Conditional growth defect of Bordetella pertussis and Bordetella bronchiseptica ferric uptake regulator (fur) mutants

Abstract

Outer-membrane vesicles (OMVs) are promising tools in the development of novel vaccines against the respiratory pathogens Bordetella pertussis and Bordetella bronchiseptica. Unfortunately, vesiculation by bordetellae is too low for cost-effective vaccine production. In other bacteria, iron limitation or inactivation of the fur gene has been shown to increase OMV production, presumably by downregulation of the mla genes, which encode machinery for maintenance of lipid asymmetry in the outer membrane. Here, we followed a similar approach in bordetellae. Whereas a fur mutant was readily obtained in B. bronchiseptica, a B. pertussis fur mutant could only be obtained in iron-deplete conditions, indicating that a fur mutation is conditionally lethal in this bacterium. The fur mutants displayed a growth defect in iron-replete media, presumably because constitutive expression of iron-uptake systems resulted in iron intoxication. Accordingly, expression of the Escherichia coli ferritin FtnA to sequester intracellularly accumulated iron rescued the growth of the mutants in these media. The fur mutations led to the constitutive expression of novel vaccine candidates, such as the TonB-dependent receptors FauA for the siderophore alcaligin and BhuR for heme. However, neither inactivation of fur nor growth under iron limitation improved vesiculation, presumably because the expression of the mla genes appeared unaffected.

Bordetella fur mutants display a growth defect in iron-replete media most likely due to iron overload and can be rescued by the heterologous expression of Escherichia coli ferritin FtnA.

Introduction

Iron is an essential nutrient metal for most bacteria. During infection, the availability of iron is very low for invading pathogens because the host produces proteins, such as transferrin and lactoferrin, which sequester free iron as a defense mechanism against infection (Hood and Skaar 2012). Bacteria deploy several mechanisms to capture iron from the environment under iron-limiting conditions. In Bordetella pertussis and Bordetella bronchiseptica, two Gram-negative bacteria that infect the respiratory tract of humans and other mammals, respectively (Mattoo and Cherry 2005), these mechanisms include the production and secretion of the siderophore alcaligin (Moore et al. 1995). After chelating iron in the environment, the ferric complex of the siderophore is taken up by the bacteria in a process that requires for the first step a TonB-dependent receptor in the outer membrane (OM). The receptor for Fe³⁺-alcaligin is FauA (Brickman and Armstrong 1999). Besides alcaligin, bordetellae can also utilize the xenosiderophores enterobactin (Beall and Sanden 1995a), ferrichrome, and desferrioxamine B (Beall and Hoenes 1997), as well as heme as iron sources (Brickman and Armstrong 2009). The uptake of enterobactin and heme requires the TonB-dependent receptors BfeA and BhuR, respectively, whilst the receptors for ferrichrome and desferrioxamine B have not been identified yet. Although iron is essential for growth, its excessive intracellular accumulation is toxic as it may catalyze Fenton chemistry resulting in the production of oxygen radicals (Imlay et al. 1988, Braun 1997). Therefore, the production of iron-acquisition mechanisms needs to be controlled (Braun 1997). The transcription of the genes for these mechanisms is usually repressed under iron-replete conditions by the ferric uptake regulator (Fur) (Braun 1997), as has also been shown to be the case in bordetellae (Brickman et al. 2007, Beall and Sanden 1995b).

