Drug screening to identify compounds to eliminate Burkholderia pseudomallei through Hcp protein

Summary

Burkholderia pseudomallei (B. pseudomallei) exhibits inherent resistance to multiple antibiotics and Strong pathogenicity, highlighting the urgent need for effective antibiotics. The type VI secretion system (T6SS) is a key virulence factor in B. pseudomallei, and its core structural protein hemolysin coregulated protein (Hcp) represents a conserved, essential target for potential intervention. In this study, Hcp was utilized as a molecular target to screen antibacterial agents from the FDA Orange Book database. Subsequent functional assays and mechanistic analyses identified netilmicin sulfate, which specifically binds to Hcp and exhibits a potent bacteriostatic effect. In vitro and in vivo studies demonstrated that it effectively eliminates B. pseudomallei and enhances host survival post-infection. Combination with ceftazidime exhibits synergism, strengthening host cell protection and reducing the required dosage. Mechanistically, it disrupts Hcp’s tubular structure, compromising T6SS functional architecture and attenuating B. pseudomallei viability. Beyond identifying a promising antibiotic candidate, this approach underscores a potential strategy for future antimicrobial drug discovery targeting T6SS components.

Graphical abstract

Highlights

• •Netilmicin sulfate targets the Hcp protein of T6SS-4 in B. pseudomallei
• •In vitro and in vivo, netilmicin sulfate clears B. pseudomallei and improves host survival
• •Netilmicin sulfate synergizes with ceftazidime, enhancing protection and lowering dosage
• •It disrupts Hcp tubular assembly, compromising T6SS function and bacterial viability

Cell biology; Disease

Introduction

Burkholderia pseudomallei (B. pseudomallei) is a zoonotic pathogen that presents serious biosafety hazards,¹ and it has natural resistance to multiple antibiotics and can easily develop additional resistance in later stages.^2,3 Currently, there is no licensed vaccine, and the clinical management of melioidosis remains challenging.^4,5 It is due to high relapse rates, the necessity for prolonged therapy, significant drug toxicity, and the emergence of antimicrobial resistance.^6,7,8

The standard first-line therapy for melioidosis consists of two phases: an intensive intravenous phase (using agents such as ceftazidime (CAZ), meropenem, or amoxicillin-clavulanic acid) followed by an eradication phase with oral trimethoprim-sulfamethoxazole (TMP-SMX).^4,5,6,7 Prolonged exposure to these common drugs selects for chromosomally encoded resistance mechanisms in B. pseudomallei.⁴ Key well-characterized pathways include^9,10,11: (1) enzymatic inactivation—mutations in the class A PenA β-lactamase (e.g., C69Y and P167S) that enhance hydrolysis of CAZ and amoxicillin-clavulanate; (2) efflux pump overexpression—upregulation of the AmrAB-OprA system conferring resistance to aminoglycosides and macrolides,⁸ and of the BpeEF-OprC pump leading to TMP-SMX resistance; (3) target loss—chromosomal deletion of penicillin-binding protein 3 (PBP3) resulting in CAZ resistance; and (4) enhanced intrinsic barriers—modifications of outer membrane porins and lipopolysaccharide that restrict drug penetration. The urgent need for alternative treatments is underscored by the inherent resistance potential of B. pseudomallei and the growing reports of resistance to first-line drugs in clinical isolates.¹² Therefore, improving public health measures for melioidosis prevention in endemic areas (e.g., safe water supply and protective equipment for high-risk occupations) and discovering therapeutic strategies are of paramount importance.⁹ Many century-old drugs, such as SCH-79797¹⁰ and suramin, have been successfully reused.¹¹ Drug repurposing, the investigation of therapeutic uses for approved drugs, represents a promising strategy that can leverage existing safety data and potentially accelerate clinical translation.^12,13

There are many challenges associated with current standard antibiotics for melioidosis, including those mentioned above, which motivates the search for therapeutic options.^9,14 The type VI secretion system (T6SS) of B. pseudomallei is key to the pathogenicity¹⁵ and virulence of B. pseudomallei,^16,17 and disrupting the structure and function of the T6SS can affect the pathogenicity of this bacterium.^18,19,20 The B. pseudomallei genome encodes six T6SS gene clusters, and each cluster is indispensable.^21,22

Hemolysin coregulated protein (Hcp) is an important structural protein in the T6SS and a virulence factor.²³ Hcp hexamers can be stacked into tubular structures to form the tail of the T6SS.^24,25,26 In addition, it can participate in secretory proteins or chaperones.^27,28 There are currently no reports on antibiotics targeting the Hcp protein of T6SS in B. pseudomallei. The strains isolated from the clinic generally have multidrug efflux pumps, membrane barriers, and the ability to isolate antibiotics.^2,29 The Hcp protein of the T6SS represents a potential drug target. Interfering with its function could provide an effective therapeutic strategy against B. pseudomallei infection.

Based on the above background, in this study, we analyzed the Hcp protein crystal structure of T6SS in B. pseudomallei, screened antibacterial drugs using the Hcp protein as a target, and studied the molecular mechanism of the identified antibacterial drug against B. pseudomallei. This study provides a research approach and suggests its potential as a targeted therapeutic for B. pseudomallei infection.

Results

Overall structure of the Hcp protein of T6SS-4 in B. pseudomallei

The Hcp protein, encoded by the T6SS-4 gene cluster in B. pseudomallei, was produced and purified. The recombinant protein was expressed as a soluble protein after induction with 1.0 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and purified by a HisTrap HP 5 mL column and a SuperdexTM 200 Increase 10/300 GL column (Figure S1A) to obtain a large amount of protein with high purity. The target protein was identified by SDS-PAGE and western blotting, and the results showed that there was a thick single band between 15 and 25 kDa according to the marker (approximately 18 kDa theoretical weight for the Hcp monomer) (Figure S1B).

The molecular weight of the fresh target protein was determined via sedimentation velocity experiments, and the results showed that the protein mainly formed particles of 105 kDa that likely corresponded to hexamers (108 kDa theoretical weight for Hcp hexamers) (Figure S1C). A comparative analysis of Hcp-4 sequences demonstrated that HNBP001’s Hcp-4 exhibits a high degree of similarity when juxtaposed against those from other B. pseudomallei strains, underscoring the presence of a conserved domain associated with oligomerization³⁰ drugs (Figure 1A). To understand the structure-function relationship of the Hcp protein in the TSS6-4 cluster, the crystal structure of the recombinant Hcp protein was determined using the single-wavelength anomalous diffraction (SAD) method. The monomer structure of the Hcp protein was found to consist of one α helix, eight β strands, and a loop, with the α helix between β3 and β4 (Figure 1B). In addition, the contact surface of the Hcp protein dimer consisted mainly of α helix and β6 from one monomer and β2 from the other monomer, with additional amino acid residues (Figure 1C). Overall, the six Hcp protein monomers aggregate into a hexameric circular structure, which can be stacked to form a tubular structure (Figure 1D and Table S1).

Figure 1.

The structure of the Hcp protein in the T6SS-4

(A) Sequence alignment of Hcp-4 of representative B. pseudomallei members, including CAH37616.1, CP038806.1, WP_041195571.1, WP_050043205.1, WP_038765868.1, WP_004529722.1, CAJ4793009.1 and WP_053565664.1 performed using MEGA 7.0. Differential residues are boxed in red and highlighted on a blank background and as red letters, respectively.

(B) Monomer structure of the Hcp protein.

(C) Analysis of the interactions between Hcp protein monomers and the key amino acid sites for these interactions is shown.

(D) Six polymer structure of the Hcp protein.

