The gac system integrates physical and chemical cues to promote plant root attachment

Rhizosphere, plant roots, biofilm, gac system signaling 11 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 2

12 Plants host complex communities of microbes that are attached to root surfaces. While many 13 studies have sampled mutant populations after prolonged incubation on roots to identify bacterial 14 genes that enable long-term colonization, the molecular mechanisms governing the early stages 15 root attachment remain less understood. Here, we developed an in vitro root culture system that 16 enables controlled and scalable investigation of bacterial attachment to root tissue. We used this 17 platform to perform a genome-wide screen for root attachment determinants in the plant-18 associated bacterium Pseudomonas protegens Pf-5. Our results reveal that the gacSA two-19 component system functions as a sensory integration hub for coordinating early root attachment. 20 Mutations that disrupt gacS or gacA cause severe root attachment defects despite having no 21 effect on abiotic surface attachment in standard biofilm assays. Mutation of flagellar assembly 22 genes enhances root attachment by mimicking surface contact and activating the gac system. 23 In parallel, chemical cues released by roots stimulate surface attachment in a gac dependent 24 manner. By integrating these signals, the gac system activates cyclic di-GMP-mediated 25 attachment programs that drive the transition from planktonic to sessile behavior required for 26 root association. We build on this model to show that manipulating flagellar surface sensing 27 enhances the competitive fitness of Pf-5 in the presence of a synthetic bacterial community, 28 suggesting a strategy to improve the competitive fitness of beneficial microbes on crops. These 29 findings establish a mechanistic framework linking surface sensing, global regulation, and root 30 attachment in a beneficial rhizobacterium. 31 32 Importance 33 Plant roots are covered with diverse microbes that strongly influence plant health. Growing in 34 association with roots has many benefits, but how bacteria attach to root tissue remains poorly 35 understood. We developed a system to study how a bacterium that improves plant growth called 36 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 3 Pseudomonas protegens Pf-5 attaches to root tissue. We found that physical contact with the 37 root surface and chemical cues released by roots both enhance attachment to root tissue. A 38 sensory system called gacSA is responsible for integrating physical and chemical cues to 39 activate a root attachment program. Variant bacteria that prematurely activate the gac system 40 compete more effectively with other bacteria on roots, suggesting that the root attachment 41 pathway we characterized could serve as a strategy to use beneficial bacteria in agriculture. 42 43

44 Surface-associated communities called biofilms represent the predominant lifestyle for 45 most bacteria (1, 2). Cells in biofilms are encased in a self-produced extracellular matrix that 46 provides resistance to stresses such as shear flow, unfavorable temperatures, antimicrobial 47 substances, and nutrient limitation (3). While biofilms are found in nearly all environments on 48 earth, their association with living tissues has particularly significant effects on humans. Biofilms 49 within human tissues are thought be the source of chronic infections, and tolerance to 50 antimicrobial treatments makes them difficult to treat in clinical settings (4). In agricultural 51 systems, biofilms that from on plant tissues are crucial determinates of crop productivity (5). 52 Despite the profound importance of host-associated biofilms, mechanistic analyses of biofilm 53 formation have historically been dominated by studies on abiotic surfaces such as plastic, glass, 54 and agar. Knowledge of host-associated biofilms is often inferred from these inanimate models 55 (6). Increasing efforts to study bacterial colonization of biotic substrates have begun to reveal 56 substantial differences in host-associated biofilm formation, but continued work is needed to 57 accelerate the development of strategies for managing biofilms in host-associated settings (7). 58 Plant health is strongly influenced by complex biofilms in the rhizosphere, a narrow zone 59 of soil surrounding root tissues (8–10). Plants interact with microbes in the rhizosphere by 60 secreting sugars, amino acids, organic acids, and secondary metabolites that are collectively 61 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 4 referred to as exudates (11, 12). Exudates create chemical gradients that attract bacteria toward 62 the root surface (Fig. 1). They also promote bacterial aggregation and provide nutrients that 63 make the root surface a favorable niche (13). Plants, in turn, rely on microbes in the rhizosphere 64 for nutrient acquisition, suppression of pathogenic organisms, and tolerance to stressors such 65 as drought. This mutualistic relationship requires microbes to physically attach to the root 66 surface. 67 Efforts to identify the genetic determinants of root attachment have largely relied on 68 inoculating mutant libraries onto plants growing in substrates such as soil, vermiculite, sand, or 69 synthetic plant media. Sampling the root-associated population using transposon sequencing 70 (Tn-Seq) has consistently identified genes involved in amino acid biosynthesis, carbon 71 utilization, nitrogen assimilation, iron acquisition, suppression of plant defenses, and competition 72 with other microbes (14–23). Such studies have been instrumental in defining metabolic 73 pathways required for persistence in the rhizosphere. However, the prolonged incubations 74 emphasize sustained colonization rather than the earliest stage of root attachment. A more 75 thorough understanding of how bacteria attach to root tissue is needed to understand biofilm 76 formation in the rhizosphere. 77 Members of the Pseudomonas fluorescens species complex are frequent residents of the 78 rhizosphere (24, 25). These bacteria often promote plant health by enhancing nutrient uptake, 79 producing growth-promoting compounds, and antagonizing pathogens. There is extensive 80 knowledge of the mechanisms that govern biofilm formation within the Pseudomonas genus, 81 particularly in the opportunistic human pathogen Pseudomonas aeruginosa (26). 82 Pseudomonads coordinate biofilm formation through a regulatory paradigm centered on the 83 second messenger cyclic-di-GMP (c-di-GMP). While the details of each signaling cascade vary 84 among individual strains, environmental cues trigger elevated c-di-GMP levels which in turn bind 85 effector proteins that promote biofilm-associated behaviors such as reduced motility, biofilm 86 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 5 matrix production, and expression of adhesins. A well-characterized output of c-di-GMP 87 signaling in P. fluorescens is the Lap system, which controls a large cell-surface adhesin LapA. 88 Elevated intracellular c-di-GMP concentrations promote the accumulation of LapA at the cell 89 surface where it mediates attachment and supports biofilm formation. Reduced c-di-GMP levels 90

in LapA cleavage and release, leading to detachment and dispersal (27). 91 In this study, we develop an in vitro root tissue culture system to investigate bacterial root 92 attachment. This approach overcomes practical limitations of studying root attachment in whole 93 plants and avoids abiotic adhesion assays that do not capture interactions with living host tissue. 94 We use Pseudomonas protegens Pf-5, a soil bacterium that associates with the roots of various 95 crops (28–32), to perform a genome-wide screen for root attachment determinants. Our results 96 demonstrate that the gacSA two-component system acts as a key integrator of signals to 97 coordinate root attachment. Both flagellum-mediated surface sensing and chemical cues 98 released by roots activate gac signaling to promote c-di-GMP-dependent attachment. 99 Furthermore, we show that manipulating flagellar surface sensing enhances the competitive 100 fitness of Pf-5 in the presence of a synthetic bacterial community, suggesting a potential strategy 101 for improving the performance of beneficial bacteria in agricultural applications. Our results shed 102 new light on how host tissues influence the earliest stages of biofilm formation. 103 104

