Development of an inhibitory monoclonal nanobody targeting Streptococcus pyogenes siderophore binding protein FtsB

TL;DR

Researchers developed a nanobody that inhibits a key iron transporter in Streptococcus pyogenes, offering a new tool to study infection and potential antibacterial therapy.

Contribution

A novel inhibitory nanobody targeting the FtsB protein in S. pyogenes was developed and characterized.

Findings

Nb1 binds to FtsB with sub-nM affinity and a slow dissociation rate.

The nanobody competitively inhibits hydroxamate siderophore binding and uptake in S. pyogenes.

Nb1 provides a new tool to study the FtsABCD transporter's role in infection.

Abstract

Due to the limited availability of metals inside the human body, pathogenic bacteria must produce multiple highly specialized metal transporters to cause infection. These transporters constitute attractive targets for developing novel antibacterial strategies. Streptococcus pyogenes possesses three iron transporters, of which the FtsABCD system is specialized in the uptake of ferric hydroxamates. The role of this transporter in infection remains unclear. In this study, we developed a monoclonal alpaca VHH, or nanobody, Nb1, targeting FtsB. Nb1 binds to FtsB with sub-nM affinity, in an enthalpy-driven manner, and with a characteristically slow dissociation rate. Solvent accessibility analysis by hydrogen/deuterium exchange coupled with mass spectrometry, mutational analyses, and X-ray crystallography revealed that the epitope of Nb1 is in the binding pocket of FtsB. The nanobody competitively inhibited the binding of multiple hydroxamate siderophores and partially inhibited the uptake of siderophores in S. pyogenes cells. The inhibitory activity of Nb1 on siderophore transport represents a new tool to study the role of the FtsABCD transporter and can be used as a potential inhibitor of S. pyogenes growth under iron-limited conditions.

Figures

Full text

First-row transition metals are essential micronutrients for most living organisms and are required for many key biological processes (1, 2, 3). In contrast to their biological significance, many of these elements are not abundant in biological environments, either because of their natural scarcity in the Earth’s crust or their poor solubility under physiological conditions (2). The availability of metals is further restricted in the human body due to natural homeostasis mechanisms and the production of proteins that sequester metals to prevent bacterial growth, such as lactoferrin, an iron-specific protein found in many secretions (4), and calprotectin, a protein secreted by neutrophiles during infections that chelates diverse transition metals, including manganese, zinc, iron, and nickel (5, 6, 7). This process is known as nutritional immunity (4, 8, 9, 10). These mechanisms make first-row transition metals in the human body difficult for pathogens to access. Therefore, efficient metal uptake is one of the main challenges that pathogenic bacteria must overcome to cause infection (11, 12).

Streptococcus pyogenes (S. pyogenes) is a strictly human-adapted Gram-positive pathogenic bacterium, often referred to as group A Streptococcus (GAS). GAS is one of the leading causes of bacterial infections in humans, responsible for a wide range of conditions, from mild pharyngitis and skin infections to life-threatening diseases like necrotizing fasciitis and septicemia (13). The rise of antibiotic-resistant strains (13, 14, 15) and the lack of a vaccine (13, 16) demand the development of alternative treatment strategies against this bacterium. In this sense, the genome of S. pyogenes encodes three iron uptake transporters (17, 18), which constitute appealing targets for developing new antimicrobials. These are the Shr/Shp/SiaABC, the MtsABC, and the FtsABCD systems, which transport heme from hemoglobin (19), ferrous iron and manganese (20), and ferric iron in complex with hydroxamate siderophores (21), respectively.

The FtsABCD system is an ABC transporter composed of four genes organized in an operon in the genome of S. pyogenes. These include (i) FtsA, which encodes for the ATPase that provides energy for the transport; (ii) FtsB, which serves as the substrate-binding protein (SBP) that binds to siderophores; (iii) and two different genes for transmembrane permeases, FtsC and D. We recently reported that FtsB binds promiscuously to hydroxamate siderophores (22), potentially allowing this bacterium to exploit xenosiderophores as iron sources (23, 24, 25). However, the role of FtsABCD in S. pyogenes is poorly understood. While almost every pathogenic bacterium possesses transporters specific for hydroxamate siderophores, only a few can produce them. These include bacteria from the genus Streptococcus (26). Despite this, KO strains of the siderophore transporters of several Streptococcus species show reduced virulence in animal models (27, 28). Therefore, inhibitors against the FtsABCD system would help to understand its biological relevance and could be explored for the treatment of infections caused by S. pyogenes.

The development of new antibiotics has often focused on small compounds of natural or synthetic origin. However, in recent years, nanobodies have gained traction in the development of therapeutics against bacteria and other infectious diseases. Beyond their applications in diagnostics, several studies have demonstrated the potential of nanobodies in the neutralization of bacterial toxins, responsible for the symptoms and lethality associated with some bacterial pathogens (29, 30, 31, 32). Additionally, bacterial surface antigens such as flagellin (33), or proteins involved in adhesion and internalization to host cells (34, 35, 36), have also shown significant potential as targets, with many of the nanobodies reported in these studies reducing the virulence of the bacteria by limiting its motility, dissemination, and ability to colonize host cells and tissues. In this context, bacterial nutrient transporters are an emerging class of therapeutic target for the development of nanobodies and other antibody fragments (37, 38, 39, 40). However, studies focusing on the SBPs of ABC transporters have suggested that the bacterial cell surface may constitute a challenge for developing effective antibody inhibitors of bacterial nutrient uptake (38, 39, 40).

Herein, we report the development of a monoclonal alpaca variable heavy domain of heavy chain (VHH) against the hydroxamate siderophore SBP FtsB from S. pyogenes. The antibody (Nb1) bound to FtsB with high specificity and sub-nM affinity. Through a combination of mutational analyses, solvent accessibility data collected by hydrogen/deuterium exchange coupled with mass spectrometry (HDX-MS), and structural analysis by X-ray crystallography, we identified that the epitope of Nb1 is the binding pocket of FtsB. Isothermal titration calorimetry (ITC) experiments in the absence or presence of Nb1 showed that the antibody inhibits the binding of several hydroxamate siderophores to FtsB. In addition, ^55^Fe uptake experiments showed a partial inhibition of siderophore uptake using Nb1. To our knowledge, this constitutes the first report and characterization of an antibody with inhibitory activity targeting S. pyogenes nutrient transporters. Nb1 represents a new tool to study and inhibit siderophore uptake through the FtsABCD system in S. pyogenes and could potentially be used to suppress bacterial growth under iron-limited conditions.