High-throughput drug screening and kinetic analysis

The Hcp-4 protein, a core component of the T6SS-4 from B. pseudomallei strain HNBP001 (isolated in 2018 from Wenchang City, Hainan Province, China), exhibits natural genetic variation characteristics consistent with those of the Hcp family proteins in B. pseudomallei when compared to the published Hcp-4 sequence of the reference strain K96243.³¹ The Hcp-4 protein features a highly conserved core β-barrel domain, with surface loops serving as functional variation hotspots. Such surface variations do not disrupt the overall tertiary structure of the protein but specifically alter molecular interaction-related functions—a pattern validated by structural alignment studies on the Hcp family proteins from Gram-negative bacteria.²² Based on the crystal structure of the Hcp protein in the T6SS-4, small-molecule drugs were virtually screened for binding to the Hcp protein, and 16 drugs (Table 1) were screened experimentally to verify binding and activity. First, the primary binding screen of 16 candidates against Hcp was performed using nano-differential scanning fluorimetry (nanoDSF) on the Prometheus NT.48 system (PR.NT.48). The assay detects interactions through ligand-induced shifts in the protein melting temperature (Tm). Among the candidates, only netilmicin sulfate (NETS) at 200 μM produced a substantial positive thermal shift (ΔTm >1°C). This increase in thermostability suggests that NETS may directly bind to and stabilize the Hcp protein (Figure 2A). Moreover, the Octet RED96 system (Pall ForteBio LLC, USA) was used to analyze the interaction of 16 drugs with purified Hcp protein by chromatography through biolayer interferometry (BLI). The results showed that only NETS interacted with the Hcp protein, and the interaction mode was slow binding and slow dissociation (Figure 2B).

Table 1.

Sixteen drugs were selected for screening

Number	Drug name	CAS
1	lapatinib ditosylate monohydrate	388082-78-8
2	pemetrexed disodium	150399-23-8
3	troxerutin	7085-55-4
4	tilmicosin	108050-54-0
5	hygromycin B	31282-04-9
6	netilmicin sulfate	56391-57-2
7	fedratinib	936091-26-8
8	ponatinib	943319-70-8
9	madecassoside	34540-22-2
10	asiaticoside	16830-15-2
11	brigatinib	1197953-54-0
12	saikosaponin D	20874-52-6
13	gilteritinib	1254053-43-4
14	pemetrexed disodium hydrate	357166-30-4
15	copanlisib	1032568-63-0
16	remdesivir	1809249-37-3

Figure 2.

High-throughput drug screening and kinetic analysis

(A) Analysis of the thermal stability of the Hcp protein in combination with other drugs. The drug corresponding to each curve is detailed in Table 1.

(B) Analysis of the interaction between the Hcp protein and drugs via BIL. Bio-layer interferometry (BLI) is a label-free technology for analyzing biomolecular interactions, based on the principle of optical interference. BLI enables real-time monitoring of binding events and provides key data such as binding affinity. The Y axis represents the response value (nm), which quantifies the spectral shift caused by changes in biomolecular layer thickness. The response-over-time curve reflects the real-time processes of molecular “association” and “dissociation.”

(C) Drugs can inhibit the growth of B. pseudomallei (HNBP001) in vitro.

(D) Kinetics analysis of the Hcp protein and netilmicin sulfate, using BLI.

We investigated whether 16 candidates could inhibit B. pseudomallei growth in vitro. The growth of B. pseudomallei in vitro was continuously monitored for 24 h by a microplate reader. The results showed that NETS could inhibit B. pseudomallei growth compared with that in the noninhibitor group (Figure 2C). NETS interacted with Hcp and inhibited B. pseudomallei growth; therefore, we further characterized NETS as a candidate inhibitor. Hcp was incubated with different concentrations of NETS, and the kinetic data were fitted using 1:1 Langmuir binding Octet RED96 software to determine the equilibrium dispersion constant (KD). The results showed that the affinity of the Hcp protein for the small-molecule inhibitor NETS was KD (M) = 1.94E-05 (Figure 2D).

Netilmicin sulfate alone and in combination exerted protective effects on an infected A549 cell model

To assess the susceptibility of B. pseudomallei to NETS, we measured the minimal inhibitory concentration (MIC) of NETS against B. pseudomallei. Moreover, gentamicin (GM), SMX/TMP (SXT), and CAZ were selected as control drugs. In addition, a combination of NETS+CAZ was designed. In this study, we defined the MIC as the concentration of inhibitor that resulted in no visible bacterial growth after 20 h of growth at 37 °C. We found that NETS demonstrated bacteriostatic activity against B. pseudomallei (HNBP001). In addition, when NETS was combined with CAZ, the fractional inhibitory concentration³² (FIC) was 1, which reflected a 2-fold decrease in the MIC of NETS and CAZ. The results indicated that their combined effect was additive (Table S2 and Figure S1D).

To evaluate the cytotoxicity of the drugs to A549 cells, we measured the cytotoxic effects of NETS, GM, SXT, CAZ, and NETS+CAZ on A549 cells. The concentration range of the drugs was 0–500 μM, and the results showed that the cytotoxicity of these drugs toward A549 cells was low (Figure 3A). To verify its activity against B. pseudomallei, NETS was experimentally applied to treat infected A549 cells, after which the survival rate of the infected cells was measured. In the B. pseudomallei-infected A549 cell model, exposure to B. pseudomallei resulted in approximately 70% A549 cell death. This effect was mitigated through treatment with NETS. In the present study, as the concentration of the NETS increased, the survival rate of the infected A549 cells also increased (Figure 3B). When the concentration of NETS was 2 × MIC, the survival rate of the infected A549 cells exceeded 50% (Figure 3B). Moreover, in the B. pseudomallei-infected A549 cell model, we tested the protective effects of GM, SXT, CAZ and NETS+CAZ on A549 cells at an inhibitor concentration of 2 × MIC. The survival rate of infected A549 cells treated with NETS was significantly greater than that of infected A549 cells treated with GM or SXT for both early and late intervention (Figure 3C). In addition, the survival rates of infected A549 cells treated with NETS+CAZ were 82.8 ± 5.3% (early intervention) and 72.2 ± 7.7% (late intervention), which were greater than those in the groups treated with NETS or CAZ alone (Figure 3C).

Figure 3.

Netilmicin sulfate alone and in combination exerted protective effects on an infected A549 cell model

(A) Cytotoxic effects of the drugs on A549 cells.

(B) Protective effects of different concentrations of NETS on A549 cells. The early interventions on the abscissa reflect the incubation of A549 cells with the drug for 2 h before B. pseudomallei HNBP001 infection. Late intervention was defined as the introduction of the drug to A549 cells 2 h after infection with B. pseudomallei HNBP001. The ordinate is the survival rate of the infected A549 cells. NC: the blank negative control group of fresh 10% DMEM. Bp: infected A549 cell group. Bp+30 μM NETS: The infected A549 cells were treated with 30 μM NETS; the other conditions were the same, but different concentrations of NETS were used.

(C) Protective effects of several drugs on infected A549 cells; the ordinate is the survival rate of infected A549 cells. NC, blank negative control group of fresh 10% DMEM. Bp: infected A549 cell group. Other: The infected A549 cells were treated with drugs (including SXT, GM, NETS, CAZ, and NETS+CAZ).

(D) Number of intracellular bacteria in infected A549 cells after drug treatment; the ordinate represents the number of bacteria in infected A549 cells. NC is the blank negative control group of fresh 10% DMEM. Bp: infected A549 cell group. Drug group: 2 × MIC drug was used to treat infected A549 cells. All drugs, including SXT, GM, NETS, CAZ, and NETS+CAZ. (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001; ns, no significant difference. Data are represented as mean ± SEM.).