105 Genome-wide screen identifies Pseudomonas protegens Pf-5 genetic determinants for 106 early root attachment 107 We developed an in vitro tissue culture system for propagating tomato roots in liquid 108 medium (Fig. 1A, B). This platform allows large volumes of root tissue with an extensive surface 109 for attachment to be cultivated independently of whole plants under highly controlled conditions. 110 We then engineered a Pf-5 strain expressing mVenus and used it to carefully optimize conditions 111 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 6 for robust colonization of root tissue. Pf-5 was inviable in standard root culture medium, but 112 adding carbon sources commonly detected in root exudates allowed Pf-5 to grow (Fig. S1A-C) 113 (33). We validated our system by comparing root association of fluorescently-tagged wild-type 114 Pf-5 to a mutant lacking the large surface adhesin LapA (27). Bacterial abundance on roots was 115 measured by imaging total emitted fluorescence from roots in plate wells (Fig. 1C), measuring 116 the fluorescence emitted by cells recovered from roots or directly quantifying viable cell counts 117 (Fig. 1D; Fig. S1D). Both strains attached to root tissue, but roots inoculated with the Δ lapA 118 mutant emitted reduced levels of fluorescence and carried less CFUs compared with wild-type 119 Pf-5, consistent with a deficiency in root attachment. 120 We used the in vitro colonization system to perform a genome-wide screen for Pf-5 genes 121 involved in attachment to roots. A barcoded transposon mutant library was inoculated into flasks 122 containing root tissue, and the mutant pool was allowed to attach to roots for 24 hours (34). We 123 predicted that mutants that retained the ability to associate with root tissue would partition out of 124 the medium as they attached to roots, enriching for attachment deficient mutants in the liquid 125 medium. We amplified these selections by inoculating the unattached population into fresh root 126 tissue for four additional passages (Fig. 3A). Barcode abundances from the initial mutant 127 population and from the unattached fraction after each root passage were used to calculate 128 fitness scores for each gene across the passages (Fig. 3B; Table S1). 129 Genes with the strongest effects on root attachment are shown in Tables 1 and 2. 130 Mutations causing the strongest defects in root attachment mapped to gacS and gacA which 131 encode a histidine kinase-response regulator pair that sits atop a large regulatory hierarchy. 132 Mutations affecting transcription factors ( PFL_2828, PFL_0502, PFL_1588), uncharacterized 133 transporters (PFL_4192, PFL_1740), the pxpABC operon encoding the allophanate hydrolase 134 enzyme (PFL_1513-1515) (35), as well as genes involved in c-di-GMP (PFL_5779, PFL_4532), 135 lipid (PFL_5687), and carbon (PFL_6157) metabolism also caused root attachment deficiencies. 136 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 7 Mutations that enhanced root attachment were concentrated in genes required for the function 137 of the flagellum, including assembly regulators, structural components and chemotaxis systems 138 (36). In-frame deletion mutants Δ gacS and ΔgacA recapitulated the strong attachment defect, 139 with root attachment reduced by ~60% relative to wild-type (Fig. 3C). The Δ fliF mutant, 140 representing flagellar structural mutants and lacking a flagellum, attached ~30% more effectively 141 than wild type. Deletion mutants of PFL_1588, PFL_4192, PFL_2828, PFL_0502, PFL_5779, 142 and PFL_4532 were hypoadhesive, consistent with the Tn-Seq results. The Δ cheY mutant was 143 hypoadhesive despite appearing hyperadhesive in the Tn-Seq results. ΔPFL_1513 did not differ 144 significantly from wild type, although it was hypoadhesive in the Tn-Seq results. All the deletion 145 mutants displayed comparable growth kinetics to wild type (Fig. S2A). 146 147 Flagellar-dependent surface sensing activates the gac system 148 We sought to understand two groups of genes with opposing effects on root colonization. 149 gacS and gacA mutants have severe root attachment defects (Fig. 3). Activation of the gacSA 150 two-component system induces expression of small RNAs called rsm, which go on to activate 151 genes involved in biofilm formation, secondary metabolite production, and exopolysaccharide 152 synthesis (37). Mutating flagellar genes in Pf-5 strongly enhances root attachment (Fig. 3). The 153 flagellum plays numerous roles throughout the process of biofilm formation including facilitating 154 productive contact with surfaces and recognizing physical cues for contact with surfaces (36, 155 38). We confirmed that the root colonization phenotypes of Δ gacS, Δ gacA, and Δ fliF were 156 complemented by ectopic expression of arabinose-inducible forms of each mutated gene (Fig. 157 S2B). We then used epistasis analysis to determine whether the gac and flagellar genes 158 contributed to a unified pathway controlling root attachment. A Δ fliF ΔgacA double mutant 159 displayed a root colonization defect comparable to the ∆ gacA single mutant, indicating that the 160 gac system is required for enhanced root attachment in flagellar mutants (Fig. 4A, Fig. S3). To 161 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 8 determine whether differences in root attachment correlate with altered c-di-GMP levels, which 162 are known to promote bacterial adhesion during responses to surface contact, we quantified c-163 di-GMP levels using a fluorescent reporter (39). The Δ fliF mutant showed elevated c-di-GMP 164 and the ΔgacA mutant exhibited lower c-di-GMP concentrations than wild-type. The ΔfliF ΔgacA 165 double mutant displayed c-di-GMP levels matching the wild-type phenotype and confirming that 166 the hyperadhesive, high-c-di-GMP state of ΔfliF depends on gacA. 167 Mutations that disrupt flagellar assembly enhance biofilm formation in several bacteria by 168 activating surface sensing pathways (36, 38). We predicted that the Δ fliF mutation enhances 169 adhesion by mimicking surface contact, which in turn activates the gac system. We monitored 170 transcriptional reporters for the three small RNAs that are directly activated by GacA~P in Pf-5 171 (rsmX, rsmY, and rsmZ) under conditions designed to induce surface sensing (Fig. 4C) (40, 41). 172 Surface sensing was stimulated by increasing medium viscosity with methyl cellulose (42), by 173 disrupting flagellar assembly using the ΔfliF mutation, or by incubating cells on an agar surface. 174 gac activity was elevated across all conditions where surface sensing was stimulated (Fig. 3D-175 G). Similar trends were observed for all three rsm reporters (Fig. S4). Together, these findings 176 support a model in which physical contact with surfaces causes the flagellum to activate the gac 177 system. 178 179 The gac system integrates physical and chemical cues to control surface attachment 180 To evaluate how the attachment phenotypes we measured on root tissue compare to 181 abiotic surface attachment, we quantified biofilm formation in microtiter plates using a crystal 182 violet (CV) staining assay. We observed a different attachment pattern on an abiotic surface 183 than we measured on root tissue. The ∆fliF mutant remained hyperadhesive on abiotic surfaces, 184 and a Δ fliF ΔgacA double mutant still lost the enhanced adhesion of Δ fliF alone (Fig. 5A). 185 However, the ∆gacA single mutant and the ΔfliF ΔgacA double mutant showed identical levels 186 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 9 of CV staining to the wild-type strain. These results indicate the gac system plays a stronger role 187 in attachment to tissue than to abiotic surfaces. 188 We predicted that differences in colonization efficiencies of gac mutants on root tissues 189 compared to abiotic surfaces could be explained by the presence of exudates that accumulate 190 in tissue culture medium (43, 44). We measured gac reporter activity in media that had been 191 conditioned by roots for 6, 16, and 24 hours. Activation of all three gac reporters increased with 192 longer root conditioning times. No activation was observed in the Δ gacA mutant (Fig. 5B, Fig. 193 S5A and C). These results demonstrate that root-secreted molecules stimulate gac signaling. 