Results

Selection of an anti-FtsB nanobody

Recombinant FtsB was expressed and purified as we recently reported (22). The protein was used to immunize an alpaca (Hokudō), and sera of 5- and 7-week postimmunization were used to prepare the immune library as previously described (37). Two rounds of selection against FtsB were conducted (Fig. 1A), and 47 single clones were isolated. The phage containing supernatants of the isolated clones were used to evaluate the binding of the selected nanobodies to FtsB by ELISA. Clones that showed higher absorbance against FtsB than against a bovine serum albumin (BSA) reference were selected for sequencing (Fig. 1B). All the sequenced single clones belonged to a common sequence cluster, with only minor differences found in the framework regions (Fig. 1C). Although there were relatively small differences between FtsB and BSA in ELISA, sequence convergence demonstrated the successful selection of a single nanobody against FtsB, which we termed Nb1.Figure 1Selection of monoclonal nanobodies against FtsB. A, schematic representation of the strategy for antibody selection. Selection was conducted against FtsB adsorbed to the surface of an immunotube, and specific phages were eluted by acidic elution. B, ELISA screening after the second round of selection. The 47 selected single clones were infected with helper phage VCSM13, and ELISA was performed with the supernatant of the culture against immobilized FtsB (white) or BSA (red), using an anti-VHH-HRP secondary antibody. “C” indicates the positive control, performed with phages displaying an anti-lysozyme VHH against immobilized hen egg lysozyme (white) or BSA (red). C, sequence logos, generated with WebLogo3 (78), of the amino acid sequence cluster that includes clones 2, 5, 6, 7, 14, 15, 16, 17, 20, 21, 25, 26, 27, 30, 31, 34, 36, 40, and 43. The CDR1, CDR2, and CDR3 regions (AbM definition) appear highlighted in yellow, blue, and green, respectively. CDR, complementarity-determining region; VHH, variable heavy domain of heavy chain.

Binding analysis between FtsB and Nb1

Nb1 was recombinantly produced in Escherichia coli BL21 strain as previously described (37). The binding between FtsB and Nb1 was evaluated by surface plasmon resonance (SPR) (Fig. 2A, Fig. S1A), Western blotting (Fig. 2B), and ITC (Fig. S1, B and C). To assess the Nb1 specificity toward FtsB, we utilized FhuD2, a close homolog of FtsB from Staphylococcus aureus (22, 41), as a control (Fig. S1). No binding of Nb1 to FhuD2 was observed, indicating high specificity of Nb1 toward FtsB.Figure 2Binding analysis between Nb1 and FtsB. A, sensorgrams of the analysis by SPR of the binding of Nb1 at 0.125, 0.25, 0.5, 1, or 2 nM against FtsB immobilized by amine coupling in the surface of a CM5 chip. Sensorgrams were measured by multicycle kinetics, in PBS-T (0.005%) with 300 s of association and 1200 s of dissociation, followed by regeneration with 1 M L-Gly pH 3.0 for 30 s. The thick colored lines represent the raw sensorgrams and the thin black lines are the fittings. The analyses were performed with the Biacore T200 Evaluation Software (Cytiva) using a 1:1 binding kinetic fitting model. B, Western blotting analysis of the binding of Nb1 to either recombinant FtsB (left) or recombinant FhuD2 (right). Transferred membranes were first treated with a 0.1 mg/ml solution of Nb1 in PBS-T (0.05%) for 1 h, followed by 45 min of incubation with a commercial goat anti-Alpaca VHH antibody conjugated with HRP (1:5000 dilution in PBS-T (0.05%)). For CBB and Western blotting, experiments with FtsB and FhuD2 were run in parallel in the same gel using a single marker lane. Therefore, the same molecular weight marker is shown for both proteins. SPR, surface plasmon resonance.

SPR analysis revealed a high affinity of 21 ± 4 pM, with an association constant of (4.9 ± 0.8)·10^6^ M^-1^ s^-1^ and a slow dissociation constant of (1.0 ± 0.1)·10^-4^ s^-1^ (Fig. 2A and Table 1). Binding of Nb1 to FtsB could be detected at low concentrations (125 pM to 2 nM). In contrast, with an injection of 16 μM Nb1 on a chip with FhuD2 immobilized, no binding was detected (Fig. S1A). A summary of the parameters of the Nb1–FtsB interaction obtained in SPR experiments is shown in Table 1. Similarly, when the recombinant protein was loaded and immunoblotted against Nb1, a specific band was observed when using FtsB, but not FhuD2 (Fig. 2B). ITC measurements revealed a strong enthalpy-driven binding with 1:1 stoichiometry, characterized by an enthalpy change of −25.0 ± 0.6 kcal/mol, and an apparent KD of 1.4 ± 0.5 nM (Fig. S1B). In ITC experiments performed with FhuD2 instead of FtsB, the enthalpy change was not observed (Fig. S1C).Table 1. Kinetic and thermodynamic parameters of the Nb1–FtsB interactionSPRaLigandR_max_ (RU)kon (M^-1^s^-1^)koff (s^-1^)KD (pM)FtsB WT81.2 ± 3.5(4.9 ± 0.8) × 10^6^(1.0 ± 0.1) × 10^-4^21 ± 4FtsBY81A75.3 ± 4.6(2.9 ± 1.4) × 10^6^(8.6 ± 2.5) × 10^-3^3100 ± 630FtsBY137A97.7 ± 0.3(3.9 ± 0.3) × 10^6^(3.2 ± 0.2) × 10^-2^8300 ± 120FtsBW204A109.5 ± 2.3(1.7 ± 0.1) × 10^6^(7.9 ± 0.3) × 10^-2^46,300 ± 2800FtsBY232A72.5 ± 1.0(4.0 ± 0.5) × 10^6^(1.8 ± 0.2) × 10^-2^4600 ± 170SPR, surface plasmon resonance.aValues in the table correspond to the experiments performed in Figure S4 and represent the averages and SDs of the parameters obtained in two multicycle kinetic assays with five concentrations per assay.

While ITC is a powerful method for characterizing protein–protein and protein–ligand interactions, it is less reliable than SPR for measuring very tight affinities (42). For interactions with affinities under nM range, the binding isotherm presents with a near-vertical transition, too steep to obtain a physically accurate measurement of the KD, and very sensitive to the experimental design (particularly, to the spacing between injections). The c value of a reaction is a parameter often used to evaluate the validity of an ITC experiment, and its typical acceptable range is 1 to 1000 (43). Using the KD obtained by SPR, the c value estimated for these ITC assays is 420,000. Therefore, we consider the KD obtained by SPR to more accurately reflect the affinity between FtsB and Nb1. Despite the limitations of ITC in measuring very tight KD values, the enthalpy and stoichiometry can be measured accurately even for reactions with an unsuitable c value. The results obtained in this study show stoichiometric and strongly exothermic binding between FtsB and Nb1, and the experiments with FhuD2 show the specificity of this interaction.