The drug protection assay is a classical method that is used to assess the invasiveness of bacteria by determining the number of surviving bacteria in the lysates of the infected cells. A lower number of surviving bacteria compared to that in the untreated control group indicates a defect in invasion or intracellular survival. The drug protection assay allows us to compare only the bacterial capacity to enter host cells. In this assay, A549 cells were infected with B. pseudomallei in vitro, and the extracellular bacteria were killed by a drug. The internalized bacteria, which were protected from the bactericidal action of the inhibitor, were recovered by lysing the A549 cells and enumerated by counting the colonies that formed on solid medium. Compared with that in the untreated control group, the number of viable bacteria in infected A549 cells treated with NETS was significantly lower for both early and late intervention (Figure 3D). In the early intervention group, the number of viable bacteria in the infected A549 cells treated with NETS+CAZ was significantly lower than that in the infected cells treated with NETS or CAZ alone, and the combination treatment increased the bacteriostatic effect (Figure 3D). In the late intervention group, there were significantly fewer viable bacteria among the A549 cells treated with NETS+CAZ than among those treated with NETS alone, but no significant difference was observed compared to CAZ alone (Figure 3D). This limited efficacy is likely due to the poor cellular penetration of aminoglycosides such as NETS, resulting in sub-inhibitory concentrations inside host cells and thus reduced activity against intracellular B. pseudomallei. A study demonstrated that aminoglycosides, such as netilmicin, have a limited ability to penetrate host cells, which reduces their effectiveness against intracellular bacteria.³³ This finding aligns with our observations on NETS. In summary, these results indicate that NETS has significantly better antibacterial activity against B. pseudomallei than GM and SXT. NETS+CAZ was better than NETS or CAZ alone, with an increased bacteriostatic effect and heightened survival rate of infected A549 cells.

Netilmicin sulfate alone and in combination exerted protective effects on an infected C. elegans model

B. pseudomallei can infect and kill C. elegans.^34,35 Using this system, several possible virulence determinants have been successfully elucidated.^4,36 In this study, compared to those in the blank control group, all the C. elegans in the experimental group were infected with bacteria, indicating that the model was stably established during the experiment. When the B. pseudomallei concentration was 6 × 10⁸ CFU/mL, the DilC18 probe-labeled bacteria were distributed throughout almost the entire digestive tract (Figure S2). We first established that incubating C. elegans with NETS at concentrations 16 times higher than the MIC did not result in greater host toxicity than incubation with the solvent-only control (C. elegans) (Figure 4A). In addition, the host toxicity in C. elegans treated with NETS+CAZ was similar to that caused by treatment with the control antibiotic CAZ (Figure 4A). These results showed that the host toxicity of NETS, CAZ, and NETS+CAZ to C. elegans was weak. Given the promising ability of NETS to kill B. pseudomallei, we sought to determine whether NETS can function as an effective antibiotic in vivo.

Figure 4.

Netilmicin sulfate alone and in combination exerted protective effects on an infected C. elegans model

(A) Drug toxicity of NETS, CAZ, and NETS+CAZ to C. elegans. The abscissa represents the duration of continuous incubation with the drug C. elegans, and the survival rate was observed every 8 h. The ordinate is the survival rate of C. elegans. C. elegans: blank negative control group without drug incubation; the others are the drug groups. 30 μM NETS: C. elegans treated with 30 μM NETS. 3.75 μM CAZ: C. elegans treated with 3.75 μM CAZ. 30 μM NETS+1.88 μM CAZ: C. elegans treated with 30 μM NETS and 1.88 μM CAZ. The other groups are labeled in the same way.

(B and C) Protection efficacy of the drug on C. elegans (B: Early intervention; that is, C. elegans was infected with the bacteria after 2 h of incubation with the inhibitor. C: Late intervention, namely, introduction of the inhibitor 2 h after the infection of C. elegans.) The abscissa shows the duration of continuous incubation with the drug C. elegans, and the survival rate was observed every 8 h. The ordinate is the survival rate of C. elegans. C. elegans: blank negative control group. Bp: the infected C. elegans group. The other groups were infected C. elegans treated with drugs. Bp+60 μM NETS: infected C. elegans were treated with 60 μM NETS. Bp+30 μM NETS+1.88 μM CAZ: C. elegans were treated with 30 μM NETS and 1.88 μM CAZ. The other groups are labeled in the same way. (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Data are represented as mean ± SEM.).

(D) Fluorescence Microscopy: The distribution and burden of DilC18-labeled Burkholderia pseudomallei within nematodes following early and late interventions with different antibiotics were investigated. NETS Group: infected C. elegans treated with BP + NETS. NETS + CAZ Group: C. elegans treated with BP + NETS + CAZ. CAZ Group: infected C. elegans treated with BP + CAZ. Untreated control group: infected C. elegans treated with BP only.

To test its antimicrobial activity in the context of an animal host infection, we focused on the utility of the B. pseudomallei-C. elegans infection model to prove that NETS can rescue worms from infection. In the untreated control group, no worms survived at 48 h post-infection. In the model of infected C. elegans, with increasing NETS concentration, the bacteriostatic effect increased, and the survival rate of C. elegans increased for both early and late intervention (Figures 4B and 4C). When the inhibitory concentration of NETS was 2 × MIC, the survival rate of C. elegans reached more than 50% for both early and late intervention (Figures 4B and 4C). In the model of infected C. elegans, compared with that in the untreated control group, the survival rate of C. elegans treated with NETS, CAZ and NETS+CAZ was significantly greater, with significant differences in both early and late intervention (Figures 4B and 4C). Similar results were recorded in the C. elegans models, indicating that NETS can inhibit B. pseudomallei and increase the survival rate of infected hosts.

To measure the gut bacterial load of C. elegans, B. pseudomallei HNBP001 were labeled with fluorescent DilC18 and used to infect C. elegans (divided into early and late intervention). The distribution and load of the fluorescently labeled DilC18 bacteria in the intestine of surviving C. elegans were observed and recorded. In both early and late intervention, no surviving fluorescently labeled bacteria were found in the intestines of C. elegans, whereas C. elegans in the untreated positive control group were filled with fluorescently labeled bacteria (Figure 4D). Detailed results are provided in Figure S3. The results showed that NETS alone or in combination with CAZ could kill B. pseudomallei.

Molecular docking analysis of netilmicin sulfate and Hcp protein

The molecular docking method was used to predict the visible binding modes and corresponding mechanisms of Hcp-NETS complexes. The best binding site information for NETS and Hcp is shown in Figures 5A and 5B. The details of the interactions between the amino acid residues and the NETS in the three-dimensional docking model and the two-dimensional schematic diagram are illustrated. The docking results showed that NETS formed hydrogen bonds with Hcp residues, including ALA51, VAL53, ILE155, LYS50, and ASP103. In addition to hydrogen bonding, carbon‒hydrogen bonds formed between NETS and ILE155, ASN158, and SER128, and van der Waals forces were also present around the amino acid residues (Figure 5B).

Figure 5.

Molecular docking analysis of netilmicin sulfate and Hcp protein

(A) A three-dimensional model of the docking of the Hcp protein monomer and NETS.

(B) The key amino acid residues and interactions involved in the docking of the Hcp protein monomer with NETS.

Molecular dynamics simulation analysis

To elucidate the inhibitory mechanisms and explore the stability of the Hcp-NETS complexes, we performed molecular dynamics simulations. The root mean square deviation (RMSD), root mean square fluctuation (RMSF), and conformational changes in the active centers of the Hcp-NETS complexes were investigated. In the presence of NETS, the RMSD reached equilibrium after approximately 30 ns of simulation and oscillated stably around ∼0.3 nm (Figure 6A). This evidence clearly shows that the Hcp-NETS complex is stable and balanced. In addition, the amino acid residues that fluctuated greatly according to the calculated the RMSF include 114Ser, 115Arg, 116Asp, 117Asp, 140Asn, 141Ala, 142Gln, 143Gly, 144Gly, 145Ser, 146Gly, 157Gly, 158Asn, and 159Lys (Figure 6B). These results showed that these amino acid residues are more flexible than the other residues and are more strongly affected by NETS. These regions may play a role in protein functions.

Figure 6.

Molecular dynamics simulation analysis

(A) RMSD value of the amino acid skeleton atom of the complex with respect to time.