194 We next tested whether the chemical signals from root exudates and physical signals 195 from the flagellum activate the gac system in an additive manner or through a sequential logic 196 in which one cue is required before the other can be detected. To distinguish these possibilities, 197 we examined gac activity, c-di-GMP levels, and surface attachment in a CV assay, under 198 conditions with surface sensing activated or chemical cues alone, or both inputs together. When 199 we combined the two signals using root-conditioned media and a Δ fliF mutant, we observed 200 higher levels of gac reporter activation, higher c-di-GMP levels, and increased adhesion, 201 compared to conditions with either input alone (Fig. 5C-E, Fig. S5B and D). This enhancement 202 indicates that the gac system integrates physical and chemical signals to promote root 203 colonization. 204 205 Activating flagellar surface sensing enhances competitive root colonization 206 Rhizosphere colonization efficiency is thought to be driven by competitive interactions 207 among soil organisms. We predicted that activating surface responses and gac signaling using 208 the ΔfliF mutation would provide Pf-5 with a competitive advantage over other bacteria during 209 root attachment. We performed a root attachment assay in the presence of THOR, a synthetic 210 bacterial community composed of three members isolated from the soybean rhizosphere (45). 211 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 10 Co-inoculation with THOR reduced the attachment efficiency of wild-type Pf-5 by ~40%, while 212 the ΔgacA mutant remained at low abundance (Fig. 6A). In contrast, the ΔfliF mutant maintained 213 elevated attachment efficiency in the presence of THOR. These results show that enhanced 214 activation of attachment programs in the Δ fliF mutant is sufficient to overcome bacterial 215 competition in the root environment. 216 217

218 Previous studies have used whole plants to identify bacterial genes that promote 219 persistent colonization of roots (7). These efforts have yielded foundational insights into plant-220 microbe interactions in the rhizosphere, but the crucial stage of early root attachment has 221 remained less characterized. In this study, we developed an in vitro root tissue culture system 222 to specifically target the early stages of bacterial root attachment. Propagating large quantities 223 of roots in liquid medium eliminates the time and logistical effort needed for whole-plant 224 experiments. The higher throughput and precise experimental control of a tissue culture system 225 allowed us to identify phenotypes that were not apparent in previous studies (7). For example, 226 only a subset of the attachment determinants we found were identified when Pf-5 mutants were 227 inoculated onto wheat and cotton seeds, grown in a sand/perlite mix, and recovered after one 228 week (21). More broadly, results of many long-duration root colonization Tn-Seq studies are 229 dominated by genes involved in metabolism, transport, cell wall biogenesis, and stress 230 adaptation rather than classical biofilm factors (7, 14, 16, 18, 20–23, 25, 46, 47). These contrasts 231 highlight the complex temporal progression of root attachment. Individual genes likely play 232 stage-dependent roles, and early advantages may later become neutral or even detrimental as 233 colonization progresses. Targeting the early attachment phase specifically led to the 234 identification of new signaling connections that influence attachment to roots. 235 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 11 The gacSA two component signaling system emerged as a crucial determinant of root 236 attachment in our study. While the gac system is known to influence a large set of behaviors in 237 Pseudomonads, there are conflicting reports about its role in biofilm formation and tissue 238 colonization. gac mutants generally show reduced adhesion to abiotic surfaces in crystal violet 239 assays (48–56). However, other studies indicate that abiotic surface adhesion phenotypes of 240 gac mutants are inconsistent and may depend on specific culture conditions (57–59; This study). 241 In root colonization assays with whole plants, the consequences of gac disruption are strongly 242 shaped by competition. gac mutants often do not show impaired root colonization and may even 243 outperform wild type in competition (60–62), reflecting a public goods tradeoff (55, 63). In 244 contrast, gac mutants show a competitive disadvantage in natural soil or mixed-community 245 environments (64–66). While previous Tn-Seq colonization studies that selected for competitive 246 growth identified gac, it was not among the strongest contributors to root colonization (31, 46). 247 In our study, gacS and gacA mutants had by far the strongest adhesion defects of any gene in 248 our early root attachment system. We propose that gac activation is crucial for promoting tissue 249 adhesion specifically during the earliest stages of root contact. 250 Our study also illuminates a novel function for flagella in the root colonization sequence. 251 Flagella have long been recognized as surface colonization factors, but previous root attachment 252 studies have identified a range of phenotypes for flagellar mutants. Root attachment defects, 253 neutral colonization phenotypes and hyper-colonization of roots have all been reported (16, 47, 254 64, 67, 68). Variations likely arise because specific experimental methods inadvertently 255 emphasize different stages of root colonization, but previous studies have mostly interpreted 256 flagellar phenotypes through the lens of motility. Colonization defects are attributed to an inability 257 to migrate efficiently to the root surface, and hyper-colonization is assumed to arise when cells 258 become “trapped” on the root because they cannot swim away. By focusing our tissue 259 colonization system specifically on the early attachment phase, we provide evidence for flagella 260 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 12 functioning as surface sensors. The enhanced adhesion and elevated c-di-GMP levels of the 261 ΔfliF mutant are consistent with studies showing that disrupting flagellar assembly mimics 262 surface contact, and we could recapitulate the effects of the ∆ fliF mutant in wild-type cells by 263 physically obstructing the flagellum (Fig. 4D-G, Fig. S4). Thus, our results provide evidence that 264 surface sensing mechanisms identified for abiotic surfaces are preserved in interactions with 265 living tissue. 266 Root exudates emerged as a key differentiator between abiotic surfaces and root tissue. 267 Root conditioned medium stimulated gac signaling, c-di-GMP production, and adhesion in a gac-268 dependent manner. While we do not yet know the exact exudate compound(s) responsible for 269 this effect, our findings underscore how gac phenotypes seem to be particularly sensitive to 270 chemical changes. Perhaps tuning the system to subtle chemical changes potentiates tissue 271 colonization. More broadly, exudates are by large viewed as chemoattractants that promote 272 motility toward the root (12). Our data supports an expanded view whereby exudates also 273 function as local signals that directly enhance root attachment (13, 69). This finding underscores 274 a fundamental difference between abiotic and tissue surfaces. Living tissues modify the chemical 275 environment through secretion in a way that cannot be reproduced on abiotic substrates, and 276 we propose that the gac system among Pseudomonas fluorescens species complex strains has 277 evolved to incorporate chemical cues from roots into the biofilm activation program. 