Rapid Screening of Bacterial Chemoeffectors of Pseudomonas parafulva Using a CheA ATPase Activity‐Based Assay Authors: Rui Cui, Jie Li, Ni‐Ye You, Tino Krell, De‐Feng Li Published: 2026 ABSTRACT ABSTRACTBacterial chemotaxis enables cells to move in gradients of environmental signals that are sensed by chemoreceptors. A major limitation in the field consists currently in the lacking information on the signal(s) or chemoeffectors recognized by the majority of chemoreceptors, which in turn hampers understanding of how the environment has shaped chemotactic capabilities. A number of different approaches for signal identification have been reported that are often labour‐intensive. Here, we developed a straightforward NADH‐coupled CheA ATPase assay to systematically screen potential chemoeffectors. Obtained results are complemented with differential scanning fluorimetry and isothermal titration calorimetry analyses, as well as structural protein modelling to study the chemoreceptors of Pseudomonas parafulva PSR09‐11288. Using this integrated approach, we identified L‐malate, citrate, L‐glutamine, and L‐asparagine as chemoattractants and revealed that chemoreceptor B2J77_03605 specifically recognizes L‐malate and citrate, whereas chemoreceptor B2J77_13170 binds L‐glutamine and L‐asparagine. The universality of this approach is shown by the fact that both chemoreceptors belong to different families and possess sensor domains that belong to the HBM and dCache families, respectively. This pipeline is a scalable workflow for chemoeffector identification, permitting systematic exploration of bacterial chemotactic capabilities.Keywords: ATPase, CheA, chemoattractant, chemoreceptor, chemorepellent, NADH Introduction Chemotaxis enables microorganisms to sense and respond to chemical gradients in their environment, and therefore plays a critical role in microbial colonization, nutrient acquisition, biofilm formation, microbial community interactions, and adaptation to complex environments (Wadhams and Armitage 2004; Matilla and Krell 2018). Environmental signals are perceived by chemoreceptors, also known as methyl‐accepting chemotaxis proteins (MCPs), which alters the autokinase activity of CheA and consequently the transphosphorylation kinetics to CheY. Only phosphorylated CheY can bind to the flagellar motor modulating its activity which ultimately results in a chemotactic response. In E. coli , chemoattractants reduce CheA autophosphorylation, whereas chemorepellents have the opposite effect (Parkinson et al. 2015; Bi and Sourjik 2018). Furthermore, the sensitivity of the system is adjusted to the current ambient signal concentration by the methylation and demethylation of glutamate residues in the chemoreceptor signalling domain (Bi and Sourjik 2018). A growing number of studies have focused on the interactions between microorganisms and their environment (Stocker 2012; Smriga et al. 2016; Tsai et al. 2025; Xu et al. 2025). Chemotaxis was found to be crucial for the interaction with a different hosts (Matilla and Krell 2018; Zhou et al. 2023; Tsai et al. 2025) and among different bacterial species (Yarrington et al. 2024; Clerc et al. 2025; Laganenka et al. 2025). A representative example is the bacterial chemotaxis toward L‐glutamine released by plant roots that mediates root colonization and the definition of the microbiome (Tsai et al. 2025). However, for the large majority of chemoreceptors the cognate signal has so far not been identified which largely limits our understanding of how environmental conditions have shaped chemotactic capabilities (Martín‐Mora et al. 2018). The identification of chemoreceptor signals is also hampered by the fact that chemoreceptor sensor domains evolve rapidly showing a large sequence divergence (Gavira et al. 2020). Frequently, sensor domains that bind the same signal show only a very modest sequence identity (Matilla and Krell 2024), which in turn impedes signal identification for uncharacterized receptors by sequence homology with characterized receptors. A number of methods have been reported to identify chemoreceptor signals. Chemoreceptor signals were identified by the screening of the chemotactic behaviour of wild‐type and chemoreceptor mutants (Liu et al. 2009; Matilla and Krell 2018; Hida et al. 2020; Gavira et al. 2021). However, such approaches may result unsuccessful in cases where there are multiple chemoreceptors for the same signal that compensate each other (Corral‐Lugo et al. 2018). In other cases, chemotaxis is not observed under standard experimental conditions due to a lack of chemoreceptor expression (Rico‐Jiménez et al. 2016; Martín‐Mora et al. 2019). Other approaches are protein‐based. Chemoreceptor sensor domains can be produced as individual recombinant proteins that preserve their ligand binding capacities (Monteagudo‐Cascales et al. 2025). Proteins can be used for high‐throughput screening using the thermal shift assay (Monteagudo‐Cascales et al. 2025) followed by binding studies using, for example, ITC (isothermal titration calorimetry) (Matilla et al. 2020) or MST (microscale thermophoresis) (Chen et al. 2023) to derive the dissociation constants. However, these approaches have all their advantages and disadvantages, and are often labour‐intensive. Here we report a straightforward method to identify chemoeffector signals by screening the signal‐induced alteration of CheA ATPase activity using purified cell membranes. We have used this method to annotate chemoreceptors from P. parafulva PSR09‐11288 with their signals. This protocol has the potential to help close the important knowledge gap that exists in the lacking information on chemoreceptor signals. Results Establishment of a NADH‐Coupled CheA ATPase Activity Screening Method CheA shows basal autokinase activity that is reduced by chemoattractants and stimulated by chemorepellents (Sourjik and Berg 2002a; Wadhams and Armitage 2004; Vaknin and Berg 2007). We established a method to screen potential chemoeffectors by monitoring the ATP hydrolysis activity of purified membrane fractions from cells harbouring chemoreceptors. Since CheY is a soluble protein, it is not present in the assay (Figure 1A). Our previous studies have demonstrated that citrate and L‐malate bind to chemoreceptor MCP2201 of Comamonas testosterone CNB‐1, triggering chemoattraction and chemorepellence, respectively (Hong et al. 2019; Guo et al. 2023). We isolated membrane fractions from strain CNB‐1 lacking all chemoreceptors (CNB‐1∆20) and a strain in which CNB‐1Δ20 was complemented with the gene encoding MCP2201 (CNB‐1∆20‐MCP2201). ATP hydrolysis rates of CNB‐1∆20 membranes were determined to be 1.44 ± 0.19 nM/s per mg protein, representing background levels that are due to the action of other phosphatases. No significant change was observed following the addition of 10 mM citrate or L‐malate, and rates of 1.57 ± 0.28 and 1.54 ± 0.19 nM/s per mg protein (Figure 1B,C) were obtained, respectively. These results showed that ATP hydrolysis in membrane fractions lacking MCP2201 is not modulated by citrate or L‐malate. NADH‐coupled ATPase activity assays to determine the CNB1Δ20‐MCP2201‐mediated ATP hydrolysis. (A) Schematic diagram illustrating the principle of the NADH‐coupled CheA ATPase activity assay. ATP hydrolysis by receptor‐regulated CheA is coupled to NADH oxidation, enabling continuous monitoring of CheA activity. (B, C) ATPase activity of CNB‐1Δ20 cells, either with or without MCP2201, in the presence of different citrate (B) and L‐malate (C) concentrations was measured using the NADH‐coupled assay. (D) Normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) for strain CNB1Δ20‐MCP2201 in the presence of buffer, citrate, L‐malate, or sucrose. Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Data in (B) and (C) were fitted with a logistic model for EC50 determination. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups; *p < 0.05, **p < 0.01, or ***p < 0.001. The “ns” indicates no significant difference. We next studied basal ATP hydrolysis using CNB‐1∆20 membranes harbouring MCP2201. The NADH absorbance changes at 340 nm were continuously monitored in the presence of different chemoeffectors, and the representative NADH depletion plots are shown in Figure 1D. For the membrane fraction of cells harbouring MCP2201, the ATP hydrolysis activity was 2.66 ± 0.27 nM/s per mg protein (Figure 1B,C). The ATP hydrolysis activities were significantly downregulated and upregulated by the addition of varying concentrations of citrate and L‐malate, and reached 2.18 ± 0.09 nM/s per mg protein at 10 mM citrate and 5.75 ± 0.02 nM/s per mg protein at 10 mM L‐malate (Figure 1B,C). The half‐maximal effective concentrations (EC50) for citrate and L‐malate were 0.27 ± 0.08 mM and 0.44 ± 0.14 mM (Figure S1A), respectively, which were comparable to the half‐maximal effective concentration (EC50) of strain CNB‐1∆20‐MCP2201 toward citrate (0.73 ± 0.03 mM) and L‐malate (0.12 ± 0.02 mM) in capillary assays (Figure S1B). These results demonstrate that the binding of citrate and L‐malate to the MCP2201 receptor decreased or enhanced the ATP hydrolysis activity of CheA in a dose‐dependent manner, respectively. To further assess the specificity of the assay, sucrose that was shown previously to not induce a chemotactic response of CNB‐1Δ20‐MCP2201 (Ma et al. 2009) was tested. The addition of sucrose at different concentrations did not cause significant changes in ATP hydrolysis activity (Figure S1A), indicating that the observed ATPase modulation was specific to MCP2201 chemoeffectors. In summary, the ATP hydrolysis activity of CheA in purified membrane fractions could be monitored to identify the potential chemoeffectors. To further verify whether the ligand‐dependent ATPase modulation was specifically associated with CheA activity, we generated the cheA‐deficient mutant CNB‐1Δ20ΔcheA‐MCP2201, and examined its membrane‐associated ATP hydrolysis activity using the same assay. In contrast to the CNB‐1Δ20‐MCP2201 strain, membrane fractions prepared from CNB‐1Δ20ΔcheA‐MCP2201 showed no significant changes in ATP hydrolysis activity upon the addition of citrate (Figure S2A), L‐malate (Figure S2B) or sucrose (Figure S2C) across the tested concentration range (0–10 mM). Consistent with the ATP hydrolysis rate measurements, the NADH depletion plots of CNB‐1Δ20ΔcheA‐MCP2201 membrane fractions showed no apparent differences after the addition of citrate or L‐malate (Figure S2D), further confirming that ligand‐dependent changes in NADH consumption require CheA. The residual ATP hydrolysis activity remained constant regardless of ligand addition, indicating that the ligand‐dependent signal changes observed in the MCP2201‐containing strain require CheA. These results demonstrate that the NADH‐coupled ATPase assay specifically monitors a chemoeffector‐induced modulation of CheA‐associated activity rather than nonspecific ATP hydrolysis from other membrane‐associated enzymes. Identification of Chemoeffectors Using NADH‐Coupled CheA ATPase Activity Screening To check whether our method could identify the chemoeffectors of uncharacterized chemoreceptors, we assayed the responses of purified membrane fractions from P. parafulva PSR09‐11288. This strain was isolated from the rice rhizosphere and produced the antifungal antibiotic phenazine‐1‐carboxylic acid to inhibit the pathogen Rhizoctonia solani responsible for rice sheath blight (Zhang et al. 2019). Its genome encodes a recently‐identified dirammox (direct ammonia oxidation) gene cluster dnfABC that mediates the heterotrophic oxidation of ammonia to nitrogen gas after the assimilation of ammonia (Wu et al. 2021; Qin et al. 2024). Therefore, we focused on testing its responses to some inorganic nitrogen sources, some carbon sources that might serve as energy sources for ammonia oxidation, and some amino acids that serve as both nitrogen sources and energy sources, including citrate, L‐malate, tartrate, succinate, α‐ketoglutarate, L‐glutamate, L‐glutamine, L‐aspartate, L‐asparagine, nitrate, nitrite, and ammonia. The ATP hydrolysis activity of purified membrane fractions from P. parafulva PSR09‐11288 was determined as 2.50 ± 0.18 nM/s per mg protein in the absence of ligand. After the addition of 10 mM L‐aspartate, L‐glutamate, tartrate, succinate, α‐ketoglutarate, nitrate, nitrite and ammonia, the kinase activities did not change significantly and were 2.39 ± 0.32, 2.44 ± 0.41, 2.460 ± 0.37, 2.41 ± 0.23, 2.48 ± 0.19, 2.52 ± 0.13, 2.58 ± 0.38, and 2.74 ± 0.19 nM/s per mg protein, indicating that these molecules did not alter CheA activity under the conditions tested (Figure 2A). Interestingly, the ATPase activities were significantly downregulated by the addition of 10 mM L‐malate, citrate, L‐glutamine, or L‐asparagine, which were 1.95 ± 0.15, 1.75 ± 0.15, 2.19 ± 0.08, or 2.21 ± 0.07 nM/s per mg protein, respectively (Figure 2A). Dose–response measurements showed that the ATP hydrolysis activity decreased with increasing L‐malate or citrate concentration, with EC50 of 2.14 ± 0.26 mM and 1.28 ± 0.45 mM, respectively (Figure S3A,B), suggesting that these chemicals might be the chemoeffectors of P. parafulva PSR09‐11288. In contrast, dose–response measurements for succinate and nitrate revealed no significant changes in ATP hydrolysis activity (Figure S3C,D), indicating that these compounds do not act as chemoeffectors for P. parafulva PSR09‐11288. Chemotactic responses of P. parafulva PSR09‐11288 to different chemoeffectors. (A) ATPase activity of membrane fractions containing chemoreceptors from P. parafulva PSR09‐11288 in response to effectors was measured using the NADH‐coupled assay. (B) Dose‐dependence studies of P. parafulva PSR09‐11288 chemotaxis to citrate (blue), L‐malate (green), L‐glutamine (yellow), and L‐asparagine (orange) using the capillary assay. Data were fitted with a logistic model for EC50 determination. (C) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant toward citrate, L‐malate, L‐glutamine and L‐asparagine using the gradient plate assay. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the bacterial spread, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values > 0.52 indicate attractant responses. Representative images from three independent experiments are shown. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups; *p < 0.05 or ***p < 0.001. The “ns” indicates no significant difference. To verify that ligand‐dependent changes detected by the NADH‐coupled assay were CheA‐dependent, we further analysed membrane fractions of the P. parafulva PSR09‐11288ΔcheA mutant. The basal ATP hydrolysis activity of ΔcheA membrane fractions was determined as 1.43 ± 0.08 nM/s per mg protein, which was reduced compared with that of the wild‐type strain (2.50 ± 0.18 nM/s per mg protein). After the addition of 10 mM citrate or L‐malate, the ATP hydrolysis activities were 1.43 ± 0.11 and 1.44 ± 0.10 nM/s per mg protein, respectively, showing no significant alterations in either activity or NADH depletion plots in the ΔcheA membrane fractions (Figure S4A,B). These results demonstrate that the ligand‐dependent modulation detected by our assay specifically reflects CheA‐mediated chemotactic signalling. We then investigated whether P. parafulva PSR09‐11288 responded to those chemicals using the capillary and gradient plate assays. In the capillary assays, the numbers of P. parafulva PSR09‐11288 cells in the capillary increased as citrate, L‐malate, L‐glutamine, and L‐asparagine concentrations increased. These experiments confirmed that citrate, L‐malate, L‐glutamine, and L‐asparagine triggered chemoattraction with EC50 values of 0.22 ± 0.08 mM, 0.28 ± 0.06 mM, 0.29 ± 0.13 mM and 0.69 ± 0.05 mM, respectively (Figure 2B). In the gradient plate assays, the P. parafulva PSR09‐11288 cells responded to citrate, L‐malate, L‐glutamine and L‐asparagine with chemotactic response indexes (RI) of 0.62 ± 0.04, 0.67 ± 0.03, 0.61 ± 0.02 and 0.67 ± 0.02, respectively (Figure 2C). RI values > 0.52 are indicative of attractant responses (see Methods for details). To further evaluate the reliability of the phenotypic assays and exclude potential false‐negative results, we also tested compounds that showed no response in the NADH‐coupled CheA ATPase assay, including α‐ketoglutarate, sodium acetate, nitrate, nitrite, and ammonia. Consistent with the ATP hydrolysis activity screening results, none of these compounds induced detectable chemotactic responses in either the gradient plate assay or the capillary assay across the tested concentration ranges (Figure S5A,B). Taken together, we conclude that citrate, L‐malate, L‐glutamine, and L‐asparagine are chemoattractants for P. parafulva PSR09‐11288. Identification of Chemoreceptors Responding to Chemoeffectors by Differential Scanning Fluorimetry Assays To identify the chemoreceptors that mediate the responses to the above attractants we first analysed the genome of P. parafulva PSR09‐11288 and found that it encodes 23 putative chemoreceptors. Of these, 18 have periplasmic sensor domains, 4 are membrane‐anchored receptors with a cytosolic PAS sensor domain that are likely involved in energy taxis or redox sensing, whereas one receptor lacked a sensor domain. The sensor domains of the 18 receptors with extracellular sensor domains were cloned, expressed, purified, and subjected to ligand binding studies using the differential scanning fluorimetry based thermal shift assay (Monteagudo‐Cascales et al. 2025). The tested compounds included L‐malate, citrate, succinate, α‐ketoglutarate, L‐asparagine, L‐glutamine, L‐aspartate, L‐glutamate, tartrate, fumarate, pyruvate, D‐malate, sodium acetate, nitrite, nitrate, and ammonia. The presence of 1 mM L‐malate and citrate increased the midpoint of protein unfolding transition (Tm) of B2J77_03605 sensor domain by 10°C and 3°C, respectively (Figure 3A,B). In addition, 1 mM L‐glutamine or L‐asparagine increased the Tm of the B2J77_13170 sensor domain by 13°C (Figure 3A,C). In contrast, the remaining chemicals caused Tm shifts inferior to 2°C. Similarly, the ΔTm values observed for the sensor domains of the remaining 16 chemoreceptors were also less than 2°C (Figure S6), indicating that these chemoreceptors do not bind tested chemicals. These results suggested that chemoreceptors B2J77_03605 and B2J77_13170 are the receptors for L‐malate/citrate and L‐glutamine/L‐asparagine, respectively. Identification of chemoeffectors for the B2J77_03605 and B2J77_13170 chemoreceptors. (A) Differential scanning fluorimetry based primary screening of ligand binding. Thermal‐shift assays were performed using purified sensor domains of 18 chemoreceptors from P. parafulva PSR09‐11288 in the presence of citrate, L‐malate, L‐asparagine, or L‐glutamine. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). (B, C) First derivatives of thermal unfolding curves and the corresponding ΔTm values for the B2J77_03605 (B) and B2J77_13170 sensor domains (C) in the absence or presence of their identified ligands (citrate, L‐malate, L‐asparagine, and L‐glutamine). Chemoreceptor B2J77_03605 Responds to L‐Malate and Citrate The sensor domain of chemoreceptor B2J77_03605 shares the highest sequence identity, 79.7%, with the sensor domain of McpS from Pseudomonas putida KT2440, a well‐studied chemoreceptor involved in the sensing of tricarboxylic acid (TCA) cycle intermediates (Lacal et al. 2010; Pineda‐Molina et al. 2012). The ligand‐binding affinities of B2J77_03605 sensor domain were determined using ITC. The results showed that the B2J77_03605 sensor domain binds citrate and L‐malate with dissociation constants (K D) of 88 ± 12 μM and 68 ± 5 μM (Figure 4A), respectively. In contrast, no binding was detected for succinate or acetate (Figure S7), indicating selectivity for specific TCA cycle intermediates despite sequence similarity to McpS, which recognizes a broader range of ligands including succinate (Pineda‐Molina et al. 2012). These results agreed with data showing that succinate and acetate did not significantly shift the Tm value (Figure S6) nor induced chemotaxis (Figure S7). Functional and structural analysis of the B2J77_03605 sensor domain. (A) Microcalorimetric titrations of 100 μM B2J77_03605 sensor domain with 1.5 mM citrate (pink) and 1.5 mM L‐malate (green). The raw data (upper panel) and integrated binding heats (lower panel) are shown. Data were fitted with a model for the binding of a single compound to a macromolecule. (B) Structural comparison of the ligand‐binding pockets of the B2J77_03605 (pink) and McpS sensor domains (cyan). Residues involved in ligand coordination are shown in stick mode. (C) Multiple sequence alignment the B2J77_03605 and McpS sensor domains. Conserved residues are shaded in red. Structural modelling using AlphaFold3 predicted that the B2J77_03605 sensor domain adopts a helix‐bundle module (HBM) fold characteristic of McpS sensor domain and related chemoreceptors. In the McpS sensor domain (Pineda‐Molina et al. 2012), two distinct ligand‐binding sites have been characterized: a malate‐binding pocket formed by helices α1 and the C‐terminal part of α6, involving residues R60, R63, R254, and T258 from one monomer and Q65 from the adjacent monomer; and an acetate‐binding site at the interface between helices α4 and α6, where R183, R187, and Y236 form hydrogen bonds with the ligand (Figure 4B). McpS recognizes multiple TCA cycle intermediates through these sites, including strong attractants (malate, succinate, fumarate, oxaloacetate) and weak attractants (citrate and isocitrate) (Pineda‐Molina et al. 2012). Those potential residues involved in L‐malate recognition in the B2J77_03605 sensor domain (R60, R63, Q65, R254, and T258) are fully conserved, consistent with the observed L‐malate‐binding activity (Figure 4C). Though we did not observe the response of B2J77_03605 toward acetate, those residues for acetate binding in McpS are also fully conserved in the B2J77_03605 sensor domain. Together, these results demonstrate that the B2J77_03605 sensor domain functions as a selective chemoreceptor for L‐malate and citrate. The conservation of the malate‐binding pocket explains its affinity for these ligands. Chemoreceptor B2J77_13170 Responds to L‐Glutamine and