(B) RMSF value fluctuation diagram of the amino acid skeleton atoms of the complex.

(C) Hydrogen bond formation diagram of the complex.

(D) Composite SASA analysis chart.

(E) Energy disassembly of key residues of the complex.

In the simulation process, the number of hydrogen bonds between molecules or groups and the distribution of hydrogen bond distances or angles were calculated by g_hbond, and the number of hydrogen bonds changed with increasing simulation time, indicating that the Hcp-NETS complex formed hydrogen bonds throughout the simulation (Figure 6C). The solvent-accessible surface area (SASA) increased with increasing simulation time, indicating that more hydrophilic amino acids were exposed (Figure 6D). MM/PBSA is a very useful method for measuring the affinity of macromolecules, especially for protein‒ligand systems. We calculated and estimated the binding free energy of the Hcp-NETS complex using the MM/PBSA method. The complex was simulated by molecular dynamics, and its binding energy was −41.233 kJ/mol ±9.86 kcal/mol, indicating that the binding of NETS to Hcp was spontaneous (Table S3). For the Hcp-NETS complex, the energy contributions of residues 55ASP, 106ILE and 107THR were positive, indicating that these residues may hinder the binding process (Figure 6E). Residues 54LYS, 70MET, 74LEU, 75THR, 76GLY, 105ILE, 108ARG, 109VAL, and 156LYS had negative values and may promote binding (Figure 6E). As shown in the previous docking diagram, these residues bind to the inhibitor NETS to form hydrophobic interactions or hydrogen bonds through intermolecular interactions. These findings showed that these proteins are the key residues involved in NETS binding to the Hcp protein and that they play an inhibitory role.

Discussion

B. pseudomallei poses a serious threat to human and animal health.¹ Currently, consensus guidelines for the treatment of melioidosis recommend the use of biphasic therapy, with an initial intensification stage of parenteral therapy (10–14 days) to reduce acute mortality, followed by continuous oral treatment for several months to eradicate B. pseudomallei and minimize the risk of recurrence.^{37,38,39,40,41} However, antibiotic resistance commonly occurs in clinically isolated strains,^29,39,42 which means that intensive treatment typically requires single-drug therapy with CAZ or carbapenems,⁴³ however, acquired resistance can develop during either treatment phase, but is a particular concern during the prolonged eradication phase. Mechanisms can include chromosomal mutations affecting drug targets (e.g., dihydropteroate synthase mutations for TMP-SMX resistance), enzymatic inactivation (e.g., β-lactamases for CAZ resistance), and upregulated efflux pumps (e.g., BpeAB-OprB),^{2,44,45,46,47,48} which may be related to the longer treatment course, the available single drugs, and the higher drug concentration. For many patients with melioidosis, particularly those with drug-resistant infections or comorbidities (e.g., diabetes), optimal treatment management remains difficult.⁴⁹ Despite numerous antibiotic classes, the number of antibiotics demonstrating consistent efficacy and acceptable safety profiles for both intensive and eradication phases of melioidosis is limited.^9,49 Therefore, the “drug repurposing NETS was approved long ago for clinical use and was marketed” strategy is an encouraging direction for the treatment of melioidosis.

The B. pseudomallei genome encodes six T6SS gene clusters, T6SS-1∼T6SS-6.^18,21 The T6SS is a key determinant of B. pseudomallei virulence.¹⁶ As a core structural and secretory component of the T6SS, the Hcp protein plays a dual role: It forms the inner tube of the apparatus and functions as a carrier for effector proteins and their cognate chaperones, contributing significantly to bacterial virulence.^23,27,28,42 The crystal structures of Hcp homologs from P. aeruginosa, enteroaggregative E. coli, E. tarda, and Y. pestis have been solved.^50,51,52 In addition, the crystal structure of the Hcp protein encoded by the T6SS-1 gene cluster on the B. pseudomallei chromosome has been solved.²² In this study, we analyzed the crystal structure of the Hcp protein in the TSS6-4 cluster, which can be stacked into a tubular structure similar to that in the T6SS-1 cluster.

The Hcp protein is not only the substrate of the T6SS but also a part of the T6SS secretory duct.^24,25,53 However, antibacterial drugs can exert antibacterial effects by binding to the Hcp protein and disrupting the function of the T6SS. Based on the crystal structure of the Hcp protein of T6SS-4, through virtual screening of drugs in the FDA Orange Book database and a series of experimental studies, we found that the antibiotic NETS binds to the Hcp protein and kills B. pseudomallei.

Specifically, NETS is a semisynthetic aminoglycoside antibiotic. It binds irreversibly to the 30S subunit of the bacterial ribosome and acts as a bactericide with a minimum bactericidal concentration (MBC) equal to or close to the minimum inhibitory concentration (MIC).^54,55 NETS is a clinically approved aminoglycoside antibiotic that has been used for decades, making it a candidate for drug repurposing for use in treating infections caused by Pseudomonas, Proteus, Enterobacteriaceae, Escherichia coli, Klebsiella, Citrobacter, Salmonella, and so forth.^56,57 In addition, compared with other currently available aminoglycoside drugs, NETS has a lower incidence of nephrotoxicity and ototoxicity in animals, and it can be taken with ototoxic drugs and aminoglycoside drugs for a long time.^54,58 GM is also an aminoglycoside antibiotic. It blocks protein synthesis by binding to the 30S subunit of the bacterial ribosome to achieve irreversible inhibition of bacterial proliferation.⁵⁹ However, in our study, we found that GM requires a high concentration to be bacteriostatic against B. pseudomallei, whereas NETS is bacteriostatic against B. pseudomallei at a relatively low concentration, and the difference between them is statistically significant. This difference in potency may be attributed to the ability of netilmicin to effectively target B. pseudomallei, potentially through specific interaction with components such as the Hcp protein of T6SS-4.^49,60 From the perspective of mechanism, we believe that NETS can not only irreversibly bind to the 30S subunit of the B. pseudomallei ribosome and act as a bactericide, but also effectively target specific interactions with components such as the Hcp protein of T6SS-4 in B. pseudomallei to achieve bactericidal effects. In addition, we are well aware of the importance of conducting additional experiments to distinguish between Hcp-specific versus ribosomal effects. Unfortunately, we tried different approaches but none obtained Hcp-deficient or 30S-deficient strains, which may be attributed to the fact that both genes are critical genes in B. pseudomallei. To further understand the mechanism of action of NETS against B. pseudomallei, we used molecular docking and molecular dynamics simulation methods to simulate the interaction mechanism between NETS and the Hcp protein of T6SS-4. We found that NETS spontaneously binds to the Hcp protein of T6SS-4, and the complex is stable and balanced. The amino acid residues 55ASP, 106ILE, and 107THR of the Hcp protein may hinder the binding process. Mutation of these residues could increase the affinity of NETS for the Hcp protein, thereby reducing the consumption of NETS and enhancing its ability to kill B. pseudomallei. In addition, the amino acid residues 54LYS, 70MET, 74LEU, 75THR, 76GLY, 105ILE, 108ARG, 109VAL, and 156LYS of the Hcp protein may promote binding. NETS plays an inhibitory role through the formation of hydrophobic interactions, hydrogen bonds, and van der Waals interactions with these amino acid residues through intermolecular forces. The function of T6SS-4 is closely related to the virulence, nutrient uptake, growth, and reproduction of B. pseudomallei,^15,18 and disrupting the structural function of T6SS-4 can affect the toxicity and survival of B. pseudomallei. In summary, we found that NETS combines with the Hcp protein monomer of T6SS-4 to destroy the functional structure of the Hcp protein, affecting the function of T6SS-4 and killing B. pseudomallei.

