AdiY acts as a cytoplasmic pH sensor via histidine protonation to regulate acid stress adaptation in Escherichia coli

TL;DR

The study reveals how Escherichia coli uses the AdiY protein to sense acidic conditions and activate survival genes.

Contribution

The novel contribution is the discovery that AdiY functions as a cytoplasmic pH sensor via histidine protonation.

Findings

AdiY's pH sensing is mediated by protonation of His34 and His60 in its N-terminal domain.

pH-dependent conformational changes in AdiY lead to DNA binding as a tetramer under acidic conditions.

Replacing key histidines in AdiY abolishes pH-dependent activation and DNA-binding activity.

Abstract

The arginine-dependent acid resistance (Adi) system is a vital component that enables Escherichia coli and other enterobacteria to withstand the extreme acidity in the human gastrointestinal tract. It consists of the proton-consuming decarboxylation of arginine, catalyzed by AdiA, and the uptake of arginine, as well as the excretion of the more alkaline agmatine, catalyzed by the antiporter AdiC. The corresponding genes adiA and adiC are induced in E. coli under acidic conditions (pH < 5.5), a process that is tightly regulated by the AraC/XylS transcriptional activator AdiY. Here, we show that the pH-sensing mechanism of AdiY functions through the protonation of two histidines (His34 and His60) in the N-terminal domain. Replacing these histidine residues with alanine, glutamine, or aspartate abolishes the pH-dependent activation of AdiY, both in vivo, as demonstrated by promoter-reporter assays, and in vitro, as indicated by the loss of DNA-binding activity detected by surface plasmon resonance spectroscopy. Biochemical analyses of purified wild-type AdiY using size-exclusion chromatography and intrinsic tryptophan fluorescence revealed a pronounced and reversible pH-dependent conformational change that does not occur in the pH-sensing–deficient AdiY variant. A model is proposed in which AdiY forms a monomer at physiological pH. At a lower intracellular pH, the protonation of histidine in AdiY causes a conformational change that leads to the binding of AdiY as a tetramer to the DNA. This work elucidates the molecular mechanism of a one-component signal transduction system that combines both sensory and responsive functions.

Throughout their life, Escherichia coli and other bacteria may encounter acidic environments, for example, when passing through the human stomach. Their chances of survival under these conditions depend on the number and efficiency of acid resistance systems. Although many acid resistance mechanisms have been extensively studied, the molecular mechanism by which bacteria sense low pH is not yet fully understood. This study demonstrates that the transcription factor AdiY acts as a direct pH sensor by using two histidines to detect intracellular acidification in E. coli. When these histidines become protonated, AdiY changes its conformation and activates genes that support cell survival under acid stress. These findings not only reveal a new way in which bacteria can perceive extremely low pH environments but also provide the basis for the development of AdiY as a pH reporter.

Figures

pH-dependent activation of the adiA, adiC, and adiY promoters in E. coli MG1655 wild-type and the adiY mutant. (A) Schematic overview of the luminescence-based reporter, in which the expression of luciferase is controlled by the adiA (PadiA), adiC (PadiC), or adiY (PadiY) promoter. E. coli MG1655 wild type and the adiY mutant were each transformed with the reporter plasmid pBBR1-MCS5-PadiA-lux (B), pBBR1-MCS5-PadiC-lux (C), and pBBR1-MCS5-PadiY-lux (D) and were grown in LB medium adjusted to different pH values ranging from 4.0 to 7.0. Data are reported as RLUs in counts per second per milliliter per OD600, and the maximal RLU after 2 hours of growth is shown. All experiments were performed three times (N = 3), and error bars represent the standard deviation of the means.

Two histidine residues are essential for AdiY-mediated activation of the adiA promoter under acidic conditions. (A) E. coli MG1655 ΔadiY was co-transformed with the reporter plasmid pBBR1-MCS5-PadiA-lux and plasmid-based adiY wild-type or variants. Cells were grown in a citrate-buffered medium adjusted to pH 4.4 (gray) or pH 7.0 (dark gray) and supplemented with 0.1% L-arabinose. Data are reported as RLUs in counts per second per milliliter per OD600, and the maximal RLU after 2 hours of growth is shown. All experiments were performed at least three times (n ≥ 3), and error bars represent the standard deviation of the means. Production of AdiY wild-type or AdiY variants was confirmed by Western blot analysis using antibodies against the His-tag. Protein bands correspond to a 29-kDa protein. The blot is shown split because the samples were initially loaded onto the gel in a different order. The bands were cut and rearranged to achieve the desired order in the graph. Statistics: Student’s unpaired two-sided t test; ns, not significant; *P = 0.0203; P = 0.0035; P < 0.0001. (B) Superposition of the AlphaFold 3 structural model of the monomer of AdiY wild-type (gray ribbon representation) and AdiY-H34A/H60A variant (pink ribbon representation). Key histidine residues at positions 34 and 60 are highlighted in golden in the AdiY wild-type structural model. (C) His34 can form a hydrogen bond with the carbonyl oxygen of Ala32 located in the beta-hairpin of the N-terminal domain (NTD). Its imidazole moiety packs on Pro91. (D) Rotation of the imidazole moiety of His60 around the Cα-Cβ or Cβ-Cγ sigma bonds allows either a hydrogen bond with Glu41 or the carbonyl oxygen of Met142. Distances (Å) are drawn as dashed lines. The figure was prepared with PyMOL (v3.1.4.1).

Comparative analysis of AdiY at pH 7.4 and pH 5.8 by SEC and intrinsic tryptophan fluorescence spectroscopy. (A) SEC elution profiles of wild-type AdiY and the AdiY-H34A/H60A variant at pH 7.4 and pH 5.8. (B) Intrinsic tryptophan fluorescence emission spectra (excitation at 280 nm) of wild-type AdiY and the AdiY-H34A/H60A variant at pH 7.4 (left panel) and pH 5.8 (right panel).

pH-dependent binding of AdiY and AdiY-H34A/H60A variant with target promoters analyzed by SPR spectroscopy. (A–C) Biotinylated DNA fragments comprising the promoters of adiA (PadiA), adiC (PadiC), or adiY (PadiY) were captured on SA sensor chips. Solutions of purified AdiY wild-type (A) or purified AdiY-H34A/H60A (B) at pH 6.0 were passed over the sensor chip. (C) Solutions of purified AdiY wild-type were passed over the adiA promoter, immobilized on the sensor chip, at different pH values ranging from 7.0 to 5.5. Sensorgram color coding corresponds to increasing protein concentrations, as indicated in the panel on the right. The dashed line at 0 RU represents the baseline, while the dashed line at 135 RU marks the maximal response of wild-type AdiY to the adiA promoter at pH 6.0. Where applicable, association rate (ka) and dissociation rate (kd) constants and the equilibrium dissociation constants (KD) are summarized in Fig. 5 and Table S1.

Kinetic parameters of AdiY binding to target promoters measured by SPR. Association rate (ka) and dissociation rate (kd) constants and the equilibrium dissociation constants (KD) were determined for AdiY binding to the adiA and adiC promoter fragments at the indicated pH values. Values at pH 5.8, 6.3, and 6.5 are from two independent experiments; values at pH 6.0 are from three replicates. Means ± SD are reported in Table S1, and representative sensorgrams are shown in Fig. S3.

pH-dependent activation mechanism of AdiY. Schematic model illustrating the proposed activation pathway of the AdiY transcription factor in E. coli under acidic stress. (1) Under neutral pH, AdiY exists predominantly as an inactive monomer (gray). (2) Low cytoplasmic pH induces protonation of key histidine residues, which triggers a conformational change. (3) Upon binding to DNA, AdiY undergoes tetramerization, which stabilizes the transcription complex. (4) This structural rearrangement enables pH-dependent activation of the target genes adiA and adiC of the Adi system. Structural predictions were generated by Alphafold 3. The figure was prepared with PyMOL (v3.1.4.1).

