uff i cient 347 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint mutation rates for coevoluti on ( Castl edine et a l., 2024; Gandon & Michal ak is, 200 2). Wider 348 work analysing coevo lutio nary int e r ac t ions wi thi n trophic lev els, inc lud i n g between cro s s-349 feeding mutualists and ex ploi tative int eractions, wil l g ive i ns i gh ts i nto how c oev olut ion oc cur s 350 in na t ural com munity contexts. The observ ed negative effe ct of commun ity complexity o n 351 evolv ed Pse u domonas dens ities might be explained by additional s electi on press ur e arising 352 from competition with ot h e r community member s . Va riov o rax be n e fi t s fro m mos t specie s in 353 the commu ni t y , while Pseudomonas competes against them leading to str on g er s electio n on 354 Pseudomonas in the commun ity treatm e n t. Thi s pot ent ial ly leads to trade- offs betw een 355 adapt ation to Variovorax and other c o mmunit y members. 356 In thi s study, we ex plicitly demons trat e antag o ni stic coev oluti on betw een naturally co-357 occurring bacteria. This a n tag on ist ic coevoluti on leads t o incr e ased ex ploitatio n but not 358 increased res istance over the ex per im ental time and s hows an important rol e of bot h AR D and 359 FS D. Furthermore , we show that pairwis e coevolut ion can be robust in the face of commu ni t y 360 complexi t y . C oev oluti onary dynamics d i d not change si g n ificantly wi th i ncreased complexi t y , 361 despite finding differences i n dens it ie s of species s u ggest ing a chang e i n s elective press ur es . 362 Und e r s t anding the i nterplay betwe en c o e voluti on and biot ic complex it y i s c rucial not onl y for 363 advanc ing evolut ionary t heory but also for applications i n medicine, bi otechnol ogy , and 364 ecology , where microbial coevolu tio n can influence antibio tic resi s t ance , pathogen evoluti on, 365 and micro bi o me stability. Th is w or k contributes t o our understanding of withi n-t rophic leve l 366 coevoluti on, wi th implicatio ns for bo th n a t ural a n d eng ineered microbial c om munities . 367 Referen ces 368 369 Alseth, E. O ., Pur s ey, E., L u j án, A. M., McL eod, I., Rollie, C ., & W es t ra, E. R. ( 2019 ). Bacterial 370 biodive r s it y drives t he evo lutio n of CRI SPR- bas e d phage resis t anc e. Nature, 574( 777 9), 371 549- 552. https://d oi .o rg/10.1038/s41 586- 0 19- 166 2-9 372 Barra clough, T. G . ( 201 5). H ow Do Spe cies In t eractions Af fec t Evolut io nary Dynamics A cross 373 Whole Co mmunit ies ? Ann ual Review o f Ecolo gy , E volutio n, and Systematics , 374 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint 46 (Volume 46, 20 15), 25- 48 . h tt ps://do i.org/htt ps:/ /doi.org/10.1146/annur e v-ecolsys-375 1124 14- 05 403 0 376 Bates, D. , Mächler , M., Bolker, B., & W alker, S . ( 2015). Fit tin g linear mix ed-effect s models usi n g 377 lme4 [sparse mat rix meth ods; li near mix ed models; penali zed leas t squares; C hole sky 378 decompositio n]. Journal of Statistic al S oft w are , 67 (1), 1- 48. 379 https:// doi.org /10.1863 7/jss . v 0 67.i0 1 380 Blazanin, M., & Turner, P. E. (2021). C o mm uni t y context matters for bacteria-phage ecology and 381 evolutio n. T he ISME Journal , 1 5 (1 1), 3 119- 3 128. ht tps://doi.org/10.1038/s41 3 96-0 21-382 0101 2-x 383 Briscoe Runquis t, R. D., Gorton, A. J., Y oder, J. B., Deacon, N. J. , G rossman, J. J., Kothari, S., 384 Lyon s , M. P., Sheth, S. N ., Tiffin, P. , & Moe ller , D. A. (2020 ). C ontext Depend ence of 385 Local A daptation to Abiot ic and Bio tic Environmen ts: A Q uant it ativ e and Q ual itat i ve 386 Synthesi s . Amer i can Natur alist , 195 (3) , 412- 43 1. ht tp s :/ / doi.org/10.1086/707 322 387 Brockhurst, M. A., & Kos kella , B. (20 13) . Experimental coev oluti on of species i nteractions. 