278 Two-component and phosphorelay systems frequently operate as signal-integration 279 nodes by enforcing coincident detection of multiple signals before committing to irreversible 280 responses (70–72). Our data place gacSA within this framework as an integrator of both 281 chemical and physical cues from roots. While many studies have examined how the sensor 282 kinase GacS is activated, there is still no consensus on the activating ligand(s) (73). Our data 283 show that both surface contact sensing through the flagella and chemical cues in exudates 284 activate the gac system on roots, but we do not yet know if these are direct activators of GacS. 285 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 13 GacS activation triggers gacA-dependent expression of rms sRNAs that go on to affect genes 286 for biofilm formation, plant growth promotion, siderophore production and antimicrobial 287 production (37, 74, 75). Our data indicates that root attachment requires a gac-depedent 288 increase in c-di-GMP levels, and we indeed identified two c-di-GMP metabolic genes that 289 promote root attachment in our screen ( PFL_4532, PFL_5779; Fig. 3). c-di-GMP binds 290 transcriptional regulators that shift cells to a sessile state by repressing motility while activating 291 adhesins (76–78). The absence of strong hits in specific adhesins in our screen likely reflects 292 redundancy among attachment factors and suggests that regulation occurs primarily at the level 293 of c-di-GMP signaling rather than direct transcriptional activation by rsm regulators. In many 294 bacteria, surface colonization follows a relatively simple pathway in which surface contact 295 elevates cyclic-di-GMP and activates adhesins and biofilm matrix components (79, 80). In 296 contrast, in the Pseudomonas fluorescens species complex this transition is embedded within 297 the global gac regulatory system. Our results suggest that this expanded regulation reflects the 298 complexity of the rhizosphere, where root colonization likely requires coordinated activation of 299 adhesion together with traits involved in competition and plant interaction (Fig. 6B). An important 300 question is whether gac activation engages these programs proportionally or can bias them 301 toward distinct outputs. One possibility is that differences in the magnitude or duration of gac 302 signaling are encoded through the kinetics of the rsm sRNAs, allowing the gac system to tune 303 the balance between adhesion and competitive behaviors depending on environmental 304 conditions during root colonization. 305 Treatments with beneficial rhizobacteria perform poorly under field conditions in part 306 because competition from native microbiota limits effective root colonization (81, 82). This 307 ecological constraint means that even large inoculum doses or recurrent applications often 308 provide limited long-term enhancement of plant performance. Current efforts to enhance 309 colonization focus largely on improving delivery through strategies that position bacteria near 310 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 14 emerging roots and by isolating new strains that may be better suited for competing in the 311 rhizosphere. While these approaches can improve early establishment, they do not directly 312 address how bacteria transition from bulk soil to the root surface. By showing that flagellar 313 surface sensing stimulates the gac signaling system, our work suggests a new route to design 314 more persistent bioinoculants. Rather than relying on genetic engineering or reformulating 315 products, it could be possible to select for naturally occurring variants with reduced flagellar 316 synthesis under the hypothesis that reduced motility will enhance root attachment at high 317 inoculation titers. Combining such surface sensing-optimized strains with existing delivery 318 technologies like seed coatings could yield inoculants that deliver more consistent yield 319 enhancement in the field. 320 321

322 Growth conditions and genetic manipulations 323 Growth media were supplemented with 1.5% (w/v) agar, 300 µM diaminopimelic acid (DAP), 50 324 µg/mL gentamycin, 100 µg/mL carbenicillin, and 50 µg/mL kanamycin when necessary. E. coli 325 was cultured in LB medium at 37°C, P. protegens in King’s B medium (KBM) at 30°C, A. 326 rhizogenes in Nutrient broth at 30°C. Plasmids were introduced into P. protegens Pf-5 by 327 biparental conjugation with WM3064 as the donor strain. Gene deletions or insertions were 328 generated using a two-step approach with a sacB-based counterselection. Mutants were 329 complemented by genomic integration of the deleted gene under the arabinose promoter using 330 the Tn7 transposon (83). Mutants were fluorescently-tagged by inserting an mVenus expression 331 cassette into the CTX att site or glmS site (83, 84). Plasmids and strains used in this study are 332 listed in Table S2 and Table S3, respectively. 333 334 Establishment of tomato hairy root cultures 335 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 15 Tomato ( Solanum lycopersicum cv. Rio-Grande) seedlings were transformed with 336 Agrobacterium rhizogenes 15834 to generate hairy roots as previously described (85). To 337 generate a root biomass, the most vigorous roots, identified by their bright white color, high 338 density of root hairs, and length, were excised and transferred to 250 mL flasks containing 50 339 mL of 0.5× MS medium (pH 5.8) with 3% sucrose and supplemented with 200 µg/mL carbenicillin 340 and 200 µg/mL cefotaxime to suppress Agrobacterium growth. Flasks were incubated in the dark 341 at 25 °C with shaking at 75 rpm, and the medium was refreshed weekly. After 2-3 weeks, when 342 roots had filled the entire flask volume, they were cut and transferred into microplate wells for 343 attachment assays. 344 345 Tn-Himar mutant library construction and mapping 346 Construction and mapping of the Pf-5 transposon library was performed based on the method 347 developed by Wetmore et al. (34). Mid-log phase cells from cultures of the APA_752 barcoded 348 transposon pool grown in LB medium supplemented with kanamycin and DAP and mid-log 349 phase Pf-5 cells from cultures grown in KBM were collected by centrifugation, washed, mixed in 350 a 1:1 ratio, and spotted on a KBM agar plate containing DAP. After incubation for 24 h at 30°C, 351 cells were scraped from the plate, resuspended in KBM medium, and spread onto 20 150-mm 352 KBM plates containing kanamycin, followed by incubation at 30°C for 3 days. Colonies were 353 scraped into KBM medium and used to inoculate a KBM culture containing 25 μg/mL kanamycin. 354 The culture was grown for three doublings, glycerol was added to 15%, and 1 mL aliquots were 355 frozen at -80°C. Mapping genomic insertion positions was performed as described previously 356 (86). The features of the obtained library are: 57,304,857 total reads, 25,081,887 mapped reads 357 in pool, 784,810 mapped barcodes in pool, 161,378 distinct insertion sites, 5,816 coding genes 358 with central insertions, 53 median strains per hit protein, 81.7 mean strains per hit protein. 359 360 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 16 Adhesion profiling of barcoded Tn-Himar mutant libraries 361 Adhesion profiling was performed based on the method developed by Hershey et al. (86). Flasks 362 containing hairy roots were prepared by refreshing the medium 48 h prior to inoculation, allowing 363 root-secreted compounds to accumulate. A 750 μL inoculum of the barcoded transposon library 364 was added to a 250 mL flask containing 75 mL of medium and hairy roots, and flasks were 365 incubated for 24 h at 25°C with shaking at 75 rpm to allow bacterial attachment. Following 366 incubation, 750 μL of the culture was transferred into a new flask with hairy roots, while cells 367 from an additional 30 mL of culture were collected by centrifugation and stored at −20°C for 368 BarSeq analysis. This procedure was performed five times for a total of five sequential passaging 369 rounds. Parallel cultures were grown in flasks containing root-conditioned medium but no hairy 370 roots, serving as a reference condition. Genomic DNA was extracted from cell pellets and served 371 as templates for amplifying the barcodes in each sample using indexed primers (34). Amplicons 372 were purified using AMPure XP magnetic beads (Beckman Coulter) and pooled for multiplexed 373 sequencing. Next, 50-bp single-end reads were collected on an Illumina NovaSeq X Plus 374 sequencer. MultiCodes.pl, combineBarSeq.pl, and BarSeqR.pl were used to determine fitness 375 by comparing the log2 ratios of barcode counts in each sample over the counts from nonselective 376 growth without hairy roots. For each mutant, the pracma R package was used to calculate the 377 area under the curve of fitness scores across passages to rank mutants according to their hypo- 378 or hyper-adhesive phenotypes. 379 380 Crystal violet staining assay 381 Overnight cultures were resuspended in 0.5× MS medium (pH 5.8) supplemented with 3% 382 sucrose and 20 mM succinate, which was either freshly prepared or conditioned with hairy roots 383 for 24 h. Resuspensions were normalized to OD 600 0.002, and 450 L was added to each well 384 of a 48-well microtiter plate. Plates were incubated shaking for 24 h at 30C and 200 rpm. OD600 385 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 17 was measured to account for growth differences. Culture broth was discarded and the wells were 386 washed by dipping the microtiter plate in a bin with tap water. 500 L of 0.1% (w/v) crystal violet 387 solution was added to each well, and the plate was shaken for 30 min before the dye was 388 discarded and wells washed in the same manner. The retained dye was dissolved in 500 L of 389 absolute ethanol by shaking for 10 min. OD 575 was measured in each well and normalized to 390 OD600. 