Tables

TABLE 1. Primer sequences used in this studya

Primer name	Primer sequence (5′ to 3′)
PadiC-lux_fw	CTCTAGAACTAGTGGATCCCCCTCTACGGTATGACGCCGCTGC
PadiC-lux_rev	GCTTGATATCGAATTCCTGCAGCCCAATTAAACTCCTGCGAAGGCGAG
PadiY-lux_fw	CTCTAGAACTAGTGGATCCCCCGGCGTAATTGTTTAAATAACATTACG
PadiY-lux_rev	GCTTGATATCGAATTCCTGCAGCCCACATGTACTCCTGAGTGCGAATAAAAG
pNPTS_adiY_FLup_F	GCCAAGCTTCTCTGCAGGATCGCGATTGTCGACAACAATGTC
adiY-OL-Flanken-F	GGAGTACATGTATGAGGATTTGCTTCTACGGTATGACGCCGCTGC
adiY-OL-Flanken-R	GCAGCGGCGTCATACCGTAGAAGCAAATCCTCATACATGTACTCC
adiY-OL-FLdown-R	GCGAATTCGTGGATCCAGATCCACAGCACCACGACGCAGGTG
adiY-KO-check-F	GAACGAAGCCTATTCCG
AdiY_check_rev	CCCAGGCCGCTGACGTTCC
pNPTS-EcoRV-fwd	ATCTGGATCCACGAATTCGCTAGCTTC
pNPTS-EcoRV-rev	ATCCTGCAGAGAAGCTTGGCGC
M13_rev	CAGGAAACAGCTATGACC
M13_uni	TGTAAAACGACGGCCAGT
adiY-H34A-fwd	GCCTTAAAGCTAACGCTATGGCG
adiY-H34A-rev	CCATAGCGTTAGCTTTAAGGC
adiY-H57A-fwd	GGTTGCTGCTATTAGTCACGAC
adiY-H57A-rev	GTCGTGACTAATAGCAGCAACC
adiY-H60A-fwd	GCTCATATTAGTGCCGACATCATC
adiY-H60A-rev	GATGATGTCGGCACTAATATGAGC
adiY-H104A-fwd	CGCGCAAGCCAGCATGATG
adiY-H104A-rev	CATCATGCTGGCTTGCGCG
adiY-H57D-fwd	GGTTGCTGATATTAGTCACGAC
adiY-H57D-rev	GTCGTGACTAATATCAGCAACC
adiY-H60D-fwd	GCTCATATTAGTGACGACATCATC
adiY-H60D-rev	GATGATGTCGTCACTAATATGAGC
adiY-H34D-fwd	GCCTTAAAGCTAACGATATGGCG
adiY-H34D-rev	CGCCATATCGTTAGCTTTAAGGC
AdiY_H34Q_F	AAGCTAACCAAATGGCGTT
AdiY_H34Q_R	AACGCCATTTGGTTAGCTT
AdiY_H60Q_F	TATTAGTCAAGACATCATCA
AdiY_H60Q_R	TGATGATGTCTTGACTAATA
PadiA_SPR-F	AGACGATATCAGTATCAGCC
PadiA_btn-SPR-R	btn-TGCTTACCCGGTTATGAAGG
PadiC_SPR-F	TCTACGGTATGACGCCGCTGC
PadiC_btn-SPR-R	btn-AATTAAACTCCTGCGAAGGCGAG
PadiY_SPR-F	GGCGTAATTGTTTAAATAACATTACG
PadiY_btn-SPR-R	btn-ACATGTACTCCTGAGTGCGAATAAAAG

TABLE 2. Strains and plasmids used in this study

Strain or plasmid	Relevant genotype or description	Reference
E. coli strains
E. coli MG1655	K-12 F– λ– ilvG– rfb-50 rph-1	(49)
E. coli DH5αλpir	endA1 hsdR17 glnV44 (= supE44) thi-1 recA1 gyrA96 relA1 φ80’lacΔ(lacZ) M15 Δ(lacZYA-argF)U169 zdg−232::Tn10 uidA::pir+	(50)
E. coli WM3064	thrB1004 pro thi rpsL hsdS lacZΔM15 RP4-1360 Δ(araBAD)567 ΔdapA1341::[erm pir]	W. Metcalf, U. Illinois, Urbana
E. coli MG1655 ΔadiY	In-frame deletion of adiY in MG1655	This work
E. coli LMG194	F- ΔlacX74 gal E thi rpsL ΔphoA (Pvu II) Δara714 leu::Tn10	(51)
Plasmids
pBAD24	Arabinose-inducible PBAD promoter, pBR322 ori, Ampᴿ	(51)
pBAD24-His6-adiY	N-terminal His6-tagged adiY in pBAD24, Ampᴿ	(22)
pBAD24-His6-adiY-H34A	N-terminal His6-tagged adiY-H34A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H57A	N-terminal His6-tagged adiY-H57A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H60A	N-terminal His6-tagged adiY-H60A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H104A	N-terminal His6-tagged adiY-H104A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H34A/H60A	N-terminal His6-tagged adiY-H34A/H60A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H57A/H60A	N-terminal His₆-tagged adiY-H57A/H60A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H34A/H57A/H60A	N-terminal His₆-tagged adiY-H34A/H57A/H60A mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H34D	N-terminal His6-tagged adiY-H34D mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H57D	N-terminal His6-tagged adiY-H57D mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H60D	N-terminal His6-tagged adiY-H60D mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H34D/H60D	N-terminal His6-tagged adiY-H34D/H60D mutant in pBAD24, Ampᴿ	This work
pBAD24-His6-adiY-H34Q/H60Q	N-terminal His6-tagged adiY-H34Q/H60Q mutant in pBAD24, Ampᴿ	This work
pNTPS138-R6KT	mobRP4 + ori R6K sacB; suicide plasmid for in-frame deletions, Kmᴿ	(52)
pNTPS138-R6KT-ΔadiY	pNPTS-138-R6KT-derived suicide plasmid for in-frame deletion of adiY in MG1655, Kmᴿ	This work
pBBR1-MCS5-TT-RBS-lux	luxCDABE and terminators lambda T0 rrnB1 T1 cloned into pBBR1-MCS5 for plasmid-based transcriptional fusions, Gmᴿ	(53)
pBBR1-MCS5-PadiA-lux	adiA promoter controlling expression of luxCDABE, in pBBR1-MCS5-TT-RBS-lux, Gmᴿ	(54)
pBBR1-MCS5-PadiC-lux	adiC promoter controlling expression of luxCDABE, in pBBR1-MCS5-TT-RBS-lux, Gmᴿ	This work
pBBR1-MCS5-PadiY-lux	adiY promoter controlling expression of luxCDABE, in pBBR1-MCS5-TT-RBS-lux, Gmᴿ	This work

Full text

INTRODUCTION

Microorganisms must constantly monitor changes in the proton concentration (pH) in their environment and respond accordingly (1). An increase in proton concentration (low pH) can lead to acid stress, which causes proteins, nucleic acids, and membranes to become protonated, disrupting their structure and function. Bacteria such as Escherichia coli that are neutralophilic face extreme acid challenges during passage through the stomach and gut of a host, where hydrochloric acid and fermentation products, such as short-chain fatty acids, lower both the extracellular and intracellular pH of the bacteria (2, 3). To survive these conditions, E. coli and other enteric bacteria have evolved robust acid resistance (AR) systems (4, 5). Acquisition of AR is indeed an essential trait to survive acidic environments; for example, survival in the gastrointestinal tract requires the ability to buffer or expel excess protons. These adaptive responses allow bacteria to maintain a viable cytoplasmic pH range despite extreme external acidity (6).

