The PgiDAH7PS-CM protein investigated in this study (UniProt ID: B2RJM7; Gene ID: PGN_RS05050) was classified into the DAH7PS-CM subfamily (PF00793-PF01817) based on domain architecture data retrieved from the Pfam database (Finn et al., 2013). Raw DAH7PS-CM protein sequences for multiple sequence alignment (MSA) were collected from Pfam and subjected to redundancy reduction using CD-HIT (Li et al., 2001), generating 697 non-redundant representative sequences. These sequences were aligned with Clustal Omega (Sievers et al., 2011), followed by targeted truncation to isolate CM domain sequences using Jalview (Waterhouse et al., 2009). Secondary structure predictions for the CM domain were generated via the JPred 4 server (Alexey et al., 2015).

The genes encoding PgiDAH7PS-CM and its variant PgiDAH7PS-CMVar were codon-optimized for expression in E. coli and commercially synthesized (Sangon Biotech, Shanghai, China). These synthetic constructs were amplified via PCR using sequence-specific primers (Supplementary Table 1), followed by restriction enzyme digestion and directional cloning into the pET28a expression vector through NdeI and XhoI restriction sites. This strategy enabled recombinant production of both enzymes featuring an N-terminal 6 × His tag followed by a TEV protease sequence to facilitate subsequent tag removal. Truncated variants PgiDAH7PS and PgiCM were generated through PCR-mediated domain isolation using PgiDAH7PS-CM as template with designed primers (Supplementary Table 1). The amplified fragments were subcloned into the pET28a backbone via the same restriction sites for constructions of corresponding expression plasmids.

The recombinant plasmids were introduced into E. coli BL21 (DE3) competent cells. Transformed cells were cultivated in lysogeny broth (LB) medium supplemented with 50 μg/mL kanamycin at 37°C with vigorous shaking until reaching mid-log phase (OD₆₀₀ of 0.4 – 0.8). Protein expression was induced by adding 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG), followed by 16-h post-induction incubation at 23°C to facilitate proper folding. Harvested cells were pelleted via centrifugation (12,000 × g, 4°C, 15 min) and resuspended in ice-cold lysis buffer containing 50 mM Bistris propane (BTP, pH 7.4), 200 mM KCl, and 2 mM dithiothreitol (DTT). Cellular disruption was achieved through probe sonication (3 s pulse/7 s interval, 20 min duration) on ice, with subsequent clarification of lysates by centrifugation (40,000 × g, 4°C, 40 min).

The supernatant was filtered and loaded onto a pre-equilibrated HisTrap HP column (Cytiva) charged with Ni²⁺ ions, using 20 mM sodium phosphate buffer (pH 7.4) containing 250 mM NaCl and 20 mM imidazole as binding buffer. Target proteins were eluted with a linear imidazole gradient up to 500 mM in the same buffer system. Pooled eluates underwent buffer exchange using PD-10 desalting columns (Cytiva) to remove imidazole prior to tag cleavage. After overnight TEV protease digestion at 4°C, cleaved products were reapplied to the HisTrap column to separate the His-tagged TEV protease, free tags, and uncleaved proteins from the untagged target proteins.

Enzyme kinetic characterization of PgiDAH7PS-CM and its variants was performed through optimized spectrophotometric quantification employing established methodologies (Schoner and Herrmann, 1976). Catalytic activities were monitored in real-time using a temperature-controlled UV-Vis spectrophotometer (37°C) with 1 cm pathlength quartz cuvettes, with DAH7PS activity quantified through PEP depletion at 232 nm and CM activity assessed via chorismate consumption at 274 nm. Reaction velocities were derived from linear absorbance changes and subsequently fitted to the Michaelis-Menten equation through nonlinear regression analysis to determine apparent kinetic parameters (K_m and k_cat).

For DAH7PS kinetic profiling, reaction mixtures contained 50 mM BTP buffer (pH 7.4), 100 μM MnSO₄ and 0.25 μM enzyme, with fixed substrate concentrations maintained at 750 μM for either PEP or E4P while systematically varying the complementary substrate concentration. CM activity assays were conducted under identical buffering conditions using 0.25 μM enzyme, initiated through chorismate addition across a concentration gradient spanning 10–300 μM. All reactions were performed in triplicate with appropriate blank corrections.