Our study revealed that, in a B. pseudomallei-A549 cell infection model, approximately 70% of A549 cells died after exposure to B. pseudomallei, but this effect was ameliorated through NETS treatment. At present, the first-line antibiotics commonly used in clinical treatment include CAZ, imipenem, meropenem, and the compound SMX.³⁸ The survival rate of infected A549 cells treated with NETS was significantly greater than that of infected A549 cells treated with SMX. These results indicate that NETS can be used clinically to replace the compound SMX in the treatment of melioidosis. The survival rates of infected A549 cells treated with NETS combined with CAZ were 82.8 ± 5.3% (early intervention group) and 72.2 ± 7.7% (late intervention group), which were greater than those of cells treated with NETS or CAZ alone. Currently, in clinical practice, antibiotics such as CAZ are commonly used for the treatment of acute and chronic melioidosis.⁴³ However, with the prolongation of treatment time and repeated use of antibiotics, drug resistance gradually occurs, which may be attributed to long-term treatment and high drug concentrations.^2,29,47 Therefore, a more comprehensive treatment regimen is also needed. At the cellular level, NETS increased the survival rate of infected host cells. In the late intervention group of this study, the efficacy of using NETS alone was not statistically different from that of cefotaxime alone (p > 0.05), but their combined use was more effective than either drug alone (p < 0.05), and it could decrease the concentration of each drug, potentially reducing the risk and extent of resistance. Clinical cases are mostly in the late intervention group, and our research results provide insights for their treatment.

To test the activity of NETS in animals, we used the B. pseudomallei-C. elegans infection model. Research has shown that B. pseudomallei can kill C. elegans.^34,35 Using the B. pseudomallei-C. elegans model, relevant antibacterial drugs have been successfully screened^4,36; therefore, using this model can preliminarily reveal the infection status of B. pseudomallei in the body. In our study, NETS had extremely low toxicity toward C. elegans. In the untreated control group, no C. elegans survived. Compared with that in the untreated control group, the survival rate of C. elegans in the treatment group was significantly greater. Infected C. elegans and A549 cell models were protected, which further confirmed that NETS can kill B. pseudomallei and that NETS combined with CAZ can reduce the drug concentration and increase the survival rate of infected hosts. The use of CAZ combined with other antibiotics is common in the treatment of bacterial infections, such as CAZ, avibactam, and aztreonam, which have a synergistic effect against gram-negative pathogens that express serine β-lactamases.^61,62 Our study proved that NETS combined with CAZ antagonizes B. pseudomallei via a synergistic effect, which provides a theoretical basis for the combined use of drugs in the clinical treatment of melioidosis. Under “therapeutic drug monitoring,” we recommend NETS combined with CAZ during the acute and severe stages of melioidosis. After the acute and severe conditions have eased, treatment should be switched to CAZ alone. This approach can quickly alleviate acute and severe symptoms and reduce the early use of high doses of CAZ, potentially further improving resistance.

In summary, Hcp was used to screen antibiotics, and it was found that NETS affects the functional structure of T6SS through binding to Hcp monomer, and this mechanism plays a role in killing B. pseudomallei (Figure 7). Further research revealed that the ability of the NETS to kill B. pseudomallei was greater than that of the current clinical drug compound SMX. In addition, the observed synergistic effect between NETS and CAZ suggests the potential to use lower doses of each drug, which in theory could reduce the selective pressure for resistance development. This study addresses a current research gap by providing a potential strategy for the development of targeted therapeutics against B. pseudomallei infection.

Figure 7.

Mechanism of action of netilmicin sulfate against B. pseudomallei

Limitations of the study

While this study identifies Hcp as a promising target and NETS as an effective agent, several limitations must be acknowledged. First, the natural sequence variation in Hcp across different B. pseudomallei strains, particularly in surface loops, pose a consideration for the universality of this targeting strategy. Although our structural and efficacy data suggest the drug binds a conserved core domain, the potential impact of extreme sequence variation on drug binding affinity cannot be ruled out. Second, this study focused specifically on the Hcp protein of the T6SS-4 cluster. Whether NETS similarly disrupts the functional architecture of Hcp proteins in other T6SS clusters remains to be investigated. Third, due to the unavailability of isogenic Hcp-deficient or 30S ribosomal subunit-deficient mutant strains, we could not definitively dissect the contribution of Hcp-specific effects from potential ribosomal effects in the bacteriostatic activity of NETS. In addition, unfortunately, due to limitations in the experimental conditions, it was not possible to observe the degree of damage to the structure of the T6SS through cryo-electron microscopy of ultrathin sections of live bacteria. Looking forward, beyond addressing these limitations, further optimization of the NETS scaffold to enhance its anti-B. pseudomallei potency and reduce effective concentrations represents a promising direction for medicinal chemistry efforts.

Resource availability

Lead contact

Further information and requests for resources and reagents may be directed to and will be fulfilled by the lead contact, Qianfeng Xia (xiaqianfeng@muhn.edu.cn).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed Materials Transfer Agreement.

Data and code availability

• •This study uses data extracted from the U.S. Food and Drug Administration (FDA) Orange Book (https://www.fda.gov/drugs/drug-approvals-and-databases/orange-book-data-files), which is a publicly available database. The Orange Book data are government-published information in the public domain and are not subject to copyright or access restrictions. No identifiable personal or patient information is involved in this study.
• •The atomic coordinates and structure factors data have been deposited at the Protein DataBank (PDB) and are publicly available as of the date of publication. Accession numbers are listed in the key resources table.
• •This article does not report any original code.
• •The authors confirm that the data supporting the findings of this study are available in the article and its supplemental information or by consulting this link https://figshare.com/s/2ffd2a6e7d1e4a63eb5c (DOI: https://doi.org/10.6084/m9.figshare.28423724).

Acknowledgments

This study was sponsored by the Hainan Provincial Natural Science Foundation of China [822QN318]; National Natural Science Foundation of China [82120387], the National Natural Science Foundation of China [81960002], the National Natural Science Foundation of China [82370018], the Major Science and Technology Program of Hainan Province [ZDKJ202003, ZDKJ2021036], and the Graduate Research Innovation Project in Hainan Province [Qhyb2023-170].

Author contributions

Q.F.X, L.P.D, B.W., and C.Z.C., conceptualization and project administration; C.Z.C., X.M.L., and L.P.D., methodology; X.M.L., L.P.D., B.W., X.L., and K.J.Y., investigation; C.Z.C., Q.F.X., B.W., visualization; C.Z.C. and Q.F.X., writing original draft; C.Z.C., Q.F.X., L.P.D., and B.W., writing review and editing. All authors have read and approved the final manuscript.