Epitope mapping of the FtsB–Nb1 interaction

To determine the binding site of Nb1 in FtsB, we conducted HDX-MS with the protein in complex with Nb1 or the unbound protein. FtsB in complex with the hydroxamate siderophore ferrichrome (FCH) was also analyzed to evaluate the changes in solvent accessibility caused by the natural ligands of the protein. To evaluate the results, we utilized a scoring algorithm, as described in Experimental procedures. A diagram of the analysis of the data collected by HDX-MS is shown in Figure S2.

We observed that the binding of both FCH (Fig. 3A) and Nb1 (Fig. 3B) caused a decrease in the deuterium uptake in peptides belonging to the globular lobes of FtsB, but not in the helix that connects them. The patterns in the change of deuterium uptake caused by both molecules were very similar. In particular, peptides falling within two regions in the N-lobe of FtsB showed a major decrease in deuterium exchange in the FCH-bound or Nb1-bound protein, compared to unbound FtsB. We termed these “region 1” and “region 2.” Moreover, we identified minor but noticeable changes in two additional regions (“region 3” and “region 4”) in the C-lobe of FtsB. Deuterium uptake kinetics of individual peptides representative of these four regions are shown in Figure 3C.Figure 3Epitope mapping of the Nb1–FtsB interaction. A and B, residual plots of the deuterium uptake scores of Fe(III)-FCH bound to FtsB (A) or Nb1-bound FtsB (B) relative to the unbound protein, obtained in HDX-MS experiments. The peptide coverage, and the domains of FtsB that each peptide and amino acid belong to, are indicated above the plot. Four regions of interest, showing changes in deuterium uptake between the unbound protein and the FCH-bound or Nb1-bound protein (regions 1–4) appear highlighted in red, blue, purple, and orange, respectively. C, deuterium uptake kinetics of individual peptides representative of region 1 (left), region 2 (middle left), region 3 (middle right), or region 4 (right). The curves represent unbound FtsB (blue), Nb1-bound FtsB (red), and FCH-bound FtsB (green). The sequence of each peptide is included above the graph, highlighting the amino acids known to participate in siderophore binding. D, sensorgrams of the analysis by SPR of the binding of Nb1 at 0.125, 0.25, 0.5, 1, or 2 nM against FtsB Y81A (left), Y137A (middle left), W204A (middle right), and Y232A (right) immobilized by amine coupling in the surface of a CM5 chip. Kinetics were measured with 300 s of association and 1200 s of dissociation. The thick colored lines represent the raw sensorgrams and the thin black lines the fittings, performed with the Biacore T200 Evaluation Software (Cytiva) using a 1:1 binding kinetic fitting model. FCH, ferrichrome; HDX-MS, hydrogen/deuterium exchange coupled with mass spectrometry; SPR, surface plasmon resonance.

All of the four regions selected for analysis contained amino acids that we previously found to be relevant for the interaction of FtsB with hydroxamate siderophores (22, 44), particularly the aromatic amino acids Tyr81 (“region 1”), Tyr137 (“region 2”), and Tyr232 (“region 3”), which are involved in the ability of FtsB to accommodate the variable backbones of different siderophores, and the HB-loop containing Phe200, Trp204, and Arg206 (“region 3”) responsible for the specificity of the protein (22, 44).

Based on these findings, we determined that the epitope of Nb1 is the siderophore binding pocket of FtsB.

To validate the results obtained by HDX-MS, we examined the interaction of Nb1 with the single mutants Y81A, Y137A, W204A, and Y232A of FtsB. Since the affinity of Nb1 is tighter than what can be confidently measured by ITC, we decided to use SPR for this mutational analysis, since we hypothesized that mutations with small, but still relevant contributions to the binding may not be reflected in ITC assays. Experiments were carried out at the same concentrations used to analyze binding to WT FtsB (Fig. 3D). Interestingly, the binding of Nb1 to the W204A mutant was almost not detectable. Nb1 bound to the other three mutants. However, the maximum response at the highest concentration was lower than that of the WT protein (Fig. 3D versus Fig. 2B). Additionally, of the three mutants for which binding was observed, the dissociation rate was faster than the WT protein (Fig. 3D versus Fig. 2B), reaching an almost complete dissociation within 5 min after the end of the injection.

To obtain more accurate parameters of the interaction of FtsB with these mutants, binding analyses were performed at higher nanobody concentrations (Fig. S4). The kinetic parameters of the interactions with the FtsB mutants are summarized in Table 1. All four mutants exhibited a KD at least 140-fold higher than the WT protein. Notably, the W204A mutant displayed a 2200-fold increase in its dissociation constant. While the kon values for all mutants were slightly slower than the WT protein, these differences were minimal. In contrast, all mutants showed a substantial increase in their koff values.

Crystal structure of the FtsB–Nb1 complex

To better understand the interaction mechanism between FtsB and Nb1, we resolved the crystal structure of the antigen–nanobody complex. Diffracting crystals were obtained using a solution of 6% tacsimate pH 6.0, 0.1 M BIS-Tris pH 6.5, 20% PEG 3350. The crystal structure of the binary complex was solved at a maximum resolution of 2.55 Å, and it showed that Nb1 binds to the cavity in FtsB that constitutes the siderophore binding pocket (Fig. 4A). The amino acids in the regions 1, 2, and 3 selected from HDX-MS analysis formed a network of interactions with the complementarity-determining region 3 (CDR3). Tyr81 in FtsB appeared to form a CH-π interaction with one of the carbon atoms in the side chain of Pro101 in the CDR3 of Nb1 (Fig. 4B). The side chain of Pro102 in CDR3 was found nestled between Tyr137 (region 2) and Trp204 (region 3). Additionally, Arg206, in region 3, formed a hydrogen bond with the carbonyl of Pro101. In region 4, Tyr232 was observed to contact the CDR1 of Nb1, specifically through a CH-π interaction with the alpha carbon of Gly30 (Fig. 4C).Figure 4Crystal structure of the Nb1–FtsB complex. A, overall structure of FtsB (white) in complex with Nb1 (gray). The binding pocket of FtsB is shown in magenta, and the CDR1, 2 and 3 regions of Nb1 in yellow, blue, and green, respectively. Boxes represent the areas shown in panels B and C. B and C, close-up view of the Nb1-FtsB contacts in the regions 1 (red) (B), 2 (blue), 3 (purple), and 4 (orange) (C) of FtsB identified by HDX-MS. Hydrogen bond between R206 (FtsB) and the carbonyl of P101 (Nb1) is shown as a black dashed line. D, percentage of the surface of interaction contributed by each CDR, relative to the total interface area. Bars represent the average and standard deviations for these values observed in the chain A–chain C interface and the chain B–chain D interface of the crystal structure. Interface analysis was conducted using PDBePISA. E, comparison of the crystal structure of the Nb1–FtsB complex (top) with the AlphaFold3 model of the FtsB–FtsA2CD complex (bottom). The bar graph shows the average and SDs of the predicted template modeling score (pTM) and interface predicted template modeling score (ipTM) of the five solutions of four independent seeds (n = 20). CDR, complementarity-determining region; HDX-MS, hydrogen/deuterium exchange coupled with mass spectrometry.