E. coli employs multiple AR systems, with the most effective enzyme-based mechanisms being the glutamate-dependent (AR2/Gad), arginine-dependent (AR3/Adi), and lysine-dependent (AR4/Cad) systems. In each system, a cytoplasmic decarboxylase consumes a proton while converting its substrate (glutamate, arginine, or lysine, respectively) to the corresponding amine (γ-aminobutyrate, agmatine, or cadaverine, respectively), and a specific antiporter (GadC, AdiC, or CadB) exports the amine in exchange for its cognate amino acid (4). This coupled decarboxylation–antiport cycle reduces the intracellular proton load, thereby increasing the cytoplasmic pH and, in the case of the Cad and Adi systems, also elevating the external pH by the secretion of the alkaline cadaverine and agmatine, respectively (7–9). A fourth AR system (AR1) operates via the Sigma factor S (RpoS) and indirectly affects the F_1_F_o_-ATPase proton pump, and AR5, the ornithine decarboxylase system, plays only a minor role in laboratory strains (4, 10). Each AR system is induced under specific conditions and pH ranges, and an intricate network of sensory and regulatory proteins regulates its expression. The Gad (AR2) system, for example, is activated by a cascade of sensors and transcription factors and is essential for survival at pH levels of 2–3. In all cases, AR predominantly relies on the proton-consuming reactions to restore intracellular pH homeostasis (11, 12). The regulation of AR systems often involves transcription factors from the AraC/XylS family (13). For example, the genes of the Gad system are controlled by the AraC-like regulators GadE, GadW, and GadX (14). Transcription factors of the same family, such as AppY and EnvY, modulate gene expression under anaerobic or starvation conditions (15, 16). AdiY, which is in the focus of this study, is the activator of the Adi AR system and also a member of the AraC/XylS family which are typically composed of a C-terminal DNA-binding domain and an N-terminal ligand-binding domain (13, 17–21).

AdiY is a transcriptional activator that consists of 253 amino acids and binds upstream of adiA and adiC. It induces their expression under low pH and anaerobic conditions, such as those found in the stomach and gut (19). The adiY gene is located immediately downstream of adiA, and adiA and adiC are maximally induced at an external pH of 4.4 when E. coli is grown in an amino acid-rich medium. The Adi system is heterogeneously activated with only a fraction of the population expressing high levels of Adi components under acid stress (22). This pattern can be shifted to a more uniform expression when the copy number of AdiY is increased (22).

Most bacterial acid stress sensors are membrane-integrated proteins with an outwardly exposed sensor domain. A classic example is CadC, the activator of the Cad system. CadC has a C-terminal periplasmic domain that functions as an external pH sensor. At low external pH (pH 5.8) and in the presence of lysine, conformational changes in CadC trigger transcription of the cadBA operon, thereby coupling the external acid stress to gene activation (23–26).

In contrast, cytosolic regulators such as AdiY lack a membrane-linked sensor domain. This raises the question of which molecular features enable a cytosolic activator to respond to acid stress.

Bacteria, in general, are able to maintain a relatively constant internal pH (pH) when grown in a wide range of media with different external pH values (reviewed in [5, 27, 28]). The cytoplasmic pH of E. coli, for example, only changes from 7.2 to 7.8 over an environmental pH range of 5.5 to 9 (29). Although the cytoplasmic membrane is impermeable to protons, some protons enter the cytoplasm through protein channels and via transient water channels or damaged membranes (4, 30). For example, the pH of the cytoplasm drops transiently to pH 6.2 after exposure of E. coli to pH 5.8. However, intracellular pH rapidly returns to neutral (<4 min) due to the intrinsic buffering capacity of the cytoplasm or alterations in the flux of other ions (27). At more severe acid stress (pH 4.4), the intracellular pH decreases by approximately one pH unit and remains unchanged. Genes coding for the Adi system are induced. The correlation between a decrease in intracellular pH and the activation of the Adi system was previously indirectly demonstrated, as the adiA promoter was strongly activated in a mutant lacking the AR4/Cad system or less activated after overproduction of the AR4/Cad system (22).

It is hypothesized that AdiY responds to the acidification of the cytoplasm. In this study, we combined genetic, biochemical, and structural analyses to find that the pH-dependent protonation of two histidines triggers a conformational change in AdiY, enabling its DNA binding and transcriptional activation. These studies offer new functional insights into a member of the AraC/XylS family transcription regulator, which may also be representative of other members, particularly those involved in virulence (13).

RESULTS

AdiY is essential for pH-dependent activation of the AR3/Adi system in E. coli

To characterize the regulatory logic underlying the activation of the Adi system, we quantified the promoter activity of the adiA, adiC, and adiY genes using luciferase-based reporter plasmids (Fig. 1A). These constructs were introduced into wild-type and ΔadiY E. coli MG1655 strains, and promoter activities were monitored across a range of pH values under both aerobic and microaerobic conditions (Fig. 1B through D).

pH-dependent activation of the adiA, adiC, and adiY promoters in E. coli MG1655 wild-type and the adiY mutant. (A) Schematic overview of the luminescence-based reporter, in which the expression of luciferase is controlled by the adiA (PadiA), adiC (PadiC), or adiY (PadiY) promoter. E. coli MG1655 wild type and the adiY mutant were each transformed with the reporter plasmid pBBR1-MCS5-PadiA-lux (B), pBBR1-MCS5-PadiC-lux (C), and pBBR1-MCS5-PadiY-lux (D) and were grown in LB medium adjusted to different pH values ranging from 4.0 to 7.0. Data are reported as RLUs in counts per second per milliliter per OD600, and the maximal RLU after 2 hours of growth is shown. All experiments were performed three times (N = 3), and error bars represent the standard deviation of the means.

The results showed that in the absence of AdiY, the adiA, adiC, and adiY promoters remained inactive under all tested conditions, regardless of pH or oxygen availability (Fig. 1B through D). In contrast, wild-type cells showed strong induction of both the adiA and adiC promoters under acidic conditions, with activation thresholds below pH 5.8 in microaerobic environments and maximal expression at pH 4.4. Under aerobic conditions, activation of both promoters was diminished and occurred only at external pH values below 4.8 (Fig. 1B and C). Notably, the activation profiles of the adiA and adiC promoters were nearly identical; however, the adiA promoter exhibited an eightfold higher level of induction across all tested conditions. In contrast, the activity of the adiY promoter remained consistently low under all tested conditions, with only a modest increase under microaerobic, acidic conditions (Fig. 1D).

These results demonstrate that activation of the adiA and adiC promoters is strictly pH-dependent and modulated by oxygen availability, with robust induction of adiA and adiC occurring under acidic, microaerobic conditions, reduced activation under aerobic conditions, and no activation in the absence of AdiY. The adiY promoter itself is only weakly responsive, which is consistent with its role as a constitutive or weakly inducible regulator within the system.

Two histidine residues are essential for AdiY activation in vivo

To elucidate the molecular mechanism of AdiY-mediated pH sensing, a systematic substitution mutagenesis approach was employed to investigate the role of histidine residues in the N-terminal domain of AdiY. Depending on the pH and protein environment, the imidazole ring of histidine can take up distinct chemical forms, which include the two neutral N^δ1^-H and N^ε2^-H tautomers and the biprotonated H^+^-form and can thus influence protein conformation and function (31, 32).