391 392 Measurement of c-di-GMP 393 Strains were transformed with the pConRef-2H12.D11 c-di-GMP reporter plasmid (39). 394 Overnight cultures were resuspended in 0.5× MS medium (pH 5.8) supplemented with 3% 395 sucrose and 20 mM succinate. MS media was either freshly prepared or conditioned with hairy 396 roots for 24 h. Cell suspensions were normalized to OD600 0.02, and 200 μL was dispensed into 397 wells of a 96-well microtiter plate. Plates were incubated shaking for 24 h at 30 °C and 200 rpm, 398 after which fluorescence was measured using a BioTek Synergy H1 microplate reader for 399 cdGreen2.1 (Excitation: 470 nm, emission: 515 nm) and normalized to the reference mScarlet-I 400 (Excitation: 575 nm, emission: 632 nm). 401 402 Measurement of gac activity reporters 403 Strains were grown overnight in 0.5× MS medium (pH 5.8) supplemented with 3% sucrose and 404 20 mM succinate. To evaluate the effects of the ΔfliF mutation and methyl cellulose on reporter 405 activity, overnight cultures were diluted 100-fold into 96-well plates containing 200 μL of the 406 indicated medium per well and incubated for 24 h. Plates were shaken when assessing the effect 407 of the ΔfliF mutation and incubated statically for methyl cellulose. Fluorescence was measured 408 using a BioTek Synergy H1 microplate reader (mVenus; Excitation: 500 nm; emission: 540 nm) 409 and normalized to OD 600. For microscopy, overnight cultures were diluted 10-fold, and 2 µL of 410 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 18 cells were placed on a 2% agarose pad to immobilize the cells. Cells were imaged immediately 411 to quantify reporter activity in planktonic cells and imaged again after 1 h to measure reporter 412 activity in surface-sensing cells. Microscopy was performed using a Nikon Ti-E inverted 413 microscope equipped with an Orca Fusion BT digital CMOS camera (Hamamatsu). 414 Fluorescence images were collected using a Prior Lumen 200 metal halide light source and a 415 YFP-specific filter set (Chroma) with identical exposure times and laser intensity. Image analysis 416 was performed with the Fiji software. 417 418 Root attachment assays 419 Hairy roots were excised into segments and placed in wells of a 48-well plate with 0.5× MS 420 medium (pH 5.8) supplemented with 3% sucrose and 20 mM succinate conditioned by roots for 421 24 h. Wells were inoculated with the indicated strains at OD 600 0.02. Plates were incubated at 422 25 °C with shaking at 75 rpm. To qualitatively assess bacterial root attachment, roots were 423 washed three times within the plate wells and imaged for fluorescence using the Invitrogen™ 424 iBright™ CL1500 Imaging System (Excitation: 455-485 nm, emission: 508-557 nm). To 425 quantitatively assess root attachment, roots were transferred to microcentrifuge tubes containing 426 1 mL of 0.5× MS medium. Root-associated bacteria were detached by vortexing for 1 min, and 427 the resulting suspensions were serially diluted and plated on LB agar for CFU enumeration, or 428 had their fluorescence measured using a BioTek Synergy H1 microplate reader (mVenus; 429 Excitation: 500 nm, emission: 540 nm). 430 431 Bacterial growth curve measurement 432 Overnight cultures were grown overnight in LB medium at 30 °C and adjusted to OD 600 0.02 in 433 indicated media. 200 µL aliquots were dispensed into 96-well plates, with five technical replicates 434 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 19 per strain. Plates were incubated at 30 °C with shaking in a BioTek Synergy H1 microplate 435 reader, and OD600 was recorded at 30-min intervals for the duration of the experiment. 436 437 Data availability: The sequence data used to map the barcoded Tn- Himar library and the 438 barcoded amplicon sequences collected after passaging in plant roots have been deposited in 439 the NCBI Sequence Read Archive (SRA) under the PRJNA1440512 project accession number. 440 Strains, plasmids, and details of their construction are available upon request. 441 442

This work was supported by a National Institutes of Health award 443 R35GM150652 to D.M.H, Beckman Young Investigator award to D.M.H., and startup funds from 444 the University of Wisconsin – Madison to D.M.H. This research was supported by BARD, the 445 United States - Israel Binational Agricultural Research and Development Fund, Vaadia-BARD 446 Postdoctoral Fellowship Award No. FI-638-2024. We thank Or Sharon for technical assistance 447 with preparing and uploading data to the NCBI SRA. 448 449

450 1. Sauer K, Stoodley P, Goeres DM, Hall-Stoodley L, Burmølle M, Stewart PS, Bjarnsholt T. 451 2022. The biofilm life cycle: expanding the conceptual model of biofilm formation. Nat Rev 452 Microbiol 20:608–620. 453 2. Flemming H-C, Wuertz S. 2019. Bacteria and archaea on Earth and their abundance in 454 biofilms. Nat Rev Microbiol 17:247–260. 455 3. Baxter KJ, Sas E, Clark KB, Walsh M, Pradeep N, Batool A, Naney C, Vargas Cruz MA, 456 Kennerdale N, Das K, Shi Z, Kelam A, Verma V, Simões MF, Neefs D, Ravichandran V, 457 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 20 Tirumalai MR, Barcenilla BB, Macori G, Gonzalez E, Sikes B, Karouia F, Brereton NJB. 458 2026. Biofilms: from the cradle of life to life support. npj Biofilms Microbiomes 12:11. 459 4. Grari O, Ezrari S, El Yandouzi I, Benaissa E, Ben Lahlou Y, Lahmer M, Saddari A, 460 Elouennass M, Maleb A. 2025. A comprehensive review on biofilm-associated infections: 461 Mechanisms, diagnostic challenges, and innovative therapeutic strategies. The Microbe 462 8:100436. 463 5. Liu Y, Shi A, Chen Y, Xu Z, Liu Y, Yao Y, Wang Y, Jia B. 2025. Beneficial 464 microorganisms: Regulating growth and defense for plant welfare. Plant Biotechnology 465 Journal 23:986–998. 466 6. Rumbaugh KP, Whiteley M. 2025. Towards improved biofilm models. Nat Rev Microbiol 467 23:57–66. 468 7. Torres M, Paszti S, Eberl L. 2024. Shedding light on bacteria–host interactions with the 469 aid of TnSeq approaches. mBio 15:e00390-24. 470 8. Pearce D, Bazin MJ, Lynch JM. 1995. The rhizosphere as a biofilm, p. 207–220. 471 Cambridge University Press, Cambridge. 472 9. Liu Y, Xu Z, Chen L, Xun W, Shu X, Chen Y, Sun X, Wang Z, Ren Y, Shen Q, Zhang R. 473 2024. Root colonization by beneficial rhizobacteria. FEMS Microbiol Rev 48:fuad066. 474 10. Bhattacharyya A, Mavrodi O, Bhowmik N, Weller D, Thomashow L, Mavrodi D. 2023. 475 Bacterial biofilms as an essential component of rhizosphere plant-microbe interactions. 476

Microbiol 53:3–48. 477 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 21 11. Robert CAM, Himmighofen P, McLaughlin S, Cofer TM, Khan SA, Siffert A, Sasse J. 478 2025. Environmental and biological drivers of root exudation. Annual Review of Plant 479 Biology 76:317–339. 480 12. Yang C-X, Chen S-J, Hong X-Y, Wang L-Z, Wu H-M, Tang Y-Y, Gao Y-Y, Hao G-F. 2025. 481 Plant exudates-driven microbiome recruitment and assembly facilitates plant health 482 management. FEMS Microbiol Rev 49:fuaf008. 483 13. Bhattacharyya A, Mavrodi O, Bhowmik N, Weller D, Thomashow L, Mavrodi D. 2023. 484 Bacterial biofilms as an essential component of rhizosphere plant-microbe interactions. 485

Microbiol 53:3–48. 486 14. Liu Z, Beskrovnaya P, Melnyk RA, Hossain SS, Khorasani S, O’Sullivan LR, Wiesmann 487 CL, Bush J, Richard JD, Haney CH. 2018. A genome-wide screen identifies genes in 488 rhizosphere-associated Pseudomonas required to evade plant defenses. mBio 489 9:10.1128/mbio.00433-18. 490 15. Torres M, Jiquel A, Jeanne E, Naquin D, Dessaux Y, Faure D. 2022. Agrobacterium 491 tumefaciens fitness genes involved in the colonization of plant tumors and roots. New 492 Phytologist 233:905–918. 493 16. do Amaral FP, Tuleski TR, Pankievicz VCS, Melnyk RA, Arkin AP, Griffitts J, Tadra-Sfeir 494 MZ, Maltempi de Souza E, Deutschbauer A, Monteiro RA, Stacey G. 2020. Diverse 495 bacterial genes modulate plant root association by beneficial bacteria. mBio 496 11:10.1128/mbio.03078-20. 497 17. Cole BJ, Feltcher ME, Waters RJ, Wetmore KM, Mucyn TS, Ryan EM, Wang G, Ul-Hasan 498 S, McDonald M, Yoshikuni Y, Malmstrom RR, Deutschbauer AM, Dangl JL, Visel A. 2017. 499 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 22 Genome-wide identification of bacterial plant colonization genes. PLOS Biology 500 15:e2002860. 501 18. Knights HE, Ramachandran VK, Jorrin B, Ledermann R, Parsons JD, Aroney STN, Poole 502 PS. 2024. Rhizobium determinants of rhizosphere persistence and root colonization. 503 ISME J 18:wrae072. 