The CDR3 of antibodies, particularly of alpaca nanobodies, often show the highest variability among the three CDRs and are often regarded as crucial for the interaction with the antigen. In accordance with this, interaction interface analysis using PDBePISA (45) showed that the CDR3 of Nb1, despite its relatively short length (10 amino acids) (46, 47), contributed over half of the total interface area between the antibody and its antigen, whereas CDR1 and CDR2 contributed only about 20% (Fig. 4D).

Since Nb1 binds to the siderophore binding pocket of FtsB, we hypothesized that it may inhibit the siderophore binding to the protein. Additionally, it has been reported that SBPs need to bind to their cognate permeases to transfer the substrate and stimulate ATP hydrolysis (48). The inhibition of this interaction disrupts the function of the transporter (40). To evaluate this hypothesis, we compared the structure of the FtsB–Nb1 complex with previously reported structures of ABC transporters (49, 50), and with the AlphaFold 3 prediction (51) of the FtsB–FtsA2CD complex (Fig. 4E). The comparison showed that Nb1 completely overlaps with the transmembrane permease, and therefore Nb1 is likely to inhibit both the binding of siderophores to FtsB and the binding of FtsB to its cognate permease complex.

Effect of Nb1 on the siderophore binding

To test the potential of Nb1 to inhibit siderophore binding to recombinant FtsB, we conducted competitive ITC binding assays in the form of titrations of the unbound protein or the protein in complex with a 1.2-fold molar excess of Nb1, with different hydroxamate siderophores in complex with Fe(III). FtsB bound to the hydroxamate siderophores ferrioxamine E, ferrioxamine B (FOB), FCH, and bisucaberin (BIS) alone (Fig. 5, A–D, left panels) with a characteristically enthalpy-driven affinity in the order of 5 to 100 nM (Fig. S5). However, no binding was observed when titrations were performed in the presence of 1.2-fold molar excess VHH (Fig. 5, A–D, right panels). These results indicated that Nb1 competitively inhibited the binding of siderophores to FtsB.Figure 5Inhibition of siderophore binding by Nb1. ITC titrations of FtsB with the ferric complexes of the hydroxamate siderophores FOE (A), FOB (B), FCH (C), or BIS (D), in the absence (left) or presence (right) of 1.2-fold molar excess of Nb1. Titrations were performed in 50 mM acetate pH 5.5 (A–C) or 20 mM Hepes pH 7.5, 150 mM NaCl (D), using FtsB concentrations of 10 (A), 20 (B) or 50 (C and D) μM in the cell, and a concentration 10 times higher of siderophore in the syringe. The top panels correspond to the titration kinetics, and the bottom panels represent the integrated binding isotherms. Titrations in the absence of Nb1 correspond to our previous data, published in structure (22). BIS, bisucaberin; FCH, ferrichrome; FOB, ferrioxamine B; FXM, foroxymithine; ITC, isothermal titration calorimetry.

Inhibition of iron uptake in S. pyogenes by Nb1

To further assess the ability of Nb1 to inhibit siderophore uptake, radioisotope uptake assays were performed using a ^55^Fe radiotracer. First, we evaluated the uptake of different iron sources by GAS through its three iron uptake pathways (Fi. 6A), using S. pyogenes JRS4 strain, a well-studied serotype M6 GAS isolate often used as a standard laboratory model (52). A KO strain of FtsB was used as a control. All assays were conducted with a final concentration of 0.75 μM ^55^Fe, and 6-fold excess of either heme or siderophores, when applicable.Figure 6Effect of Nb1 on siderophore uptake by S. pyogenes JRS4. A, schematic representation of the three iron uptake pathways of S. pyogenes: the Heme transporters Shr/Shp/SiaABC (red), the free Fe(II) transporter MtsABC (blue), and the hydroxamate siderophore transporter FtsABCD (yellow). B, uptake of different iron sources by GAS JRS4 (white bars) or GAS JRS4 ΔftsB (gray bars). Bars in the graph represent the average c.p.m. (counts per minute) measured from three aliquots of 0.6 ml taken from a single 2 ml culture tube, and error bars represent the SD. The c.p.m. values from each independent technical replicate are shown overlayed on top of the bars. C, uptake of 0.75 μM ^55^Fe complexed with 5-fold molar excess FXM, FCH, or FOB by WT JRS4 (white bars), or WT JRS4 in the presence of 7.5 μM Nb1 (gray bars). Bars represent the mean c.p.m. from three independent biological replicate assays, each calculated from the average of three technical replicates. Error bars indicate SD. Statistical significance was determined using a paired two-tailed t test (∗ = p < 0.05; ∗∗ = p < 0.01). Average c.p.m. values from each independent biological replicate are shown overlayed on top of the bars. D, diagram of the hypothetical explanations for the lack of complete inhibition observed in uptake assays. If the bacterial surface was permeable to Nb1 (i), complete inhibition could be achieved. However, when the diffusion of Nb1 through the cell wall (ii) or capsule (iii) is limited, or if Nb1 could be recognized by some bacterial antibody binding protein (iv), inhibition efficacy would decrease. Additionally, it is possible that Nb1 is susceptible to proteolysis by secreted (v) or surface-anchored proteases. E, uptake of 0.75 μM ^55^Fe complexed with 5-fold molar excess FOB by WT JRS4 (left), or a ΔHasA KO strain (right) in the absence (white bars) or presence of 7.5 μM Nb1 (gray bars). Bars represent the average c.p.m measured from three aliquots of 0.6 ml taken from a single 2 ml culture tube, across three independent replicate assays, and error bars represent the SD. Statistical significance was determined using a paired two-tailed t test (∗ = p < 0.05; ∗∗ = p < 0.01). Average c.p.m. values from each independent biological replicate are shown overlayed on top of the bars. FCH, ferrichrome; FOB, ferrioxamine B; FXM, foroxymithine; GAS, group A Streptococcus.

The results showed that S. pyogenes JRS4 transports free iron (MtsABC system), Heme-bound ^55^Fe (Shr/Shp/SiaABC system), and FCH-bound ^55^Fe (FtsABCD system) (Fig. 6, A and B). Interestingly, we observed that GAS JRS4 could utilize not only FCH but also FOB and the hydroxamate siderophore foroxymithine (FXM) as iron sources (Fig. 6B). The binding between FXM and FtsB is shown in Figure S5A. The KO strain of FtsB utilized Heme or free ^55^Fe but not the three hydroxamate siderophores (Fig. 6B). Importantly, we found a correlation between the amount of uptake of each hydroxamate siderophore and the strength of its binding to FtsB (Fig. S5B).

