Type IV pilus length determines virulence by regulating a hidden subpopulation of non-contributing filaments

7 Many clinically important bacterial pathogens, including Pseudomonas, Vibrio, Neisseria, 8 and Acinetobacter species, employ dynamic extracellular appendages called type IV pili 9 (T4P) to facilitate virulence through cyclical extension and retraction of pilus filaments. T o 10 dissect how T4P dynamics govern pathogenesis, we engineered a genetic system to 11 precisely tune pilus length across a continuum. We demonstrate that pilus length critically 12 determines four major T4P -dependent virulence traits in Pseudomonas aeruginosa 13 (motility, surface sensing, biofilm formation, and phage infection) and reveal a hidden 14 subpopulation of pili that are unable to interact with environmental substrates or host cells, 15 rendering them non- contributing to any T4P -mediated function. Integrating molecular 16 dynamics simulations, we show that low inner -membrane abundance of the major pilin 17 forces the extension mechanism into transient idle states, restricting both velocity and 18 final length. Molecularly, t his finding reveals how two key biophysical parameters , pilin 19 abundance and diffusion, impose a fundamental physical constraint on T4P assembly , 20 and that regulating pilin abundance presents a strong lever over regulating pilus count for 21 controlling the amount of functionally contributing filaments . Contrary to the prevailing 22 view that retraction force generation primarily dictates T4P -mediated behaviors, our 23

establish extension dynamics as the overlooked bottleneck constraining all 24 retraction-enabled virulence traits , with population heterogeneity in length enabling 25 adaptive bet-hedging. 26 27 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint

28 Type IV pili (T4P) are dynamic extracellular filaments that enable bacteria to interact with 29 their environment and host tissues. These fibers assemble from a pool of major pilin 30 subunit at a multiprotein envelope-spanning machine that allows extension from the inner 31 membrane into the extracellular space [1-5]. Through repeated cycles of extension and 32 retraction, T4P power key virulence behaviors, including twitching motility for surface 33 exploration and dispersal [1, 6, 7], biofilm formation that promotes colony protection 34 against host defenses and antibiotics [8-12], DNA uptake for horizontal gene transfer of 35 resistance and virulence trait s [13-16], and surface sensing to coordinate pathogenic 36 responses [17-20]. Consequently, T4P-equipped pathogens are notoriously difficult to 37 treat, as their adaptability drives rapid colonization and systemic infection. 38 T4P-mediated virulence traits are widespread across Gram -negative pathogens 39 (Pseudomonas, Vibrio, Neisseria, Acinetobacter, and others) [17, 19, 21 -28], but 40 Pseudomonas aeruginosa stands out as a unique and powerful model system: it robustly 41 employs four major T4P-dependent behaviors (twitching motility, surface sensing, biofilm 42 formation, and phage infection), making it ideal for dissecting how T4P dynamics control 43 pathogenesis [9, 19, 29-36]. 44 In P. aeruginosa, the major pilin subunit PilA is polymerized to extend the pilus fiber and 45 depolymerized during retraction. PilA expression is regulated by the two- component 46 system PilSR. When PilA levels are low, the sensor kinase PilS auto-phosphorylates and 47 activates the response regulator PilR, which drives pilA transcription in a rpoN-dependent 48 manner [36-38]. This feedback is thought to maintain a constant inner-membrane pool of 49 PilA for pilus assembly [39]. However, the functional consequences of changes in PilA 50 abundance and how they affect T4P dynamics and virulence remain poorly understood. 51 The T4P assembly complex includes the PilMNOP alignment subcomplex linking the 52 cytoplasmic platform PilC to the outer -membrane secretin PilQ [40-42]. Hexameric 53 ATPases PilB and PilT engage PilC to drive extension and retraction, respectively [41-54 45]. ATP hydrolysis by these motors induces conformational changes in PilC that 55 incorporate or remove PilA subunits from the inner membrane [2, 4, 46]. Recent studies 56 suggest that the switch between extension and retraction is primarily driven by stochastic 57 and mutually exclusive binding of PilB and PilT to PilC [5, 47, 48] , however, 58 mechanosensitive effects triggered by surface contact may also contribute to modulating 59 this switch [19, 33, 49, 50]. 60 Retraction generates a large range of forces from 100 pN per pilus or nN in 61 bundles for different species, making T4P one of the most powerful known molecular 62 motors [13, 17, 18, 51-57]. Optical tweezers studies have revealed detailed insights into 63 the molecualr mechanisms driving retraction, including bimodal velocities, force-64 dependenta switching to extension, and two-dimensional tug-of-war in twitching [58, 59]. 65 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint However, the flexibility of T4P fibers limits optical tweezers for extension studies, leaving 66 the molecular processes governing pilus extension far less explored. 67 T4P research has primarily emphasized retraction as this is the force- generating step, 68 with extension often viewed as a prerequisite rather than a regulatory bottleneck. Yet 69 retraction depends on prior extension, which determines pilus length and the fraction of 70 pili that is available for functional interactions. Our incomplete understanding of extension 71 mechanisms thus limits insight into T4P -mediated virulence. Here, using P . aeruginosa 72 as a model, we developed a genetic tool to precisely tune pilus length while visualizing 73 dynamics in real time. This approach reveals that pilus length critically governs virulence 74 traits, uncovers a hidden subpopulation of non- contributing short pili, and identifies 75 diffusion-limited PilA supply as the biophysical constraint that biases extension and 76 makes length regulation more efficient than count regulation for controlling functional 77 output. 78 79

80 Clonal populations display a broad heterogeneity of pilus activity and PilA levels 81 To investigate the molecular mechanisms underlying pilus extension, we quantified pilus 82 number and length in individual cells using cysteine-maleimide-based fluorescent labeling 83 of pili in a pilA -A86C background [18, 48, 60]. Consistent with prior studies, pilus count 84 and length distributions were exponential, with a strong bias toward cells producing few 85 and short pili [13, 48] (Figure 1A,B). A typical cell extended 1 –5 pili within a 30 -second 86 observation window, with lengths ranging from 0.2 to 1.5 µm, although some cells 87 produced up to 13 pili reaching lengths of up to 4 µm. 88 We also observed marked variation in cell body fluorescence: some appeared dark while 89 others were significantly brighter (Figure 1C). As this labeling targets the major pilin PilA, 90 cell body brightness reflects intracellular PilA levels. Quantification revealed an 91 exponential distribution of brightness from ~2 to 20 arbitrary units (a.u.), with most cells 92 at 3 –6 a.u. ( Figure 1D). To rule out non -specific membrane labeling artifacts, we 93 confirmed similar heterogeneity in single cells using a transcriptional reporter PpilA::yfp, 94 showing pilA promoter activity spanning roughly one order of magnitude ( Figure 1E,F). 95 Together, these data demonstrate substantial cell-to-cell heterogeneity in PilA abundance 96 within the population. 97 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 98 Figure 1. Pilus dynamics and major pilin expression display a 10-fold difference between individual 99 cells in a clonal population. A) Total count of individual pili per cell, quantified from 30- second live-cell 100 movies. B) The maximum extension lengths of individual pili within the same 30-second live-cell movies as 101 in (A). C) Representative fluorescence still image of cells with Alexa488 labeled PilA-A86C cysteine knock-102 in mutant. D) Brightness of individual cell bodies with A lexa488 labeled PilA-A86C. E) Representative 103 fluorescence still image of pilA transcription using a PpilA::yfp reporter in single cells. F) Brightness of 104 PpilA::yfp reporter in single cells. Box plots: Boxes represent median and interquartile range (25th –75th 105 percentiles) of N = 150 cells/pili from three independent biological replicates. Images: White arrows: bright 106 cell bodies, indicating high PilA protein levels (C) or gene expression (E). Green arrows: dim cell bodies, 107 indicating low PilA protein levels (C) or gene expression (E). Scale bars: 2 µm. 108 109 Abundance of the major pilin PilA tunes pilus dynamics 110 The observed cell-to-cell variation in PilA abundance was unexpected, as PilA expression 111 is thought to be tightly regulated by the PilSR two- component system, which provides 112 WT 0 5 10 15 20 PPilA::yfp TranscriptionalActivity (a.u.) WT 0 5 10 15 20PilA-A86C Cell Brightness (a.u.) WT 0 5 10 15# of Pili per Cell WT 0 1 2 3 4 Pilus Length (μm) A B C D E F was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint feedback between protein levels and gene expression [37, 38, 61]. To directly test how 113 PilA abundance influences pilus extension, we constructed an arabinose-inducible allele 114 (Pbad::pilA-A86C) integrated into the chromosome in a pilA background. Western blotting 115 confirmed a tunable range that recapitulates the natural PilA heterogeneity observed in 116 individual cells: no inducer (0% arabinose) yielded ~10% of wild- type (WT) PilA levels, 117 while concentrations above 0.4 –0.5% plateaued at ~130% of WT . Induction of 0.3% 118 arabinose closely matched WT levels (Figure 2A, Supplementary Figure 1). 