The oligomeric state and hydrodynamic properties of PgiDAH7PS-CM and its variants were investigated through analytical size exclusion chromatography (SEC) coupled with calibrated molecular dimension analysis (Le Maire et al., 1986; Uversky, 1993). Chromatographic separations were performed using an ENrich SEC 650 column (Bio-Rad, CA) pre-equilibrated with 50 mM BTP buffer (pH 7.4) at 4°C. Molecular mass determination employed a six-component calibration set containing Vitamin B12 (1.35 kDa), myoglobin (17 kDa), ovalbumin (44 kDa), γ-globulin (158 kDa), thyroglobulin (670 kDa), and blue dextran (2,000 kDa), with each standard injected at 0.5 mg/mL. For Stokes radius (R_s) calculations, an alternative calibration series was implemented using thyroglobulin (8.6 nm), aldolase (4.8 nm), ovalbumin (2.8 nm), myoglobin (1.9 nm), and cytochrome C (1.7 nm), following established methodologies.

Sample proteins were loaded at 1 mg/mL in SEC buffer and eluted isocratically at 0.2 mL/min, with elution volumes (V_e) recorded relative to the void volume (V₀) determined by blue dextran migration. The partition coefficient (K_av) was calculated as (V_e - V₀)/(V_c - V₀), where Vc represents the total column volume. Molecular mass estimation involved plotting log-transformed standard masses (lgMW) against their corresponding K_av values to generate a linear regression model, enabling extrapolation of experimental protein masses from elution profiles. Experimental molecular masses were normalized against theoretical monomeric masses derived from sequence analysis to deduce oligomeric states. While hydrodynamic dimension analysis required transformation of the calibration strategy, with K_av values plotted against the logarithmic Stokes radii (lgR_s) of the alternative standard series. This secondary calibration permitted calculation of experimental R_s values through inverse interpolation of sample retention characteristics.

The three-dimensional architecture of PgiDAH7PS-CM was predicted de novo using AlphaFold3 (Abramson et al., 2024), employing its advanced deep learning framework for multi-chain protein structure modeling. Predicted models were visualized in PyMOL 4.6.0. Oligomeric interface characterization was performed through complementary computational approaches: PDBePISA (EMBL-EBI) for buried surface area quantification and interaction thermodynamics (Krissinel and Henrick, 2007), alongside PLIP (Adasme et al., 2021) for polar and hydrophobic contacts mapping. Comparative structural biology was conducted using the jFATCAT-flexible algorithm within the RCSB PDB Pairwise Alignment Tool (Zhanwen et al., 2020; Sebastian et al., 2024), aligning the predicted PgiDAH7PS-CM structure against the crystallographically resolved the DAH7PS from Pyrococcus furiosus (PfuDAH7PS, PDB ID: 4c1k) with iterative rigid-body and flexible loop adjustments. Structural similarity was quantified through root-mean-square deviation (RMSD) calculations for Cα atoms and template modeling score (TM-score) analysis.

Pfam analysis revealed that PgiDAH7PS-CM comprises 366 amino acid residues, with the DAH7PS domain spanning residues 17 - 253 and the CM domain folded by residues 265 - 356. These two domains are connected by an 11-residue flexible linker (Supplementary Figure 1). To elucidate functional and structural heterodomain interdependencies, the DAH7PS and CM domains (PgiDAH7PS and PgiCM) were split between R258 and I259, and expressed and purified individually, followed by comparative kinetic and quaternary structural analyses.

Kinetic characterization demonstrated significant activity loss upon domain separation. For PgiDAH7PS, the k_cat decreased from 6.3 ± 0.6 s^–1 in the full-length enzyme to 0.58 ± 0.05 s^–1, while substrate affinities for DAH7PS substrates, PEP and E4P, were severely impaired, with K_m values increasing from 96 ± 8 μM to 515 ± 72 μM and 115 ± 10 μM to 718 ± 66 μM, respectively (Table 1). Similarly, the K_m for CM substrate, chorismite, rose from 19.1 ± 1.3 μM to 207 ± 25 μM in separated PgiCM, though its k_cat remained largely unaffected (Table 1). These results indicate that domain separation disrupts catalytic efficiency for both enzymes, underscoring a profound functional interdependence between the DAH7PS and CM domains.