504 19. Torres M, Price MN, Khasanova A, Kosina SM, Zhalnina K, Northen TR, Deutschbauer 505 AM. 2025. Bacterial fitness for plant colonization is influenced by plant growth substrate. 506 New Phytologist 248:3168–3190. 507 20. Sivakumar R, Ranjani J, Vishnu US, Jayashree S, Lozano GL, Miles J, Broderick NA, 508 Guan C, Gunasekaran P, Handelsman J, Rajendhran J. 2019. Evaluation of INSeq to 509 identify genes essential for Pseudomonas aeruginosa PGPR2 corn root colonization. G3 510 Genes|Genomes|Genetics 9:651–661. 511 21. Ghaly TM, Fabian BK, Vick SHW, Foster C, Asher AJ, Hassan KA, Elbourne LDH, 512 Paulsen IT, Tetu SG. 2025. Genetic drivers of plant root colonisation by the biocontrol 513 agent Pseudomonas protegens Pf-5. Environmental Microbiology Reports 17:e70179. 514 22. Cheng X, Etalo DW, van de Mortel JE, Dekkers E, Nguyen L, Medema MH, Raaijmakers 515 JM. 2017. Genome-wide analysis of bacterial determinants of plant growth promotion and 516 induced systemic resistance by Pseudomonas fluorescens. Environmental Microbiology 517 19:4638–4656. 518 23. Pranav PS, Sivakumar R, Suvekbala V, Rajendhran J. 2024. Genome-wide identification 519 of root colonization fitness genes in plant growth promoting Pseudomonas asiatica 520 employing transposon-insertion sequencing. Ann Microbiol 74:40. 521 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 23 24. Taylor TB, Silby MW, Jackson RW. 2025. Pseudomonas fluorescens. Trends in 522 Microbiology 33:250–251. 523 25. Zboralski A, Filion M. 2020. Genetic factors involved in rhizosphere colonization by 524 phytobeneficial Pseudomonas spp. Computational and Structural Biotechnology Journal 525 18:3539–3554. 526 26. Krell T, Matilla MA. 2024. Pseudomonas aeruginosa. Trends in Microbiology 32:216–218. 527 27. Collins AJ, Smith TJ, Sondermann H, O’Toole GA. 2020. From input to output: The Lap/c-528 di-GMP biofilm regulatory circuit. Annual Review of Microbiology 74:607–631. 529 28. Loper JE, Kobayashi DY, Paulsen IT. 2007. The genomic sequence of Pseudomonas 530 fluorescens Pf-5: Insights into biological control. Phytopathology® 97:233–238. 531 29. Paulsen IT, Press CM, Ravel J, Kobayashi DY, Myers GSA, Mavrodi DV, DeBoy RT, 532 Seshadri R, Ren Q, Madupu R, Dodson RJ, Durkin AS, Brinkac LM, Daugherty SC, 533 Sullivan SA, Rosovitz MJ, Gwinn ML, Zhou L, Schneider DJ, Cartinhour SW, Nelson WC, 534 Weidman J, Watkins K, Tran K, Khouri H, Pierson EA, Pierson LS, Thomashow LS, Loper 535 JE. 2005. Complete genome sequence of the plant commensal Pseudomonas 536 fluorescens Pf-5. Nat Biotechnol 23:873–878. 537 30. Jing X, Cui Q, Li X, Yin J, Ravichandran V, Pan D, Fu J, Tu Q, Wang H, Bian X, Zhang Y. 538 2020. Engineering Pseudomonas protegens Pf-5 to improve its antifungal activity and 539 nitrogen fixation. Microbial Biotechnology 13:118–133. 540 31. Ghaly TM, Fabian BK, Vick SHW, Foster C, Asher AJ, Hassan KA, Elbourne LDH, 541 Paulsen IT, Tetu SG. 2025. Genetic drivers of plant root colonisation by the biocontrol 542 agent Pseudomonas protegens Pf-5. Environmental Microbiology Reports 17:e70179. 543 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 24 32. Zhao Q, Wang R, Song Y, Lu J, Zhou B, Song F, Zhang L, Huang Q, Gong J, Lei J, Dong 544 S, Gu Q, Borriss R, Gao X, Wu H. 2024. Pyoluteorin-deficient Pseudomonas protegens 545 improves cooperation with Bacillus velezensis, biofilm formation, co-colonizing, and 546 reshapes rhizosphere microbiome. npj Biofilms Microbiomes 10:145. 547 33. Kamilova F, Kravchenko LV, Shaposhnikov AI, Makarova N, Lugtenberg B. 2006. Effects 548 of the tomato pathogen Fusarium oxysporum f. sp. radicis-lycopersici and of the biocontrol 549 bacterium Pseudomonas fluorescens WCS365 on the composition of organic acids and 550 sugars in tomato root exudate. MPMI 19:1121–1126. 551 34. Wetmore KM, Price MN, Waters RJ, Lamson JS, He J, Hoover CA, Blow MJ, Bristow J, 552 Butland G, Arkin AP, Deutschbauer A. 2015. Rapid quantification of mutant fitness in 553 diverse bacteria by sequencing randomly bar-coded transposons. mBio 6:e00306-15–15. 554 35. Niehaus TD, Elbadawi-Sidhu M, Crécy-Lagard V de, Fiehn O, Hanson AD. 2017. 555 Discovery of a widespread prokaryotic 5-oxoprolinase that was hiding in plain sight. 556 Journal of Biological Chemistry 292:16360–16367. 557 36. Laventie B-J, Jenal U. 2020. Surface Sensing and adaptation in bacteria. Annual Review 558 of Microbiology 74:735–760. 559 37. Song H, Li Y, Wang Y. 2023. Two-component system GacS/GacA, a global response 560 regulator of bacterial physiological behaviors. Engineering Microbiology 3:100051. 561 38. Dufrêne YF, Persat A. 2020. Mechanomicrobiology: how bacteria sense and respond to 562 forces. Nat Rev Microbiol 18:227–240. 563 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 25 39. Kaczmarczyk A, Van Vliet S, Jakob RP, Teixeira RD, Scheidat I, Reinders A, Klotz A, 564 Maier T, Jenal U. 2024. A genetically encoded biosensor to monitor dynamic changes of 565 c-di-GMP with high temporal resolution. Nat Commun 15:3920. 566 40. Kay E, Dubuis C, Haas D. 2005. Three small RNAs jointly ensure secondary metabolism 567 and biocontrol in Pseudomonas fluorescens CHA0. Proceedings of the National Academy 568 of Sciences 102:17136–17141. 569 41. Zha D, Xu L, Zhang H, Yan Y. 2014. The two-component GacS-GacA system activates 570 lipA translation by RsmE but Not RsmA in Pseudomonas protegens Pf-5. Applied and 571 Environmental Microbiology 80:6627–6637. 572 42. Pijper A. 1947. Methylcellulose and bacterial motility. J Bacteriol 53:257–269. 573 43. Joller C, Waelchli J, Schlaepfer J, Schlaeppi K. 2026. Paralleled dynamics of Arabidopsis 574 root exudation and syncom assembly in a controlled environment. bioRxiv 575 https://doi.org/10.64898/2026.01.29.702624. 576 44. McLaughlin S, Zhalnina K, Kosina S, Northen TR, Sasse J. 2023. The core metabolome 577 and root exudation dynamics of three phylogenetically distinct plant species. Nat Commun 578 14:1649. 579 45. Lozano GL, Bravo JI, Garavito Diago MF, Park HB, Hurley A, Peterson SB, Stabb EV, 580 Crawford JM, Broderick NA, Handelsman J. 2019. Introducing THOR, a model 581 microbiome for genetic dissection of community behavior. mBio 10: pp.10-1128. 582 46. Cole BJ, Feltcher ME, Waters RJ, Wetmore KM, Mucyn TS, Ryan EM, Wang G, Ul-Hasan 583 S, McDonald M, Yoshikuni Y, Malmstrom RR, Deutschbauer AM, Dangl JL, Visel A. 2017. 584 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 26 Genome-wide identification of bacterial plant colonization genes. PLOS Biology 585 15:e2002860. 586 47. Wallner A, Busset N, Lachat J, Guigard L, King E, Rimbault I, Mergaert P, Béna G, Moulin 587 L. 2022. Differential genetic strategies of Burkholderia vietnamiensis and 588 Paraburkholderia kururiensis for root colonization of Oryza sativa subsp. japonica and O. 589 sativa subsp. indica, as revealed by transposon mutagenesis sequencing. Applied and 590 Environmental Microbiology 88:e00642-22. 591 48. López-Sánchez A, Leal-Morales A, Jiménez-Díaz L, Platero AI, Bardallo-Pérez J, Díaz-592 Romero A, Acemel RD, Illán JM, Jiménez-López J, Govantes F. 2016. Biofilm formation-593 defective mutants in Pseudomonas putida. FEMS Microbiol Lett 363:fnw127. 594 49. Zhang L, Shi Y, Wu Z, Tan G. 2018. Characterization of response regulator GacA 595 involved in phaseolotoxin production, hypersensitive response and cellular processes in 596 Pseudomonas syringae pv. actinidiae A18. Physiological and Molecular Plant Pathology 597 103:137–142. 598 50. Takeuchi K. 2018. GABA, A primary metabolite controlled by the Gac/Rsm regulatory 599 pathway, favors a planktonic over a biofilm lifestyle in Pseudomonas protegens CHA0. 600 MPMI 31:274–282. 601 51. Marutani M, Taguchi F, Ogawa Y, Hossain MdM, Inagaki Y, Toyoda K, Shiraishi T, 602 Ichinose Y. 2008. Gac two-component system in Pseudomonas syringae pv. tabaci is 603 required for virulence but not for hypersensitive reaction. Mol Genet Genomics 279:313–604 322. 605 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 27 52. Kidarsa TA, Shaffer BT, Goebel NC, Roberts DP, Buyer JS, Johnson A, Kobayashi DY, 606 Zabriskie TM, Paulsen I, Loper JE. 2013. Genes expressed by the biological control 607 bacterium Pseudomonas protegens Pf-5 on seed surfaces under the control of the global 608 regulators GacA and RpoS. Environmental Microbiology 15:716–735. 609 53. Lalaouna D, Fochesato S, Sanchez L, Schmitt-Kopplin P, Haas D, Heulin T, Achouak W. 610 2012. Phenotypic switching in Pseudomonas brassicacearum involves GacS- and GacA-611 dependent Rsm Small RNAs. Applied and Environmental Microbiology 78:1658–1665. 