Each of the four histidine residues in the N-terminal domain was individually replaced with alanine, and the resulting AdiY variants (expressed from a pBAD24 plasmid) were analyzed for their ability to activate the adiA promoter in vivo under acidic conditions, using the E. coli MG1655 ΔadiY background for comparison to wild-type AdiY. Only wild-type AdiY complemented the adiY deletion strain and restored full activation of the adiA promoter at pH 4.4 (Fig. 2A). In contrast, the histidine-to-alanine substitutions at positions 34 and 60 resulted in a significant loss of function, substantially impairing AdiY’s ability to respond to low pH and activate the adiA promoter. Substitution of histidine at position 57 (AdiY-H57A) also led to a significant reduction in promoter activation at pH 4.4, but the effect was smaller compared to the AdiY-H34A or AdiY-H60A variants (Fig. 2A). The double histidine variant AdiY-H34A/H60A displayed a synergistic loss of function and retained only about 5% of the wild-type AdiY activity (Fig. 2A), while the AdiY-H57A/H60A double variant retained activity comparable to that of the AdiY-H60A single variant. The AdiY-H34A/H57A/H60A triple variant was comparable in activity to the AdiY-H34A/H60A double variant (Fig. 2A). Similarly, a loss of function occurred when these two positions were exchanged to glutamine or aspartate (Fig. S1). All variants and wild-type AdiY were produced upon induction with L-arabinose at comparable levels, as assessed by Western blotting. Furthermore, neither these variants nor wild-type AdiY showed activation of the adiA promoter at physiological pH (pH 7.0) (Fig. 2A; Fig. S1). This finding suggests that the histidine residues at positions 34 and 60 are critical for the ability of AdiY to activate the adiA promoter in response to acidic pH.

*Two histidine residues are essential for AdiY-mediated activation of the adiA promoter under acidic conditions. (A) E. coli MG1655 ΔadiY was co-transformed with the reporter plasmid pBBR1-MCS5-PadiA-lux and plasmid-based adiY wild-type or variants. Cells were grown in a citrate-buffered medium adjusted to pH 4.4 (gray) or pH 7.0 (dark gray) and supplemented with 0.1% L-arabinose. Data are reported as RLUs in counts per second per milliliter per OD600, and the maximal RLU after 2 hours of growth is shown. All experiments were performed at least three times (n ≥ 3), and error bars represent the standard deviation of the means. Production of AdiY wild-type or AdiY variants was confirmed by Western blot analysis using antibodies against the His-tag. Protein bands correspond to a 29-kDa protein. The blot is shown split because the samples were initially loaded onto the gel in a different order. The bands were cut and rearranged to achieve the desired order in the graph. Statistics: Student’s unpaired two-sided t test; ns, not significant; *P = 0.0203; **P = 0.0035; ***P < 0.0001. (B) Superposition of the AlphaFold 3 structural model of the monomer of AdiY wild-type (gray ribbon representation) and AdiY-H34A/H60A variant (pink ribbon representation). Key histidine residues at positions 34 and 60 are highlighted in golden in the AdiY wild-type structural model. (C) His34 can form a hydrogen bond with the carbonyl oxygen of Ala32 located in the beta-hairpin of the N-terminal domain (NTD). Its imidazole moiety packs on Pro91. (D) Rotation of the imidazole moiety of His60 around the Cα-Cβ or Cβ-Cγ sigma bonds allows either a hydrogen bond with Glu41 or the carbonyl oxygen of Met142. Distances (Å) are drawn as dashed lines. The figure was prepared with PyMOL (v3.1.4.1).

To gain structural insight, AlphaFold 3 models were generated for both wild-type AdiY and the double histidine variant AdiY-H34A/H60A (Fig. 2B). The predicted monomeric structure of wild-type AdiY (Fig. 2B, gray) displays the expected AraC-family fold. His60, one of the essential histidine residues (highlighted in golden), is strategically positioned at the interface between the N-terminal regulatory and C-terminal DNA-binding domain. By contrast, the AdiY-H34A/H60A model (Fig. 2B, pink) shows a different arrangement of the domains, with the N-terminal domain slightly spread out relative to the C-terminal DNA-binding domain around the H60 hinge. This structural difference suggests that these histidine residues are essential for maintaining the interdomain architecture necessary for effective transcriptional activation. His34 is located at a β-turn in the N-terminal domain, where its protonation state also likely impacts the structure and dynamics of its environment (Fig. 2C). Depending on the protonation state, His60 might form hydrogen bonds to Glu41 and via rotation of its imidazole moiety around the C^α^-C^β^ and/or C^β^-C^γ^ bonds, with the carbonyl oxygen of Met142 (Fig. 2D).

Collectively, these data demonstrate that two histidine residues in AdiY function as pH-sensitive molecular switches. Their protonation likely induces conformational changes that facilitate DNA binding and enable transcriptional activation. Substitution of these residues abolishes AdiY activity in vivo, highlighting their critical role in acid-responsive gene regulation.

pH-Dependent conformational dynamics of AdiY in vitro

To corroborate the in vivo finding that only wild-type AdiY, but not the AdiY-H34A/H60A mutant, can activate the adiA promoter in a pH-dependent manner in vivo (Fig. 2A), both proteins were purified via their N-terminal His_6_-tags (Fig. S2) and subjected to in vitro analysis of their conformational dynamics using size-exclusion chromatography (SEC) and intrinsic tryptophan fluorescence spectroscopy at pH 7.4 and pH 5.8. SEC analysis revealed that both purified proteins show a distinctive peak corresponding to the size of a monomer, but at pH 5.8, wild-type AdiY exhibited a shift of approximately 0.3 mL toward lower retention volumes compared to pH 7.4 (Fig. 3A), suggesting an increase in hydrodynamic radius and/or changes in interdomain interactions. Calibration with molecular mass standards indicated that AdiY adopts a more compact conformation at pH 7.4. In contrast, the AdiY-H34A/H60A variant showed only minor shifts between the two pH conditions, suggesting a reduced sensitivity to pH-induced conformational rearrangements. Intrinsic tryptophan fluorescence measurements (excitation at 280 nm, emission recorded between 300 and 400 nm) supported these findings (Fig. 3B through D). It is noteworthy that the fluorophores are distributed unevenly across the domains: the N-terminal region contains two Trp and two Tyr, whereas the C-terminal region contains one Trp and seven Tyr. This enables domain-aware interpretation of the intrinsic fluorescence signals.

Comparative analysis of AdiY at pH 7.4 and pH 5.8 by SEC and intrinsic tryptophan fluorescence spectroscopy. (A) SEC elution profiles of wild-type AdiY and the AdiY-H34A/H60A variant at pH 7.4 and pH 5.8. (B) Intrinsic tryptophan fluorescence emission spectra (excitation at 280 nm) of wild-type AdiY and the AdiY-H34A/H60A variant at pH 7.4 (left panel) and pH 5.8 (right panel).

At pH 7.4, wild-type AdiY showed an emission maximum at approximately 335–336 nm. At lower pH (5.8), we observed a red shift (~2 nm) of the emission maximum and a decrease in fluorescence intensity, suggesting increased solvent exposure of tryptophan residues due to pH-induced structural rearrangements (Fig. 3B). The AdiY-H34A/H60A variant showed minimal spectral shifts and intensity changes compared to wild-type AdiY at pH 7.4 (Fig. 3B, left panel). At pH 5.8, its emission maximum remained at approximately 335–336 nm (Fig. 3B, right panel), similar to that of wild-type AdiY at pH 7.4, further supporting a reduced conformational flexibility in the mutant relative to the wild-type protein.