Several studies have suggested that outer-membrane vesicles (OMVs) could aid in iron uptake (Schwechheimer and Kuehn 2015, Orench-Rivera and Kuehn 2016). OMVs are released from Gram-negative bacteria as a result of bulging of the OM. The role of OMVs in metal acquisition is supported by the abundance of TonB-dependent receptors in OMVs from Neisseria meningitidis, Porphyromonas gingivalis, and B. pertussis (Lappann et al. 2013, Veith et al. 2014, Gasperini et al. 2017). Indeed, OMVs from B. pertussis were reported to be able to load iron and deliver it to bacterial cells (Gasperini et al. 2017). Furthermore, OMV production was found to be increased in Haemophilus influenzae under iron-limiting conditions and in fur mutants of Vibrio cholerae, Escherichia coli, and H. influenzae (Roier et al. 2016), indirectly supporting a role for OMVs in iron acquisition. This increased OMV production was related to the downregulation of genes coding for the maintenance of lipid asymmetry (Mla) system (Roier et al. 2016). The OM is an asymmetric bilayer, with phospholipids and lipopolysaccharides being present in the inner and outer leaflet, respectively. The Mla system transports phospholipids that appear in the outer leaflet back to the inner membrane, thereby maintaining OM lipid asymmetry (Malinverni and Silhavy 2009). According to the proposed model, downregulation of the Mla system under iron limitation results in the accumulation of phospholipids in the outer leaflet of the OM and, consequently, the OM bulges and OMVs pinch off (Roier et al. 2016). We have previously demonstrated that disruption of the Mla system in B. pertussis resulted in only a slight increase in vesiculation (de Jonge et al., manuscript submitted for publication). However, when also the pldA gene was inactivated, a gene that encodes OM phospholipase A, which degrades phospholipids in the outer leaflet of the OM, OMV production was strongly enhanced.

We are investigating the possibility of using OMVs in the development of novel Bordetella vaccines. Previously developed whole-cell vaccines against B. pertussis appeared to be too reactogenic (Cody et al. 1981), and they were replaced by acellular pertussis (aP) vaccines consisting of one to five purified antigens. However, these aP vaccines induce only short-term immunity and do not protect against colonization and transmission of B. pertussis (Warfel et al. 2014, Wilk et al. 2019). Probably as a consequence, pertussis is resurging in the last decades (Esposito et al. 2019). Vaccines against B. bronchiseptica have also been developed but the duration of vaccine-induced immunity remains unclear (Ellis 2015). Thus, novel vaccines against both B. pertussis and B. bronchiseptica need to be developed. OMVs are attractive nanostructures for vaccine development because of their content of a wide variety of OM proteins, their adjuvant properties, and their uptake by antigen-presenting cells (van der Pol et al. 2015). Immunization studies with B. pertussis OMVs in mice demonstrated the induction of a mixed systemic T helper (Th)1/Th2/Th17 response with reduced pro-inflammatory activity compared to immunization with whole cells (Raeven et al. 2016). Efficient bacterial clearance of the lungs and the nasal cavity was achieved after intranasal immunization (Raeven et al. 2020). Together, these data indicate that Bordetella OMVs are promising vaccine candidates. Unfortunately, spontaneous OMV production by Bordetella species is relatively low (Hozbor et al. 1999). In this study, we investigated the effect of iron limitation on OMV production in B. pertussis and B. bronchiseptica. Although we have already reported several methods to increase the production of native OMVs (de Jonge et al. 2021; de Jonge et al., manuscript submitted for publication), the increased production of OMVs under iron limitation could have the additional advantage of the expression of novel relevant antigens. TonB-dependent receptors are considered attractive vaccine candidates, since they are exposed at the cell surface, upregulated in vivo, and essential for successful infection (Wang et al. 2021). Besides, iron limitation also leads to increased expression of other vaccine candidates, such as IRP1-3 and the type III secretion system in Bordetella (Alvarez Hayes et al. 2011, Brickman et al. 2011, Kurushima et al. 2012, Fasciano et al. 2019). Since iron limitation is also expected to restrict bacterial growth, which again could negatively impact OMV yield, we have also constructed fur mutants for their potential application.