612 54. Martínez-Gil M, Ramos-González MI, Espinosa-Urgel M. 2014. Roles of cyclic di-GMP 613 and the Gac system in transcriptional control of the genes coding for the Pseudomonas 614 putida adhesins LapA and LapF. J Bacteriol 196:1484–1495. 615 55. Driscoll WW, Pepper JW, Pierson LS, Pierson EA. 2011. Spontaneous Gac Mutants of 616 Pseudomonas Biological Control Strains: Cheaters or Mutualists? Applied and 617 Environmental Microbiology 77:7227–7235. 618 56. Cheng X, de Bruijn I, van der Voort M, Loper JE, Raaijmakers JM. 2013. The Gac regulon 619 of Pseudomonas fluorescens SBW25. Environmental Microbiology Reports 5:608–619. 620 57. Luo Y, Srinivas A, Guidry C, Bull C, Haney CH, Hamilton C. 2025. GacA regulates 621 symbiosis and mediates lifestyle transitions in Pseudomonas. mSphere 10:e00277-25. 622 58. Kim JS, Kim YH, Park JY, Anderson AJ, Kim YC. 2014. The global regulator GacS 623 regulates biofilm formation in Pseudomonas chlororaphis O6 differently with carbon 624 source. Can J Microbiol 60:133–138. 625 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 28 59. Li J, Yang Y, Dubern J-F, Li H, Halliday N, Chernin L, Gao K, Cámara M, Liu X. 2015. 626 Regulation of GacA in Pseudomonas chlororaphis strains shows a niche specificity. PLOS 627 ONE 10:e0137553. 628 60. Natsch A, Keel C, Pfirter HA, Haas D, Défago G. 1994. Contribution of the global 629 regulator gene gacA to persistence and dissemination of Pseudomonas fluorescens 630 biocontrol strain CHA0 introduced into soil microcosms. Applied and Environmental 631 Microbiology 60:2553–2560. 632 61. Zuber S, Carruthers F, Keel C, Mattart A, Blumer C, Pessi G, Gigot-Bonnefoy C, 633 Schnider-Keel U, Heeb S, Reimmann C, Haas D. 2003. GacS sensor domains pertinent 634 to the regulation of exoproduct formation and to the biocontrol potential of Pseudomonas 635 fluorescens CHA0. MPMI 16:634–644. 636 62. Kim JS, Kim YH, Park JY, Anderson AJ, Kim YC. 2014. The global regulator GacS 637 regulates biofilm formation in Pseudomonas chlororaphis O6 differently with carbon 638 source. Can J Microbiol 60:133–138. 639 63. Song C, Kidarsa TA, van de Mortel JE, Loper JE, Raaijmakers JM. 2016. Living on the 640 edge: Emergence of spontaneous gac mutations in Pseudomonas protegens during 641 swarming motility. Environmental Microbiology 18:3453–3465. 642 64. Li E, Zhang H, Jiang H, Pieterse CMJ, Jousset A, Bakker PAHM, de Jonge R. 2021. 643 Experimental-evolution-driven identification of Arabidopsis rhizosphere competence 644 genes in Pseudomonas protegens. mBio 12:10.1128/mbio.00927-21. 645 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 29 65. Martínez-Granero F, Rivilla R, Martín M. 2006. Rhizosphere selection of highly motile 646 phenotypic variants of Pseudomonas fluorescens with enhanced competitive colonization 647 ability. Applied and Environmental Microbiology 72:3429–3434. 648 66. Chancey ST, Wood DW, Pierson EA, Pierson LS. 2002. Survival of GacS/GacA mutants 649 of the biological control bacterium Pseudomonas aureofaciens 30-84 in the wheat 650 rhizosphere. Applied and Environmental Microbiology 68:3308–3314. 651 67. Ishizawa H, Kuroda M, Inoue D, Ike M. 2022. Genome-wide identification of bacterial 652 colonization and fitness determinants on the floating macrophyte, duckweed. Commun 653 Biol 5:68. 654 68. Li J, Zhang Y, Jiang W, Zhang L-Q. 2025. Experimental evolution of plant rhizobacteria 655 reveals emerging adaptive mutations. mBio 16:e01023-25. 656 69. O’Banion BS, Jones P, Demetros AA, Kelley BR, Knoor LH, Wagner AS, Chen J-G, 657 Muchero W, Reynolds TB, Jacobson D, Lebeis SL. 2023. Plant myo-inositol transport 658 influences bacterial colonization phenotypes. Current Biology 33:3111-3124.e5. 659 70. Schultz D, Wolynes PG, Jacob EB, Onuchic JN. 2009. Deciding fate in adverse times: 660 Sporulation and competence in Bacillus subtilis. Proceedings of the National Academy of 661 Sciences 106:21027–21034. 662 71. O'Malley MR, Dearing HN, Zheng X, Kretschmer AN, Cho TH, Raivio TL, Parsek MR. 663 2026. The Pseudomonas aeruginosa Cpx system provides a cyclic-di-GMP independent 664 link between cell envelope stress and surface sensing. mBio 17.1: e02726-25. 665 72. Mao M, He L, Yan Q. 2025. An updated overview on the bacterial PhoP/PhoQ two-666 component signal transduction system. Front Cell Infect Microbiol 15:1509037. 667 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 30 73. Ali-Ahmad A, Fadel F, Sebban-Kreuzer C, Ba M, Pélissier GD, Bornet O, Guerlesquin F, 668 Bourne Y, Bordi C, Vincent F. 2017. Structural and functional insights into the periplasmic 669 detector domain of the GacS histidine kinase controlling biofilm formation in 670 Pseudomonas aeruginosa. Sci Rep 7:11262. 671 74. Liang F, Zhang B, Yang Q, Zhang Y, Zheng D, Zhang L, Yan Q, Wu X. 2020. Cyclic-di-672 GMP Regulates the Quorum-Sensing System and Biocontrol Activity of Pseudomonas 673 fluorescens 2P24 through the RsmA and RsmE Proteins. Applied and Environmental 674 Microbiology 86:e02016-20. 675 75. Zhang Y, Zhang B, Wu H, Wu X, Yan Q, Zhang L-Q. 2020. Pleiotropic effects of RsmA 676 and RsmE proteins in Pseudomonas fluorescens 2P24. BMC Microbiol 20:191. 677 76. Collins AJ, Pastora AB, Smith TJ, O’Toole GA. 2020. MapA, a Second Large RTX 678 adhesin conserved across the Pseudomonads, contributes to biofilm formation by 679 Pseudomonas fluorescens. J Bacteriol 202:10.1128/jb.00277-20. 680 77. Pastora AB, O’Toole GA. The regulator FleQ both transcriptionally and post-681 transcriptionally regulates the level of RTX adhesins of Pseudomonas fluorescens. J 682 Bacteriol 205:e00152-23. 683 78. Martínez-Granero F, Navazo A, Barahona E, Redondo-Nieto M, Rivilla R, Martín M. 2012. 684 The Gac-Rsm and SadB signal transduction pathways converge on AlgU to downregulate 685 motility in Pseudomonas fluorescens. PLOS ONE 7:e31765. 686 79. Brinkley DM, Bertolli SK, Gallagher LA, Tan Y, Silva MM de, Brockman A, Zhang D, 687 Peterson SB, Mougous JD. 2025. Pseudomonads coordinate innate defense against 688 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 31 viruses and bacteria with a single regulatory system. Cell Host & Microbe 33:1333-689 1346.e7. 690 80. Liu C, Shi R, Jensen MS, Zhu J, Liu J, Liu X, Sun D, Liu W. 2024. The global regulation of 691 c-di-GMP and cAMP in bacteria. mLife 3:42–56. 692 81. Azarbad H, Junker RR. 2024. Biological and experimental factors that define the 693 effectiveness of microbial inoculation on plant traits: a meta-analysis. ISME Commun 694 4:ycae122. 695 82. Lopes MJ dos S, Dias-Filho MB, Gurgel ESC. 2021. Successful plant growth-promoting 696 microbes: Inoculation methods and abiotic factors. Front Sustain Food Syst 5. 697 83. Gheorghita AA, Wolfram F, Whitfield GB, Jacobs HM, Pfoh R, Wong SSY, Guitor AK, 698 Goodyear MC, Berezuk AM, Khursigara CM, Parsek MR, Howell PL. 2022. The 699 Pseudomonas aeruginosa homeostasis enzyme AlgL clears the periplasmic space of 700 accumulated alginate during polymer biosynthesis. Journal of Biological Chemistry 298. 701 84. Wang Z, Xiong G, Lutz F. 1995. Site-specific integration of the phage ΦCTX genome into 702 the Pseudomonas aeruginosa chromosome: characterization of the functional integrase 703 gene located close to and upstream of attP. Molec Gen Genet 246:72–79. 704 85. Morcillo RJL, Zhao A, Tamayo-Navarrete MI, García-Garrido JM, Macho AP. 2020. 705 Tomato root transformation followed by inoculation with Ralstonia Solanacearum for 706 straightforward genetic analysis of bacterial wilt disease. JoVE 60302. 707 86. Hershey DM, Fiebig A, Crosson S. 2021. Flagellar perturbations activate adhesion 708 through two distinct pathways in Caulobacter crescentus. mBio 12:e03266-20. 709 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 32 Figures and Tables 710 711 Figure 1. Bacterial attachment to root surfaces 712 Roots release a mixture of exudates into the surrounding soil, generating a concentration 713 gradient that is highest near the root surface (illustrated as a pink gradient). Rhizosphere bacteria 714 actively move toward roots based on the exudate gradient, attach to the root surface via secreted 715 adhesins, and subsequently form a biofilm. 