Declaration of interests

All authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-His-tag mAb-HRP-DirecT	MBL	Code No. D291-7
Bacterial and virus strains
Burkholderia pseudomallei HNBP001	This paper	N/A
Escherichia coli BL21 (DE3)	Thermo Fisher Scientific	Cat#C600003
Escherichia coli OP50	Caenorhabditis Genetics Center	Cat#OP50
Pseudomonas aeruginosa ATCC 27853 (MIC quality control)	ATCC	Cat#27853
Biological samples
C. elegans N2 (wild-type)	Caenorhabditis Genetics Center	WormBase ID: N2
Chemicals, peptides, and recombinant proteins
Netilmicin sulfate (NETS)	TargetMol	Cat#T1036; CAS: 56391-57-2
Gentamicin (GM)	TargetMol	Cat#T25447; CAS: 1403-66-3
Sulfamethoxazole/Trimethoprim (SXT)	TargetMol	Cat#T24899; CAS: 8064-90-2
Ceftazidime (CAZ)	TargetMol	Cat#T1305; CAS: 72558-82-8
Recombinant Hcp protein (C-terminal 6×His tag)	This paper	N/A
IPTG (isopropyl-β-d-thiogalactopyranoside)	Aladdin	Cat#206-703-0
DilC18 dye (DiI)	Invitrogen	Cat#D3911
CCK-8 reagent	Beyotime	Cat#C0038
Deposited data
Hcp protein X-ray crystallographic structure	This paper	PDB: 8YUN
PDB validation report	This paper	Submitted as separate file
FDA Orange Book drug dataset	U.S. FDA	https://www.fda.gov/drugs/drug-approvals-and-databases/orange-book-data-files
Raw data for figures	This paper; Figshare	DOI: https://doi.org/10.6084/m9.figshare.28423724;;https://figshare.com/s/2ffd2a6e7d1e4a63eb5c
Hcp structure model (template for molecular replacement)	PDB	PDB: 4TV4
Experimental models: Cell lines
A549 human lung carcinoma cell line	Procell	Cat#CL-0016
Experimental models: Organisms/strains
C. elegans: Strain N2 (wild-type)	Caenorhabditis Genetics Center	WormBase ID: N2
Recombinant DNA
pET30a(+)-Hcp (C-terminal 6×His tag)	This paper	N/A
Software and algorithms
HKL2000	HKL Research	http://www.hkl-xray.com
CCP4 program suite	CCP4	http://www.ccp4.ac.uk
Phaser	CCP4	http://www.ccp4.ac.uk/html/phaser.html
Coot	Coot	https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot
Phenix	Phenix	https://www.phenix-online.org
MolProbity	Duke University	http://molprobity.biochem.duke.edu
PyMOL	Schrödinger	https://pymol.org
Dock6	UC San Francisco	http://dock.compbio.ucsf.edu
AutoDock Vina	NIH	http://vina.scripps.edu
AutoDock Tools 1.5.6	NIH	http://mgltools.scripps.edu
Discovery Studio Client 2019	Dassault Systèmes	https://www.3ds.com/products-services/biovia
GROMACS 2019	GROMACS	http://www.gromacs.org
AmberTools	Amber	https://ambermd.org
Yinfu Cloud Platform	Yinfotek	https://cloud.yinfotek.com
MEGA 7.0	MEGA	https://www.megasoftware.net
Octet RED96 software	Pall ForteBio	Instrument software, no public download link (https://www.sartorius.com.cn/products/protein-analysis/biolayer-interferometry/octet-red96e)
SEDFIT	NIH	https://sedfitsedphat.nibib.nih.gov
GraphPad Prism 7.0	GraphPad Software	https://www.graphpad.com
Other
Crystallization kits (PEGRx, PEG/Ion, SaltRx, Crystal Screen, Index, Natrix, Structure Screen, Stura Foot Print, MacroSol, PCAT, JCSG Plus, Proplex, Morpheus, PGA Screen, MIDAS, Wizard Classic)	Hampton Research, Molecular Dimensions	See Table S4 for complete catalog numbers
AKTA Pure	GE healthcare	Cat#29-0182-24
His-Trap HP 5 mL columns	GE Healthcare	Cat#17-5248-05
Superdex 200 Increase 10/300 GL	Cytiva	Cat#28990944
Super-Streptavidin (SSA) biosensor	Pall ForteBio	Cat#18-5057
10 kDa-cutoff Millipore Ultra centrifugal filter	Millipore	Cat#UFC801024

Experimental model and study participant details

Cell line model

The A549 human lung carcinoma epithelial cell line was purchased from Procell (Wuhan, China; Cat# CL-0016). The cell line was authenticated by the supplier using short tandem repeat (STR) profiling. Cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37°C in a 5% CO₂ humidified incubator. The cell line is of male origin. For experiments, cells were seeded in 96-well plates and cultured overnight prior to treatment. Cells were regularly tested for mycoplasma contamination using a PCR-based method and were confirmed to be negative throughout the study.

Bacterial strain

Burkholderia pseudomallei HNBP001 was used for all infection experiments. Bacteria were cultured in LB broth at 37°C with shaking until reaching mid-log phase, then diluted to approximately 5 × 10⁷ CFU/mL in fresh medium for cell and worm infection assays.

C. elegans model

The Caenorhabditis elegans (C. elegans) wild-type strain N2 was obtained from the Caenorhabditis Genetics Center (CGC, University of Minnesota, USA). Worms were maintained on nematode growth medium (NGM) agar plates seeded with Escherichia coli OP50 at 20°C under standard conditions. Age-synchronized L4-stage hermaphrodites were used for all infection experiments. For each experimental group, 20 L4-stage worms were examined per biological replicate.

Experimental groups

Seven experimental groups were included: untreated control, blank negative control, and five drug treatment groups (NETS, GM, SXT, CAZ, and NETS+CAZ). For both cell and worm models, infections were divided into two intervention time points: early intervention (treatment 2 h before bacterial infection) and late intervention (treatment 2 h after bacterial infection).

Sex information

The A549 cell line used in this study is of male origin. All C. elegans used are hermaphrodites, the default sex of the N2 strain under standard culture conditions. As the study focused on responses to bacterial infection and drug treatment without comparison between sexes, sex was not considered as a biological variable.

Ethical approval

This study was approved by the Ethics Committee of Hainan Medical University (Approval No. HYPLL-2025-004). All experiments involving A549 cells (male origin) and Caenorhabditis elegans were performed in compliance with the relevant institutional and national guidelines and regulations.

Method details

Protein expression and purification

The coding sequence of the Hcp protein in T6SS-4 was synthesized, codon-optimized, inserted into the expression vector pET30a (+), which encodes a noncleavable C-terminal 6×His fusion tag, and subsequently expressed in the E. coli expression system. The recombinant protein was expressed in the E. coli strain BL21 (DE3) as a soluble protein after induction with 1.0 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at OD₆₀₀ = 0.8 and expression at 16 °C for 18 h. The cells were lysed by sonication in lysis buffer (20 mM Tris, 150 mM NaCl, pH 8.0). The cell debris was discarded via centrifugation (10,000 rpm, 4 °C, 30 min), and a 0.22 μm filter was used to collect the supernatant. The protein solution was purified by affinity chromatography using His-Trap HP 5 mL columns (GE Healthcare), and the target protein was eluted using buffer A (20 mM Tris, 150 mM NaCl, pH 8.0) supplemented with 300 mM imidazole. The eluted fractions were pooled and concentrated using a 10 kDa-cutoff Millipore Ultra centrifugal filter for further purification by size exclusion equilibrated with PBS or binding buffer (20 mM Tris-HCl, 150 mM NaCl, pH 8.0). The eluted peak fractions were collected and confirmed by SDS–PAGE and western blotting.

Sedimentation velocity experiments

Sedimentation velocity experiments were performed in a Proteome Lab XL-I analytical ultracentrifuge (Beckman Coulter, Brea, CA) equipped with an AN-60Ti rotor (4-holes) and conventional double-sector aluminum centerpieces with a 12 mm optical path length. The samples were loaded with 380 μL of sample and 400 μL of buffer (20 mM Tris, 150 mM NaCl, pH 8.0). Before the run, the rotor was equilibrated for approximately 1 h at 20 °C in the centrifuge. Then, the experiments were carried out at 20 °C and 42,000 rpm using continuous scan mode and a radial spacing of 0.003 cm. Scans were collected at 3 min intervals at 280 nm. The fitting of absorbance versus cell radius data was performed using SEDFIT software (https://sedfitsedphat.nibib.nih.gov/software/default.aspx) and the continuous sedimentation coefficient distribution c(s) model, covering the range of 0–20S. The biophysical parameters of the buffer were as follows: density ρ = 1.006 g/cm3, viscosity η = 0.01031P, and protein: partial specific Volume V-bar = 0.73000 cm³/g.

Crystallization

Commercially available kits from Hampton Research, Molecular Dimensions Ltd (Table S4). and Emerald Biosystems were used to screen for initial crystallization of the target protein. The purified Hcp-His crystals were screened in 96-well crystallization plates. The purified fresh recombinant protein was concentrated to 5 mg/mL and 10 mg/mL by using a 10 kDa-cutoff Millipore Ultra centrifugal filter. All the crystals were obtained by using the sitting drop vapor diffusion method, in which 1 μL of protein and 1 μL of reservoir solution were mixed and then allowed to equilibrate against 100 μL of reservoir solution at 18 °C. After 3 days, the Hcp protein was crystallized with good quality in 20% v/v 2-propanol, 0.1 M Tris (pH 8.0), and 5% w/v polystyrene glycol (8,000). Crystals were exchanged from crystallization solution to cryoprotectant, harvested and immersed in liquid nitrogen for rapid cooling and temporary preservation.