L‐Asparagine The inspection of the SwissProt database showed that the B2J77_13170 sensor domain shared the highest sequence identities with several well‐studied amino acid chemoreceptors, including 53.6% with PctB, 52% with PctA, and 41.9% with PctC from Pseudomonas aeruginosa (Taguchi et al. 1997; Gavira et al. 2020), as well as 45.6% with CtaA from Pseudomonas fluorescens Pf0‐1 (Ud‐Din et al. 2020) (Figure S8). All of these homologous proteins contain a dCache (double calcium channels and chemotaxis receptor) sensor domain, characterized by a long N‐terminal helix followed by two stacked α/β modules. These domains sense different amino acids at the membrane‐distal module of the dCache domain. Microcalorimetric titrations showed that the B2J77_13170 sensor domain binds L‐glutamine with a K D of 3.7 ± 0.1 μM and L‐asparagine with a K D of 3.3 ± 0.1 μM (Figure 5A). In contrast, no binding was observed for L‐aspartate or L‐glutamate (Figure S9), which agrees with the observation that P. parafulva PSR09‐11288 or B2J77_13170 did not respond to L‐glutamate or L‐aspartate in NADH‐coupled assays, differential scanning fluorimetry assays, or chemotaxis assays (Figure S9). In contrast, PctA exhibits a broad ligand spectrum responsive to most proteinogenic amino acids, and PctB and PctC show narrower specificity, with strong preferential binding for L‐glutamine and γ‐aminobutyrate, respectively (Gavira et al. 2020). This result indicated the specificity of B2J77_13170 for the amide‐containing amino acids L‐glutamine and L‐asparagine, a ligand profile similar to that of PctA or McpB of Bacillus subtilis (Glekas et al. 2010). Functional and structural analysis of the B2J77_13170 sensor domain. (A) Microcalorimetric titrations of 100 μM B2J77_13170 sensor domain with 1.5 mM L‐glutamine (orange) and 1.5 mM L‐asparagine (blue). The raw data (upper panel) and integrated integrate heats (lower panel) are shown, and are fitted with a model for the binding of a single compound to a macromolecule. (B) Structural comparison of the ligand‐binding pockets of the B2J77_13170 (cyan) with the PctB sensor domain (pink). Residues involved in ligand coordination are shown in stick mode. (C) Multiple sequence alignment of the B2J77_13170 and PctB sensor domains. Conserved residues are shaded in red. To obtain structural insights, we modelled the B2J77_13170 sensor domain using AlphaFold3. The predicted structure adopts the characteristic dCache fold. Structural comparison with PctB revealed local differences between the two ligand binding sites (Figure 5B,C). In PctB, the amino and carboxyl groups of bound L‐glutamine are coordinated by residues Y121, R126, W128, Y144, E146, and D173. These residues are fully conserved in the B2J77_13170 sensor domain. In PctB, the L‐glutamine sidechain interacts with residues Y109, S115, and D116. Of these, only D116 is conserved in the B2J77_13170 sensor domain, whereas Y109 and S115 of PctB are replaced by residues F107 and A113. These differences in side‐chain coordination may explain differences in the PctB and B2J77_13170 ligand specificity: whereas B2J77_13170 binds L‐glutamine and L‐asparagine, PctB binds preferentially L‐glutamine but does not interact with L‐asparagine (Rico‐Jiménez et al. 2013). Discussion In this study, we established a NADH‐coupled CheA ATPase activity assay to screen for potential chemoeffectors and integrated this method into a pipeline for identifying and characterizing receptor‐ligand pairs. By combining a NADH‐coupled CheA ATPase assay for high‐throughput screening with differential scanning fluorimetry (Pantoliano et al. 2001; Wang et al. 2025), capillary assay (Cui et al. 2025), and ITC (Matilla et al. 2020; Guo et al. 2023; Cui et al. 2025) for validation, we successfully identified L‐malate, citrate, L‐glutamine, and L‐asparagine as chemoeffectors recognized by two previously uncharacterized chemoreceptors from P. parafulva PSR09‐11288. B2J77_03605, an HBM‐type chemoreceptor (Ortega and Krell 2014), selectively binds L‐malate and citrate, whereas B2J77_13170, a dCache‐type receptor (Ortega et al. 2017), exhibits high affinity for L‐glutamine and L‐asparagine. Sequence comparison suggested that the P. parafulva receptors shared the residues that were essential for signal binding by the TCA cycle intermediate chemoreceptor McpS (Pineda‐Molina et al. 2012) and the L‐glutamine‐sensing chemoreceptor PctB (Gavira et al. 2020). Both B2J77_03605 and B2J77_13170 had a narrower ligand spectrum than their McpS and PctB homologues. The structural rationale for the narrower ligand specificity of both B2J77_03605 and B2J77_13170 requires further studies. The NADH‐coupled CheA ATPase assay (Garrity and Ordal 1997; Inclán et al. 2008) of purified membrane fractions served as the primary screening tool in our workflow. This method offers several practical advantages that make it particularly suitable for initial chemoeffectors identification. First, it is a cell‐free system, allowing application to both aerobic and anaerobic bacteria and potentially complex microbial systems without the need for culture maintenance or genetic manipulation. Second, it avoids the use of radioactive compounds. Third, the assay requires only small amounts of candidate ligands, substantially reducing reagent costs when screening large chemical libraries. Fourth, its compatibility with microplate formats provides the potential for high‐throughput screening of candidate chemoeffectors and may facilitate future large‐scale screening efforts. These features collectively establish the NADH‐coupled ATPase assay as a promising platform for initial chemoeffector discovery across diverse bacterial species, with future applications in larger‐scale screening requiring, which, however, would require further optimization and validation. Both the NADH‐coupled CheA ATPase assay and the widely used FRET‐based chemotaxis assay monitor CheA kinase activity, although through different experimental principles (Sourjik and Berg 2002b; Vaknin and Berg 2004). The FRET assay provides a readout of CheA activity by monitoring phosphorylation‐dependent interactions between CheY and CheZ in intact cells, which depend on CheY phosphorylation by CheA, thereby enabling high‐resolution analysis of signalling dynamics and receptor adaptation. (Sourjik and Berg 2002b; Vaknin and Berg 2004). In contrast, the NADH‐coupled assay provides a biochemical readout of apparent CheA ATPase activity by measuring NADH oxidation coupled to ATP regeneration following ATP hydrolysis in isolated membrane fractions. Because the assay lacks the soluble adaptation proteins CheB and CheR, the measured response primarily reflects receptor‐dependent modulation of CheA activity rather than CheB/CheR‐mediated adaptation‐associated signalling dynamics (Sourjik 2004). Therefore, this approach captures the initial biochemical regulation of CheA activity while avoiding the complexity introduced by downstream adaptation processes. Moreover, unlike the FRET assay, which relies on fluorescent reporters and specialized, relatively expensive instrumentation, the NADH‐coupled assay requires only a standard spectrophotometer or microplate reader, making it a simple, scalable, and readily accessible platform for high‐throughput chemoeffector screening. This method has several inherent limitations that should be acknowledged. Further optimization of assay performance, including