Together, these data indicate that AdiY experiences significant pH-dependent conformational transitions that do not occur in the AdiY-H34A/H60A variant (Fig. 2B).

AdiY-DNA interaction is pH-dependent in vitro

Given the pH-dependent conformational changes observed in AdiY wild-type (Fig. 3), we investigated whether these structural shifts influence the ability of AdiY wild-type to bind the target promoters in vitro, in comparison to the AdiY-H34A/H60A variant. We used purified AdiY wild-type and AdiY-H34A/H60A proteins to quantitatively determine their binding affinity to the adiA, adiC, and adiY promoters using surface plasmon resonance (SPR) spectroscopy at different pH values (Fig. 4). At pH 6.0, wild-type AdiY binds preferentially to the adiA and adiC promoters, with affinities of 170 ± 23.2 nM and 120 ± 34.4 nM, respectively (Fig. 4A; Table S1). In both cases, the stoichiometry was 1:4 (DNA:protein), indicating that AdiY assembles as a tetramer on the promoter DNA (Fig. 4A; Table S1). Furthermore, AdiY bound the adiA and adiC promoters with similar association rate constants (ka), averaging 1.01 × 10^4^ M^−1^ s^−1^ and 1.60 × 10^4^ M^−1^ s^−1^, respectively. The dissociation rates (kd) were also comparable at 1.70 × 10^3^ s^−1^ for the adiA and 1.80 × 10^3^ s^−1^ for the adiC promoter, indicating that once bound, the complexes had similar stability (Table S1). No binding of wild-type AdiY to the adiY promoter was detected under the same conditions or at any tested pH values (Fig. 4A; Fig. S3). The AdiY-H34A/H60A mutant failed to exhibit any detectable DNA binding across all tested conditions (Fig. 4B; Fig. S4 ), underscoring the critical role of histidine residues at positions 34 and 60 in mediating the pH-dependent conformational changes required for promoter binding.

pH-dependent binding of AdiY and AdiY-H34A/H60A variant with target promoters analyzed by SPR spectroscopy. (A–C) Biotinylated DNA fragments comprising the promoters of adiA (PadiA), adiC (PadiC), or adiY (PadiY) were captured on SA sensor chips. Solutions of purified AdiY wild-type (A) or purified AdiY-H34A/H60A (B) at pH 6.0 were passed over the sensor chip. (C) Solutions of purified AdiY wild-type were passed over the adiA promoter, immobilized on the sensor chip, at different pH values ranging from 7.0 to 5.5. Sensorgram color coding corresponds to increasing protein concentrations, as indicated in the panel on the right. The dashed line at 0 RU represents the baseline, while the dashed line at 135 RU marks the maximal response of wild-type AdiY to the adiA promoter at pH 6.0. Where applicable, association rate (ka) and dissociation rate (kd) constants and the equilibrium dissociation constants (KD) are summarized in Fig. 5 and Table S1.

To assess pH dependence in vitro, the binding of AdiY wild-type to its target promoter was tested across a pH range from 7.0 to 5.5. Binding of AdiY wild-type to the adiA and adiC promoters was strongest at acidic pH values (6.0, 5.8, and 6.3), weaker at pH 6.5, and undetectable at pH 5.5 or physiological pH values above 6.7. However, tetramerization of AdiY was only achieved at pH values of 5.8 and 6.0 on these two promoters (Fig. 4C; Fig. S3; Table S1), showing a pH-dependent change in stoichiometry. Across the tested pH range, ka values for the adiA promoter varied from 5.86 to 8.17 × 10^3^ M^−1^ s^−1^ at pH 6.3 to 1.46−1.62 × 10^4^ M^−1^ s^−1^ at pH 6.5, whereas the value for the adiC promoter showed a slightly higher maximum ka (2.15–3.50 × 10^4^ M^−1^ s^−1^ at pH 5.8), which was lower at pH 6.5 (ka 2.36–4.70 × 10^3^ M^−1^ s^−1^) (Fig. 5; Table S1). For both promoters, the dissociation rate was lowest and binding stability was highest at mildly acidic pH values (5.8–6.3). At pH 5.8, the adiA and adiC promoters exhibited kd values of 1.34–2.11 × 10^3^ s^−1^ and 2.64–9.73 × 10^3^ s^−1^, respectively. At pH 6.5, the dissociation rates increased three- to sevenfold for the adiA promoter (4.59–10.4 × 10^3^ s^−1^) and about 10-fold for the adiC promoter (1.52–2.03 × 10^2^ s^−1^) compared to pH 6.0 (Table S1; Fig. 5), which primarily accounts for the lower affinity at pH 6.5. This faster dissociation likely reflects the loss of tetramer-mediated stabilization and reduced protonation of key DNA-contact residues at the higher pH. Overall, both promoters exhibited the fastest association and slowest dissociation at moderately acidic pH values, with maximal binding between pH 5.8 and 6.3. These results demonstrate that AdiY binds specifically and cooperatively to the adiA and adiC promoters, with DNA-binding activity sharply tuned to a narrow range of moderately acidic intracellular pH values. Outside this range, binding was almost completely abolished, indicating that AdiY is selectively activated under these conditions.

Kinetic parameters of AdiY binding to target promoters measured by SPR. Association rate (ka) and dissociation rate (kd) constants and the equilibrium dissociation constants (KD) were determined for AdiY binding to the adiA and adiC promoter fragments at the indicated pH values. Values at pH 5.8, 6.3, and 6.5 are from two independent experiments; values at pH 6.0 are from three replicates. Means ± SD are reported in Table S1, and representative sensorgrams are shown in Fig. S3.

DISCUSSION

This study identified AdiY as a direct pH sensor and transcriptional activator of adiA and adiC of the AR3/Adi system in E. coli. AdiY is activated by protonation of two histidine residues at positions 34 and 60 in the N-terminal domain, which triggers a conformational change of the position of the C-terminal domain, leading to DNA-binding and tetramerization. Substitution of His34 and His60 not only abolished the pH sensitivity but also the pH-dependent conformational rearrangement of AdiY. Based on the AlphaFold prediction of the AdiY structure, His34 and His60 are positioned such that their protonation state likely impacts the overall arrangement and dynamics of the two domains (Fig. 2B and C). The imidazole ring of histidine can adopt distinct chemical forms, which include the two neutral N^δ1^-H and N^ε2^-H tautomers and the biprotonated H^+^-form. Moreover, the neutral N^δ1^-protonated and N^ε2^-protonated tautomers differ in their pKa values of 6.8 and 6.4, respectively (33). In addition, different rotameric structures are possible due to rotation around the C^α^-C^β^ and/or C^β^-C^γ^ bonds. Chemical shift experiments by NMR using different model proteins have shown that the distribution of the different chemical forms of histidines is determined by the protein environment and pH, although the protein environment largely impacts the pH-dependent protonation (32). Protonation of His60 at pH 6.5 presumably facilitates the interaction with Glu41, thus impacting domain rearrangements, and hence DNA binding (Fig. 2C).

SPR analyses demonstrated that wild-type AdiY binds specifically and cooperatively to the adiA and adiC promoters in vitro, with maximum affinity and tetramerization observed at mildly acidic pH values between pH 5.8 and 6.0 (Fig. 4; Fig. S3; Table S1). For both promoters, AdiY binding becomes weaker at higher pH values. The large change between pH 6.3 and 6.5 is accompanied by large standard deviations, especially at pH 6.5, where the signals were low and near the detection limit. Therefore, although binding of AdiY is clearly weaker at pH 6.5, we avoid over-interpreting the magnitude of this change.