Quaternary structural analysis via analytical SEC revealed distinct assembly patterns between PgiDAH7PS-CM and its two truncated variants. The full-length PgiDAH7PS-CM whose theoretical monomer MW is 42 kDa, was eluted at 13.70 ± 0.14 mL (Figure 2A), corresponding to a molecular weight (MW) of 68 ± 6 kDa (Figure 2B), consistent with a homodimeric assembly. In contrast, PgiDAH7PS (theoretical monomer MW: 29 kDa) eluted at 14.79 ± 0.06 mL (Figure 2A), aligning with a monomeric state (28 ± 2 kDa) (Figure 2B), while PgiCM (theoretical monomer MW: 13 kDa) eluted at 14.87 ± 0.07 mL (Figure 2A), matching a homodimer (26 ± 2 kDa) (Figure 2B). This suggests that the separated DAH7PS barrels lack homo-interactions intensive enough for oligomerization, whereas the CM domain adopts a canonical homodimeric quaternary structure. Therefore, the homodimeric assembly of full-length PgiDAH7PS-CM appears to be predominantly maintained by the dimerization of CM domain. It should also be noted that the so-called disintegration of DAH7PS barrel homodimerization strictly refers to the dynamic equilibrium between monomer and dimer being dramatically shifted toward the monomer, thereby significantly decreasing the content of homodimeric species to levels undetectable by SEC under laboratory conditions.

To elucidate the molecular mechanism underlying the functional interplay between the DAH7PS and CM domains, we generated a structural model for the PgiDAH7PS-CM homodimer using AlphaFold3 open sever (Figure 3A). In this model, the CM domains from both protomers adopt a triple-helix bundle fold and assemble into a homodimeric configuration, as anticipated (Figure 3A). In contrast, the DAH7PS domains exhibit the canonical TIM barrel architecture, with two barrels forming a dimer-like assembly positioned adjacent to the α1 helices of the CM dimer. The overall architecture of the PgiDAH7PS-CM homodimer displays a centrosymmetric arrangement (Figure 3A).

The AlphaFold3 prediction yielded high confidence scores for the majority of the model, with the exception of the helix α3 and the unstructured C-terminal region in the CM domain, which showed lower confidence values of 70–90% and 50–70%, respectively (Figure 3B). Secondary structure analysis of PgiCM via JPred corroborated AlphaFold3’s prediction of a triple-helix bundle topology in the CM domain. Residue segments Q256–E305 and E334–L354 were reliably predicted to adopt helices α1 and α3 (confidence scores > 0.9), while the C-terminal 15 residues (∼ 330 to 345) displayed no stable secondary structural elements (Figure 3B and Supplementary Figure S2). Based on this concordance between computational predictions, we adopted the AlphaFold3 model as a reliable framework for probing intramolecular interactions within the PgiDAH7PS-CM homodimer.

The structural model of PgiDAH7PS-CM revealed two functionally distinct interaction interfaces critical for its architecture and catalytic regulation. The first interface emerges between two DAH7PS barrels (D-D interface), exhibiting an extensive interaction area of ∼ 1451 Ų with a calculated free energy gain at complexation (ΔG^s) of –18.6 kcal/mol (Figure 4A). This interface is stabilized through complementary polar and hydrophobic interactions involving conserved structural motifs: loop β2-α2 (P55, R56, T57), loop β4-α4 (R110, N114, F116), loop β5-α5 (I139 - L143), helix α5 (D144, L145, E151, R152), and loop β6-α6 (S169, Y171) (Figure 4A and Table 2). Notably, loops β2-α2, β4-α4, and β6-α6 do not only contribute to the DAH7PS homo-interaction but also form essential components of the DAH7PS’s active site responsible for substrate binding (PEP and E4P). Particularly, the direct participation of E4P-binding residues (P55, R56, T57) in interface formation further suggests a structural coupling between dimer stabilization and active site conformation (Figure 4A; Table 2; Supplementary Figure 2). This spatial overlap implies that D-D interactions may dynamically modulate the catalytic microenvironment.

A second interface bridges the DAH7PS barrels and the CM dimer (D-CM interface), spanning around 1,560 Ų with a ΔG^s of –16.6 kcal/mol (Figure 4B). Polar interactions dominate this interface, organized into two distinct sub-regions (Table 2). The intra-chain sub-interface involves DAH7PS structural elements (loop β6-α6: T173, T175; helix α6: K189, R190; loop α7-β8: N222) and the CM domain’s helix α1 (E269, E270 and E287) within the same protomer (Figure 4B and Table 2). In contrast, the inter-protomer sub-interface engages DAH7PS helix α6 (Q183) and CM helix α1 (R291) from the adjacent protomer (Figure 4B; Table 2). Strikingly, loop β6-α6, a critical active site element of DAH7PS as aforementioned, serves multiple roles by contributing to substrate coordination and both D-D and D-CM interface formations (Figures 4A,B). Structural alignment further identified that the residue R291 electrostatically coupled with the end of loop β6-α6 (Q183) is positioned adjacent to the conserved active-site residue R292 of the CM domain, suggesting a direct functional crosstalk between the two enzymatic domains potentially (Figure 4C and Supplementary Figure 3).