119 We next examined the impact of PilA levels on pilus dynamics. The fraction of piliated 120 cells increased with PilA abundance ( Figure 2B): at 0% induction, only ~5% of cells 121 produced visible pili, rising steadily to WT levels at 0.3– 0.4% arabinose. Similarly, the 122 number of pili per cell in a 30-second window scaled with induction (Figure 2C): uninduced 123 cells rarely produced pili (mostly 0, occasionally 1– 2), while higher PilA levels restored 124 distributions similar to WT . To better understand the mechanism that correlates pilin 125 abundance to pilus number, we specifically analyzed the lengths of individual pili (Figure 126 2D). At 0% induction, observable pili were very short (<200 nm). The median lengths 127 increased progressively with induction and reached WT values at 0.4% arabinose. These 128

establish that the abundance of PilA protein in the cell is a key determinant of pilus 129 dynamics. 130 131 Pilin abundance sets the maximum length of pilus fibers but not the count of pili 132 The sharp decrease in pilus length at low PilA levels prompted us to ask whether the 133 apparent reduction in piliated cells under low induction was a true biological effect or an 134 artifact of fluorescence microscopy resolution (where pili <200 nm would be 135 undetectable). To resolve this at the molecular level, we employed computer simulations 136 of pilus extension dynamics. Following our previous modeling framework [17, 48], we built 137 a simulation that combines Brownian dynamics for the diffusion of PilA monomers in the 138 inner membrane with a stochastic model of the PilB extension motor. PilB binds to the 139 pilus base for random durations following an exponential distribution (median ~2– 4 140 seconds) [48], during which it can drive extension; monomer addition is then rate-limited 141 by the ATPase's cycle time (~3 ms per PilA subunit under saturating conditions). In 142 discrete 1 ms time steps, the model evaluates whether PilB is primed for a cycle, and 143 whether a diffusing PilA monomer is close enough to the extension complex to be 144 incorporated into the growing fiber. If both conditions are met, a monomer is added; 145 otherwise, extension pauses for that step. 146 We then varied PilA membrane concentrations from 10% to 130% of WT levels and 147 performed simulations for 10,000 individual pili per condition. Maximum and median pilus 148 lengths increased with PilA availability, closely matching our experimental data ( Figure 149 2E). At 10% PilA, the median length fell below 100 nm and medians exceeded 200 nm 150 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint only above ~30% PilA, the practical detection threshold in our microscopy setup. Based 151 on this observation, we calculated the ratio of pili that were longer than 200 nm for each 152 concentration of PilA (Figure 2F). As expected, the fraction of simulated pili longer than 153 200 nm rose from <10% at the lowest PilA concentration to ~80% at WT levels. 154 These results explain our experimental pattern: a t low PilA levels, the number of pili 155 produced per cell does not decrease, the underlying count remains similar (governed by 156 the availability of T4P complexes). However, a larger proportion of pili are too short to be 157 visible by fluorescence microscopy. Thus, PilA abundance primarily controls pilus length 158 (setting maximum and median values), while the number of pili per cell is largely 159 independent of PilA levels. Consequently, a significant fraction of pili at low PilA levels 160 remain too short to be observed, revealing a hidden subpopulation of filaments within the 161 population. This also shows that controlling pilin abundance provides a reliable tool to 162 tune median pilus length across a population without altering pilus count. 163 164 0 5 10 15 # of Pili per Cell ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ns ns ✱✱ 0 1 2 3 Pilus Length (μm) ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱✱C D E F ns 0 25 50 75 100 125 % of Pili Forming Cells ✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱ BA 100 50 0 WT 0.0 0.1 0.2 0.3 0.4 0.5 Arabinose Concentration(%) 150 200 PilA Protein Levels (% of WT) ✱ ✱✱✱✱ ✱ ns ns ✱ PilA Levels (% WT) Fraction of Pili >200nm (%) 0 20 40 60 80 100 WT 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration (%) WT 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration (%) WT 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration (%) Length of Pili (Logμm) 0.005 0.01 0.05 0.1 0.5 1 1.5 PilA Levels (% WT) was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Figure 2. PilA protein levels determine the activity and dynamics of pili in individual cells. A) PilA 165 protein titration using Pbad::pilA in a pilA background determined by Western blotting using a polyclonal PilA 166 antibody. Bars represent mean ± SEM of three biological replicates (indicated as black dots). B) Fraction 167 of cells forming at least one pilus in a 30 second window as a function of PilA titration (same strain as in 168 panel A). Bars represent mean ± SD of N = 12 technical replicates from three biological replicates. C) Total 169 number of individual pili in a 30 s window in single cells as a function of PilA titration (same strain as in 170 panel A). Boxes show median and 25th–75th percentiles of N = 150 cells from three independent biological 171 replicates. D) Maximum pilus extension lengths for individual pili as a function of PilA titration (same strain 172 as in panel A). E) Pilus lengths for 10,000 simulations of individual pilus extension as a function of PilA 173 levels in the inner membrane. Violin plots show median (horizontal line) and 5% to 95% perc entiles. F) 174 Fraction of simulated pili that reached a maximum length of >200 nm from the 10,000 simulated pili in E as 175 a function of PilA levels. A) Statistical tests were performed using unpaired t -test. B) – D) Statistical tests 176 were performed using one-way ANOVA. Significance: (ns) not significant, P > 0.05; (*) P < 0.05 ; (**) P < 177 0.01; (***) P < 0.001; (****) P < 0.0001. 178 Low pilin abundance forces extension into transient idle states that limit extension 179 velocity 180 Motivated by our finding that PilA induction can reliably set the median pilus length, we 181 sought to determine how pilin abundance modulates length at the molecular level. Pilus 182 length is the product of extension velocity and the total duration of extension events. We 183 therefore examined whether pilin levels influence one or both parameters. 184 Analysis of extension duration across induction conditions showed relatively little variation 185 in the median time pili spent extending between 0.2% and 0.4% arabinose compared to 186 WT levels (Figure 3A). Only at 0.1% induction did we observe a moderate reduction of 187 30–40%, with median extension times dropping to approximately 2 seconds. In contrast, 188 extension velocity of individual pili increased steadily and significantly (by 77%) from 0.2% 189 to 0.4% arabinose, with minimal change between 0.1% and 0.2% ( Figure 3B). This 190 velocity increase closely matches the 88% rise in median pilus length over the same 191 range. These trends indicate that pilin abundance primarily affects extension velocity 192 rather than the total duration of motor engagement (extension time) . Our biophysical 193 simulations recapitulated this pattern, showing a similar steady rise in extension velocity 194 with increasing PilA levels that aligned well with the experimental measurements (Figure 195 3C). Importantly, retraction velocity remained constant across all conditions, confirming 196 that monomer abundance specifically influences the extension mechanism and does not 197 affect retraction (Figure 3D). 198 To uncover the molecular basis for the velocity dependence on PilA, we examined the 199 simulation trajectories in greater detail ( Figure 3E). When focusing on a short 250 ms 200 window of extension at 100% PilA levels, the time between consecutive monomer 201 incorporations was dominated by the ATPase cycle time (~3 ms, corresponding to ~300 202 nm/s velocity in WT, assuming 1 nm per monomer), as expected. However, occasional 203 longer intervals between incorporations were present. At 20% PilA, these delays became 204 much more pronounced, with clear extended gaps between monomer additions. 