716 717 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 33 718 Figure 2. Establishing an in vitro root tissue culture system to study root colonization 719 (A) Seedlings grown under sterile tissue culture conditions were inoculated with Agrobacterium 720 at an incision at the base of the hypocotyl to generate hairy roots. (B) Excised root tissue 721 transferred to liquid medium and propagated in shaking culture, resulting in root biomass suitable 722 for downstream colonization assays. C and D, Root tissue placed in plate wells was inoculated 723 with a bacterial suspension of mVenus-tagged Pf-5 wild-type or Δ lapA. Root colonization was 724 evaluated by measuring fluorescence from roots (C) or quantified by recovering the root-725 attached bacterial fraction and measuring its emitted fluorescence (D). Bold lines within each 726 box represent median values. Top and bottom sides of the boxes represent the third and first 727 quartiles of the value distribution, respectively. Lines extending from the boxes denote the 728 extreme values within 1.5 times the interquartile range. Statistical significance was determined 729 by a pairwise t-test. Asterisks indicate P < 0.001. Data are pooled from three independent 730 experiments. 731 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 34 732 Figure 3. A genome-wide screen identifies genetic determinants for root attachment 733 (A) A pooled transposon mutant library carrying mutants impaired in root association (red) was 734 inoculated into flasks containing hairy roots. After incubation, the unattached (planktonic) fraction 735 was collected and serially passaged into fresh root-containing flasks for five rounds. The 736 resulting population is enriched for adhesion-defective mutants and sequenced. (B) Fitness 737 trajectories for the 15 most hypo- and 15 most hyper-adhesive mutants identified in the root 738 passaging Tn-Seq screen. Lines connect fitness scores for each gene across passages 0-5. 739 Positive fitness scores indicate mutants that increased in abundance in the unattached (media) 740 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 35 fraction, consistent with reduced root adhesion, whereas negative scores reflect depletion from 741 the media and therefore enhanced adhesion to roots. Genes highlighted in purple represent the 742 gacS and gacA genes, orange denotes flagellar assembly-associated genes, and gray indicates 743 genes outside these categories. (C) Root colonization phenotypes for in-frame deletion mutants 744 of genes identified in the screen. Root colonization was quantified by recovering the root-745 attached fraction of cells and measuring its emitted fluorescence. Bold lines within each box 746 represent median values. Top and bottom sides of the boxes represent the third and first 747 quartiles of the value distribution, respectively. Lines extending from the boxes denote the 748 extreme values within 1.5 times the interquartile range. Statistical significance was determined 749 by a pairwise t-test between wild-type and each strain. Asterisks indicate P < 0.001, “ns” no 750 significant difference. Data are pooled from three independent experiments. 751 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 36 752 Figure 4. Flagellar-dependent surface sensing activates the gac system 753 (A) Root colonization efficiencies of key mutants. Top and bottom sides of the boxes represent 754 the third and first quartiles of the value distribution, respectively. Lines extending from the boxes 755 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 37 denote the extreme values within 1.5 times the interquartile range. Data are pooled from three 756 independent replicates. (B) Intracellular c-di-GMP levels of strains from panel A. Reporter values 757 were normalized to wild type. (C) Overview of the gac signaling system in Pseudomonads. GacS 758 activates the response regulator GacA, which in turn directly induces expression of the small 759 RNAs rsmX, rsmY, and rsmZ. In D-G, the effect of surface sensing on gac activity was evaluated 760 via PrsmY expression by increasing medium viscosity with methyl cellulose (MC) (D), disrupting 761 flagellar assembly (fliF) (E), or comparing planktonic and surface-attached cells (F and G). In 762 F, scale bars represent 5 μm. In B, D, and E, a representative replicate of three individual 763 experiments is shown. Data are means ± SD. In all panels letters represent statistical 764 significance determined by one-way ANOVA and Tukey's posthoc test (P < 0.05). 765 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 38 766 Figure 5. The gac system integrates physical and chemical cues to control surface 767 attachment 768 (A) Crystal violet (CV)-based adhesion assay evaluating attachment to an abiotic surface. (B) 769 gac system activity quantified using a fluorescent reporter for rsmY evaluating the effect of 770 conditioning time of media by roots on gac system activity. (C–E) Integration of chemical and 771 mechanical inputs by the gac system. Indicated strains were analyzed under conditions 772 providing mechanical input alone (ΔfliF), chemical input alone (root-conditioned media), or both 773 inputs combined. (C) gac system activity quantified via the rsmY fluorescent reporter. (E) 774 Intracellular c-di-GMP levels. (F) CV-based adhesion assay. gac system activity levels, c-di-775 GMP levels, and adhesion were normalized to wild-type in unconditioned (Fresh) media. A 776 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 39 representative replicate of three independent experiments is shown. Data represent means ± 777 SD. In all panels, different letters indicate statistically significant differences determined by one-778 way ANOVA with Tukey’s post hoc test ( P < 0.05). In panel B, asterisks indicate statistically 779 significant differences determined by pairwise t-tests comparing each time point to the preceding 780 time point within the same strain. 781 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 40 782 Figure 6. Activating flagellar surface sensing enhances competitive root colonization 783 (A) Root colonization phenotypes in the presence of a synthetic bacterial community. Root 784 colonization was quantified by recovering the root-attached fraction of cells and measuring its 785 emitted fluorescence. Bold lines within each box represent median values. Top and bottom sides 786 of the boxes represent the third and first quartiles of the value distribution, respectively. Lines 787 extending from the boxes denote the extreme values within 1.5 times the interquartile range. 788 Data are pooled from three independent replicates. (B) The gac system additively integrates 789 physical and chemical cues to coordinately enhance adhesion and reduce motility through c-di-790 GMP, while simultaneously activating competitive traits, together enabling effective root 791 attachment. 792 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 41 Table 1: Top 15 protein-encoding hits enriched in the non-root-associated fraction 793 (reduced attachment). 794 Gene Locus ID Gene annotation Area under curve gacS PFL_4451 Sensor protein GacS 19.20 gacA PFL_3563 Response regulator GacA 18.13 PFL_2828 LysR family transcriptional regulator 8.27 PFL_4192 ABC transporter substrate-binding protein 7.27 adrA PFL_4532 Diguanylate cyclase 6.58 PFL_0502 DeoR family transcriptional regulator 6.58 pxpB PFL_1514 Allophanate hydrolase subunit 1 5.84 PFL_1588 AraC family transcriptional regulator 5.61 pxpA PFL_1513 LamB/YcsF family protein 5.48 pxpC PFL_1515 Allophanate hydrolase subunit 2 5.35 PFL_5687 Acyl-CoA dehydrogenase 5.13 PFL_1740 TonB-dependent outermembrane receptor 5.13 PFL_5779 response regulator/sensory box/GGDEF domain/EAL domain-containing protein 5.07 pycA PFL_6157 Pyruvate carboxylase subunit B 4.79 PFL_2692 Hypothetical protein 4.61 795 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint 42 Table 2: Top 15 protein encoding hits depleted from the non-root-associated fraction 796 (enhanced attachment). 797 Gene Locus ID Gene annotation Area under curve fleE PFL_1637 Flagellar hook-basal body complex protein FliE 11.63 cheA PFL_1670 Chemotaxis protein CheA 10.58 PFL_2045 Amino acid aminotransferase 9.93 fliF PFL_1638 Flagellar basal-body MS-ring/collar protein FliF 9.81 flgF PFL_1613 Flagellar basal-body rod protein FlgF 9.74 fliI PFL_1641 Flagellar protein export ATPase FliI 9.64 fliG PFL_1639 Flagellar motor switch protein FliG 9.53 fliA PFL_1667 RNA polymerase sigma factor FliA 9.43 flgE PFL_4477 Flagellar hook protein FlgE 9.43 fliK PFL_1646 Flagellar hook-length control protein FliK 9.42 flgG PFL_1614 Flagellar basal-body rod protein FlgG 9.39 flgC PFL_4479 Flagellar basal body rod protein FlgC 9.16 cheY PFL_1668 Chemotaxis response regulator CheY 9.13 fleR PFL_1636 Sigma-54 dependent transcriptional regulator 9.11 PFL_1752 HAMP domain-containing sensor histidine kinase 8.85 798 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted April 12, 2026. ; https://doi.org/10.64898/2026.04.10.717875doi: bioRxiv preprint