systematic evaluation of reproducibility, signal‐to‐background ratio, and screening parameters, will be necessary for future large‐scale applications. Although it detects ligand‐induced changes in apparent CheA ATPase activity (Rule et al. 2016), it does not directly identify the specific chemoreceptor responsible for the observed response. Receptor assignment therefore requires complementary approaches, including the construction and analysis of gene knockout mutants (Datsenko and Wanner 2000), or direct biophysical characterization using purified ligand‐binding domains, such as ITC or surface plasmon resonance (Patching 2014). In addition, because the assay measures the integrated activity of the membrane‐associated chemosensory array, multiple receptors responding to the same ligand may contribute to the observed signal, potentially generating composite responses that require further receptor‐level validation. Furthermore, the assay captures only the initial biochemical event in chemoreceptor signalling and does not reflect downstream behavioural responses in living cells. Potential background activity from other membrane‐associated enzymes, including basal activities of sensor histidine kinases, may contribute to the measured signal (Parkinson and Kofoid 1992; Urao et al. 2000). Therefore, it should be viewed as a screening tool rather than a standalone functional assay, and positive hits require validation through orthogonal methods and ultimately bacterial chemotaxis assays (Adler 1966, 1969; Mao et al. 2003). Beyond the methodological implications, the four attractants identified in this study, citrate, L‐malate, glutamine, and asparagine, are not only ligands recognized by two chemoreceptors but also ecologically relevant signals that may influence the ecological behaviour of P. parafulva PSR09‐11288 in its native rice rhizosphere habitat. Root exudates contain diverse organic acids and amino acids, including citrate, malate, glutamine, and asparagine, which serve as important nutrient sources and signalling molecules mediating plant–microbe interactions (Badri and Vivanco 2009; Sasse et al. 2018). These compounds can also act as chemotactic cues that guide rhizosphere‐associated bacteria toward plant roots (de Weert et al. 2002; Oku et al. 2012). Chemotaxis toward root‐derived metabolites has been shown to contribute to root colonization and ecological fitness in several Pseudomonas species (de Weert et al. 2002; Oku et al. 2012). For P. parafulva PSR09‐11288, that exhibits biocontrol activity against soilborne fungal pathogens (Zhang et al. 2019), the ability to sense and respond to these root‐derived compounds may facilitate root localization, persistence in the rhizosphere, and interactions with surrounding microbial communities. Thus, our findings advance our knowledge on how P. parafulva PSR09‐11288 perceives ecologically relevant signals, linking chemoreceptor‐mediated sensing with rhizosphere colonization and biocontrol potential. Overall, the integrated workflow presented here combines high‐throughput biochemical screening, biophysical validation, structural modelling, and sequence analysis to establish a scalable platform for chemoreceptor signal identification. By enabling the discovery of novel receptor‐ligand interactions and providing mechanistic insights into ligand recognition, this approach offers a useful framework for investigating the molecular basis of bacterial chemotaxis, environmental adaptation, and interspecies interactions. Conclusion In conclusion, we established a simple, cost‐effective, and scalable NADH‐coupled CheA ATPase assay for the systematic identification of bacterial chemoeffectors. By integrating biochemical screening with differential scanning fluorimetry, isothermal titration calorimetry, and structural modelling, we identified four chemoattractants of P. parafulva PSR09‐11288 and assigned their recognition to two chemoreceptors belonging to the HBM and dCache sensor‐domain families. The successful application of this workflow to structurally distinct receptor families demonstrates its broad applicability for decoding chemosensory functions. Because the assay requires only standard spectrophotometric instrumentation and directly reports receptor‐dependent modulation of CheA activity, it provides a practical platform for high‐throughput chemoeffector discovery. This approach should facilitate systematic mapping of chemoreceptor–ligand interactions and advance our understanding of how environmental signals shape bacterial chemotactic behaviour and ecological adaptation. Materials and Methods Strains, Plasmids, Media, and Cultivation All bacterial strains and plasmids used in this study are listed in Table S1. Genetic complementation in Comamonas testosteroni CNB‐1 was conducted using pBBR1MCS2 (Wu et al. 2006). The cheA deletion mutants of C. testosteroni CNB‐1Δ20 and P. parafulva PSR09‐11288 were constructed by homologous recombination and the suicide plasmid pK18mobsacB. Briefly, homologous fragments flanking the cheA gene were amplified by PCR and cloned into pK18mobsacB. The resulting recombinant plasmids were introduced into the corresponding strains, and allelic exchange mutants were obtained through two‐step homologous recombination. The cheA deletion mutants ( C. testosteroni CNB‐1Δ20ΔcheA and P. parafulva PSR09‐11288ΔcheA) were verified by PCR amplification and sequencing. All strains used in this study, including wild‐type strains, deletion mutants, and complemented strains, were cultivated in Luria—Bertani (LB) broth or on LB plates with 1.5% (w/v) agar (Lablead, Beijing, China) at 30°C. Capillary Assay Capillary assays were performed as described previously with minor modifications (Liu et al. 2020; Cui et al. 2025). Overnight cultures of bacteria were grown in LB medium, and 1% of the culture was transferred to fresh LB medium and incubated at 30°C with shaking at 180 rpm until the OD600 reached 1.0. The cultures were harvested by centrifugation at 5000 rpm for 1 h, washed three times with MSB (1 g/L Na2HPO4·12H2O, 0.5 g/L KH2PO4, 0.03 g/L MgSO4·7H2O and 1 g/L NH4Cl pH 8.0) medium, and resuspended in MSB medium to a final OD600 of 0.8. The bacterial suspension was placed into a 5 mL tube (ACMEC, Shanghai, China) to ensure consistent OD values in each well of a multi‐channel pipette. Then, 100 μL of L‐malate, citrate, L‐asparagine or L‐glutamine were added to a 1 mL syringe (STEEMA, Beijing, China), and the syringe needle was inserted into one end of the pipette. After incubation at room temperature for 1 h, the pipette containing the bacterial suspension was removed, and the contents of the syringe were injected into 900 μL of MSB medium. Serial dilutions were plated onto LB plates and incubated overnight prior to colony‐forming unit (CFU) counting. Each treatment was performed using three independent biological replicates. Gradient Soft‐Agar Swim Plate Assay Semisolid‐agar assay was performed as previously described with small changes (Yamamoto and Imae 1993; Pham and Parkinson 2011; Guo et al. 2023; Cui et al. 2025). Briefly, agar plugs containing chemical effectors were prepared by mixing molten 1.5% (wt/vol) agar with 10 mM of either L‐malate, citrate, L‐asparagine, or L‐glutamine (each effector was tested separately) and placing them into the centre of a semisolid agar Petri dish containing 0.25% (wt/vol) agar in MSB medium. Growth of C. testosteroni CNB‐1 or P. parafulva PSR09‐11288 cells were grown in LB medium to an OD600 of 0.8, washed and resuspended in 50 μL of MSB medium. 