Data collection, processing and refinement

Diffraction data were collected at the Shanghai Synchrotron Radiation Facility (SSRF) BL17U1 (wavelength, 0.979183 Å), and a 2.7 Å resolution dataset was collected. HKL2000 software was used to process the datasets. The structure of Hcp (PDB entry 4TV4) from Burkholderia pseudomallei 1710b was used as a model. The structure of the target protein was solved by molecular replacement using the Phaser⁶³ package in the CCP4 program suite.^64,65 The atomic model was completed with Coot and refined with phenix.refine in Phoenix and the stereochemical quality of the final model was evaluated with MolProbity. The Hcp protein X-ray crystallographic coordinates and structure factor files were deposited in the RCSB Protein Data Bank (PDB). The coordinates⁶⁶ (https://www.rcsb.org/) with accession number 8YUN. Table S1 summarizes the data collection, processing and improvement of the statistics. PyMOL was used to analyze and prepare structural diagrams.⁶⁷

Virtual screening

Virtual screening (VS) was conducted using the Dock6 protocol in the Yinfu Cloud Platform (https://cloud.yinfotek.com/). The crystal structure of the Hcp protein (PDB ID: 8YUN). The U.S. Food and Drug Administration (FDA) Orange Book dataset of approximately 2,858 compounds was used. The crystal ligand was used to define the binding pocket. The DOCK 6^68,69 program was utilized to execute semiflexible docking, and output poses were evaluated by the Grid scoring function. Finally, the binding modes of the top 100 docked compounds were visually analyzed, and hits were manually selected.

High-throughput inhibitor screening and assay

The Octet platform is based on biolayer interferometry (BLI) and is used to detect and analyze the interactions of biomolecules. This platform can be used to measure the interactions between proteins and biomolecules using only microsamples. In this study, the interactions between the Hcp protein (random biotinylation) and the inhibitor (purchased from TargetMol) were analyzed via the Octet RED96 System (Pall ForteBio LLC, USA). The assays were performed in the same buffer, which was PBS supplemented with 0.02% Tween 20 and 0.1% BSA buffer, using a volume of 200 μL for all the incubations. The Super-Streptavidin (SSA) biosensor was loaded with 20 μg/mL random biotinylated Hcp protein and incubated with drugs at different concentrations (7.41 μM, 22.2 μM, 66.7 μM and 200 μM). Using Octet RED96 software, the kinetic datasets were processed using Double Reference and fitted using 1:1 Langmuir binding to yield the equilibrium dissociation constants (KD).

In addition, the interactions between the Hcp protein and the drug(purchased from TargetMol) were analyzed by Prometheus NT.48 (PR.NT.48). The assays were performed by incubating Hcp protein (1 mg/mL) with various drugs (200 μM) in the same buffer, 1×PBS, using a volume of 10 μL for all the incubations.

Growth of B. pseudomallei in vitro

Burkholderia pseudomallei is HNBP001 preserved by our laboratory. The samples were divided into an experimental group (NETS, GM, SXT, CAZ and NETS+CAZ), a positive control group (DMSO), and a negative control group (MH culture medium). The antibiotics for susceptibility testing were selected to include a first-line therapeutic (ceftazidime, CAZ), the standard eradication phase drug (trimethoprim-sulfamethoxazole, SXT), and an aminoglycoside control against which B. pseudomallei is typically resistant (gentamicin, GM).⁷⁰ Prior to these experiments, the HNBP001 strain was confirmed to exhibit the typical intrinsic resistance profile to gentamicin (MIC >32 μg/mL).⁹ The final concentration of the inhibitor was 200 μM. The final concentration of the bacterial solution was 1 × 10⁴ CFU/ml. Experiments were performed in six independent assays. The OD₆₀₀ values in the presence of different drugs were continuously monitored for 24 hours by a microplate reader.

MIC measurement by the 96-well dilution method

The strains were divided into an experimental group (NETS, GM, SXT, CAZ and NETS+CAZ), a positive control group infected with B. pseudomallei (HNBP001) and untreated, a negative control group cultured in MH culture medium, and a quality control group infected with a standard strain of P. aeruginosa (ATCC27853). The final concentrations of the inhibitor were 0.94 μM, 1.88 μM, 3.76 μM, 15 μM, 30 μM, 60 μM, 120 μM and 240 μM. The experiments were performed in triplicate. For this purpose, 100 μL of fresh inhibitor-containing medium was added directly to each well of a 96-well inhibitor-sensitive plate. HNBP001 and ATCC27853 single clones were taken and mixed in MH broth culture medium, and the bacterial suspension was added to a 0.5 Maxwell standard at approximately 1× 10⁸ CFU/mL. Then, the bacterial solution was diluted 100 times (approximately 1× 10⁶ CFU/mL) with MH broth culture medium, and 100 μL of the bacterial solution was added to the above inhibitor-sensitive plate well so that the final concentration of the bacterial solution was 5 × 10⁵ CFU/mL. The bacteriostasis effects of different concentrations of each inhibitor were observed and recorded after the samples were sealed and incubated for 20 h at 37 °C.

Cytotoxicity assay

For cytotoxicity assessment against A549 cells, the protocol included druggroups (NETS, GM, SXT, CAZ and NETS+CAZ) and blank negative controls. First, we used 10% DMEM to prepare media with different inhibitor concentrations: 31.5 μM, 62.5 μM, 125 μM, 250 μM, and 500 μM. A549 cells were cultured overnight in 96-well plates and washed three times with PBS. The medium was removed, and fresh drug-containing medium was added to the cells. Fresh 10% DMEM was added to the A549 cells as a blank negative control. After 24 h, the medium was removed from the 96-well plate, the wells were washed three times with PBS, and the cells were observed and recorded under a microscope. Next, 100 μL of CCK8 reagent (C0038, Beyotime, China) was added directly to each well. The OD₄₅₀ was measured by using a multifunction automatic microplate reader after 25 min of incubation at 37 °C. Experiments were performed in six independent assays. The OD₄₅₀ value in the presence of different concentrations of compound was divided by the OD₄₅₀ value of the negative control to calculate the percentage cytotoxicity. Statistical analysis was performed using GraphPad Prism 7.0 software.

Cell viability assay

In this study, there were drug groups (NETS, GM, SXT, CAZ and NETS+CAZ), an untreated control group and a blank negative control group. In the drug groups, A549 cells infected with B. pseudomallei were treated with different drugs at different concentrations. A549 cells were purchased from Procell with the product number CL - 0016. A549 cells were seeded in 96-well plates and cultured overnight. The medium was removed, and 10 μL of fresh B. pseudomallei HNBP001 culture (approximately 5 × 10⁷ CFU/mL) was added to the cells. Fresh 10% DMEM was added to A549 cells as a blank negative control. B. pseudomallei infection was divided into early and late intervention (early intervention: A549 cells incubated with an inhibitor 2 hours before infection with B. pseudomallei; late events: A549 cells incubated with an inhibitor 2 hours after infection with B. pseudomallei). After 24 h, the medium was removed from the 96-well plate, the wells were washed three times with PBS, and the cells were observed and recorded under a microscope. Then, 100 μL of CCK-8 reagent (C0038, Beyotime, China) was added to each well directly. The OD₄₅₀ was measured by using a multifunction automatic microplate reader after 25 min of incubation at 37 °C. Experiments were performed in six independent assays. The OD₄₅₀ value in the presence of different concentrations of compound was divided by the OD₄₅₀ value of the negative control to calculate the percentage cell viability. The inhibitor concentrations of NETS were 30 μM, 60 μM, 120 μM, 240 μM and 480 μM. The concentration of SXT was 60 μM. The concentration of GM was 240 μM. The concentration of CAZ was 3.75 μM. The concentration of NETS+CAZ was 60 μM NETS and 3.75 μM CAZ.