Materials and methods

Growth conditions

Bordetella pertussis and B. bronchiseptica strains were grown on Bordet-Gengou (BG) agar (Difco) plates containing 15% (v/v) defibrinated sheep blood (bioTRADING) or 10 mg/mL of bovine serum albumin (BSA) (Hiramatsu et al. 2019) at 35°C. For liquid cultures, bacteria were scraped from BG plates and grown in Verwey medium (Verwey et al. 1949), which was supplemented with 0.01 g/L of FeSO₄·7H₂O where indicated, or in Stainer-Scholte (SS) medium (Stainer and Scholte 1971) at 35°C while shaking at 175 rpm. For growth of B. pertussis in SS medium, 1 g/L of heptakis(2,6-di-O-methyl)-β-cyclodextrin (Sigma-Aldrich) was added. E. coli strains were grown on lysogeny broth (LB) agar plates at 37°C or in liquid LB while shaking at 200 rpm. For strain selection or plasmid maintenance, 100 µg/mL (for B. pertussis and E. coli) or 200 µg/mL (for B. bronchiseptica) of ampicillin, 10 µg/mL of gentamicin, 300 µg/mL of streptomycin, 50 µg/mL of nalidixic acid, or 5 µg/mL of cefotaxime were added to the medium. To induce gene expression, media were supplemented with 1 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG). Bacterial growth in liquid cultures was monitored by measuring the optical density at 600 nm (OD₆₀₀) using a Novaspec III + spectrophotometer (Biochrom).

Construction of mutants and plasmids

Details of genetic constructions are provided in the Supplementary Materials, where all strains, primers, and plasmids used are listed in Supplementary Tables S1, S2, and S3, respectively.

Siderophore production

Production of siderophores was determined with the chrome azurol S (CAS) assay (Schwyn and Neilands 1987). Bacterial cells were pelleted from cultures by low-speed centrifugation, and 0.5 mL of supernatant, not normalized to OD₆₀₀, was mixed with 0.5 mL of CAS solution. After incubation for 1 h at room temperature, absorbance at 630 nm (A₆₃₀) was measured.

OMV isolation and quantification

Bacterial strains were pre-grown for 24 h in Verwey medium. Cultures were then diluted to an OD₆₀₀ of 0.05 in 50 mL Verwey medium, supplemented with FeSO₄·7H₂O where indicated, and grown in 250-mL baffled flasks. After 48 h of growth, OMVs were isolated as described (de Jonge et al. 2021) and resuspended in phosphate-buffered saline. OMV yield was determined based on protein content using the Lowry DC protein assay (Bio-Rad) following the instructions of the manufacturer.

SDS-PAGE and Western blotting

SDS-PAGE and Western blotting were performed as described (de Jonge et al. 2021). Primary antibodies used were a rabbit anti-FauA antiserum (de Jonge et al. 2021) and horseradish peroxidase-conjugated anti-FLAG-tag monoclonal antibodies (Sigma Aldrich).

RNA isolation and reverse transcription-quantitative PCR (RT-qPCR)

Bacterial strains were grown in Verwey medium, and cells were harvested during exponential growth. Harvested cells were stored at -80°C in RNAprotect Bacteria Reagent (Qiagen) until RNA isolation. RNA was isolated using the NucleoSpin RNA kit (Macherey-Nagel), and residual DNA was removed with the TURBO DNA-free kit (Ambion). Next, cDNA was generated from 1 µg of RNA using the Transcriptor High Fidelity cDNA Synthesis Kit (Roche). RT-qPCR was performed using the SYBR Green master mix (ThermoFisher) and a ViiA 7 Real-Time PCR System (Applied Biosystems) with primers listed in Supplementary Table S2 with three technical replicates per sample. Data was analyzed with QuantStudio v1.3, and relative gene expression was determined using the 2^–∆∆Ct method with rpoB for normalization. Gene expression was related to one sample of the wild-type strains per gene.

Results

Isolation of B. pertussis and B. bronchiseptica fur mutants

To investigate the influence of iron availability on OMV production in Bordetella, we first attempted to construct fur mutants of B. pertussis and B. bronchiseptica via allelic exchange. However, whilst a fur mutant was readily obtained in B. bronchiseptica, the same selection procedure failed for B. pertussis, suggesting that fur is an essential gene in the latter species. The selection procedure was performed on BG agar plates supplemented with blood, which is rich in iron resources. We considered the possibility that the anticipated constitutive expression of iron-uptake systems in a fur mutant could lead to excessive iron uptake, which could potentially be toxic for the cells. To get around this potential problem, mutant selection was next performed on BG plates containing BSA instead of blood. Following this strategy, we succeeded to also obtain a B. pertussis fur mutant.