205 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Histograms of the dwell times between individual incorporation events confirmed this shift 206 (Figure 3F). At 100% PilA, roughly half of the intervals were limited by the ATPase cycle 207 time, and no pauses exceeded 20 ms. At 20% PilA, only about 10% of events matched 208 the ATPase cycle time, while pauses showed a broad distribution extending up to ~80 ms. 209 These results demonstrate that extension velocity is ultimately limited by the local 210 availability of PilA monomers in the inner membrane. When pilin levels are low, monomers 211 are not always present at the extension complex precisely when the ATPase is ready for 212 the next cycle. This forces the extension machinery into transient idle states, reducing the 213 effective incorporation rate and thereby constraining overall velocity. 214 215 Figure 3. Changes in pilin levels modulate the extension velocity but not extension duration. A) Total 216 duration individual pili spent extending until maximum length was reached as a function of PilA titration 217 0.0 0.5 1.0 1.5 Retraction Velocity (μm/Sec.) ns ✱✱✱✱ ns ns ns 0.0 0.2 0.4 0.6 Extension Velocity (μm/Sec.) ✱✱✱✱ ns ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ Extension Time (Sec.) 0 2 4 6 8 10 ✱✱✱✱ ✱✱✱✱ ✱ ns ns D A B C 0.01 0.1 1 PilA Levels (% WT) Extension Velocity (Log µm/Sec.) 0 250 50 100 150 200 Total Time Extended (ms) 0 10 20 30 40 50 PilA Protein Incorporated (#) 20% PilA 100% PilA 0 10 20 30 40 50 0 10 20 30 40 50 60 70 80 Pause Between Monomers (ms) Fraction of Events (%) 20% PilA 100% PilA FE WT 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration (%) WT 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration (%) WT 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration (%) was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint (Pbad::pilA in a pilA background). B) Extension velocity of individual pili as a function PilA titration (same 218 strain as in panel A). C) Extension velocity for 10,000 simulated pilus extension events as a function of PilA 219 levels in the inner membrane. Violin plots show the median (horizontal line) and 5% to 95% percentiles. D) 220 Retraction velocity of individual pili as a function of PilA titration (same strain as in panel A). E) Number of 221 PilA proteins incorporated into the pilus fiber for 20% (purple) and 100% (gray) levels of total PilA from 222 biophysical simulations. F) Percentage of events as a function of dwell times between individual PilA 223 incorporations for 20% (purple) and 100% (gray) PilA levels. A,B,D) Boxes show median and 25th– 75th 224 percentiles of N = 150 pili from three independent biological replicates . Statistical tests were performed 225 using one-way ANOVA: (ns) not significant, P > 0.05; (*) P < 0.05; (**) P < 0.01; (***) P < 0.001; (****) P < 226 0.0001. 227 228 PilSR Regulates Pilus Dynamics Through Modulation of Pilin Levels 229 Our finding that pilin monomer abundance limits pilus extension velocity suggests that the 230 PilSR two -component system regulates T4P dynamics by controlling PilA levels. To 231 determine whether PilSR exerts any additional, direct effects on pilus dynamics 232 independent of PilA abundance, we repeated the inducible PilA titration experiment in a 233 pilA pilSR background and quantified the same key parameters: number of pili per cell 234 (Figure 4A), the length of individual pili ( Figure 4B), the extension duration ( Figure 4C), 235 and the extension velocity ( Figure 4D). Across all four parameters, the trends with 236 increasing arabinose induction were nearly identical to those observed in the pilA 237

alone. The number of pili per cell and median pilus lengths increased 238 progressively with induction, reaching WT levels at ~0.3– 0.4% arabinose. Extension 239 duration showed a similar plateau between 0.2% and 0.4% arabinose, with a comparable 240 30–40% reduction at 0.1% induction. Extension velocity also increased with induction, 241 though the pattern differed slightly: unlike the pilA background (where velocity was flat 242 between 0.1% and 0.2% before rising), velocity in the pilA pilSR background increased 243 steadily from 0.1% to 0.4% arabinose. 244 These highly similar behaviors indicate that deleting PilSR does not substantially alter the 245 relationship between PilA abundance and pilus dynamics. Collectively, the data 246 demonstrate that PilSR modulates T4P dynamics primarily through control of pilin levels, 247 with no strong evidence for additional direct mechanisms acting on extension or 248 retraction. 249 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 250 Figure 4 Pilus dynamics are not affected by PilSR directly. A) Total number of individual pili in a 30 s 251 window in single cells as a function of PilA titration (Pbad::pilA in a pilApilSR background). B) Maximum pilus 252 extension lengths for individual pili as a function of PilA titration (same strain as in panel( A). C) Total 253 duration individual pili spent extending until maximum length was reached as a function PilA titration (same 254 strain as in panel (A). D) Extension velocity of individual pili as a function PilA titration (same strain as in 255 panel (A). A - D) Boxes show median and 25th–75th percentiles of N = 150 cells (A) or pili (B-D) from three 256 independent biological replicates . Statistical tests were performed using one-way ANOVA : (ns) not 257 significant, P > 0.05; (*) P < 0.05; (**) P < 0.01; (***) P < 0.001; (****) P < 0.0001. 258 259 Pilus length limits TFP-mediated virulence behaviors 260 Our observation that a substantial fraction of pili are too short to be detected by 261 fluorescence microscopy suggests that they might also be too short for functional 262 contribution, for example to different virulence traits. This prompted us to investigate how 263 pilus length influences T4P -dependent virulence traits in P. aeruginosa. To test this, we 264 used the inducible PilA system to titrate pilus length and assessed twitching motility, 265 surface sensing, phage infection, and biofilm formation across a range of arabinose 266 concentrations. 267 Twitching Motility 268 Twitching motility requires pili to make physical contact with a surface and retract to pull 269 the cell forward. In a standard stab- agar twitch plate assay, twitching zone diameter 270 increased gradually with arabinose induction, mirroring the trends in pilus dynamics 271 (Figure 5A). Wild-type twitching levels were reached only at 0.5% arabinose, higher than 272 # of Pili per Cell 0 5 10 15 WT 0.0 0.1 0.2 0.3 0.4 % Arabinose Concentration ✱✱✱✱ ✱✱✱✱ ✱✱ ns ns 0 1 2 3 WT 0.0 0.1 0.2 0.3 0.4 % Arabinose Concentration Pilus Length (μm) ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ns WT 0.0 0.1 0.2 0.3 0.4 % Arabinose Concentration 0 2 4 6 8 10 Extension Time (Sec.) ✱✱✱✱ ✱✱✱✱ ✱ ns ns 0.0 0.2 0.4 0.6 WT 0.0 0.1 0.2 0.3 0.4 % Arabinose Concentration Extension Velocity (μm/Sec.) ✱✱✱✱ ns ✱✱ ✱✱✱✱ ns A B C D was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint the 0.3–0.4% needed for wild-type PilA protein levels. This delay is consistent with pili at 273 0.4% being approximately 15% shorter than in WT cells. Further increasing arabinose to 274 0.6% produced no additional gain in twitching. In the pilSR background, twitching 275 increased more rapidly with induction and plateaued at 0.3% arabinose, with maximum 276 twitching being 20% higher than in the WT background ( Figure 5B). These results 277 demonstrate that pilus length directly limits twitching capacity: shorter pili reduce the 278 efficiency of surface contact and retraction. The enhanced twitching in the pilSR strain 279 further suggests that PilSR negatively regulates twitching motility through targets outside 280 the T4P system, independent of its effect on PilA levels. 281 Surface Sensing 282 Similar to twitching, surface sensing in P. aeruginosa relies on pili detecting mechanical 283 cues from substrate contact, triggering cAMP production and Vfr -dependent gene 284 expression via the Pil-Chp pathway [19, 20, 33, 36, 62, 63]. We monitored this response 285 using the PaQa transcriptional fluorescent reporter for Vfr activity on 1.5% agarose pads 286 over a 7-hour time course. WT cells showed a progressive increase in Vfr activity, peaking 287 at approximately four -fold upregulation relative to liquid- grown cells after five hours 288 (Figure 5C). At 0% arabinose, the response was completely abolished. Increasing 289 induction gradually restored the response, with amplitudes rising from 1.5-fold at 0.2% to 290 2.5-fold at 0.4% arabinose. Similar to twitching, WT PilA protein levels (0.3–0.4%) did not 291 fully restore WT sensing; even at 0.6% induction, the response reached only 3-fold. In the 292 pilSR background, surface sensing was abolished across all inducer concentrations 293 (Figure 5D). These findings indicate that both pilus length and the ability to modulate PilA 294 levels via PilSR are essential for efficient surface sensing through the Pil-Chp pathway. 