0.5 μL aliquots of bacterial suspensions were inoculated at 2 cm away from the centre of the ligand plug. Plates were incubated at 30°C for 24 h and the distance from the inoculation sites to the edges of the colony closest (D1) and furthest (D2) to the centre of the agar plug were measured. The response index (RI) values were determined as follows: RI = D1/(D1 + D2), as described previously (Pham and Parkinson 2011). RI values > 0.52 and lower than 0.48, respectively, are indicative of attractant and repellent responses. respectively. Preparation of Bacterial Membranes Containing Methyl‐Accepting Chemotaxis Receptor Proteins C. testosteroni CNB‐1Δ20‐MCP2201 and C. testosteroni CNB‐1Δ20ΔcheA‐MCP2201 were cultured overnight at 30°C in LB medium supplemented with 50 μg/mL kanamycin (Coolaber, Beijing, China), whereas P. parafulva PSR09‐11288 and P. parafulva PSR09‐11288ΔcheA were cultured in LB medium without antibiotics. Subsequently, 1 mL of the overnight culture was inoculated into 1 L of fresh LB medium and incubated at 30°C with shaking at 180 rpm for 18 h. Cells were harvested by centrifugation at 5000 rpm for 1 h and resuspended in lysis buffer containing 50 mM Tris/HCl, pH 7.5, 200 mM NaCl, and 20 mM imidazole. Cell lysis was performed on ice using sonication (SCIENTZ, Ningbo, China). The lysate was clarified by centrifugation at 15,000 rpm for 0.5 h. The resulting supernatant was subjected to ultracentrifugation at 41,000 rpm for 2 h using a 70‐Ti rotor in an ultracentrifuge (Beckman, Indianapolis, IN, USA). The membrane pellet containing the methyl‐accepting chemotaxis proteins was resuspended in HEPES/KCl buffer and lysed again on ice by sonication (SCIENTZ, Ningbo, China). Protein concentration was determined and the sample stored for further use (Garrity and Ordal 1997). NADH‐Coupled ATPase Activity Assays A reaction master mix (2 mL) was prepared containing 1928 μL of 0.1 mol/L HEPES/KOH, 20 μL of 0.36 mol/L magnesium acetate, 12 μL of 100 mM phosphoenolpyruvate (Solarbio, Beijing, China, Cat. No. P8800‐100 mg), 24 μL of lactate dehydrogenase (Solarbio, Beijing, China, Cat. No. 9001‐60‐9), and 16 μL of pyruvate kinase (Solarbio, Beijing, China, Cat. No. P8820‐1KU). Then, 173 μL of this master mix was aliquoted into a 200 μL quartz cuvette. Subsequently, 10 μL of effector, 1 μL of chemoreceptors membrane fraction suspension were added, and the mixture was gently pipetted to homogenize. It was then incubated on ice for 5 min. After incubation, 4 μL of 100 mM NADH, 2 μL of 79 mM ATP (Solarbio, Beijing, China, Cat. No. IA1310‐20 mg), and 10 μL of 0.1 mol/L HEPES/KOH pH 7.5 were added, bringing the total reaction volume to 200 μL. The solution was mixed gently by pipetting and immediately transferred to a spectrophotometer. The absorbance of NADH at 340 nm was monitored continuously for 360 s, with measurements taken every second (Inclán et al. 2008). The specific activity was calculated using the following equation: where: ΔA340 = A360 − A0 represents the difference at the end (t = 360 s) of the reaction with respect to the start (t = 0 s). Because NADH is consumed during the assay, absorbance decreases over time, making ΔA340 negative; thus, the calculation uses (A0 − A360) to yield a positive value; Δt = 360 s, the total monitoring duration; ε = 6.22 mM−1 cm−1,the molar extinction coefficient of NADH at 340 nm (with a 1 cm path length); Cprotein is the final concentration of membrane protein in the reaction mixture (mg/mL). Protein Expression and Purification The coding sequence for the 18 chemoreceptors sensor domain was cloned into the pET‐22b expression vector (Merck, Darmstadt, Germany), each fused with a C‐terminal His6‐tag. The plasmids were transformed into E. coli BL21 (DE3) (Vazyme, Beijing, China). The cells were cultured in LB medium containing 50 μg/mL ampicillin (Coolaber, Beijing, China) at 37°C. When the OD600 reached 0.8, protein expression was induced by the addition of Isopropyl β‐d‐1‐thiogalactopyranoside at a final concentration of 0.2 mM, followed by incubation at 16°C for 20 h. The cells were harvested by centrifugation at 5000 rpm for 1 h, resuspended in lysis buffer (50 mM Tris/HCl pH 7.5, 200 mM NaCl, and 15 mM imidazole), lysed on ice by sonication (SCIENTZ, Ningbo, China), and centrifuged at 15,000 rpm for 1 h. The supernatant was loaded on nickel affinity resin columns (Ni‐NTA, Qiagen, Dusseldorf, Germany), and the flow‐through was reloaded for a total of five cycles. The resin was washed five times with washing buffer (50 mM Tris/HCl, pH 7.5, 300 mM NaCl, 25 mM imidazole). The protein was eluted using elution buffer containing 50 mM Tris/HCl, pH 7.5, 300 mM NaCl, and 300 mM imidazole. The eluate was further purified using a Superdex 75 10/300 GL size‐exclusion chromatography column (GE Healthcare, Pittsburgh, PA, USA) and eluted with buffer containing 50 mM Tris/HCl, pH 7.5, 150 mM NaCl. Fractions containing the target protein were concentrated for Differential scanning fluorimetry and Isothermal titration calorimetry assays. Differential Scanning Fluorimetry Differential scanning fluorimetry measurements were performed using a Bio‐Rad CFX96 Real‐Time PCR Instrument (Roche, USA). Ligands (citate, L‐malate, D‐malate, succinate, α‐ketoglutarate, tartrate, fumarate, pyruvate, sodium acetate, sodium nitrate, sodium nitrite, and ammonium chloride, L‐aspartate, L‐asparagine, L‐glutamine, and L‐glutamate) were dissolved in buffer (50 mM of Tris/HCL, 150 mM of NaCl, pH 7.5) to a final concentration of 10 mM. Screening was performed using a 96‐well plate with a total reaction volume of 20 μL per well. Each well contained 1 mM of the small‐molecule compound, 0.1 mg/mL protein, and a final 5× concentration of the differential scanning fluorimetry dye (SYPRO Orange protein gel stain, Thermo Fisher Scientific, Waltham, MA, USA, Cat. No. S6650). Control wells contained protein without added small molecules and small molecules without protein. The samples were heated from 4°C to 95°C at a scanning rate of 1°C/min while monitoring the change in fluorescence (FRET 480 nm–620 nm) (Pantoliano et al. 2001; Wang et al. 2025). Isothermal Titration Calorimetry ITC experiments were conducted using an Affinity ITC instrument (TA Instruments, New Castle, DE, USA) at 25°C. To assess ligand binding affinity, 100 μM of B2J77_03605 sensor domain protein or 50 μM of B2J77_13170 sensor domain protein were placed into the sample cell and titrated with aliquots of 1.5 mM of either citrate/L‐malate or L‐glutamine/L‐asparagine dissolved in 50 mM of Tris/HCL, 150 mM of NaCl, pH 7.5. Data were analysed by the NanoAnalyze software (Versuion 3.11.0) package (TA Instruments, New Castle, DE, USA) (Matilla et al. 2020; Cui et al. 2025). Protein Sequence Alignment and Conservation Analysis Multiple sequence alignment of protein sequences was performed using MEGA11 (Molecular Evolutionary Genetics Analysis) with the ClustalW algorithm (Thompson et al. 1994). The resulting alignments were visualized and annotated using ESPript 3.2 to display sequence conservation and secondary structure features (Robert and Gouet 2014). Structural Modelling and Visualization The Sensor Domain Model of B2J77_03605 or B2J77_13170 Were Predicted Using AlphaFold3 (Abramson et al. 2024) With the Multiple Sequence