Growth characteristics of the fur mutants

To test the effect of the fur mutations on growth, wild-type B. pertussis and B. bronchiseptica and their fur mutant derivatives were grown in two liquid media routinely used for growth of B. pertussis, Verwey medium and SS medium. We previously reported that Verwey medium is limited in iron and that growth of wild-type Bordetella in this medium results in the induction of the synthesis of the alcaligin receptor FauA (de Jonge et al. 2021). In contrast, standard SS medium contains iron sulfate as an iron source. Compared to the wild types, growth of both fur mutants was reduced in both media, but growth restriction was most severe in the SS medium, where particularly the B. pertussis fur mutant failed to grow at all (Fig. 1A). As the composition of the media used also differs in other respects besides the iron content, we next wanted to determine directly the influence of iron availability. To this end, growth was also studied in Verwey medium supplemented with iron sulfate. Indeed, whilst iron supplementation, as expected, improved the growth of the wild-type strains considerably, it impacted the growth of both mutants in this medium (Fig. 1A). The residual growth of the B. pertussis fur mutant in iron-supplemented Verwey medium could be due to the presence of starch in this medium, which has the capacity to bind iron (Thomas et al. 1976). Large variability was noticeable in the final growth yield of the B. bronchiseptica fur mutant in Verwey medium supplemented with iron sulfate (Fig. 1A). This could be the result of suppressor mutations arising in the cultures, allowing for a better growth of the fur mutant in iron-rich conditions. Accordingly, whilst colonies of the wild-type B. bronchiseptica strain streaked on blood-agar plates were rather uniform in size, the size of the fur mutant colonies was heterogeneous with large colonies appearing in a background of small colonies, indicating the selection of suppressor mutants with improved growth on these plates (Supplementary Fig. S1). Together, these results indicate a growth defect of the fur mutants in iron-replete conditions, presumably because of an excessive uptake of iron.

Figure 1.

Influence of fur inactivation on growth. (A) Wild types (WT) and fur mutants of B. pertussis (Bp) and B. bronchiseptica (Bb) were grown in Verwey medium, either supplemented with iron (+Fe) or not, and in SS medium, and growth was monitored by measuring the OD₆₀₀. (B) The fur mutants, either containing pFtnA or not, as indicated, were grown in SS medium. The pFtnA-containing strains were grown either with, or where indicated (w/o IPTG), without IPTG, and these cultures were inoculated from precultures on BG blood agar plates containing IPTG. Growth was monitored by measuring the OD₆₀₀. (A,B) Graphs show mean values with standard deviations of three independent experiments, except for the growth of the B. bronchiseptica fur mutant in Verwey medium supplemented with iron, where each individual result of four replicates is depicted and for the B. pertussis fur mutant containing pFtnA grown in the absence of IPTG (n = 1).

Expression of E. coli FtnA rescues Bordetella fur mutants from iron toxicity

To further substantiate the idea that the growth defect of the Bordetella fur mutants under iron-replete conditions is due to the toxic accumulation of iron within the cells, we considered the possibility of sequestering intracellular iron by expressing an iron-storage protein. In E. coli, excessive intracellular iron can be stored in ferritin (FtnA) and bacterioferritin (Bfr). Of these two, FtnA seems to play the major role in iron storage (Abdul-Tehrani et al. 1999). Bordetella pertussis and B. bronchiseptica do contain homologues of Bfr (locus tags BP0174 and BB4918 in reference strains Tohama I and RB50, respectively), but not of FtnA. Thus, we expressed codon-optimized ftnA of E. coli in the fur mutant strains of B. pertussis and B. bronchiseptica, and expression was confirmed by Western blotting targeting the engineered C-terminal FLAG tag on the protein (shown for B. pertussis in Supplementary Fig. S2). Subsequently, growth was assessed in iron-replete SS medium. The growth curves indeed showed that expression of ftnA drastically improved the growth of the fur mutants in this medium (Fig. 1B). Thus, we conclude that the growth defect of the fur mutants is due to the toxic intracellular accumulation of iron.