295 Phage Infection 296 Unlike twitching and surface sensing, phage infection does not require pili to contact a 297 solid substrate. Instead, m any phages bind along the pilus fiber and use retraction to 298 reach the surface of the cell. We tested infection efficiency using phage JBD68 (which 299 targets the tip protein FimU) in a liquid growth assay, monitoring optical density (OD) over 300 12 hours after infection (Supplementary Figure 2). Infected WT cultures showed an initial 301 rise in OD followed by a sharp decline due to lysis, with maximum OD dropping to ~0.1 302 compared to uninfected controls (~1.4) or pilA (~1.2) (Figure 5E, Supplementary Figure 303 2A,B). Increasing PilA induction caused a progressive decrease in maximum OD, 304 indicating more efficient phage infection with longer pili. The pattern was similar in the 305 pilSR background, with no significant difference across conditions ( Figure 5F, 306 Supplementary Figure 2C,D). These results show that pilus length enhances phage 307 infection efficiency, likely by increasing the probability of productive tip binding and 308 retraction, but PilSR does not appear to play a major role in this process. 309 Biofilm Production 310 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Lastly, T4P contribute to biofilm formation through adhesion and microcolony spreading, 311 however, the mechanisms of how T4P dynamics contribute to more robust biofilms 312 remains largely unknown [11, 32, 62, 64]. We quantified the formation of biofilm biomass 313 on peg lids in a 96- well microtiter format after 18 hours of growth, using crystal violet 314 staining and acetic acid solubilization of bound dye (Figure 5G). The pilA control produced 315 ~40% less biofilm mass than WT and gradual PilA induction restored biofilm biomass to 316 WT levels by 0.4% arabinose ( Figure 5G). The pilSR background showed a similar 317 restoration pattern (Figure 5H). These data indicate that robust biofilm formation depends 318 on pilus length, with shorter pili limiting overall biomass accumulation. 319 320 5 4 3 2 1 0 VFR Activity (YFP/mKate2, a.u.) 100 200 300 400 500 Time Elapsed (min.) WT% 0.6% 0.4% 0.2% 0.0% 0 0 50 100 150Twitching Area (% of WT) A C G Control JBD68 0.0 0.5 1.0 Phage Infection (Max O.D.600nm) ns ns ns ns ns ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ns ✱✱✱✱ WT ΔpilA 0.0% 0.1% 0.2% 0.3% 0.4% E 1.5 Arabinose Concentration(%) 0.50.40.30.20.10.0 0.6 0 50 100 150Twitching Area (% of WT) B D 0.0 0.5 1.0 ns ns ns ns ✱ 1.5 ✱✱✱✱ WT ΔpilA 0.0% 0.1% 0.2% 0.3% 0.4% ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ✱✱✱✱ ns ✱✱✱✱ F 0 0 1 2 3 5 4 100 200 300 400 500 Time Elapsed (min.) WT% 0.6% 0.4% 0.2% 0.0% Control JBD68 H Arabinose Concentration(%) 0.50.40.30.20.10.0 0.6 Phage Infection (Max O.D.600nm) VFR Activity (YFP/mKate2, a.u.) Biofilm Formation (O.D.600nm) WT ΔpilA 0.0 0.1 0.2 0.3 0.4 4 3 2 1 0 nsZZZZ ns ✱✱✱✱ ✱✱ZZZ ✱ ns Biofilm Formation (O.D.600nm) 0 1 2 3 4 ✱✱zzz ns ZZZ ns ✱✱✱✱ ✱✱✱✱ ✱✱ WT ΔpilA 0.0 0.1 0.2 0.3 0.4 Arabinose Concentration(%) Arabinose Concentration(%) was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Figure 5. The length of pili governs T4P-dependent virulence traits. A,B) The area of twitching motility 321 as percentage of WT on 1% LB agar plates as a function of PilA titration (Pbad::pilA) in pilA (A) and pilA 322 pilSR (B) backgrounds. The horizontal dotted line indicates the mean twitching area of the WT strain under 323 identical conditions. Bars represent mean ± SD of three biological replicates (indicated as black dots). C,D) 324 Surface sensing activity over time for cells on a 1.5% agarose pad. Surface sensing was measured by the 325 activity of the transcription factor Vfr using the fluorescent reporter PaQa as a function of PilA titration 326 (Pbad::pilA) in pilA (C) and pilA pilSR (D) backgrounds. Data points represent mean ± SEM of three biological 327 replicates. E,F) Maximum bacterial growth (O.D. = optical density) after infection with the pilus tip- binding 328 phage JBD68 as a function of PilA titration ( Pbad::pilA) in pilA (E) and pilA pilSR (F) backgrounds (see 329 Supplementary Information for individual growth curves). Bars represent mean OD600 ± SEM (n = 8 technical 330 replicates across 4 biological replicates). G,H) Biomass of biofilms grown on liquid- submerged pegs as a 331 function of PilA titration (Pbad::pilA) in pilA (G) and pilA pilSR (H) backgrounds. Bars represent mean OD600 332 ± SD of N = 18 technical replicates across three biological replicates. G-H) Statistical tests were performed 333 using unpaired t-test. E – F) Statistical tests were performed using one-way ANOVA: (ns) not significant, P 334 > 0.05; (*) P < 0.05; (**) P < 0.01; (***) P < 0.001; (****) P < 0.0001. 335 336

337 Type IV pili (T4P) enable bacteria to interact with their environments through dynamic 338 cycles of extension and retraction. Much of the literature has focused on retraction as the 339 critical rate-limiting step in T4P function, highlighting the high forces (>100 pN per pilus) 340 generated by retraction ATPases (PilT/PilU) that power twitching motility, surface sensing, 341 phage uptake, natural transformation, and biofilm maturation [9, 13, 17-19, 29-36, 51-57]. 342 Extension on the other hand is typically seen as a mere preparatory step. Here, we 343 challenge this perspective by showing that extension dynamics , specifically pilus length 344 governed by PilA monomer availability , represent a critical overlooked bottleneck that 345 constrains all retraction-enabled virulence traits. 346 We demonstrated that PilA levels vary ~10- fold between cells in natural populations 347 despite PilSR regulation. Our inducible titration system, combined with biophysical 348 simulations, revealed that PilA abundance primarily tunes extension velocity and pilus 349 length, not pilus count per cell. At low PilA concentrations, diffusion- limited monomer 350 supply in the 2D inner membrane increases waiting times for PilA to reach the PilB 351 extension complex, forcing transient idle states where PilB is primed but cannot 352 incorporate subunits. This slows net extension velocity and caps length, while retraction 353 velocity remains unaffected. Because extension events terminate stochastically via PilB 354 unbinding, the resulting pilus lengths follow an exponential distribution inherently biased 355 toward short filaments [13, 48]. This bias produces a large hidden subpopulation of non-356 contributing pili that cannot meaningfully interact with substrates, host cells, or phages. 357 Binding to DNA, phages, or surfaces during twitching requires pili to extend sufficiently 358 far, often beyond the cell’s extracellular polysaccharide coat or to reach a distant target, 359 much like a fishing line must be long enough to reach the water to catch fish. 360 For example, the mean pilus length at low PilA induction (20%) is ~160 nm and increases 361 4.7-fold to ~750 nm for WT -like cells (100% , simulations). Assuming the functional 362 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint threshold needed for effective substrate interaction is similar to the mean length of T4P 363 in WT cells (750 nm), this increase in mean length by 4.7- fold raises the fraction of 364 contributing pili (the long tail of the exponential distribution) from ~0.9% to 37%, yielding 365 approximately 40- fold more pili capable of contributing to adhesion or motility 366 (Supplementary Figure 3). This represents a strong non-linear lever that boosts the ratio 367 of contributing pili disproportionately to the increase in pilin abundance. Achieving the 368 same gain by increasing pilus count alone (length distribution unchanged at 160 nm ) 369 would require scaling total pilus number ~40-fold, which is biologically challenging. Cells 370 typically maintain only ~5 pilus machines per cell (each comprising ~50 subunits of 371 PilCMNOPQ, plus ~10× more PilB/T/U hexamers), limited by inner-membrane space and 372 assembly energetics. Scaling to 40-fold more machines would demand enormous protein 373 synthesis, membrane insertion, and coordinated assembly, far exceeding the modest cost 374 of elevating PilA from ~2000 molecules (20% induction) to ~10,000 ( WT levels). 375 Moreover, such expansion would likely overwhelm downstream signaling control via Pil -376 Chp, PilZ, and other regulators, risking dysregulation and interference with other 377 membrane functions [19, 20, 33, 36- 38, 62, 63, 65, 66]. In contrast, increasing PilA 378 abundance is transcriptionally simple, energetically modest, and avoids these 379 bottlenecks, enabling precise, low -cost shifts in the contributing fraction via the 380 exponential tail of the distribution . This makes length regulation far more efficient , and 381 likely biologically favored, than count regulation for enhancing T4P -mediated virulence 382 traits, and presents another strong non- linear functional lever in the efficiency of 383 increasing the ratio of contributing pili relative to an increase in pilus machine count. 