Outside this pH range, AdiY did not bind to the DNA, suggesting its very restricted activation. This observed pH dependence is consistent with the activation profile of the Adi system, which reaches its maximum at a low external pH between 5.0 and 4.4, a condition that causes a decrease of the cytoplasmic pH to about 5.8 to 6.0 (27). Under these conditions, AdiY senses the intracellular acidification, tetramerizes and binds to its target promoters, and induces expression of adiA and adiC. Based on these findings, we propose the following model (Fig. 6): (i) at neutral pH, AdiY is predominantly monomeric and adopts a conformation that is not able to bind DNA. (ii) Acidification of the cytoplasm leads to protonation of key histidines (His34 and His60), which induces a rearrangement of the N-terminal regulatory domain relative to the C-terminal DNA-binding domain. (iii) Upon promoter recognition, two AdiY dimers cooperatively assemble into a tetramer on the adiA and adiC promoters, as supported by the pH-dependent 1:4 DNA:protein stoichiometry and reduced dissociation rates observed by SPR. (iv) Tetramer formation stabilizes the promoter-bound complex and enables activation of adiA and adiC transcription within a narrow acidic pH window. This mechanism is consistent with other AraC/XylS regulators, where ligand- or promoter-induced oligomerization reinforces DNA engagement (34, 35). The activity of the arginine decarboxylase AdiA is also strongly affected by the intracellular pH, which regulates its oligomerization. At pH 5.2, AdiA forms active decamers, which dissociate in inactive dimers at pH >6.0 (36). It is noteworthy that the catabolic genes for arginine are downregulated under severe acid stress (37), presumably to preserve arginine for decarboxylation. In summary, all components of the Adi system are finely tuned to be most active under severe acid stress within a very narrow range and become immediately inactive when the intracellular pH rises.

pH-dependent activation mechanism of AdiY. Schematic model illustrating the proposed activation pathway of the AdiY transcription factor in E. coli under acidic stress. (1) Under neutral pH, AdiY exists predominantly as an inactive monomer (gray). (2) Low cytoplasmic pH induces protonation of key histidine residues, which triggers a conformational change. (3) Upon binding to DNA, AdiY undergoes tetramerization, which stabilizes the transcription complex. (4) This structural rearrangement enables pH-dependent activation of the target genes adiA and adiC of the Adi system. Structural predictions were generated by Alphafold 3. The figure was prepared with PyMOL (v3.1.4.1).

As H34 and H60 are functionally important in vivo (Fig. 2A), we investigated the pKa values of the imidazole rings of these histidine residues using PypKa Server (https://pypka.org/ [38]), which yielded pKₐ values of 6.08 for H34 and 7.00 for H60. When the pH is lowered from 7 to 6, the theoretical protonation at the imidazole rings increases from 10.7% to 54.6% for H34 and from 50.0% to 90.9% for H60. His34 is located at a β-turn in the N-terminal domain, where its protonation state likely impacts the structure and dynamics of its environment (Fig. 2B). Protonation of His60 at pH 6.5 presumably facilitates interaction with Glu41, thus impacting domain dynamics and hence DNA binding (Fig. 2D).

The protonation of amino acids whose side chain pKa values are around pH 6.0 (histidine) or below (aspartic acid, pKa 3.88; glutamic acid, pKa 4.25) is a well-documented molecular mechanism in bacterial acid stress sensing (26, 39, 40), exemplified by the histidine kinases SsrA (five histidines), ArsS (one histidine, four glutamic acids), and the ToxR-like regulator CadC (three aspartic acids, two glutamic acids) (5, 6, 40–43). A difference between the pH sensors lies in the compartment they monitor. AdiY responds to cytoplasmic acidification, where the protonation of histidine residues at positions 34 and 60 causes a conformational change that enables DNA binding (Fig. 5). In contrast, SsrA, AsrS, and CadC detect acidity in the periplasm, for example, CadC via protonation of a cluster of negatively charged amino acids (Asp198, Asp200, Glu461, Glu468, and Asp471) in its periplasmic domain, which is located at the interface between two monomers (26). CadC, as a periplasmic sensor, is activated by extracellular moderate acid stress (pH 5.8), before the cytosol becomes acidic. In contrast, AdiY, a cytosolic sensor activated under severe acid stress (pH 4.4), senses intracellular acidification. This ensures hierarchical activation of functionally similar systems depending on the severity of acid stress in E. coli (22). Cytoplasmic regulators such as AphB of Vibrio cholerae (44) and Vibrio campbellii (45) also integrate intracellular pH into the regulation of virulence and stress responses. However, unlike AdiY, V. cholerae AphB relies on cysteine- and lysine-mediated redox and protonation changes (44). V. campbellii AphB achieves regulation through rapid, high-affinity promoter binding (45), while AdiY relies on cooperative, tetramerization-driven DNA binding with sub-micromolar affinity, which is restricted to a narrow pH range. The requirement for tetramerization to achieve stable promoter occupancy is consistent with the mechanisms described for AraC and LacI family regulators, where oligomerization is essential for precise DNA binding and transcriptional control (46, 47). AdiY belongs to the large family of AraC/XylS family of transcription regulators. Most of them consist of a characteristic and conserved DNA-binding domain and a “companion” domain (48), also referred to as an “effector” or “regulatory” domain (13). In many proteins, this domain has been shown to sense specific signaling molecules, such as fatty acids (VirF), arabinose (AraC), or xylose (XylS). When these molecules are bound, the transcription factors bind as a dimer to the corresponding promoter, inducing transcription (13). AdiY is an exception in this family, as it undergoes protonation/deprotonation and forms a tetramer (Fig. 5). Further work will be needed to determine whether small molecules also modulate this response. The Adi system is also found in other enterobacteria, such as Shigella, Salmonella, and Citrobacter (6). E. coli AdiY is highly similar to its orthologs in Escherichia albertii and Shigella, all of which retain His34 and His60 (Fig. S5). AdiY in representative strains of Salmonella and Citrobacter has His34 but no His60, where an arginine is located instead. It is worth noting that Salmonella completely lacks the Gad system, and Citrobacter lacks many regulatory components of the Gad system. In comparison, E. coli and Shigella have four acid-inducible amino acid–dependent decarboxylase systems and the highest resistance under extreme acid stress (6). Therefore, we hypothesize that the histidine-controlled protonation of AdiY in E. coli has evolved as a highly restrictive mechanism that responds to minute pH changes in the cytosol and confers maximal resistance for E. coli, for example, during passage through the strongly acidic environment of the stomach.

From an applied perspective, understanding this histidine-dependent pH-sensing mechanism opens up new avenues for developing AdiY in a pH-sensor that is useful in biosensors for microbiological, clinical, and biotechnological applications.

MATERIALS AND METHODS

Construction of plasmids and strains

Molecular methods were performed according to standard protocols or as instructed by the manufacturer. Kits for the isolation of plasmids and the purification of PCR products were purchased from Süd-Laborbedarf (SLG; Gauting, Germany). Enzymes and HiFi DNA Assembly Master Mix were purchased from New England BioLabs (Frankfurt, Germany). Strains, plasmids, and primers used in this study are summarized in Table 1 and Table 2.