Structural analysis employing AlphaFold-predicted models demonstrated that polar contacts predominantly stabilize both the D-D and D-CM interfaces of PgiDAH7PS-CM, indicating their critical role in maintaining catalytic competence. Building on prior studies demonstrating that Na⁺ ions act as kosmotropic agents to disrupt protein-protein hydrogen bonds through water competition and attenuate electrostatic interactions via charge screening at elevated concentrations (Wiggins, 1996; Zhang and Cremer, 2006), we systematically evaluated the structural-functional dependence on polar contacts by quantifying catalytic parameters and monitoring conformational alteration under increased NaCl concentrations.

In the absence of NaCl, PgiDAH7PS-CM exhibited robust DAH7PS activity, whereas this activity was nearly abolished under 300 mM NaCl. For the CM domain, detectable activity persisted but was severely attenuated. Kinetic analysis revealed a 10-fold increase in the K_m value for chorismate (from 19.1 ± 1.3 μM to 199 ± 15 μM) and a marginal reduction in k_cat (from 1.12 ± 0.03 to 1.04 ± 0.09 s^–1), indicating impaired substrate binding without significant alteration of catalytic turnover (Table 1).

The AlphaFold model predicts that disruption of the D-D and D-CM interfaces would destabilize the two DAH7PS barrels, leading to their dissociation from the CM dimer. This structural separation is hypothesized to increase the overall molecular dimensions of the protein. Supporting this notion, prior studies on PniDAH7PS-CM revealed significant conformational polymorphism, with radius of gyration (R_g) values ranging from 30 to 46 Å, reflecting its structural flexibility (Bai et al., 2019; Bai and Parker, 2021). To probe the conformational changes in PgiDAH7PS-CM under high ionic strength, we performed analytical SEC in the presence of 300 mM NaCl. Retention volumes were converted to Stokes radius, R_s, defined as the radius of a hypothetical sphere with equivalent hydrodynamic properties to the biomolecule (Le Maire et al., 1986; Uversky, 1993). Under NaCl-free conditions, PgiDAH7PS-CM exhibited an R_s value of 3.4 ± 0.2 nm, consistent with a compact conformation stabilized by extensive interdomain interactions (Figure 5). In contrast, at 300 mM NaCl, the R_s increased to 4.4 ± 0.1 nm, indicating a markedly extended conformation (Figure 5). This expansion aligns with the predicted dissociation of DAH7PS barrels from the CM dimer and suggests that high Na⁺ concentrations destabilize domain interfaces, disfavoring the DAH7PS-CM assembly required for optimal catalytic function.

Structural analysis identified four polar contact-forming residues (E269, E270, E287, R291) in helix α1 of CM domain that interact with DAH7PS barrels at the D-CM interface. To investigate their functional roles, the four residues were all mutated to Ala (E269A/E270A/E287A/R291A), generating variant PgiDAH7PS-CM^Var.

Enzyme kinetics demonstrated substantial impairment of CM activity in PgiDAH7PS-CM^Var, with chorismate binding affinity decreasing 4.2-fold (K_m = 81.1 ± 6.5 μM vs wild-type 19.1 ± 1.3 μM), accompanied by a modest 7% reduction in turnover number, from a k_cat value of 1.12 ± 0.03 s^–1 to 1.04 ± 0.08 s^–1 (Table 1). Strikingly, DAH7PS activity was concurrently compromised, showing ∼ 4.5- and 4-fold increases in apparent K_m values for PEP (436 ± 46 μM) and E4P (511 ± 43 μM), respectively, along with reduced k_cat from 6.3 ± 0.6 s^–1 to 1.28 ± 0.11 s^–1 (Table 1).

SEC analysis revealed pronounced structural perturbations, with PgiDAH7PS-CM^Var exhibiting decreased retention volume of 13.12 ± 0.16 mL compared to the 13.70 ± 0.14 mL for wild-type enzyme (Figure 6). This change in retention volumes was correspondent to an increased R_s from 3.4 ± 0.2 to 4.0 ± 0.2 nm due to the quadruple mutations (Figure 6B), being indicative of a transition from a compact to an extended conformation of PgiDAH7PS-CM.

Collectively, the observed disruption in both CM and DAH7PS catalytic functions, coupled with altered structural compactness of PgiDAH7PS-CM, validated that residues E269, E270, E287, and R291, through their interdomain polar contacts, are critical for preserving the structural integration and heterodomain functional interdependence of PgiDAH7PS-CM.