384 Our experiments support this conclusion across T4P-mediated behaviors: longer pili 385 consistently restored function more effectively than equivalent changes in pilus count 386 alone would imply. Twitching motility and surface sensing (via Pil -Chp/Vfr) showed the 387 strongest length dependence, with PilSR providing additional independent regulation, 388 negative for twitching (likely via unknown targets such as surfactants or adhesion 389 components) and essential for surface sensing. Phage infection and biofilm formation 390 scaled primarily with length and showed little PilSR sensitivity, suggesting partial 391 decoupling from mechanosensing pathways. 392 We like to note a recent study in Vibrio cholerae that reported lowering of PilA levels 393 increased pilus number, in contrast to our observations in P. aeruginosa [67]. T hese 394 complementary findings highlight how PilA titration effects can be context -dependent 395 (species, mutant background) and reinforce that monomer availability critically shapes 396 T4P dynamics and functional output. It will be interesting to uncover how context -397 dependent and other confounding biological factors regulate T4P dynamics and their 398 associated virulence traits in future studies. 399 Together, our study reframes the importance and functional consequences of T4P 400 dynamics: extension velocity, limited by diffusion-driven monomer supply and idle states, 401 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint governs the effective fraction of pili that contribute to function and thus constrains 402 retraction-powered virulence. The exponential length distribution, combined with ~10-fold 403 cell-to-cell variation in PilA levels, generates substantial phenotypic heterogeneity in pilus 404 length within a clonal population. This heterogeneity enables bet-hedging: subpopulations 405 with predominantly short pili may persist under conditions where long pili are costly or 406 risky (e.g., phage exposure or energy limitation), while those with longer pili are poised 407 for rapid colonization when opportunities arise. Such diversity increases the overall 408 resilience of the population to fluctuating environments without requiring every cell to 409 commit to a single strategy. This mechanism may extend to other T4P pathogens, where 410 length heterogeneity influences pathogenesis. Future studies on length thresholds in vivo 411 and regulatory feedback loops will further clarify how extension emerges as a governing 412 step in T4P biology. 413 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint

414 We would like to thank Joanne Engel and Yuki Inclan for the gift of the polyclonal PilA 415 antibody, Karen Maxwell and Veronique Taylor for the gift of phage JBD68, Joseph Sorg 416 and Morgan Osborne for training and use of the LiCOR Odyssey M. We would like to 417 thank Joseph Sanfilippo and the entire Koch lab for stimulating discussion, and the 418 Department of Biology at Texas A&M for its supportive environment. 419 This work was supported by grant R35GM155280 from the National Institute of Health 420 and startup funds from the College of Arts and Sciences , Division of Research, and 421 Department of Biology at Texas A&M University to M.D.K. 422 423 Author contributions 424 Z.M., J.R., and M.D.K. designed research. Z.M. and M.D.K. wrote the manuscript. Z.M., 425 J.R., A.O.Y, and M.D.K. provided reagents and constructs. Z.M., T.P., N.C, and S.D. 426 performed experiments and analyzed data. 427 428 Competing Interest Statement 429 The authors declare no competing finical interests. 430 431

and Protocols 432 Strains and growth conditions: P. aeruginosa PAO1 was grown in liquid lysogeny broth 433 (LB) Miller (Difco) in a floor shaker, or on LB Miller agar (1.5% Bacto Agar), on Vogel -434 Bonner minimal medium agar (200 mg/L MgSO4 7H2O, 2 g/L citric acid, 10 g/L K2HPO4, 435 3.5 g/L NaNH4HPO4 4 H2O, and 1.5% agar), and on no-salt LB agar (10 g/L tryptone, 5 436 g/L yeast extract, and 1.5% agar) at 30 °C (for cloning) or at 37 °C. E. coli S17 was grown 437 in liquid LB Miller (Difco) in a floor shaker and on LB Miller agar (1.5% Bacto Agar) at 438 30 °C or at 37 °C. Arabinose- supplemented media were prepared by adding filter -439 sterilized 4% (w/v) arabinose stock solution to LB or autoclaved medium to achieve the 440 desired final concentration. Antibiotics were used at the following concentrations: 441 100μg/mL or 200 μg/mL carbenicillin in liquid (300 μg/mL on plates) or 10 μg/mL 442 gentamycin in liquid (30 μg/mL on plates) for P. aeruginosa, and 100 μg/mL carbenicillin 443 in liquid (100 μ g/mL on plates) or 30 μ g/mL gentamycin in liquid (30 μ g/mL on plates) 444 for E. coli. Western blotting medias were formulated as follows: 4x loading Dye ( 4ml of 445 Glycerol, 2.5 mL of 1 M Tris -HCL pH6.8, 20 mg Bromophenol blue, 2 mL β-446 mercaptoethanol, 0.8 g SDS, 1.5 ml H20), 10x Tris-Buffer Saline ph7.6 (24.23g of Tris, 447 80g NaCl, adjusted to 1 l wtih H2O), 1x Tris-Buffer Saline with Tween-20 (1 ml Tween 20, 448 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 100 ml of 10x TBS, adjusted to 1 l withH2O ), 1x Running Buffer (12 g Tris, 57.6 g of 449 Glycine, 4g SDS, adjusted to 4 l with H2O ), Blocking Buffer ( 2.5g of powdered milk, 50 450 ml of 1x TBS). 451 Generation of arabinose -inducible pilA construct (pMK119): The mini-Tn7 delivery 452 vectors pMK119 harboring either a native copy of pilA or the A86C allele were constructed 453 by Gibson assembly. The backbone was obtained by linearizing pUC18-mini-Tn7T-LAC 454 with Eco53kI and NsiI restriction enzymes . The fragment containing araC and Pbad was 455 PCR amplified from the iRFP670 plasmid [68] using primers pMK119_f1.For/Rev and 456 digested with AvrII. The pilA fragments were PCR amplified using primers 457 pMK119_F2.For/Rev from genomic DNA. All fragments were assembled using NEBuilder 458 HiFi DNA Assembly Master Mix (New England Biolabs) according to the manufacturer's 459 protocol. Assembly products were electroporated into electrocompetent E. coli S17: 50 460 µL competent cells were mixed with the assembly product, transferred to a chilled 461 electroporation cuvette, and pulsed. Immediately after electroporation, 1 mL LB was 462 added, and cells recovered at 37°C with shaking for 1–1.5 h. Cells were concentrated by 463 centrifugation, resuspended in 100 µL LB, and plated on LB agar supplemented with 100 464 µg/mL carbenicillin. Plates were incubated overnight at 37°C. Constructs were verified by 465 colony PCR and confirmed by Sanger sequencing using primers pTn7_Ver1/2. Verified 466 strains were grown overnight in LB + 100 μg/ml carbenicillin, and glycerol stocks ( 25% 467 v/v) were prepared and stored at −80°C. 468 Chromosomal integration of pMK119: Plasmids pMK119 with either the native of A86C 469 allele of pilA and helper plasmid pTNS2 were purified using the Qiagen miniprep kit. 470 Electrocompetent recipient strain were prepared by growing cultures overnight in LB, 471 diluting 1:1000 into fresh LB, and harvesting at early stationary phase. Cells were washed 472 three times in ice-cold 300 mM sucrose and resuspended in 50 μl of the same buffer. For 473 each electroporation, 50 μl competent cells were mixed with 300 ng each of pMK119 and 474 pTNS2 plasmid DNA, electroporated, and recovered in 1 ml LB at 37°C with shaking for 475 1–1.5 h. Cells were concentrated, plated on LB agar + 10 μ g/ml gentamycin, and 476 incubated overnight at 37°C. Successful insertion at the attTn7 site was confirmed by 477 colony PCR using primers pTN7_Ver1/2, which produce a diagnostic band shift. 478 Confirmed integrands were grown overnight in LB + 10 μ g/ml gentamycin and stored as 479 50% glycerol stocks at −80°C. 480 Generation of pilSR recipient strain: In-frame deletions were generated using allelic 481 exchange with the suicide vector pEXG2 following [69]. Flanking regions upstream of pilS 482 and downstream of piLR were amplified with PCR using primers pilS_P1/P2 and 483 pilR_P3/P4. Fragments were joined by overlap extension PCR using the outermost 484 primers, digested with HindIII -HF, and ligated into HindIII -digested pEXG2. Constructs 485 were transformed into E. coli S17, verified by PCR and Sanger sequencing with 486 pEXG2_Ver1/Ver2 primers, and introduced into PAO1 by conjugation. For mating, 1.5 ml 487 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint donor containing the vector were grown to OD 0.5. The PA01 parental strain was grown 488 overnight, and 0.5 ml culture was diluted 1:2 into fresh LB and incubated for 3 hours at 489 42 °C. Both cultures were concentrated into 100 µl and spotted onto an LB agar plate and 490 incubated overnight at 30 °C. The puddle was scrapped off, resuspended into 150 µl PBS 491 and spread on VBMM + 30 μ g/ml gentamicin at 37°C for 24 h. For counter-selection, 492 single colonies from the VBMM plate were struck onto NSLB and incubated for 24 hours 493 at 30 °C. Colonies from the NSLB plate were screened for the correct deletion mutation 494 using PCR amplification with the flaking primers and confirmed using sanger sequencing. 