Constitutive synthesis of siderophores and TonB-dependent receptors in fur mutants

To confirm the expected phenotype of the constructed fur mutants, the biosynthesis of the alcaligin receptor FauA, which is repressed by Fur under iron-replete conditions (Brickman and Armstrong 2009), was assessed by Western blot analysis of whole-cell lysates grown in iron-poor (Verwey) or iron-replete (Verwey medium supplemented with iron and SS) media. As expected, FauA was detected after growth of the wild-type strains in the iron-deficient Verwey medium but not after growth in the iron-replete media (Fig. 2). In contrast, FauA was detected in the fur mutants, independently of the culture conditions used. We also assessed the production of the siderophore alcaligin in iron-rich SS medium using the CAS assay. As expected, siderophore production was drastically increased in the B. bronchiseptica fur mutant relative to the wild type (Supplementary Fig. S3). However, such increase was not observed in the B. pertussis fur mutant. This is probably due to the very poor growth of the B. pertussis fur mutant in SS medium (Fig. 1A). Thus, these data show the constitutive production of FauA in the fur mutants of B. pertussis and B. bronchiseptica, independent of the growth medium used.

Figure 2.

FauA production in fur mutants. Wild-type (WT) B. pertussis (Bp) and B. bronchiseptica (Bb) and their fur mutants were grown in Verwey medium, either supplemented with iron (+Fe) or not, or in SS medium. Whole-cell lysates from equal amounts of cells (based on OD₆₀₀) were analyzed by Western blotting using antiserum directed against FauA. Only the relevant part of the blot is shown. Molecular weight markers are indicated on the left.

The expression of fauA was further evaluated by RT-qPCR, a method, which also allowed us to assess the expression of genes for other antigens for which no antisera were available. For this purpose, the heme receptor BhuR and another TonB-dependent receptor, BfrD, which functions as a receptor for catecholamines (Brickman et al. 2015), were elected. The synthesis of both of them was expected to be regulated by iron availability in a Fur-dependent manner, since a putative Fur-binding site is present in the promoter regions of their genes (Brickman and Armstrong 1999, Passerini de Rossi et al. 2003, 2009). Wild-type B. pertussis and B. bronchiseptica and their fur mutant derivatives were grown in Verwey medium as SS medium could not be used due to poor growth of the fur mutants. Verwey is an iron-poor medium, and the limited iron sources available are consumed during growth. RNA was isolated at time points early in the growth when FauA was still absent in the wild-type strains, that is before the limited iron resources available were depleted (Supplementary Fig. S4). Expression of fauA and bhuR was higher in the fur mutants compared to the wild types, although the difference for fauA was not significant in B. pertussis (P = 0.07) (Fig. 3A and B). Remarkably, a large variation in bhuR expression was noticeable between the biological replicates of wild-type B. pertussis, where high and low expression correlated with a relatively high and low OD₆₀₀ (0.81 ± 0.01 versus 0.39 ± 0.05), respectively, at the time of harvest. Presumably, reaching a higher OD implicates that the limited iron resources available were more exhausted than in the cultures with lower OD. Expression of bfrD was not upregulated in the fur mutants or perhaps even downregulated in the fur mutant of B. bronchiseptica, although the difference was not significant (P = 0.17) (Fig. 3A and B).

Figure 3.

Expression of genes for relevant antigens in fur mutants. Relative gene expression of selected vaccine candidates was determined by RT-qPCR in wild-type (WT) B. pertussis (Bp) (A) and B. bronchiseptica (Bb) (B) and their fur mutant derivatives. The mean value of each data set of four biological replicates is depicted with a horizontal line. Significant differences were determined with multiple t tests using GraphPad Prism 6 and are indicated by asterisks (*P < 0.05; **P < 0.01; ^****P < 0.0001).