We then performed the ^55^Fe uptake assays of FXM, FCH, and FOB in the presence or absence of Nb1 at 7.5 μM (100 μg/ml, with an inhibitor: ligand ratio of 10:1). The bacteria grown in the presence of Nb1 showed a decrease in the uptake of hydroxamate siderophores of ∼40 to 50% (Fig. 6C). This incomplete inhibition contrasted with the total competition observed in ITC (Fig. 5). We hypothesized that there could be three possible explanations for this observation (Fig. 6D): it is possible that diffusion through either the cell wall (Fig. 6D, ii) or capsule (Fig. 6D, iii) of S. pyogenes is restricted, limiting the access of Nb1 to its antigen. Another possible explanation could be the sequestration of Nb1 by streptococcal antibody binding proteins (Fig. 6D, iv). Finally, we hypothesized that Nb1 could be susceptible to streptococcal proteases (Fig. 6D and v).

We believed that the physical barrier imposed by the capsule of S. pyogenes would be the most likely explanation, as previous studies have suggested it as a major limitation for antibodies (39, 53). To test this hypothesis, we generated a capsule-deficient strain through the deletion of the hyaluronic synthase gene HasA. Uptake inhibition assays with Nb1 and FOB were conducted, comparing this JRS4 ΔHasA mutant with its parent strain (Fig. 6E). We observed that the inhibition caused by Nb1 was substantially increased in this capsule-deficient strain, from 47% to 82%. It is important to note that the radiation detected in the Nb1-treated ΔHasA strain (270 ± 80 c.p.m., Fig. 6E) was found to be above the background radiation observed in the ΔFtsB strain (50 ± 15 c.p.m., Fig. 6B).

Discussion

In this study, we focused on the ferric hydroxamate transporter FtsB from S. pyogenes, one of the components of the FtsABCD system, responsible for binding to siderophores. We successfully isolated a nanobody, Nb1, which binds to the siderophore binding pocket of FtsB with sub-nM affinity. The specific targeting of this functional pocket allowed Nb1 to directly interfere with the protein's iron-scavenging activity. In experiments with live bacteria, Nb1 was able to cause a significant, though incomplete, reduction in Fe uptake. Inhibition was greatly enhanced in a capsule-deficient strain, revealing the capsule to be a physical barrier for nanobody-based therapies against this bacterium.

Over the last 2 decades, alpaca nanobodies, characterized by high affinity and low molecular weight, have demonstrated significant potential in the treatment of infectious diseases. Several studies have reported the development of nanobodies against viral glycoproteins (54, 55, 56) and bacterial surface adhesins (33, 34, 35, 36) that limit the pathogen’s ability to invade host cells and colonize tissues. At least one nanobody-based antibacterial therapeutic against Campylobacter jejuni has been tested in phase I clinical trials (33), and is currently undergoing phase II clinical trials (NCT04182490), proving their establishment as a new class of antibacterial therapeutics.

Bacterial nutrient transporters, and particularly metal transporters, constitute attractive targets for developing novel inhibitors. Our group has recently shown that nanobodies against the cell wall-anchored heme transporters IsdH and IsdB from S. aureus could inhibit the transport of heme and subsequently the growth of the bacterium in iron-restricted medium supplemented with hemoglobin (37). Similarly, previous reports have suggested the potential of nanobodies and other antibody fragments as inhibitors of bacterial ABC transporters (38, 40). To our knowledge, Nb1 is the first nanobody to achieve inhibition of nutrient transport targeting an ABC transporter in bacteria. Importantly, the biological effect observed was only partial, and the concentrations required to achieve it exceeded those reported by studies targeting other classes of antigens (34, 37). It has been reported that the capsule of S. pyogenes plays a major role in its resistance to antibodies (39, 53). In this sense, we observed that the removal of the capsule by knocking out HasA enhanced the inhibition caused by Nb1. These findings highlight the need to engineer mechanisms to improve the penetration of nanobodies through the capsule of S. pyogenes in order to apply them as therapies against this bacterium. One example of how this could be achieved would be the combination of nanobody-based therapies with capsule-degrading enzymes, such as the hyaluronate lyases expressed by many S. pyogenes-infecting bacteriophages (57).

The biological role of the FtsABCD system itself presents an intriguing paradox. While hydroxamate siderophore transporters are ubiquitous among pathogenic bacteria (22, 23, 58, 59), only a few species produce siderophores of this family. It is hypothesized that bacteria use these transporters to exploit xenosiderophores from other organisms during coinfections (60, 61) or in cases where siderophores are used as drugs to treat other diseases (23, 24, 25). It is also plausible that they are a remnant of the evolutionary history of these bacteria. However, the knock out of the hydroxamate transporters of several Streptococcus species decreases their virulence in animal models (27, 28). Similarly, a recent report identified that nonpathogenic strains of S. pyogenes isolated in Japan were characterized by a mutation in FtsC, which reduced the ability of the bacterium to grow in human blood, but not in culture media (62).

The universal presence of these transporters amongst pathogenic bacteria, together with evidence of their impact on virulence, demonstrates that hydroxamate siderophore transporters play a critical but poorly understood role in bacterial pathogenesis. In this context, Nb1 constitutes a valuable tool to study the role of hydroxamate siderophore transporters, using S. pyogenes as a model organism. For example, beyond its canonical role in hydroxamate siderophore transport, the FtsABCD system has been suggested to be involved in the uptake of acyl-homoseryl lactones (AHLs) (63), quorum sensing molecules produced by Gram-negative organisms (64). While Gram-positive bacteria do not produce AHLs, S. pyogenes has been shown to respond phenotypically to them, modulating its pathogenicity and virulence in response to these compounds (63, 65, 66). We believe the noncanonical interaction between the FtsABCD system and AHLs to be one of the examples where Nb1 could constitute a potent tool for the study of this transporter.

Moreover, the inhibitory activity of Nb1 could be exploited therapeutically to limit the growth of S. pyogenes under the iron-restricted conditions found during an infection, as the FtsABCD system is important for the ability of S. pyogenes to cause infections (27), and grow in human blood (62). However, given the redundancy of iron uptake pathways, a successful anti-infective strategy would likely require combining Nb1 with inhibitors against the MtsABC (67) and Shr/Shp/SiaABC (19) systems. Assessing the potential of Nb1 as a therapeutic tool requires a detailed investigation of the relative relevance of the three transporters in iron uptake, as well as the importance of the recently proposed noncanonical roles of FtsABCD in S. pyogenes virulence. Nevertheless, our work serves as a proof of principle for nanobodies as inhibitors of ABC transporters in bacteria, and positions Nb1 not only as a new tool for analyzing the role of FtsABCD in S. pyogenes pathogenesis but also as a promising model for the optimization and improvement of this novel therapeutic strategy.