495 Construction of PpilA::yfp transcriptional reporter plasmid 496 The PpilA::yfp reporter plasmid was constructed using the established PPaQa::yfp 497 PrpoD::mKate2 dual-reporter vector that includes the constitutive PrpoD promoter for 498 normalization [19]. The vector was linearized by digestion with XhoI and BsiWI. The native 499 pilA promoter region and yfp coding sequence were PCR -amplified as separate 500 fragments using primers PpilA_F1.For/REV (yfp) and PpilA_F2.For/REV (pilA promoter). 501 All fragments were assembled using NEBuilder HiFi DNA Assembly Master Mix (New 502 England Biolabs) according to the manufacturer's protocol . electroporated into 503 electrocompetent E. coli S17, and 50 µL competent cells were mixed with the assembly 504 product, transferred to a chilled electroporation cuvette, recovered in 1 mL LB and grown 505 at 37°C with shaking for 1– 1.5 h. Cells were then concentrated by centrifugation, 506 resuspended in 100 µL LB, and plated on LB agar supplemented with 100 µg/mL 507 carbenicillin. Plates were incubated overnight at 37°C. Successful assembly was 508 confirmed by colony PCR using primers PpilA_ Ver1/2, which flank the insertion site and 509 produce a diagnostic band shift. Verified clones were grown overnight in LB + 100 μg/ml 510 carbenicillin at 37°C with shaking, mixed 1:1 with 50% (v/v) glycerol, and stored as frozen 511 stocks at −80°C. 512 This reporter plasmid was then isolated using a Qiagen miniprep kit and introduced into 513 recipient strains using electroporation. Electrocompetent cells were prepared by growing 514 cultures overnight in LB, diluting 1:1000 into fresh LB, harvesting at early stationary 515 phase, washing three times in ice-cold 300 mM sucrose, and resuspending in 50 μl of the 516 same buffer. For each electroporation, 50 µL competent cells were mixed and with 100ng 517 of donor plasmid, transferred to a chilled cuvette, electroporated, recovered in 1 mL LB 518 at 37°C shaking for 1 –1.5 h. Cells were concentrated, resuspended in 100 µL LB, and 519 plated on LB agar containing 300 µg/mL carbenicillin. Plates were incubated overnight at 520 37°C. Successful plasmid maintenance was confirmed by colony PCR using priemrs 521 PpilA_Ver1/2. Confirmed transformants were grown overnight in LB + 2 00 μg/ml 522 carbenicillin and stored as 25% glycerol stocks at −80°C. 523 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Twitch Plate Assay 524 LB broth supplemented with 1% (w/v) agar was autoclaved and allowed to cool to 525 approximately 50–60°C. Specified concentrations of L-arabinose were added to separate 526 100 mL aliquots of the cooled molten agar, mixed thoroughly, and poured evenly into Petri 527 dishes. Plates were allowed to solidify and were stored overnight at room temperature to 528 dry. Bacterial strains were recovered from frozen glycerol stocks by streaking onto 529 standard LB agar plates (1.5% agar) and incubating overnight at 37°C. The following day, 530 single colonies from these plates were used to inoculate designated areas on the surface 531 of the arabinose-supplemented 1% LB agar plates by spotting through the agar layer to 532 ensure contact with the agar -plastic interface. Plates were incubated at 37°C for 3 days 533 to permit development of twitching motility zones at the agar -Petri dish interface. 534 Following incubation, the agar was carefully removed and discarded. A 1% (w/v) crystal 535 violet solution was added to stain the biofilm matrix on the Petri dish surface for 10– 15 536 min. Excess stain was gently rinsed off with deionized water, and plates were air -dried. 537 The total area of twitching motility zones was measured using digital imaging. Mutant 538 twitching zones TM were normalized to WT (TWT, positive control) and pilA (TpilA negative 539 control) to yield percentage of WT twitching according to (TM – TpilA) / (TWT – TpilA). 540 Phage Infection Growth Assay 541 Phage infection assays were carried out in 96- well plates liquid culture formats as 542 described previously [70-72]. In brief, Bacterial strains were inoculated into LB broth and 543 grown overnight at 37°C with shaking (250 rpm). The following day, overnight cultures 544 were diluted 1:333 into 2 mL of fresh LB medium supplemented with the appropriate 545 concentration of L- arabinose and grown at 37°C with shaking until an optical density 546 OD₆₀₀ = 0.5 was reached. These mid- log phase cultures were then diluted 1:100 into 2 547 mL of fresh arabinose- supplemented LB medium for a concentration of 5*10 6 CFU/mL. 548 For the infection assay, 25 µL of each phage dilution ( total PFU: 3.87*106 PFU/mL) was 549 added to designated wells of a clear flat -bottom 96-well microtiter plate. Subsequently, 550 225uL (Total CFU: 1.12*10 6 CFU/mL) of the diluted bacterial cultures were added to 551 designated wells, resulting in a final volume of 250 uL per well (MOI ~3.44). Bacterial 552 growth was monitored by measuring OD600 every 15 min overnight at 37C with 553 continuous orbital shaking in a microplate reader. 554 Biofilm Assay 555 Biofilm assays were carried out in 96- well plates liquid culture formats as described 556 previously [73]. In brief, Bacterial strains were inoculated into LB broth and grown 557 overnight at 37°C with shaking (250 rpm). The following day, overnight cultures were 558 diluted 1:333 into 2 mL of fresh LB medium supplemented with the appropriate 559 concentration of L- arabinose and grown at 37°C with shaking until an optical density 560 OD₆₀₀ = 0.5 was reached. These mid- log phase cultures were then diluted 1:100 into 2 561 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint mL of fresh arabinose -supplemented LB medium. 180 µL of the resulting dilution were 562 dispensed into designated wells of a 96-well microtiter plate. A Nunc-Immuno™ TSP 96-563 peg lid was placed onto the plate, and biofilms were allowed to form during overnight 564 incubation at 37 °C. The following day, the peg lid was gently washed by immersion in a 565 96-well plate containing 180 µL of sterile phosphate- buffered saline (PBS) per well. 566 Biofilms were then stained by transferring the peg lid to a separate 96- well plate 567 containing 180 µL of crystal violet solution per well and incubating for 15 min at room 568 temperature. Excess stain was removed by washing the peg lid in new 96- well plate 569 containing 180 µL of PBS per well. For biofilm solubilization, the peg lid was transferred 570 to a new 96-well plate containing 180 µL of 100% acetone per well and incubated for 15 571 min at room temperature. The peg lid was subsequently removed, and biofilm biomass 572 was quantified by measuring the absorbance at 600nm using a microplate reader. 573 Western Blots 574 Bacterial strains were inoculated into LB broth and grown overnight at 37°C with shaking 575 (250 rpm). 800 µL of overnight culture was pelleted by centrifugation at 13,000 rpm for 2 576 min, the supernatant was discarded, and the pellet was resuspended in 800 µL of fresh 577 LB medium. This washed culture was used to inoculate 4 mL of fresh arabinose-578 supplemented LB medium at a 1:333 dilution and grown at 37 °C with shaking until an 579 optical density OD ₆₀₀ = 0.8 was reached. Cells were harvested from 800– 2400 µL of 580 culture (adjusted based on protein expression levels) by centrifugation. Supernatants 581 were removed, and pellets were washed by resuspension in 200 µL of phosphate-582 buffered saline (PBS) followed by centrifugation. Final pellets were resuspended in 20 µL 583 of SDS sample loading buffer and heated at 95 °C for 20 min. 584 Protein samples (5 µL) and LI-COR Chameleon Duo protein ladder (0.8 µL) were loaded 585 into 20% Mini -PROTEAN TGX precast gels (Bio -Rad). Electrophoresis was performed 586 using the Mini -PROTEAN system (Bio -Rad) in pre- chilled running buffer with the gel 587 apparatus surrounded by ice until the dye front reached the bottom of the gel. 588 Nitrocellulose membranes and filter papers were pre-equilibrated in transfer buffer for at 589 least 1h prior to transfer. Proteins were transferred to nitrocellulose membranes using the 590 Trans-Blot Turbo system (Bio- Rad) according to the manufacturer’s instructions. 591 Following transfer, total protein was visualized using Revert™ 520 Total Protein Stain (LI-592 COR) for loading normalization. Membranes were washed in Tris -buffered saline (TBS), 593 dried at 37 °C for 20 min, rehydrated in TBS, and rinsed in deionized water. Membranes 594 were incubated in Revert 520 stain until protein bands were visible, then washed twice in 595 Revert Wash Solution followed by a wash in deionized water. Total protein signal was 596 imaged using an Odyssey CLX imager (LI-COR) in the 520 nm channel. Membranes were 597 subsequently rinsed in ultrapure water for 5 min, destained using Revert Destaining 598 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Solution until background signal was cleared, and rinsed again in ultrapure water for 5 599 min while shaking. 