The influence of iron limitation on OMV production

To determine if iron availability affects OMV production, as has been reported in several other Gram-negative bacteria (Roier et al. 2016), OMV production by wild-type B. bronchiseptica and B. pertussis grown in either Verwey medium or Verwey medium supplemented with iron and by the fur mutants was evaluated. To verify iron limitation of the wild types grown in Verwey medium, FauA and siderophore production in the cultures used for OMV isolation was confirmed (Supplementary Fig. S5). Quantification of isolated OMVs based on protein content indicated no significant effect of iron limitation or the fur mutations on OMV production (Fig. 4A).

Figure 4.

Effect of iron limitation on OMV production and mlaF expression. Wild-type (WT) B. pertussis (Bp) and B. bronchiseptica (Bb) and their fur mutant derivatives were grown in Verwey medium, either supplemented with iron (+Fe) or not. (A) OMVs released were isolated and quantified based on protein content using a Lowry assay. The yield is expressed as the amount of protein per liter of bacterial culture per OD₆₀₀ unit. Graph shows mean values with standard deviations of three independent experiments. Significant differences were determined using one-way ANOVA followed by Tukey's multiple comparisons test using GraphPad Prism 6. No statistically significant differences were found. (B) Relative expression of mlaF was determined using RT-qPCR. The mean value of each data set of four biological replicates is depicted with a horizontal line. Significant differences between the wild types and their fur mutant derivates were determined with unpaired t tests using GraphPad Prism 6. No statistically significant differences were found.

Increased OMV production by the fur mutants of V. cholerae, H. influenzae, and E. coli was related to downregulation of the synthesis of the Mla system, which leads to an accumulation of phospholipids in the outer leaflet of the OM and bleb formation (Roier et al. 2016). Therefore, we analyzed whether the fur mutations affected expression of the mla operon in Bordetella. RT-qPCR analysis indicated an increased, rather than the anticipated decreased mlaF expression in the B. pertussis fur mutant, although this difference was not significant (P = 0.17) (Fig. 4B). Inactivation of fur did not change mlaF expression in B. bronchiseptica.

Discussion

To study the effect of iron limitation on OMV production in B. pertussis and B. bronchiseptica, we chose to inactivate the fur gene to mimic iron limitation independent of iron availability. Inactivation of fur has been successful in many other Gram-negative bacteria (Roier et al. 2016, Lee et al. 2017). Also for B. bronchiseptica, fur mutants have been described before (Brickman and Armstrong 1995), and such mutants were readily obtained in the present study. Inactivation of fur in B. pertussis, however, proved to be more challenging and, to our knowledge, such mutants have not been described before. Indeed, fur has previously been classified as an essential gene in B. pertussis based on high-throughput transposon sequencing (Tn-seq) experiments both in vitro and in vivo in an intranasal murine infection model in which fur mutants were not obtained (Gonyar et al. 2019, Belcher et al. 2020). Nevertheless, we could isolate B. pertussis fur mutants on plates containing BSA instead of blood, presumably because fur mutants are not viable on blood agar due to excessive iron uptake. Accordingly, growth of the fur mutants of both B. pertussis and B. bronchiseptica was severely restricted in iron-rich liquid media. In a few other species, Fur has been described to be essential, including Pseudomonas aeruginosa and Chromobacterium violaceum (Pasqua et al. 2017, Santos et al. 2020). However, in C. violaceum, essentiality of Fur appeared to be conditional (Santos et al. 2020), as, like in B. pertussis, fur mutants could eventually be obtained in iron-deplete conditions. Growth of the fur mutants in iron-rich medium could be rescued by the expression of iron-storage protein FtnA of E. coli, demonstrating that the growth defect is due to intracellular iron accumulation. Although B. pertussis and B. bronchiseptica contain a homologue of another iron-storage protein, Bfr, sequestration of iron by this protein in these conditions is apparently not sufficient to sustain growth. Expression of bfr in B. pertussis is induced under iron-rich conditions (Brickman et al. 2011) and is presumably positively regulated by Fur, as has been described in E. coli (Massé and Gottesman 2002). Consequently, bfr expression in the Bordetella fur mutants is probably low and insufficient to avoid iron intoxication.