388 Trends in Ecolo gy & E volution , 2 8 ( 6), 3 67-3 75. 389 https:// doi.org /https:/ / doi.org/10.1016 /j.tree .2013.02.00 9 390 Buckling, A ., & R aine y, P. B. ( 200 2). An t agonist i c c oevolut ion betw een a bacte rium and a 391 bacteriophage. Proce edings of the Ro y al Society of L on do n. Ser i es B: Biologica l 392 Scien ces , 269(1 494), 93 1-9 36. h ttps:// doi.or g / doi:10.1098/rspb.2001.1945 393 Ca stledine, M., Padfield, D. , & Bucklin g, A . (2 020). Experimental (co)evo l u tio n i n a mul ti-394 specie s microbial commun ity res u l t s i n local maladapta t ion. Ec o l og y L e tt e r s , 2 3 (1 1), 395 1673 -16 81. https://doi.or g / https:/ /doi. o rg/10.1111/ele.13599 396 Ca stledine, M., Padfield, D. , Sierocinsk i, P., Soria Pascual, J., H ughes, A., M äkinen, L., Fr i m an, 397 V.-P ., Pir na y, J.-P., Mera bishvili, M., de Vos, D., & Buckling, A . (20 22). Parallel ev o l u tio n 398 of Pseudomonas a er ug ino sa phage r e s istance and virulence l os s in respons e to phage 399 treatm e nt in vi v o and i n vitro. eLi fe , 11 , e7367 9. ht tps:// doi.org /10.7554/eLife.7367 9 400 Ca stledine, M., Pennyc o ok, J., New bury , A., Lear, L., Er dos, Z., Lew is, R., Kay, S . , Sande r s , D . , 401 Sünderha uf, D . , Buckli ng, A., Hes se, E. , & Padfield, D. (2024). C haracter i zing a stable 402 five-spec ies m icrobial communi ty for use in e xperimental evolut ion and ecol ogy. 403 Microbiology , 170 (9). https://d oi .o rg/ht tps:/ / doi.org/10.1099/mic.0.00148 9 404 Ca stledine, M., Sierocinski, P., Inglis, M., Kay , S. , Hay ward, A., Buckling, A ., & Padfiel d , D . 405 (202 2). G r e at er Phag e Geno typi c D i versity C on strains Arms-Race Coevolut io n [Ori gina l 406 Research]. Frontiers in Cellular and In f ection Microbiology , Vol ume 12 - 2 022 . 407 https:// doi.org /10.3389/fcimb.2022.83 4406 408 Chang , C.-Y., O sbor ne, M. L., Bajic, D ., & Sanchez , A. (2 020). Artificially select ing bacter i al 409 communities using propagule s t rategi e s†. Ev ol ut i on , 7 4 (1 0), 2 392 -2403. 410 https:// doi.org /10.1111/evo.14092 411 Chen, N., & Zhang, Q.-G. ( 202 4). Suf fer ing makes y ou weak er : Limi ted evolut i onary adaptation 412 in c ompetiti v ely inferior populatio ns. Ec ol og y L e t t e r s , 27 ( 6) , e144 57. 413 https:// doi.org /https://doi.org/10.1111 /ele.14457 414 Friman, V.-P., & B uck ling, A. ( 201 3). Effects of predat i o n on real-t i m e h ost–parasite 415 coevoluti onary dynamic s. Ecolog y L ett ers , 1 6 (1), 39- 46. 416 https:// doi.org /https://doi.org/10.1111 /ele.12010 417 Friman, V. P., Laa kso, J., Koivu ‐ orava, M., & Hiltune n, T . (20 11). Pulsed ‐ resource dynamics 418 increase t he as ymmetry of a n t ag on ist i c coev olut i on between a pr edatory pr o tist and a 419 prey bac t erium. Journal of Evolut i o nar y Biolog y , 2 4 (12), 25 63- 257 3. 420 https:// doi.org /10.1111/j.1420- 910 1.2 011.02 37 9.x 421 Gaba , S. , & Eber t, D . (2009). T ime- s h ift experiments as a t ool to study antago n istic c o e v olutio n. 422 Trends in Ecolo gy & E volution , 2 4 ( 4), 2 26-2 32. 