Experimental procedures

Antibody library construction

An alpaca was immunized with recombinant FtsB. Immunization was conducted for 7 weeks with a unique dose of a mix of different target antigens, including FtsB, per week (Hokudō). After confirmation of antigenic reaction by ELISA, peripheric blood B lymphocytes were collected, and total mRNA was extracted using TriZol (Thermo Fisher Scientific) according to the manufacturer’s specifications. Total mRNA was used as a template for cDNA synthesis by retro transcription using superscript III (Thermo Fisher Scientific), followed by amplification of the cDNA of VHHs by PCR, using the KOD1 PCR Master Mix (Toyobo). Primers used for library construction are shown in Table S1. The cDNAs were cloned in a phagemid pLuck vector using the 2x HiFi DNA Assembly Master Mix (NEB). Plasmids were purified with AMPure magnetic beads and transformed into electrocompetent TG-1 E. coli cells by electroporation using 100 ng of DNA, 1800 V pulse, in a 1 mm gap electroporation cuvette. Transformed bacteria were grown in TYE-Agar plates (20 g/L tryptone, 6 g/L yeast extract, 10 g/L NaCl, 10 g/L D-(+)-glucose, and 15 g/L Agarose) containing 100 μg/ml of ampicillin, at 37 °C for 20 h, and harvested with a cell spreader. The final nanobody library was prepared in liquid 2x

TY (16 g/L tryptone, 10 g/L yeast extract, 5 g/L NaCl), containing 10% glycerol and stored at −80 °C.

Antibody selection by phage display

To select specific anti-FtsB nanobodies two rounds of selection were conducted against the recombinant protein. A schematic representation of the selection process is depicted in Figure 1A.

The antibody selection was conducted following a previous study (37) with some modifications. For each round of selection, E. coli from the previous round library were grown and infected with VCSM13 helper phage. The infected E. coli were cultured overnight and nanobody-displaying phages were purified from supernatant by PEG precipitation as previously described (68). Panning was conducted in a 4 ml plastic immunotube coated with 10 μg/ml FtsB and blocked with PBS containing 3% skim milk. Phages were added to a negative selection tube, which was coated with only blocking buffer, incubated for 1 h at room temperature, subsequently transferred to the FtsB tube, and incubated for 1 h at room temperature. The tube was washed 10 times with PBS containing 0.05% Tween-20. Following washing, FtsB-bound phages were eluted by incubation with 0.1 M glycine buffer pH 2.2 and collected into a tube containing 0.1 ml of 1.0 M Tris pH 9.1 to neutralize. Eluted phages were used to infect XL1-Blue E. coli and infected bacteria were grown in TYE plates supplemented with 100 μg/ml of ampicillin, at 37 °C for 20 h, and harvested to produce the next round of phagemid library.

Library screening to identify specific anti-FtsB nanobodies

Library screening by ELISA was performed after the second round of selection. Bacteria from the second-round selection library were diluted and grown in TYE plates supplemented with 100 μg/ml of ampicillin at 37 °C for 20 h to obtain single colonies. Forty-seven individual clones from the plates were inoculated in 1 ml 2xTY tubes and grown overnight at 37 °C. The next day, 50 μl of the culture were diluted with 950 μl of fresh media, storing the remaining 950 μl of culture at 4 °C. The new cultures were grown at 37 °C until absorbance at 600 nm reached 0.4. Bacteria were infected with VCSM13 helper phage and allowed to grow at 30 °C overnight to produce VHH-displaying phages. Culture supernatants were applied into the wells of a flat-bottom 96-well plate, which had been coated with PBS containing either 10 μg/ml of FtsB or BSA. After incubation with the phage culture supernatants for 1 h, the plate was washed, and a secondary anti-VHH HRP antibody was added. The plate was washed once more, and 50 μl/well of TMB solution were added. After color developed, the reaction was stopped with 50 μl/well of TMB stop solution, and the result was evaluated by measuring the absorbance at 450 nm on a PHERAstar plate reader. Clones that showed specific binding against FtsB in ELISA were selected, and DNA from the remaining 950 μl of the original culture was purified by miniprep and sent for sequencing.

Recombinant proteins expression and purification

Expression and purification of recombinant FtsB were conducted as previously described (22). For the expression of FhuD2, cDNA spanning amino acids 19 to 302 was amplified from the S. aureus Mu50 genome and cloned in the pCold-SUMO vector. Expression and purification of FhuD2 were conducted in the same manner as FtsB.

For the expression of the anti-FtsB nanobody Nb1, cDNA was amplified by PCR from a pLuck plasmid purified from a single clone from the second-round library, and cloned into a pRA2 expression vector, with an N-terminal PelB signal peptide and a C-terminal 6xHis tag. Primers used for the cloning of Nb1 are shown in Table S1.

Recombinant Nb1 was expressed in E. coli BL21 (DE3) grown in LB broth supplemented with 100 μg/ml of ampicillin, at 37 °C. When absorbance at 600 nm reached 0.8, protein expression was induced by IPTG to a 0.5 mM final concentration, and the cells were further incubated at 25 °C for 16 h. Bacteria were harvested by centrifugation at 4 °C and 8000×g for 10 min and resuspended in binding buffer (20 mM Tris pH 8.0, 500 mM NaCl) with 5 mM imidazole. Cells were disrupted by sonication on ice and centrifuged at 40,000×g, 4 °C, for 30 min. The soluble fraction of the cell lysate was cleared by sequentially filtering with 0.8 μm and 0.2 μm syringe filters. Cleared lysate was loaded onto an immobilized metal affinity chromatograhy open column with 2 ml of bed volume, which had been pre-equilibrated with binding buffer. The column was washed with 10 column volumes of binding buffer and eluted with 15 ml of the same buffer containing imidazole at a concentration of 200 mM. Nb1 was further purified by size-exclusion chromatography using an AKTA Purifier system (GE Healthcare) with a HiLoad 16/600 Superdex 75 column equilibrated with PBS containing 200 mM L-Arg. Nb1-containing fractions were pooled, divided in aliquots of 1 ml at 1 mg/ml, flash-frozen with liquid nitrogen, and stored at −20 °C. Protein concentration was evaluated based on its absorbance at 280 nm on a NanodropOne (Thermo Fisher Scientific), using a molar extinction coefficient of 23,045 M^-1^ cm^-1^, and a molecular weight of 15.72 kDa.