Inactivation of fur was expected to result in the constitutive production of iron-limitation-inducible receptors in the OM, which are potentially relevant vaccine antigens (Wang et al. 2021). So far, 12 and 16 putative TonB-dependent receptors have been identified in B. pertussis and B. bronchiseptica, respectively (Brickman et al. 2007). At least four of these receptors, i.e. FauA, BfeA, BhuR and BfrD, are involved in iron acquisition (Brickman et al. 2007, Brickman et al. 2015). We examined expression of the fauA, bhuR, and bfrD genes. Expression of fauA and bhuR was indeed upregulated in the fur mutants, but expression of bfrD did not increase, even though a putative Fur box was identified upstream of bfrD (Passerini de Rossi et al. 2003). Previous studies gave conflicting results in this respect and showed either an increase (Passerini de Rossi et al. 2003) or a decrease (Brickman et al. 2015) of bfrD expression during iron limitation. BfrD was previously selected as a promising vaccine candidate, and proteomic analysis demonstrated the presence of BfrD in OMVs (Gasperini et al. 2017, 2018). As these OMVs were isolated from cultures grown in iron-rich SS medium, BfrD synthesis is apparently already high under those conditions.

We expected that fur inactivation and iron limitation would increase OMV production in bordetellae as has been reported in other bacteria (Roier et al. 2016). However, such an increase was not observed. The improved OMV production in E. coli, V. cholerae, and H. influenzae in the absence of Fur was reported to be related to the downregulation of the synthesis of the Mla system (Roier et al. 2016). In the absence of the Mla system, phospholipids accumulate in the outer leaflet of the OM which results in bulging of the OM and subsequent OMV formation (Roier et al. 2016). Fur-binding sites were predicted upstream of the mla operons in E. coli, V. cholerae, and H. influenzae, indicating a role for Fur in positive regulation of mla transcription (Roier et al 2016). We have previously shown that OMV production is increased in an Mla-deficient B. pertussis mutant also lacking the OM phospholipase A (de Jonge et al., manuscript submitted for publication), consistent with the proposed model of OMV biogenesis (Roier et al. 2016). However, expression of mlaF was not significantly decreased in the Bordetella fur mutants (Fig. 4B), which is in agreement with data of a previous transcriptomics study that did not show altered expression of the B. pertussis mla genes in iron-limiting conditions (Brickman et al. 2011). Accordingly, we could not identify an obvious Fur-binding site in the promoter region of the mla operon. Thus, a different regulation of the expression of the mla genes could explain why iron limitation and fur mutations do not increase OM vesiculation in Bordetella.

OMVs are promising tools in the development of new vaccines for B. pertussis and B. bronchiseptica (Raeven et al. 2016, 2018, 2020). Since spontaneous OMV production by bordetellae is low, we are studying ways to increase OMV release. Unfortunately, neither iron limitation nor inactivation of fur increased OMV production. However, since genes encoding relevant antigens, such as FauA and BhuR, are upregulated in the fur mutant strains, these strains could still be useful in the development of OMV-based vaccines. OMV production by the fur mutants could then be increased, e.g. by applying a heat shock which was previously demonstrated to increase vesiculation (de Jonge et al. 2021). Either expression of E. coli FtnA or suppressor mutations could improve the growth of the fur mutants under iron-replete conditions, which is needed for cost-effective vaccine production.

Funding

This work was supported by the domain Applied and Engineering Sciences (TTW) of The Netherlands Organization for Scientific Research (NWO) (TTW Perspectief Grant number 14921), which received financial contributions for this grant from GlaxoSmithKline Biologicals SA and PULIKE Biological Engineering Inc.

Conflict of interest statement. None declared.