To construct the reporter plasmids pBBR1-MCS5-PadiC-lux and pBBR1-MCS5-PadiY-lux, 200 bp and 300 bp fragments from the regions upstream of adiC or adiY, respectively, were amplified by PCR using the appropriate primer pairs (PadiC-lux_fw and PadiC-lux_rev or PadiY-lux_fw and PadiY-lux_rev). Genomic DNA from E. coli strain MG1655 served as the template. After purification, the PCR products were each assembled into SmaI-digested pBBR1-MCS5-TT-RBS-lux plasmid using Gibson assembly (55), resulting in the plasmids pBBR1-MCS5-PadiC-lux and pBBR1-MCS5-PadiY-lux. Colony PCR and sequencing verified the correct insertion.

To construct a marker-less in-frame deletion of adiY in E. coli MG1655, firstly, the suicide plasmid pNPTS138-R6KT-ΔadiY was generated as previously described in (22). Briefly, flanking regions of each 600 bp upstream and downstream of adiY were amplified by PCR from E. coli MG1655 genomic DNA using the appropriate primer pairs (pNPTS_adiY_FLup_F and adiY-OL-Flanken-R or adiY-OL-Flanken-F and adiY-OL-FLdown-R). After purification, the PCR products were assembled into the PCR linearized pNPTS138-R6KT plasmid, previously amplified by PCR using the primer pair pNPTS-EcoRV-fwd and pNPTS-EcoRV-rev, via Gibson assembly (55). The resulting plasmid, pNPTS138-R6KT-ΔadiY, was verified by sequence analysis using primers M13_rev and M13_uni. To generate the deletion of adiY in E. coli MG1655, at first, the suicide plasmid pNPTS138-R6KT-ΔadiY was introduced into E. coli MG1655 by conjugative mating using E. coli WM3064 as a donor in Luria-Bertani (LB) medium containing 2,3-diaminopropionic acid (DAP). Single-crossover integration mutants were selected on LB plates containing kanamycin but lacking DAP. Single colonies were grown for a day without antibiotics and plated onto LB plates containing 10% (wt/vol) sucrose but lacking NaCl to select for plasmid excision. Kanamycin-sensitive colonies were checked for targeted deletion by colony PCR using primers bracketing the site of the insertion. Deletion of adiY was verified by colony PCR and sequencing using the primers adiY-KO-check-F and AdiY_check_rev, resulting in the strain E. coli MG1655 ΔadiY.

Site-directed mutagenesis of the adiY gene was performed by PCR amplification using primers containing the desired mutations within their overhangs. The plasmid pBAD24-His_6_-adiY served as the template (22). A list of the mismatch-containing oligonucleotide primers is provided in Table 1. The resulting PCR products, encoding the intended amino acid substitutions, were assembled into SmaI-digested pBAD24 using Gibson assembly. Successful construction of the mutant plasmids was confirmed by colony PCR and Sanger sequencing.

Structural predictions

Protein structure predictions were performed using AlphaFold 3 (56–58). The complete amino acid sequence was submitted to the AlphaFold 3 web server using default parameters. AlphaFold 3 applies a transformer-based architecture to capture residue–residue relationships and employs iterative diffusion-based refinement to generate convergent three-dimensional structures. Model outputs include atomic coordinates for all non-hydrogen atoms and confidence metrics such as the predicted Local Distance Difference Test (pLDDT) scores and Predicted Aligned Error maps. The model with the highest overall confidence was selected for further analysis.

For visualization, the final models were loaded into PyMOL (v3.1.4.1) to inspect domain organization, substructure elements, and potential interaction interfaces. Model reliability was assessed using per-residue pLDDT scores, color-coded to indicate prediction confidence (blue: very high; cyan: high; yellow: low; orange: very low). The AdiY-H34A/H60A model displayed high to very high confidence across most secondary structure elements, with reduced reliability restricted to flexible loop regions (see Fig. S6). AlphaFold 3 was also applied in multimeric mode to predict AdiY–DNA interactions using the AdiY sequence and the adiA promoter fragment.

In vivo AdiY activity assay

To evaluate the pH-dependent activity of AdiY and its histidine variants in vivo, we employed a bioluminescent reporter system based on a plasmid carrying the adiA promoter fused upstream of the luxCDABE operon. E. coli MG1655 cells were transformed with the reporter construct and subsequently co-transformed with expression plasmids encoding either wild-type AdiY or variants under the control of an arabinose-inducible promoter (pBAD24-derived vector backbone). Strains were grown overnight at 37°C in LB medium supplemented with the appropriate antibiotics, pH 7.0. The next day, the cells were adjusted to an OD_₆₀₀_ of 0.05 in fresh citrate buffer adjusted to pH 4.4 or pH 7.0 and grown until an OD_₆₀₀_ of approximately 0.5 was reached. They were then incubated in 96-well plates under either aerobic (shaking at 300 rpm) or microaerobic (no shaking) conditions. Induction of the production of AdiY or its variants was achieved by adding L-(+)-arabinose to a final concentration of 0.1% (wt/vol) at the time of inoculation into the 96-well plates. Bioluminescence was recorded at 10-minute intervals over 20 hours using a Tecan Infinite F500 plate reader (Tecan, Crailsheim, Germany), and promoter activity was quantified as the maximum relative light units (RLUs) measured after 2 hours.

To examine how external pH influences AdiY-dependent promoter activation, E. coli MG1655 wild-type and an adiY deletion mutant were each transformed with the reporter plasmids pBBR1-MCS5-PadiA-lux, pBBR1-MCS5-PadiC-lux, or pBBR1-MCS5-PadiY-lux. Cultures were grown in LB medium adjusted to pH values ranging from 4.0 to 7.0. Luminescence and growth (OD_₆₀₀_) were measured every 10 minutes using a Tecan Infinite F500 plate reader (Tecan, Crailsheim, Germany). Data were calculated as RLU, calculated as counts per second per milliliter per OD_₆₀₀_. Data analysis was conducted using GraphPad Prism (version 10.4.1).

AdiY production and purification

N-terminally His6-tagged wild-type AdiY and the AdiY-H34A/H60A variant were expressed in E. coli LMG194 transformed with the respective pBAD24-His_6_-adiY or pBAD24-His_6_-adiY-H34A/H60A plasmids. Cultures were grown in 1 L of terrific broth (TB; 20 g Bacto Tryptone, 24 g Bacto Yeast Extract, 4.65 mL 86% vol/vol glycerol, 89 mM potassium phosphate buffer pH 7.4) at 37°C to an OD_600_ of 0.6 and then induced with 0.2% w/v L-arabinose for 6 hours at 37°C. Cells were harvested by centrifugation, and pellets were stored at –80°C until use. Cell pellets were resuspended in 100 mL of lysis buffer (300 mM NaCl, 50 mM sodium phosphate pH 7.4, 2 mM DTT, 0.2 mM PMSF, and DNase I) and lysed using a high-pressure cell disrupter. The lysate was clarified by sequential centrifugation (5,000 × g for 30 minutes at 4°C to remove debris, followed by 70,000 × g for 1 hour at 4°C to collect the soluble fraction). Proteins were purified by Ni-IDA affinity chromatography (Protino, MACHEREY-NAGEL): after binding, the resin was washed with 10 column volumes of wash buffer (same as lysis buffer with 30 mM imidazole), and proteins were eluted in 1.5 column volumes of elution buffer (50 mM sodium phosphate pH 5.8, 300 mM NaCl, 2% vol/vol glycerol, 2 mM DTT, 5 mM EDTA). Fractions containing AdiY were pooled and further purified by SEC using a HiLoad 16/600 Superdex 75 pg column (Cytiva) equilibrated in 50 mM sodium phosphate pH 5.8 (or 7.4), 300 mM NaCl, 2% (vol/vol) glycerol, 2 mM DTT, and 2 mM EDTA. The column was calibrated with molecular weight standards (aprotinin, ribonuclease A, carbonic anhydrase, ovalbumin, and conalbumin; Cytiva) under both buffer conditions (pH 5.8 and pH 7.4) to ensure accurate calibration at each pH value. Only the fractions corresponding to the AdiY monomer, as determined by comparison to the calibration curve, were collected for subsequent analysis. Final protein samples were concentrated using Vivaspin 20 (10 kDa cutoff, Merck), quantified, aliquoted, and stored at –20°C. Protein production and purification were monitored by SDS–PAGE (12% polyacrylamide gels; 10 µg total protein per lane; PageRuler Prestained Protein Ladder, Thermo Scientific) and Western blotting. Proteins were transferred onto nitrocellulose membranes (0.45 µm, Millipore) and probed with anti-His monoclonal antibodies (1:5,000 dilution; Sigma-Aldrich), followed by anti-mouse alkaline phosphatase-conjugated antibody (Rockland Immunochemicals) as secondary antibody. Localization of the secondary antibody was visualized using colorimetric detection of alkaline phosphatase activity with 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium chloride. Representative results are shown in Fig. S2.