600 For immunodetection, membranes were washed in TBS and blocked in blocking buffer 601 for 1 h at room temperature with shaking. Membranes were then incubated overnight at 602 4 °C with shaking in primary antibody (rabbit anti -PilA polyclonal antibody, 1:50,000 603 dilution in blocking buffer). The following day, membranes were washed three times in 604 TBST and incubated for 1 h at room temperature with shaking in IRDye- conjugated 605 secondary antibody (1:50,000 dilution in blocking buffer). Membranes were washed three 606 additional times in TBST and imaged using an Odyssey CLX imager (LI -COR) in the 607 appropriate fluorescence channels. Band intensities for PilA were quantified and 608 normalized to total protein signal obtained from Revert 520 staining. 609 Fluorescence microscopy imaging 610 Fluorescence imaging was performed using a Nikon Ti2- Eclipse inverted microscope 611 controlled by NIS -Elements software (Nikon), equipped with a 40x/0.95NA air and a 612 100x/1.45NA oil objective (Nikon), an ORCA -Fusion BT sCMOS camera (Hamamatsu), 613 488/514/561 nm lasers and appropriate filters and dichroic optimized for imaging of 614 Alexa488, YFP, and mKate2 channels. 615 Imaging of PaQa and PilA transcriptional reporters: 616 Surface sensing assays were carried out in 96- well plates liquid culture formats as 617 described previously [17, 19]. Bacterial strains were inoculated into LB broth with Carb100 618 and grown overnight at 37°C with shaking (250 rpm). The following day, cultures were 619 diluted 1:333 into fresh LB medium supplemented with arabinose and grown to OD ₆₀₀ = 620 0.5. 300 µL aliquots of cultures were transferred to microcentrifuge tubes, pelleted by 621 centrifugation at 13,000 rpm for 2 min, and the supernatant was removed. Cell pellets 622 were resuspended in 300 µL of fresh LB medium. 1 uL of each of the resuspended 623 cultures were spotted onto the center of an agarose pad (1%) supplemented with the 624 designated concentration of arabinose. Agarose pads were then inverted onto glass -625 bottom dishes for imaging. For PaQa, time-lapse images were acquired every 15min for 626 a total duration of 7h. Each experiment consisted of four independent fields of view (FOV) 627 of each agarose pad. FOVs containing approximately 10- 40 cells were selected for 628 imaging. Individual cells were segmented using a custom-written, threshold-based image 629 analysis algorithm based on fluorescence from a constitutively expressed reporter 630 (PropD::mKate2) [17]. This segmentation approach enabled estimation of mean 631

and fluorescence intensities for both reporter channels. Transcriptional 632 reporter activity was quantified as the ratio of YFP to mKate2 fluorescence. For each time 633 point, the median fluorescence ratio for all cells was calculated. For the PilA 634 transcriptional reporter, the imaging setup and analysis was identical but only one still 635 image was taken. 636 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint Imaging of Pilus Dynamics 637 Pilus dynamics microscopy assays were carried out agarose gel pad from liquid culture 638 as described previously [48] . Bacterial strains were inoculated into L) broth and grown 639 overnight at 37°C with shaking (250 rpm). The following day, overnight cultures were 640 diluted 1:333 into fresh EZ Rich Defined Medium supplemented with the specified 641 concentration of L- arabinose. Cultures grew at 37°C with shaking until mid- log phase, 642 approximately OD 600 = 0.5. 200 µL of each culture was transferred to a 1.5 mL 643 microcentrifuge tube and 1uL of Alexaflour -488 maleimide dye (dissolved in anhydrous 644 DMSO at 2.5ug/uL) was added. The mixture was incubated for 45 min at 37C in the dark. 645 Labeled cells were gently washed twice to remove unbound dye by pelleting and 646 resuspension in 200 µL of fresh EZ rich media. After the second wash, cells were finally 647 resuspended in 50 µL of fresh EZ rich media. A 2 µL of the resuspended cells was spotted 648 onto the center of an agarose pad, the pad was then inverted onto a glass -bottom dish. 649 Time-lapse imaging of type IV pilus dynamics was performed every 200ms for 30s at 650 minimal laser power to reduce phototoxicity and photobleaching. Single- cell TFP 651 dynamics (extension, retraction, count, and length) were analyzed manually in 652 Fiji/ImageJ. 653 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint

654 1. Burrows, L.L., Pseudomonas aeruginosa twitching motility: type IV pili in action. 655 Annual review of microbiology, 2012. 66(1): p. 493-520. 656 2. Ochner, H., et al., Structure of the Pseudomonas aeruginosa PAO1 Type IV pilus. 657 PLoS pathogens, 2024. 20(12): p. e1012773. 658 3. Ellison, C.K., G.B. Whitfield, and Y . V . Brun, Type IV Pili: dynamic bacterial 659 nanomachines. FEMS Microbiology Reviews, 2022. 46(2): p. fuab053. 660 4. Craig, L., K.T. Forest, and B. Maier, Type IV pili: dynamics, biophysics and functional 661 consequences. Nature reviews microbiology, 2019. 17(7): p. 429-440. 662 5. Chang, Y .-W., et al., Architecture of the type IVa pilus machine. Science, 2016. 663 351(6278): p. aad2001. 664 6. Gibiansky, M.L., et al., Bacteria use type IV pili to walk upright and detach from 665 surfaces. Science, 2010. 330(6001): p. 197-197. 666 7. Semmler, A.B., C.B. Whitchurch, and J.S. Mattick, A re-examination of twitching 667 motility in Pseudomonas aeruginosa. Microbiology, 1999. 145(10): p. 2863-2873. 668 8. O'Toole, G.A. and R. Kolter, Flagellar and twitching motility are necessary for 669 Pseudomonas aeruginosa biofilm development. Molecular microbiology, 1998. 670 30(2): p. 295-304. 671 9. Klausen, M., et al., Biofilm formation by Pseudomonas aeruginosa wild type, flagella 672 and type IV pili mutants. Molecular microbiology, 2003. 48(6): p. 1511-1524. 673 10. Qin, S., et al., Pseudomonas aeruginosa: pathogenesis, virulence factors, antibiotic 674 resistance, interaction with host, technology advances and emerging therapeutics. 675 Signal transduction and targeted therapy, 2022. 7(1): p. 199. 676 11. Pohlschroder, M. and R.N. Esquivel, Archaeal type IV pili and their involvement in 677 biofilm formation. Frontiers in microbiology, 2015. 6: p. 190. 678 12. Floyd, K.A., et al., c-di-GMP modulates type IV MSHA pilus retraction and surface 679 attachment in Vibrio cholerae. Nature communications, 2020. 11(1): p. 1549. 680 13. Ellison, C.K., et al., Retraction of DNA-bound type IV competence pili initiates DNA 681 uptake during natural transformation in Vibrio cholerae. Nature microbiology, 2018. 682 3(7): p. 773-780. 683 14. Johnston, C., et al., Bacterial transformation: distribution, shared mechanisms and 684 divergent control. Nature Reviews Microbiology, 2014. 12(3): p. 181-196. 685 15. Aas, F .E., C. Løvold, and M. Koomey, An inhibitor of DNA binding and uptake events 686 dictates the proficiency of genetic transformation in Neisseria gonorrhoeae: 687 mechanism of action and links to Type IV pilus expression. Molecular microbiology, 688 2002. 46(5): p. 1441-1450. 689 16. Cehovin, A., et al., Specific DNA recognition mediated by a type IV pilin. Proceedings 690 of the National Academy of Sciences, 2013. 110(8): p. 3065-3070. 691 17. Koch, M.D., et al., Pseudomonas aeruginosa distinguishes surfaces by stiffness 692 using retraction of type IV pili. Proceedings of the National Academy of Sciences, 693 2022. 119(20): p. e2119434119. 694 18. Ellison, C.K., et al., Obstruction of pilus retraction stimulates bacterial surface 695 sensing. Science, 2017. 358(6362): p. 535-538. 696 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 19. Persat, A., et al., Type IV pili mechanochemically regulate virulence factors in 697 Pseudomonas aeruginosa. Proceedings of the National Academy of Sciences, 2015. 698 112(24): p. 7563-7568. 699 20. Luo, Y ., et al., A hierarchical cascade of second messengers regulates 700 Pseudomonas aeruginosa surface behaviors. MBio, 2015. 6(1): p. 10.1128/mbio. 701 02456-14. 702 21. Christman, N.D. and A.B. Dalia, The molecular basis for DNA-binding by 703 competence T4P is distinct in a representative Gram-positive and Gram-negative 704 species. PLoS Pathogens, 2025. 21(4): p. e1013128. 705 22. Chlebek, J.L., et al., Fresh extension of Vibrio cholerae competence type IV pili 706 predisposes them for motor-independent retraction. Applied and Environmental 707 Microbiology, 2021. 87(14): p. e00478-21. 708 23. Merz, A.J., M. So, and M.P . Sheetz, Pilus retraction powers bacterial twitching 709 motility. Nature, 2000. 407(6800): p. 98-102. 710 24. Rudel, T., et al., Interaction of two variable proteins (PilE and PilC) required for pilus-711 mediated adherence of Neisseria gonorrhoeae to human epithelial cells. Molecular 712 microbiology, 1992. 6(22): p. 3439-3450. 713 25. Harding, C.M., et al., Acinetobacter baumannii strain M2 produces type IV pili which 714 play a role in natural transformation and twitching motility but not surface-715 associated motility. MBio, 2013. 4(4): p. 10.1128/mbio. 00360-13. 