The number of intracellular bacteria in CFU/mL at different drugs

The design of the experiment, preparation of bacteria, preparation of A549 cells and procedure of infection were the same as those used for the cell viability assay. After incubation with the drug, the wells were washed three times with PBS. The PBS was aspirated from the wells, and 200 μL of sodium 0.2% X-Triton100 solution was added to lyse the A549 cells (the bacteria resisted 0.2% X-Triton100) for 30 min. Using a 1 mL pipette, the cells were scraped off the surface and pipetted up and down to lyse the cells effectively and homogenize the lysate. Log serial dilutions of this lysate were prepared in sterile PBS as follows: 450 μL of sterile PBS was dispensed into 1.5 mL microcentrifuge tubes. Then, 50 μL of cell lysate was added to one tube, marked as the 10^-1dilution. The tube was vortexed for a few seconds, after which 50 μL of suspension was transferred from the tube to the next tube containing 450 μL of sterile PBS. This new tube was marked as the 10^-2 dilution. This procedure was repeated to obtain further dilutions of the lysate. Sectors were marked on the bottom of agar plates. A 5 μL drop of diluted suspension (10^-1/10^-2/10^-3/10^-4 as labeled on the plate) was added to one of the sectors on the MH plate and allowed to dry. At least six drops of each dilution were added (representing technical replicates). On another plate, sectors were marked in a similar way (on the bottom of the plate), and drops of suspension prepared from the lysate of a duplicate well were added. This plate served as another technical replicate. The plates were incubated overnight at 37 °C.

The colonies in the drop were counted. Only the dilutions in which the number of colonies was within the countable range (3-30 colonies) were considered. From the raw colony count, the CFU/mL was determined using the following formula:

CFU/mL = No. of colonies × dilution factor × 100

We used GraphPad Prism 7.0 to plot the number of bacteria expressed in CFU/mL for each inhibitor to be analyzed, and we performed statistical analysis using the same software. For comparing only two sample values at a fixed time point, Student’s t test was used. The inhibitor concentration of NETS was 60 μM. The CAZ concentration was 3.75 μM. The concentration of NETS+CAZ was 60 μM or 3.75 μM. Experiments were performed in six independent assays.

Host toxicity assays

For host toxicity assessment in C. elegans, the protocol included inhibitor groups (NETS, CAZ and NETS+CAZ) and blank negative control groups. The experiments were performed in triplicate. L4 stage C. elegans were picked into microcentrifuge tubes with a pick device, the tubes were washed three times with M9 medium, and the C. elegans were added to a 96-well plate. Then, 100 μL of fresh inhibitor-containing medium was added directly to each well. For the blank negative control group, 100 μL of fresh M9 medium was added. The number of dead worms, identified through the response to gentle touch, was counted every 8 h for 3 days. Statistical analysis was performed using GraphPad Prism 7.0 software. The concentrations of NETS used were 30 μM, 60 μM, 120 μM, 240 μM, and 480 μM. The concentrations of CAZ used were 3.75 μM and 15 μM. The concentrations of NETS+CAZ were 30 μM NETS+1.88 μM CAZ and 120 μM NETS+7.5 μM CAZ.

In vivo inhibitor efficacy assessment

The C. elegans-B. pseudomallei infection model was established as described by Gan et al.³⁵ B. pseudomallei infection was divided into early and late intervention groups (early intervention: C. elegans incubation with a drug 2 hours before infection with B. pseudomallei; late intervention: C. elegans incubation with an inhibitor 2 hours after infection with B. pseudomallei). One hundred microliters of freshly prepared B. pseudomallei HNBP001 culture (approximately 5 × 10⁷ CFU/mL) or drug was placed on a nematode growth medium (NGM) agar plate at 37 °C for 2 h, and 20 × L4 stage C. elegans (N2 strain) was washed thrice with M9 medium and subsequently transferred to a bacterial lawn. C. elegans groups treated with only B. pseudomallei, the drug, or M9 buffer were used as controls. The dead worms, identified through the response to gentle touch, were counted every 8 h for 56 h at 20 °C. Moreover, fluorescently labeled B. pseudomallei HNBP001, which was previously stained with DilC18 dye (D3911; Invitrogen, Carlsbad, CA, USA) for 24 h, was utilized to repeat the test. Individual worms were anesthetized using isoflurane, and microbial colonization was visualized via fluorescence microscopy. The concentrations of NETS used were 60 μM, 120 μM, 240 μM, and 480 μM. The concentrations of CAZ used were 3.75 μM and 15 μM. The concentrations of NETS+CAZ were 30 μM NETS+1.88 μM CAZ and 120 μM NETS+7.5 μM CAZ. The experiments were performed in triplicate.

Molecular docking

The interaction between Hcp and NETS was simulated through a molecular docking approach performed by AutoDock Vina. The structure of NETS was obtained from PubChem (http://pubchem.ncbi.nlm.nih.gov/; CAS: 56391-57-2). The crystal structure of the Hcp protein (PDB ID: 8YUN) was obtained from the Protein Data Bank (http://www.rcsb.org/). For the preparation of Hcp, we first used Discovery Studio Client 2019 to minimize the energy needed and then used Autodock Tool 1.5.6 to apply hydrogen atoms, net charges and other processing to the Hcp. The 3D structure of NETS was constructed via energy minimization via DFT calculations. Autodock Vina was utilized to execute semiflexible docking, and output poses were evaluated by the Grid scoring function, with the lowest score selected as the predicted protein–ligand complex. Interactions between the compound molecules and the amino acid residues in the binding cavity can be observed. Subsequently, we further used Discovery Studio Client 2019 software to map the interactions in different forms, including hydrogen bonds, hydrophobic interactions, and ionic interactions.

Molecular dynamics simulation

GROMACS 2019⁷¹ was used for molecular dynamics (MD) simulation. We used the amber99sb-ildn⁷² force field generated by using the PDB2GMX tool for the receptor. The partial charge of each atom of the NETS inhibitor was first calculated by the bcc charge fitting method via the amber tool. In the molecular dynamic simulation, we first embedded the protein–ligand complex into the TIP3P water box, setting its boundary not less than 10 Å. Then, we added an appropriate number of Na⁺/Cl^- ions to keep the system neutral. The energy of the constructed system was minimized to eliminate unrealistic contacts between atoms in the system and to prevent collapse of the dynamic simulation process. After minimizing the energy of the system, NVT and NPT prebalancing are carried out. After balancing, the system is set to simulate the NPT ensemble. The temperature was set to 300 K, the pressure was set to 1.01325 bar, and the SHAKE algorithm was used to limit the stretching vibration of all the hydrogen atom chemical bonds in the system. The time step was set to 2 fs, and the total simulation time was 100 ns. Thirty frames of dynamic trajectory files were extracted from the stable dynamic trajectory of 70∼100 ns, and the combined free energy of the stabilized system was calculated using the MM/PBSA⁷³ method. In the process of molecular dynamics simulation, the energy of the system was monitored, the RMSD of the Hcp structure and NETS structure fluctuations were measured, the persistence of nonbonding interactions between Hcp and NETS molecules was assessed, and the RMSF in the protein structure was evaluated to determine the changes in relevant properties of the system over time during the molecular dynamics process.

Quantification and statistical analysis

Data were analyzed using GraphPad Prism 7.0 and presented as mean ± SEM. Multiple group comparisons were performed using Kruskal–Wallis test with Dunn‘s post hoc test. p < 0.05 was considered significant. Significance levels: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant. See figure legends for detailed statistical information.

Published: March 13, 2026