Conformational analysis of AdiY by SEC

The pH-dependent conformational changes of AdiY were analyzed by SEC using a Superdex 75 Increase 10/300 GL column (Cytiva), pre-equilibrated with 50 mM sodium phosphate buffer (pH 5.8 or 7.4) containing 300 mM NaCl, 2% (vol/vol) glycerol, 2 mM DTT, and 2 mM EDTA. SEC was performed on an ÄKTA Pure system with a flow rate of 0.5 mL/minute, using a 0.5 mL protein sample at a concentration of 0.5 mg/mL. The column was calibrated with molecular weight standards as described above. Elution profiles were analyzed using UNICORN 7.0 software, and the apparent molecular weight of AdiY was determined by comparing its elution volume to the standard calibration curve.

Conformational analysis of AdiY by intrinsic tryptophan fluorescence spectroscopy

Intrinsic tryptophan fluorescence measurements were conducted using a FluoroMax-3 (Horiba) at 25°C. The excitation wavelength was set to 280 nm to capture the intrinsic fluorescence from both tryptophan and tyrosine residues, thereby reporting on the global tertiary structure of AdiY. Emission spectra were recorded from 300 to 400 nm. Spectra were collected at 1 nm intervals with an integration time of 1 s per increment and slit widths of 5 nm for both excitation and emission. Purified wild-type AdiY and the AdiY–H34A/H60A were prepared at a final concentration of 1 μM in 50 mM sodium phosphate buffer at pH 5.8 or 7.4, supplemented with 300 mM NaCl, 2% (vol/vol) glycerol, and 2 mM DTT. Fluorescence spectra were recorded for each protein under both buffer conditions to assess pH-dependent structural rearrangements. For comparative analysis, bovine serum albumin (5 μM; Sigma-Aldrich) was used as a control under identical experimental conditions. To reduce inner filter effects, fluorescence intensities were corrected for absorption of the exciting light and reabsorption of the emitted light using the equation Fcorr=Fobs×10^(Aexc+Aem)/2^, where Fcorr is the corrected fluorescence, Fobs the measured fluorescence, Aexc the absorbance at the excitation wavelength, and Aem the absorbance at each emission wavelength; absorbance values were obtained from UV-Vis spectra of the same samples measured under identical conditions (41, 59). Fluorescence data were analyzed in GraphPad Prism 10.4.1 to quantify spectral shifts and changes in intensity. All experiments were performed in triplicate to ensure reproducibility.

SPR spectroscopy

SPR experiments were performed on a Biacore T200 instrument at 25°C using a carboxymethyl dextran sensor chip precoated with streptavidin (SA Sensor Chip Series S; Cytiva). All experiments were performed at 25°C using a running buffer composed of 50 mM sodium phosphate (pH 6.0–7.0), 250 mM NaCl, 2% (vol/vol) glycerol, 2 mM DTT, and 0.005% (vol/vol) detergent P20. For measurements at pH 5.5 and pH 5.8, sodium phosphate was replaced with 50 mM citrate, while all other buffer components and conditions were kept identical. pH 6.0 was used as the standard condition and was included in all experiments to facilitate direct comparison between SPR runs. To immobilize DNA fragments comprising the promoters of adiA, adiC, and adiY (each 200 bp), these promoter regions were amplified by PCR using E. coli MG1655 genomic DNA as the template with the following primers to incorporate a biotin (btn)-tagged PadiA_SPR-F and PadiA_btn-SPR-R, PadiC_SPR-F and PadiC_btn-SPR-R, and PadiY_SPR-F and PadiY_btn-SPR-R, respectively. Before immobilizing the DNA fragments comprising the btn-tagged promoters of adiA, adiC, and adiY, the chip was equilibrated by three injections of 1 M NaCl/50 mM NaOH applied at a flow rate of 10 mL/minute. Next, the biotinylated promoter fragments (10 nM) were injected at a flow rate of 10 mL/minute for a total contact time of 240 s on flow cells 2, 3, and 4, respectively. The chips were then washed by injecting 1 M NaCl, 50 mM NaOH, and 50% (vol/vol) propan-2-ol. About 150 response units (RUs) of the promoter fragments were bound per flow cell. Kinetic analyses of the interactions between purified wild-type His_6_-AdiY and the His_6_-AdiY-H34A/H60A variant with promoter fragments were conducted at 25°C in running buffer at various pH values, using a flow rate of 30 mL/minute. Seven concentrations of either wild-type AdiY or the AdiY-H34A/H60A variant (twofold serial dilutions from 2,000 nM down to 31.25 nM with a repetition of 62.5nM), dissolved in each running buffer, were passed over the flow cells for 120 s, and the complexes formed were allowed to dissociate for 360 s before the next cycle started. After each cycle, the surface was regenerated by injection of 2.5 M NaCl for 30 s, followed by 0.5% (wt/vol) SDS for 30 s, at a flow rate of 30 mL/minute. Sensorgrams were recorded using Biacore T200 Control software version 2.0.2 and analyzed with Biacore T200 Evaluation software version 3.2.2. The surface of flow cell one was not coated and was used to obtain blank sensorgrams for subtraction of the bulk refractive index background. The referenced sensorgrams were normalized to a baseline of 0. Peaks in the sensorgrams at the beginning and end of the injection are due to the run-time difference between the flow cells for each chip. To calculate the association and dissociation rate constants as well as the equilibrium dissociation constants (KD), the sensorgram curves were used, and the kinetics were fitted assuming a Langmuir 1:1 binding model using Biacore T200 Evaluation software version 3.2.2. Visualization of the sensograms was performed in R (R Core Team, 2024 [60]) using the tidyverse and ggplot2 (61) packages within the RStudio IDE (62).

Identification and alignment of AdiY homologs

AdiY homologs were identified with BLASTp (63) against the NCBI RefSeq Select (64) protein database (downloaded September 2025), using the E. coli K-12 MG1655 AdiY sequence as the query. Hits were retained if they met an e-value ≤ 1 × 10⁻100^⁻100^ and were within ±10% of the query length. Partial sequences were excluded. To avoid redundancy, we maintained a single representative per species, resulting in 17 non-redundant orthologs. Sequences were aligned in CLC Main Workbench v24.0.1 (QIAGEN) with the progressive “High accuracy” setting (gap-open cost = 10; gap-extension cost = 1; default protein matrix), treating terminal gaps as internal (65). No manual editing was performed. Accession numbers and organism names are provided in Table S2, and residue positions in the figure are numbered relative to E. coli AdiY.