Alignment (MSA) Option Enabled. The Experimentally Resolved Structures of McpS and PctB Were Retrieved From the Protein Data Bank (PDB). Three‐Dimensional Structural Representations and Analyses Were Performed Using PyMOL (http://pymol.org). Statistical Analysis All statistical analyses were performed using GraphPad Prism 9 (GraphPad Software Inc., USA) unless otherwise stated. Data are means and standard deviations derived from three independent biological replicates. Statistical significance *p < 0.05; **p < 0.01; ***p < 0.001. Dose–response plots from capillary assays were fitted using a logistic function in OriginPro 2026 (OriginLab Corporation, Northampton, MA, USA). All of the statistical details of experiments can be found in the figure legends. Author Contributions Rui Cui: conceptualization, methodology, investigation, data curation, formal analysis, writing – original draft, writing – review and editing, validation, visualization. Ni‐Ye You: investigation, data curation. Jie Li: investigation, writing – original draft, formal analysis, data curation, validation, visualization. Tino Krell: conceptualization, supervision, funding acquisition, writing – original draft, writing – review and editing. De‐Feng Li: conceptualization, supervision, funding acquisition, writing – original draft, writing – review and editing, methodology. Funding This work was supported by Strategic Priority Research Program of the Chinese Academy of Sciences, XDB0810000. National Natural Science Foundation of China, 31870037. Agencia Estatal de Investigación, PID2023‐146216NB‐I00. China Postdoctoral Science Foundation, 2026M793112. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1: Chemotactic responses of CNB‐1∆20‐MCP2201 to citrate, L‐malate and sucrose. (A) Dose‐dependence of ATP hydrolysis by CNB‐1∆20‐MCP2201 in response to increasing citrate, L‐malate and sucrose concentrations. (B) Dose‐dependence of chemotactic responses of CNB‐1∆20‐MCP2201 toward increasing concentrations of citrate, L‐malate and sucrose using the quantitative capillary chemotaxis assay. Data are fitted with a logistic model for EC50 determination. Error bars indicate the standard errors of three independent replicates, shown as mean ± SD. Figure S2: ATP hydrolysis by CNB‐1∆20∆cheA‐MCP2201 monitored by the NADH‐coupled ATPase activity assay. (A–C) The ATPase activity of CNB‐1∆20∆cheA‐MCP2201 to increasing concentrations of citrate (A), L‐malate (B) and sucrose (C) in the NADH‐coupled ATPase activity assay. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups; The “ns” indicates no significant difference. (D) The normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) of strain CNB1∆20∆cheA‐MCP2201 treated with buffer, citrate, L‐malate, or sucrose. Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Figure S3: ATP hydrolysis by the Pseudomonas parafulva PSR09‐11288 using the NADH‐coupled ATPase activity assay. (A) The ATPase activity of P. parafulva PSR09‐11288 toward different concentrations of citrate and L‐malate as determined by the NADH‐coupled ATPase activity assay. (B) Dose–response plots for the ATPase activity of P. parafulva PSR09‐11288 in response to citrate and L‐malate. Data from (A) were fitted with a logistic function to determine the EC50 values. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). (C) The ATPase activity of P. parafulva PSR09‐11288 toward different concentrations of succinate and nitrate using the NADH‐coupled ATPase activity assay. (D) The normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) for strain P. parafulva PSR09‐11288 in the presence of buffer and 10 mM each of citrate, L‐malate, succinate, L‐glutamine, L‐asparagine and nitrate. Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups; *p < 0.05 or **p < 0.01. The “ns” indicates no significant difference. Figure S4: ATP hydrolysis by the P. parafulva PSR09‐11288∆cheA mutant using NADH‐coupled ATPase activity assay. (A) The ATPase activity of P. parafulva PSR09‐11288∆cheA in the presence of different concentrations of citrate and L‐malate using the NADH‐coupled ATPase activity assay. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups. The “ns” stands for not significant. (B) The normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) for strain P. parafulva PSR09‐11288∆cheA (WT∆cheA) in the presence of buffer, citrate (10 mM) and L‐malate (10 mM). Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Figure S5: Chemotactic responses of P. parafulva PSR09‐11288 toward different chemoeffectors. (A) Dose‐dependence studies of P. parafulva PSR09‐11288 chemotaxis to α‐ketoglutarate, sodium tartrate, ammonia, nitrate and nitrite using the capillary assay. Error bars indicate the standard deviation (SD) from five independent replicates (shown as mean ± SD). (B) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant toward buffer (control), α‐ketoglutarate, sodium tartrate, ammonia, nitrate and nitrite using the gradient soft‐agar plate assay. Representative images from three independent experiments are shown. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the colony, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values of 0.52–0.48 indicate no significant response. Figure S6: Differential scanning fluorimetry screening of ligand binding by purified sensor domains of 18 chemoreceptors from P. parafulva PSR09‐11288 toward succinate, α‐ketoglutarate, L‐aspartate, L‐glutamate, tartrate, fumarate, pyruvate, D‐malate, sodium acetate, ammonia, nitrate and nitrite. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Figure S7: Chemotactic responses of P. parafulva PSR09‐11288 to succinate and sodium acetate. (A, B) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant to succinate (A) and sodium acetate (B) on gradient soft‐agar plate. Representative images from three independent experiments are shown. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the colony, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values 0.52–0.48 indicate no significant response. (C, D) Microcalorimetric titration of 100 μM B2J77_03605 sensor domain with aliquots of 1.5 mM succinate (C) and sodium acetate (D). The raw data (upper panel) and integrated binding heats (lower panel) are shown. In panel C, the buffer control for the succinate titration is shown in blue. Figure S8: Multiple sequence alignment of the B2J77_13170 sensor domain with the homologous domains from the PctA, PctB and PctC chemoreceptors of P. aeruginosa (Gavira et al. 2020), CtaA from Pseudomonas fluorescens (Ud‐Din et al. 2020), and McpA from Bacillus velezensis (Feng et al. 2022). The PDB ID of the corresponding 3D structures is provided in brackets. Conserved residues (shaded in red) include E67, Y99, G105, T108, P111, P116, P121, R124, W126, Y127, A130, T139, P141, G163, G167, V168, G191, A193, F194, and P206 (numbering according to the sensor domain B2J77_13170). Figure S9: Chemotactic responses of P. parafulva PSR09‐11288 to L‐aspartate and L‐glutamate. (A, B) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant to L‐aspartate (A) and L‐glutamate (B) on gradient soft‐agar plate. Representative images from three independent experiments are shown. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the bacterial spread, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values 0.52–0.48 indicate no significant response. (C, D) Microcalorimetric titration of 100 μM B2J77_13170 sensor domain with aliquots of 1.5 mM L‐aspartate (C) and L‐glutamate (D). The raw data (upper panel) and integrated binding heats (lower panel) are shown. Table S1: Bacterial strains and plasmids used in this study. Acknowledgements This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB0810000), the National Natural Science Foundation of China (31870037). Work was also supported by grant PID2023‐146216NB‐I00 to T.K. from the Spanish Ministry for Science, Innovation and Universities/Agencia Estatal de Investigación https://doi.org/10.13039/501100011033. R.C. is supported by the China Postdoctoral Science Foundation (2026M793112). We thank Dr. Wei Zhang of the Institute of Microbiology, CAS, for assisting with the ITC experiment. Contributor Information Tino Krell, Email: tino.krell@eez.csic.es. De‐Feng Li, Email: lidefeng@im.ac.cn. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References Associated Data Supplementary Materials Figure S1: Chemotactic responses of CNB‐1∆20‐MCP2201 to citrate, L‐malate and sucrose. (A) Dose‐dependence of ATP hydrolysis by CNB‐1∆20‐MCP2201 in response to increasing citrate, L‐malate and sucrose concentrations. (B) Dose‐dependence of chemotactic responses of CNB‐1∆20‐MCP2201 toward increasing concentrations of citrate, L‐malate and sucrose using the quantitative capillary chemotaxis assay. Data are fitted with a logistic model for EC50 determination. Error bars indicate the standard errors of three independent replicates, shown as mean ± SD. Figure S2: ATP hydrolysis by CNB‐1∆20∆cheA‐MCP2201 monitored by the NADH‐coupled ATPase activity assay. (A–C) The ATPase activity of CNB‐1∆20∆cheA‐MCP2201 to increasing concentrations of citrate (A), L‐malate (B) and sucrose (C) in the NADH‐coupled ATPase activity assay. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups; The “ns” indicates no significant difference. (D) The normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) of strain CNB1∆20∆cheA‐MCP2201 treated with buffer, citrate, L‐malate, or sucrose. Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Figure S3: ATP hydrolysis by the Pseudomonas parafulva PSR09‐11288 using the NADH‐coupled ATPase activity assay. (A) The ATPase activity of P. parafulva PSR09‐11288 toward different concentrations of citrate and L‐malate as determined by the NADH‐coupled ATPase activity assay. (B) Dose–response plots for the ATPase activity of P. parafulva PSR09‐11288 in response to citrate and L‐malate. Data from (A) were fitted with a logistic function to determine the EC50 values. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). (C) The ATPase activity of P. parafulva PSR09‐11288 toward different concentrations of succinate and nitrate using the NADH‐coupled ATPase activity assay. (D) The normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) for strain P. parafulva PSR09‐11288 in the presence of buffer and 10 mM each of citrate, L‐malate, succinate, L‐glutamine, L‐asparagine and nitrate. Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups; *p < 0.05 or **p < 0.01. The “ns” indicates no significant difference. Figure S4: ATP hydrolysis by the P. parafulva PSR09‐11288∆cheA mutant using NADH‐coupled ATPase activity assay. (A) The ATPase activity of P. parafulva PSR09‐11288∆cheA in the presence of different concentrations of citrate and L‐malate using the NADH‐coupled ATPase activity assay. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Statistical significance was determined by two‐tailed unpaired Student's t‐test comparing ligand‐treated groups with the non‐treated groups. The “ns” stands for not significant. (B) The normalized time‐course plots of NADH absorbance at 340 nm (0–360 s) for strain P. parafulva PSR09‐11288∆cheA (WT∆cheA) in the presence of buffer, citrate (10 mM) and L‐malate (10 mM). Absorbance at each time point was normalized to the initial reading at t = 0. Representative images from three independent experiments are shown. Figure S5: Chemotactic responses of P. parafulva PSR09‐11288 toward different chemoeffectors. (A) Dose‐dependence studies of P. parafulva PSR09‐11288 chemotaxis to α‐ketoglutarate, sodium tartrate, ammonia, nitrate and nitrite using the capillary assay. Error bars indicate the standard deviation (SD) from five independent replicates (shown as mean ± SD). (B) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant toward buffer (control), α‐ketoglutarate, sodium tartrate, ammonia, nitrate and nitrite using the gradient soft‐agar plate assay. Representative images from three independent experiments are shown. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the colony, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values of 0.52–0.48 indicate no significant response. Figure S6: Differential scanning fluorimetry screening of ligand binding by purified sensor domains of 18 chemoreceptors from P. parafulva PSR09‐11288 toward succinate, α‐ketoglutarate, L‐aspartate, L‐glutamate, tartrate, fumarate, pyruvate, D‐malate, sodium acetate, ammonia, nitrate and nitrite. Error bars indicate the standard deviation (SD) from three independent replicates (shown as mean ± SD). Figure S7: Chemotactic responses of P. parafulva PSR09‐11288 to succinate and sodium acetate. (A, B) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant to succinate (A) and sodium acetate (B) on gradient soft‐agar plate. Representative images from three independent experiments are shown. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the colony, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values 0.52–0.48 indicate no significant response. (C, D) Microcalorimetric titration of 100 μM B2J77_03605 sensor domain with aliquots of 1.5 mM succinate (C) and sodium acetate (D). The raw data (upper panel) and integrated binding heats (lower panel) are shown. In panel C, the buffer control for the succinate titration is shown in blue. Figure S8: Multiple sequence alignment of the B2J77_13170 sensor domain with the homologous domains from the PctA, PctB and PctC chemoreceptors of P. aeruginosa (Gavira et al. 2020), CtaA from Pseudomonas fluorescens (Ud‐Din et al. 2020), and McpA from Bacillus velezensis (Feng et al. 2022). The PDB ID of the corresponding 3D structures is provided in brackets. Conserved residues (shaded in red) include E67, Y99, G105, T108, P111, P116, P121, R124, W126, Y127, A130, T139, P141, G163, G167, V168, G191, A193, F194, and P206 (numbering according to the sensor domain B2J77_13170). Figure S9: Chemotactic responses of P. parafulva PSR09‐11288 to L‐aspartate and L‐glutamate. (A, B) Chemotaxis assays of wild‐type P. parafulva PSR09‐11288 and its ∆cheA mutant to L‐aspartate (A) and L‐glutamate (B) on gradient soft‐agar plate. Representative images from three independent experiments are shown. The response index (RI) was calculated as RI = D1/(D1 + D2), where D1 and D2 are the distances from the inoculation site to the nearest and farthest edges of the bacterial spread, respectively, relative to the centre of the agar plug, as described in Section 5.3 of Materials and Methods. RI values 0.52–0.48 indicate no significant response. (C, D) Microcalorimetric titration of 100 μM B2J77_13170 sensor domain with aliquots of 1.5 mM L‐aspartate (C) and L‐glutamate (D). The raw data (upper panel) and integrated binding heats (lower panel) are shown. Table S1: Bacterial strains and plasmids used in this study. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.