423 https:// doi.org /https://doi.org/10.1016 /j.tree .2008.11.00 5 424 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint Gandon, S ., B uck lin g, A., Deca estecker, E., & Day, T. (2008). Host–parasite co ev olutio n and 425 patterns of adaptation across t i m e and space. J ournal of Ev olutionary Biolog y , 21 (6 ) , 426 1861 -18 66. https://doi.or g / 10.1111/j.1420- 910 1.200 8.01598.x 427 Gandon, S ., & Michalakis, Y. ( 200 2). Lo ca l adaptation, evolutio nary potent ial and host–parasite 428 coevoluti on: i nt e r actions between m igr ation, m utation, populati on s i ze and g eneration 429 time. Journal o f E volutio nary Biolo gy , 1 5 (3), 451- 4 6 2. ht tp s :/ / doi.org/10.1046/j.1420 -430 9101.2 00 2.00402.x 431 Garbutt, J., Bonsall, M. B ., W right, D. J., & Raymond, B. ( 20 11). A ntagon ist ic co mpetit i o n 432 moderates vi r ulence in Bacillus t huri n g iens is. Ecolo g y Letters , 1 4 ( 8) , 7 65- 77 2. 433 https:// doi.org /10.1111/j.1461- 0248.2 011.01 63 8.x 434 Gómez , P., & Buckli ng, A. ( 2011). B a ct er ia - Phage Ant ag on ist ic C oev olutio n in S oi l. Science , 435 332( 602 5), 1 06- 109. ht tps://doi.org/do i:10.112 6/science.1198 76 7 436 Gómez , P., & Buckli ng, A. ( 2013). Coevolut io n wi th phage s does n ot i nf lue n c e the evolut ion of 437 bacterial mutation rates i n soil. The ISME Jour na l , 7 (1 1), 2 242 -22 44. 438 https:// doi.org /10.1038/ismej.2 013.1 05 439 Hal l, A. R., Scanlan, P. D ., Morgan, A. D ., & Buckling, A. (20 11). Ho st–parasi t e coevoluti onary 440 arms races give way t o fluct ua t in g sele ction. Ecolo g y Le t ters , 1 4 (7) , 6 35- 642. 441 https:// doi.org /https://doi.org/10.1111 /j.1461-0 248.2 011.0 162 4.x 442 Hal l, J. P. J., Harrison, E., & Brockhurst, M. A. (20 1 8) . Competi tive species in t er ac t ions co nstrain 443 abiotic adaptation in a bacterial soi l co mmunity. Ev o l L e t t , 2 (6), 58 0-5 89. 444 https:// doi.org /10.1002/evl3.83 445 Hans en, S. K., Rainey, P. B. , Haage nsen, J. A. J., & Molin, S. (2 007). Evolut ion of spec ies 446 interactions i n a biofilm communit y. Nature, 445 (712 7), 5 33- 536. 447 https:// doi.org /10.1038/nature0551 4 448 Hart, S. P., Tu r c ot te, M . M., & Levi n e , J. M. (2019 ). Effects o f r api d evolu tio n o n species 449 coex istence. Proceedings o f the National Academy of Scie n c es , 116(6), 21 12 -211 7. 450 https:// doi.org /doi:10.1073/pnas.1816 2981 16 451 Hartig, F. ( 201 8). DHARMa: Residual Diagnost ic s for H ierarchical (Multi- L evel / M ixed) 452 Regression Models. In R P ac kag v er s ion 0 20 . 453 Johnke, J., Baron, M., de Leeuw, M., Kus hm a r o, A ., J urkev itch, E., Harms, H ., & Chatzino tas, A. 454 (201 7). A G eneralist Protis t Predator E nables C o e x istence in Multi troph i c Pre dator -Prey 455 Sys t ems Con taini ng a Phage a n d th e B acter i al Pred at or Bdellov ibr i o. Fro ntie r s in 456 Ecology and Ev olution , 5 . h tt ps :/ /doi.or g/10.3389/fevo.201 7.001 24 457 Kim, W., Racimo, F., Schluter , J., Lev y, S. B., & Fos t er, K. R . ( 2014) . Impor tanc e of pos iti oni ng for 458 microbial ev olut i on . Pr ocee dings of th e National Acade my of Sciences , 111(16), E1 639 -459 E1647. ht tps:// doi.or g / doi:10.1073 / pnas . 1 3236 321 11 460 Koch, G ., Yepes, A., Förstner, Konrad U., W e r mser , C ., Stengel, Stephanie T., Modamio, J., 461 Ohls en, K., Foster, Kevi n R ., & Lopez, D. (201 4). Evo l u ti on of Resi stance t o a L ast-Resort 462 Antibio tic in Staphyl o c occus aureus v i a Bacterial C omp e t it ion. Cell, 1 5 8 ( 5) , 1 060- 107 1. 463 https:// doi.org /10.1016/j.cell.2014.0 6.046 464 Law r enc e, D., F iegna, F., Behrends, V., Bundy, J . G., Phillim ore, A. B., Bell, T ., & Bar racloug h, T. 465 G. (2 012 ). Species Int eraction s A lt er Evolutio nary Responses t o a N o vel E nviro nment. 