716 26. Vesel, N. and M. Blokesch, Pilus production in Acinetobacter baumannii is growth 717 phase dependent and essential for natural transformation. Journal of Bacteriology, 718 2021. 203(8): p. 10.1128/jb. 00034-21. 719 27. Ellison, C.K., et al., Acinetobacter baylyi regulates type IV pilus synthesis by 720 employing two extension motors and a motor protein inhibitor. Nature 721 communications, 2021. 12(1): p. 3744. 722 28. Leighton, T.L., et al., Biogenesis of P seudomonas aeruginosa type IV pili and 723 regulation of their function. Environmental microbiology, 2015. 17(11): p. 4148-4163. 724 29. Nolan, L.M., et al., A global genomic approach uncovers novel components for 725 twitching motility-mediated biofilm expansion in Pseudomonas aeruginosa. 726 Microbial genomics, 2018. 4(11): p. e000229. 727 30. Tuon, F .F ., et al., Pathogenesis of the Pseudomonas aeruginosa biofilm: a review. 728 Pathogens, 2022. 11(3): p. 300. 729 31. Barnshaw, R., et al., Separation of Pseudomonas aeruginosa type IV pilus-730 dependent twitching motility and surface-sensing responses. Mbio, 2025. 16(11): p. 731 e02521-25. 732 32. Rossy, T., et al., Pseudomonas aeruginosa type IV pili actively induce mucus 733 contraction to form biofilms in tissue-engineered human airways. PLoS biology, 734 2023. 21(8): p. e3002209. 735 33. Inclan, Y .F ., et al., A scaffold protein connects type IV pili with the Chp 736 chemosensory system to mediate activation of virulence signaling in Pseudomonas 737 aeruginosa. Molecular Microbiology, 2016. 101(4): p. 590-605. 738 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 34. Laventie, B.-J., et al., A surface-induced asymmetric program promotes tissue 739 colonization by Pseudomonas aeruginosa. Cell host & microbe, 2019. 25(1): p. 140-740 152. e6. 741 35. Lee, C.K., et al., Multigenerational memory and adaptive adhesion in early bacterial 742 biofilm communities. Proceedings of the National Academy of Sciences, 2018. 743 115(17): p. 4471-4476. 744 36. Tran, V .N., et al., The repertoire of resistance mutations selected by a 745 Pseudomonas aeruginosa type IV pilus-targeting lytic bacteriophage. mBio. 746 0(0): p. e00243-26. 747 37. Hobbs, M., et al., PilS and PilR, a two-component transcriptional regulatory system 748 controlling expression of type 4 fimbriae in Pseudomonas aeruginosa. Molecular 749 microbiology, 1993. 7(5): p. 669-682. 750 38. Kilmury, S.L. and L.L. Burrows, The Pseudomonas aeruginosa PilSR two-component 751 system regulates both twitching and swimming motilities. Mbio, 2018. 9(4): p. 752 10.1128/mbio. 01310-18. 753 39. Boyd, J.M. and S. Lory, Dual function of PilS during transcriptional activation of the 754 Pseudomonas aeruginosa pilin subunit gene. Journal of bacteriology, 1996. 178(3): 755 p. 831-839. 756 40. Tammam, S., et al., PilMNOPQ from the Pseudomonas aeruginosa type IV pilus 757 system form a transenvelope protein interaction network that interacts with PilA. 758 Journal of bacteriology, 2013. 195(10): p. 2126-2135. 759 41. Ayers, M., et al., PilM/N/O/P proteins form an inner membrane complex that affects 760 the stability of the Pseudomonas aeruginosa type IV pilus secretin. Journal of 761 molecular biology, 2009. 394(1): p. 128-142. 762 42. Roberge, N.A. and L.L. Burrows, Building permits— control of type IV pilus assembly 763 by PilB and its cofactors. Journal of Bacteriology, 2024. 206(12): p. e00359-24. 764 43. Craig, L., M.E. Pique, and J.A. Tainer, Type IV pilus structure and bacterial 765 pathogenicity. Nature Reviews Microbiology, 2004. 2(5): p. 363-378. 766 44. McCallum, M., L.L. Burrows, and P .L. Howell, The dynamic structures of the type IV 767 pilus. Microbiology spectrum, 2019. 7(2): p. 10.1128/microbiolspec. psib-0006-768 2018. 769 45. Guo, S., et al., PilY1 regulates the dynamic architecture of the type IV pilus machine 770 in Pseudomonas aeruginosa. Nature Communications, 2024. 15(1): p. 9382. 771 46. Ayers, M., P .L. Howell, and L.L. Burrows, Architecture of the type II secretion and 772 type IV pilus machineries. Future microbiology, 2010. 5(8): p. 1203-1218. 773 47. Jain, R., et al., Type IV Pilus Assembly in Pseudomonas aeruginosa over a Broad 774 Range of Cyclic di-GMP Concentrations. Journal of Bacteriology, 2012. 194(16): p. 775 4285-4294. 776 48. Koch, M.D., et al., Competitive binding of independent extension and retraction 777 motors explains the quantitative dynamics of type IV pili. Proceedings of the 778 National Academy of Sciences, 2021. 118(8): p. e2014926118. 779 49. Kühn, M.J., et al., Mechanotaxis directs Pseudomonas aeruginosa twitching motility. 780 Proceedings of the National Academy of Sciences, 2021. 118(30): p. e2101759118. 781 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 50. Patino, R., et al., Spatial control of sensory adaptation modulates mechanosensing 782 in Pseudomonas aeruginosa. bioRxiv, 2024: p. 2024.02. 27.582188. 783 51. Maier, B., et al., Single pilus motor forces exceed 100 pN. Proceedings of the 784 National Academy of Sciences, 2002. 99(25): p. 16012-16017. 785 52. Biais, N., et al., Cooperative retraction of bundled type IV pili enables nanonewton 786 force generation. PLoS biology, 2008. 6(4): p. e87. 787 53. Clausen, M., et al., High-force generation is a conserved property of type IV pilus 788 systems. Journal of Bacteriology, 2009. 191(14): p. 4633-4638. 789 54. Gurung, I., et al., Functional analysis of an unusual type IV pilus in the Gram-positive 790 Streptococcus sanguinis. Molecular microbiology, 2016. 99(2): p. 380-392. 791 55. Ribbe, J., et al., Role of cyclic di-GMP and exopolysaccharide in type IV pilus 792 dynamics. Journal of bacteriology, 2017. 199(8): p. 10.1128/jb. 00859-16. 793 56. Wolfgang, M., et al., PilT mutations lead to simultaneous defects in competence for 794 natural transformation and twitching motility in piliated Neisseria gonorrhoeae. 795 Molecular microbiology, 1998. 29(1): p. 321-330. 796 57. Ng, D., et al., The Vibrio cholerae minor pilin TcpB initiates assembly and retraction 797 of the toxin-coregulated pilus. PLoS pathogens, 2016. 12(12): p. e1006109. 798 58. Maier, B., M. Koomey, and M.P . Sheetz, A force-dependent switch reverses type IV 799 pilus retraction. Proceedings of the National Academy of Sciences, 2004. 101(30): p. 800 10961-10966. 801 59. Marathe, R., et al., Bacterial twitching motility is coordinated by a two-dimensional 802 tug-of-war with directional memory. Nature communications, 2014. 5(1): p. 3759. 803 60. Ellison, C.K., et al., Real-time microscopy and physical perturbation of bacterial pili 804 using maleimide-conjugated molecules. Nature protocols, 2019. 14(6): p. 1803-805 1819. 806 61. Francis, V .I., E.C. Stevenson, and S.L. Porter, Two-component systems required for 807 virulence in Pseudomonas aeruginosa. FEMS Microbiology Letters, 2017. 364(11). 808 62. Chang, C.-Y ., Surface sensing for biofilm formation in Pseudomonas aeruginosa. 809 Frontiers in microbiology, 2018. 8: p. 2671. 810 63. Fulcher, N.B., et al., The Pseudomonas aeruginosa Chp chemosensory system 811 regulates intracellular cAMP levels by modulating adenylate cyclase activity. 812 Molecular Microbiology, 2010. 76(4): p. 889-904. 813 64. Wang, S., et al., A spider web strategy of type IV pili-mediated migration to build a 814 fibre-like Psl polysaccharide matrix in P seudomonas aeruginosa biofilms. 815 Environmental microbiology, 2013. 15(8): p. 2238-2253. 816 65. Hendrix, H., et al., PlzR regulates type IV pili assembly in Pseudomonas aeruginosa 817 via PilZ binding. Nature Communications, 2024. 15(1): p. 8717. 818 66. Taylor, V .L., et al., Prophages block cell surface receptors to preserve their viral 819 progeny. Nature, 2025: p. 1-9. 820 67. Christman, N.D., et al., The stoichiometry of minor-to-major pilins regulates the 821 dynamic activity of the type IVa competence pilus in Vibrio cholerae. bioRxiv, 2026: 822 p. 2026.01. 17.700090. 823 68. Piatkevich, K.D., et al., Near-infrared fluorescent proteins engineered from bacterial 824 phytochromes in neuroimaging. Biophysical journal, 2017. 113(10): p. 2299-2309. 825 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint 69. Hmelo, L.R., et al., Precision-engineering the Pseudomonas aeruginosa genome 826 with two-step allelic exchange. Nature protocols, 2015. 10(11): p. 1820-1841. 827 70. Qaderi, I., et al., Structural conservation and functional role of TfpY-like proteins in 828 type IV pilus assembly. Journal of Bacteriology, 2025. 207(2): p. e00343-24. 829 71. Fitzpatrick, A.D., et al., Phage reprogramming of Pseudomonas aeruginosa amino 830 acid metabolism drives efficient phage replication. MBio, 2025. 16(3): p. e02466-24. 831 72. Sztanko, K.M., et al., A prophage-expressed type IV pilus component provides anti-832 phage defense. Cell Reports, 2026. 45(1): p. 116759. 833 73. Marko, V .A., et al., Pseudomonas aeruginosa type IV minor pilins and PilY1 regulate 834 virulence by modulating FimS-AlgR activity. PLoS pathogens, 2018. 14(5): p. 835 e1007074. 836 837 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 4, 2026. ; https://doi.org/10.64898/2026.04.02.716181doi: bioRxiv preprint