Binding analysis by SPR

The kinetic parameters of the binding of FtsB (WT and mutants) and FhuD2 to Nb1 were measured by SPR performed on a Biacore T200 (GE Healthcare) instrument. FtsB (WT or mutants) and FhuD2 were immobilized in the channels of a CM5 Biacore sensor chip at 250 response units in 10 mM acetate pH 5.5 using the amine-coupling reaction, according to the manufacturer’s instructions. After immobilization, sensorgrams were collected by multicycle kinetics, by injection of Nb1 into the sensor chip at a flow rate of 30 μl/min. Association and dissociation were measured for 300 and 1200 s, respectively, followed by regeneration of the surface by injection of 1 M L-Gly pH 3.0 for 30 s. The assays were carried out in duplicate in PBS-T (0.005%) at 25 °C. Analysis was conducted using the Biacore T200 Evaluation Software (Cytiva) (https://www.cytivalifesciences.com/en/us/support/software/biacore-downloads/biacore-t200-software) using a 1:1 binding kinetic fitting model.

Binding analysis by ITC

ITC experiments were performed by ITC using the Microcal VP-ITC instrument (Malvern) at a constant temperature of 25 °C.

For the analysis of the direct binding of Nb1 to FtsB or FhuD2, the proteins were placed in the cell and Nb1 in the syringe, at concentrations of 10 and 100 μM, respectively. Direct binding assays were conducted in PBS. For the competitive assays, siderophores in complex with ferric iron were placed in the syringe at a concentration of 200 μM and injected in a cell containing either 20 μM FtsB or 20 μM FtsB in the presence of 1.2-fold molar excess Nb1. Competitive assays were carried out in 50 mM acetate buffer pH 5.5 for FCH, FOE, FOB, and FXM, or HBS for BIS3-Fe2. Results were analyzed with the MicroCal ORIGIN software (Malvern) (https://www.malvernpanalytical.com/en/learn/knowledge-center/user-manuals/man0577en), excluding the first injection for the analysis, and using a “one set of sites” fitting model.

Binding analysis by Western blotting

To analyze the binding of Nb1 to FtsB and FhuD2 by Western blotting, SDS-PAGE was performed to separate 1, 5, or 20 μg of recombinant protein. Proteins were then transferred to a Nitrocellulose membrane. Recombinant proteins were subjected to electrophoresis, transferred, and treated with a 0.1 mg/ml solution of Nb1 in PBS-T (0.05%) for 1 h, followed by 45 min of incubation with a goat anti-Alpaca VHH conjugated with HRP, at 1:5000 dilution in PBS-T (0.05%).

Epitope analysis by HDX-MS

Protein samples of unbound FtsB, FtsB in complex with iron-bound FCH, and FtsB in complex with Nb1 were prepared in PBS buffer at a final concentration of 60 μM (1.9 mg/ml for unbound FtsB and FtsB bound to FCH and 2.9 mg/ml for the FtsB-Nb1 complex). Each protein solution was diluted 10-fold with PBS buffer in deuterium water (D_2_O). The diluted solutions were then aliquoted and incubated separately at 10 °C for 60, 120, 240, 480, 960, 1920, or 3840 s. Deuterium-labeled samples were quenched by diluting 2-fold with 8 M urea and 1 M Tris (2-carboxyethyl) phosphine hydrochloride at pH 3.0 (Quenching buffer). All the processes were performed automatically using HDx-3 PAL (LEAP Technologies). After quenching, the solutions were subjected to online pepsin digestion followed by liquid chromatography coupled with mass spectromtetry analysis using an UltiMate3000RSLCnano (Thermo Fisher Scientific) connected to a Q Exactive plus mass spectrometer (Thermo Fisher Scientific). Online pepsin digestion was performed with a protease type XIII/pepsin column (w/w, 1:1; 2.1 × 30 mm; NovaBioAssays) in formic acid solution (pH 2.5) for 3 min at 8 °C, at a flow rate of 50 μl/min. Desalting and analytical processes after pepsin digestion were performed using Acclaim PepMap 300 C18 column (1.0 × 15 mm, Thermo Fisher Scientific) and Hypersil GOLD (1.0 × 50 mm, Thermo Fisher Scientific) columns, using a mobile phase consisting of 0.1% formic acid solution (buffer A) and buffer A containing 90% acetonitrile (buffer B). The deuterated peptides were eluted at a flow rate of 45 μl/min, with a gradient of 10% to 90% of buffer B for 9 min. The conditions of the mass spectrometer were as follows: electrospray voltage, 3.8 kV; positive ion mode, sheath, and auxiliary nitrogen flow rate at 20 and two arbitrary units; ion transfer tube temperature at 275 °C; auxiliary gas heater temperature at 100 °C; and a mass range of m/z 200 to 2000. Data-dependent acquisition was performed using a normalized collision energy of 27 arbitrary units. The mass spectrometry and tandem mass spectrometry spectra were subjected to a database search analysis using the Proteome Discoverer 2.2 (Thermo Fisher Scientific) (https://knowledge1.thermofisher.com/Software_and_Downloads/Chromatography_and_Mass_Spectrometry_Software/Proteome_Discoverer/Proteome_Discoverer_User_Guides/Proteome_Discoverer_2.2_overview). Analysis of the deuteration levels of the peptide fragments was performed based on the mass spectrometry raw files using the HDExaminer software ver. 3.0 (Sierra Analytics).

To compare the spectra of deuterated samples with those of nondeuterated samples, we utilized a custom algorithm. Data from peptides from the unbound FtsB dataset and either the Nb1-bound or FCH-bound dataset was collected. For each peptide, the % of deuterium in the unbound protein was subtracted from the ligand-bound protein (ΔD%). Subsequently, for every amino acid of FtsB, deuterium uptake data of all the peptides that contain it was used to perform a linear regression between the length of the peptide (x-axis) and ΔD% (y-axis). This linear regression was extrapolated to a peptide length of n = 1 to assign a deuterium uptake score to the amino acid. A schematic representation of this process is shown in Figure S2. The error in the deuterium uptake score was estimated using the following formula:

$$[eqn]$$

where SER represents the standard error of the regression, Sxx is the sum of the squares of the difference between each x and the mean x, and n is the number of peptides used for the regression. This deuterium uptake score takes positive values when the amino acid is more solvent exposed (gains more deuterium) in the ligand-bound protein, negative values if it is more exposed in the unbound protein, or close to 0 when the ligand does not affect the solvent accessibility of the amino acid. This analysis was performed only for amino acids that had a peptide coverage of 3 or more, and at all of the time points collected during the HDX-MS experiment (1, 2, 4, 8, 16, 32, and 64 min of exposure to D_2_O solution). Examples of the linear regressions obtained from the data collected between Nb1-bound FtsB and unbound FtsB using this scoring algorithm are shown in Figure S3, including Tyr81 (A), Tyr137 (B), Trp204 (C), and Tyr232 (D) (amino acids that participate in Nb1 binding and experience changes in deuterium uptake), as well as Val115 (E) and Lys155 (F) (amino acids that do not participate in Nb1 binding and do not show changes in deuterium uptake).