466 PLoS Biolog y , 1 0 (5), e 100 133 0. ht tps:// doi.org/10.1371/jour nal.pbio.1001330 467 Lenth, R. V. (2 023). e m means : Estimat ed Marginal Means, ak a Least-S quar es Mea n s . In 468 https:// C RA N .R- project.org /pack age=emmeans 469 Loreau, M. , Jarne, P ., & Ma r tiny, J . B . H. (20 23). Opport unitie s t o advance t he s ynthes i s of 470 ecology and ev o lut ion . Ecolo gy Lett e r s , 26 (S1), S11-S 15. 471 https:// doi.org /https://doi.org/10.1111 /ele.14175 472 Luján, A. M., Pater s on, S., Hes se, E., S ommer, L. M., Ma r v ig, R. L., Shar ma, M . D . , Als et h, E. O., 473 Ci of u, O., Smania, A. M., Molin, S., Johansen, H . K., & Buckli ng, A . ( 202 2). Pol ymicrobial 474 infection s c an select ag ai nst Pseudomonas a er ug ino sa mutator s because of quor um-475 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint sensing trade-offs. N ature Ecology & E v olut i on , 6 (7), 979-988. 476 https://doi.org/10.1038/s41559-022-01768-1 477 Manriquez, B., Muller, D., & Prigent-Combaret, C. (2021). Experimental Evolution in Plant-478 Microbe Systems: A Tool for Deciphering the Functioning and Evolution of Plant-479 Associated Microbial Communities. Front Microbiol , 1 2 , 619122. 480 https://doi.org/10.3389/fmicb.2021.619122 481 Paterson, S., Vogwill, T., Buckling, A., Benmayor, R., Spiers, A. J., Thomson, N. R., Quail, M., 482 Smith, F., Walker, D., Libberton, B., Fenton, A., Hall, N., & Brockhurst, M. A. (2010). 483 Antagonistic coevolution accelerates molecular evolution. Nature, 464(7286), 275-278. 484 https://doi.org/10.1038/nature08798 485 Pearl Mizrahi, S., Goyal, A., & Gore, J. (2023). Community interactions drive the evolution of 486 antibiotic tolerance in bacteria. P roce e dings of the N ational Academy of Scien ces , 487 120(3), e2209043119. https://doi.org/doi:10.1073/pnas.2209043119 488 Piccardi, P., Ulrich, E., Garcia-Garcerà, M., Martino, R. D., Testa, S. E. A., & Mitri, S. (2024). The 489 evolution of reduced facilitation in a four-species bacterial community. E volu tion 490 Letters. https://doi.org/10.1093/evlett/qrae036 491 R Core Team. (2025). A language and environment for statistical computing. Vienna, Austria: R 492 Foundation for Statistical Computing. { https://www .R-project.org/ } 493 Yin, W., Zhou, L., Yang, K., Fang, J., Biere, A., Callaway, R. M., Wu, M., Yu, H., Shi, Y., & Ding, J. 494 (2023). Rapid evolutionary trade-offs between resistance to herbivory and tolerance to 495 abiotic stress in an invasive plant. Ecol ogy L ett ers , 2 6 (6), 942-954. 496 https://doi.org/https://doi.org/10.1111/ele.14221 497 498 F i g u r e 1 E x p e r i m e n t a l d e s i g n t o t e s t t h e i m p a c t o f c o m m u n i t y c o m p l e x i t y o n c o e v o l u t i o n b e t w e e n f o c a l s p e c i e s 499 Ps e udomonas ( P ) a nd V ariovorax (V ). Th e ev o lution exper i m en t consist ed o f thr e e tr e a t m en t s : monoc u lt ure, 500 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint co c ul t ur e a nd community . E ach trea tm e n t has b een pass a g e d ( 1% ) weekly fo r 1 0 we e k s (A). Coevol u ti o n was 501 asses sed by growth ass a y s comb in in g Ps e udo mo n as a nd Va ri ovo rax f rom d if f e re nt evolu t ion ary ti m ep o in ts fro m th e 502 same r e pli c at e line wi th i n each tr e a tm e nt ( B). a n c = anc e s tor, 6 = 6-week isola t es, 10 = 1 0-w eek is o la te s . 