Protein crystallization, data collection, and refinement

To produce crystals of FtsB in complex with Nb1, 6 mg of FtsB were mixed with a 1.5-fold molar excess of antibody. The mix was incubated on ice for 1 h, and the complex was isolated from the excess of unbound nanobody by size-exclusion chromatography using an AKTA purifier system connected to a 16/600 Superdex 75 column and equilibrated with PBS. Fractions containing the complex were collected and used for crystallization at a final concentration of 20 mg/ml. Crystals of the FtsB–Nb1 complex were obtained by the vapor diffusion method (hanging drop) at 20 °C using a solution of 6% tacsimate pH 6.0, 0.1 M BIS-Tris pH 6.5, 20% PEG 3350. Suitable crystals were harvested, incubated with crystallization solution supplemented with 20% glycerol, and flash-frozen in liquid nitrogen for storage until data collection. Diffraction datasets from single crystals were collected in beamline BL5A at the Photon Factory (Tsukuba) under cryogenic conditions (100 K). Diffraction images were processed with the program MOSFLM and merged and scaled with the program SCALA (69) of the CCP4 suite (70). The structure of the Nb1–FtsB complex was determined by the Molecular Replacement method using the coordinates of FCH-bound FtsB (PDB: 8XEU (22)) and the anti-ERP3 nanobody in chain B of the PDB entry 6QUP (71) as models for PHASER (72). The model was refined with the program REFMAC5 (73) and built manually with COOT (74). Validation was carried out with MOLPROBITY (75). Data collection and refinement statistics are given in Table 2.Table 2. Data collection and refinement statisticsParameterFtsB + Nb1 nanobodyaData collection Space groupP 2_1_ 2_1_ 2_1_ Unit cell a, b, c (Å)38.4, 60.5, 318.1 α, β, γ (°)90.0, 90.0, 90.0 Resolution (Å)48.1–2.55 (2.66–2.55) Wavelength1.0000 Observations134,720 (14,667) Unique reflections24,247 (2703) Rmerge.0.127 (0.609) Rp.i.m.0.058 (0.272) CC_1/2_0.995 (0.748) *I/σ (I)*10.2 (2.7) Multiplicity5.6 (5.4) Completeness (%)96.0 (91.0)Refinement statistics Resolution (Å)48.1–2.55 Rwork/Rfree (%)25.3/30.7 No. complexes2 No. atoms Nb1 nanobody1752 FtsB4366 Water81 B-factor (Å^2^) Nb1 nanobody40.7 FtsB40.8 Water21.9 Ramachandran plot Preferred (%)95.4 Allowed (%)4.4 Outliers (%)0.3 RMSD bond (Å)0.004 RMSD angle (°)1.14 PDB entry code9ULPaStatistical values given in parentheses refer to the highest resolution bin.

AlphaFold3 modeling of the FtsB–FtsA2CD complex

To model the structure of the FtsB–FtsA2CD complex, the amino acid sequences of FtsA (amino acids 1–260), FtsB (amino acids 27–310), FtsC (amino acids 1–306) and FtsD (amino acids 1–345) were collected from S. pyogenes strain SSI1 genome and used as input for AlphaFold3 server (51) with stoichiometry 2:1:1:1. Structure prediction was conducted over four independent runs with different seeds, and 5 output solutions per seed, for a total of 20 output solutions. The predicted template modeling score and the interface predicted template modeling score of each solution were used to evaluate the accuracy of the prediction of the structure of the individual models, as well as the relative positions of each subunit, respectively.

S. pyogenes JRS4 KO strain preparation

The preparation of JRS4 KO strains for FtsB and HasA was conducted as previously described (76). Briefly, two-step allele exchange was conducted using the thermosensitive vector pSET4s (77) carrying 5′ and 3′ flanking regions (800–1000 bp) of the corresponding target gene (amplified from GAS JRS4 strain genomic DNA). The plasmid was electro-transformed into WT JRS4 cells and plated on 100 μg/ml spectinomycin in THY agar plates at a permissive temperature of 28 °C. To generate chromosomal single-crossover mutants, selected colonies were grown at a nonpermissive temperature of 37 °C with spectinomycin. To induce the second crossover event, single-crossover mutants confirmed to exchange a chromosomal allele were then subcultured at a permissive temperature of 28 °C without spectinomycin. Spectinomycin-sensitive colonies were screened to distinguish between deletion mutants or clones with restored WT genotype using colony PCR.

Radioisotope uptake assays

The uptake of different iron sources by S. pyogenes JRS4 (WT, ΔFtsB KO or ΔHasA knock out) was evaluated using a ^55^Fe radiotracer. ^55^FeCl_3_ was purchased from Eckert and Ziegler with specific activity 590 MBq/mg (0.225 mM). For the uptake assays, ^55^Fe was used in its free form, or incorporated into Heme or siderophores. A 15 μl aliquot of the ^55^Fe stock solution was mixed with 35 μl of water, or with 35 μl of either 0.6 mM siderophores in water or 0.6 mM protoporphyrin IX in 5% dimethyl sulfoxide. The sample was incubated at room temperature for 30 min, then mixed with 170 μl of Todd-Hewitt broth supplemented with 2% yeast extract (THY) and incubated for 30 min at 37 °C. Finally, the samples were centrifuged at 15,000g, for 10 min, and the supernatant was collected.

S. pyogenes JRS4 was grown at 37 °C, 600 rpm in 1.9 ml of THY. For inhibition assays, when absorbance at 600 nm reached 0.55, 0.1 ml of culture were discarded, and 0.1 ml of either PBS or PBS containing 150 μM Nb1 were added instead. Uptake assay was started when absorbance at 600 nm reached 0.60 by the addition of 100 μl of the working ^55^Fe solution to a final volume of 2 ml. Three aliquots of 0.6 ml were taken after 30 min of uptake at 37 °C, 600 rpm, filtered through cellulose nitrate filters (pore size, 0.45 μm; Pall Life Sciences), and washed three times with 1 ml of 0.1 M LiCl. Filters were transferred to vials containing 3 ml of Ultima Gold LCS Cocktail, and the radioactivity of each sample was determined in a Tir-Carb 4810 liquid scintillation counter for 1 min.

Data availability

The coordinates and structure factors of FtsB in complex with Nb1 have been deposited in the PDB with the entry code 9ULP. All remaining data are contained within the article.

Supporting information

This article contains supporting information (22).

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.