503 504 Fi g u re 2 D e ns ity o f Ps e ud o monas (A) and V ariovorax (B) in t h e ir c o nt em p orary c o mb i na t i on s of a nces t o r a nd evolved 505 li ne a g e s i n mo n o c ult ure, cocult u r e o r c om m uni ty c u ltu re . Pop u la t io n densit ie s a re plott ed agains t evolutiona ry t i me. 506 Boxpl o ts sh owi n g me di a ns , f i rs t a nd th i rd q u arti les, w h isk ers ar e 1.5 * i n terquartil e r an g e (IQR). I nd i vi dua l poi n ts 507 rep r esen t individual tr ea t ment r e plicates . 508 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint 509 Fi g u re 3 R e lati v e proportion of V a r i o vo r a x w h en c ultu r ed with Ps eudo m o n as from diffe r ent ti m e -poi nts . 510 Pseudom o na s and V a ri ovora x evolved in (A ) m on o cu ltur e (ea ch sp ec i e s in i s olatio n ), (B) i n co c ultur e ( P s eudo m on as 511 an d Va r i ovor a x ) and in ( C ) co m m unity ( P se u do m o na s and V a r i ovor a x e mb e dde d wi t h 3 other b ac t eria l speci e s )). We 512 fou nd a significan t in te raction betwe e n e v o lutionary time of Ps e udomonas and Va r i ovo r a x w h en s p ec i es ev o l v e d a s 513 co - cul tur es (B) or with i n a c o mmunit y (C), but n ot wh e n t h es e had e v o lved i n isolatio n (A ) . I nd ependent treat m e n t 514 rep li c at e s ar e sma ll p o ints, la rg e p oin ts a re mean pr o portion of Va r i ovor a x , b a r s r e p r e s e n t ± S E . T h i n l i n e s a r e 515 conn e cting the r e plicat e s t hat a r e tracked thr ough ti me (ex cep t fo r th e a n ce s tor), thick lin e s repre s e nt th e mean. 516 A n c e s t o r - a n c es t or co m bi n at i o ns b e t w e e n t h e p an el s a r e t h e s a me 8 r ep l i c a t es p r e s e nt e d t o a i d v i s ua l iz at i o n. 517 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint 518 Fi g u re 4 Re lati v e p r opo rti on of Pse u domo na s when c ultur e d with Va r i ovor a x fro m diffe r en t t im e -points. 519 Pseudom o na s and V a ri ovora x evolved in (A ) m on o cu ltur e (ea ch sp ec i e s in i s olatio n ), (B) i n co c ultur e ( P s eudo m on as 520 an d Vari ovora x ) and i n ( C) community ( Psedu om o na s a nd V a r i ovor a x e mbedded with 3 other b a ct e ri al speci e s ). 521 Pa tter ns of coevolut ion pr esen t for (B ) a n d (C) b ut not ( A) obse rved as s i g n ifi c a nt int e r a c tion betw e e n evolut ionary 522 tim e of Ps eu d om o na s and Var i ovor ax . In d ep e nd e n t tre atm e n t r epli c a tes ar e sma ll po i nts, la rg e poin ts a r e m e a n 523 pro portion o f P s eu d om o na s , b a r s r e pr e s e n t ± S E. T h i n l i ne s a r e c o nn e c t i ng t h e r ep l i c a t es th a t ar e t r ac k e d t h r o ug h 524 tim e (exc ep t fo r th e a n c estor ) , thi ck lines rep r e sent the m ean. A ncestor- anc e s to r c o mbinati o ns b etwe e n the p an e ls 525 are t h e sa m e 8 re plicates pre s e nt ed to aid visual ization. 526 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint 527 Fi g u re 5. T he density o f indivi dua l sp ecie s a nd their co m bi n ed d en s ity wh e n a s se m ble d a s t wo-spe c ie s cocult u r es 528 (Va ri ovora x and Ps eudo m o n as) fol l owi n g e v o lution in (A ) monoc u lt ur e (each s pecies evolved alon e ), ( B) c o c ul t ur e 529 (Pse u domona s a nd Vari o vo rax only) , and (C ) i n a c o m mu n i ty (wit h t h r ee ot he r s pe c i es). Po p ula t ion de n s ities are 530 pl ott ed ag ai ns t evol ut io n ary t i m e . B o x p lots sho win g m ed i an s , f irst and t hi rd q u ar tiles, w hisker s are 1.5 * int erqua r til e 531 ran g e (IQR) . I ndi vi d ua l p o ints repr e s ent indi vi d ua l tr e a tm e n t rep li c a te s . 532 533 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 4, 2025. ; https://doi.org/10.1101/2025.07.01.662545doi: bioRxiv preprint