Selective Enhancement of the Cell-Permeabilizing Activity of Adenylate Cyclase Toxin Does Not Increase Virulence of Bordetella pertussis

The whooping cough agent, Bordetella pertussis, secretes an adenylate cyclase toxin–hemolysin (CyaA, ACT, or AC-Hly) that catalyzes the conversion of intracellular ATP to cAMP and through its signaling annihilates the bactericidal activities of host sentinel phagocytes. In parallel, CyaA permeabilizes host cells by the formation of cation-selective membrane pores that account for the hemolytic activity of CyaA. The pore-forming activity contributes to the overall cytotoxic effect of CyaA in vitro, and it has previously been proposed to synergize with the cAMP-elevating activity in conferring full virulence on B. pertussis in the mouse model of pneumonic infection. CyaA primarily targets myeloid phagocytes through binding of their complement receptor 3 (CR3, integrin αMβ2, or CD11b/CD18). However, with a reduced efficacy, the toxin can promiscuously penetrate and permeabilize the cell membrane of a variety of non-myeloid cells that lack CR3 on the cell surface, including airway epithelial cells or erythrocytes, and detectably intoxicates them by cAMP. Here, we used CyaA variants with strongly and selectively enhanced or reduced pore-forming activity that, at the same time, exhibited a full capacity to elevate cAMP concentrations in both CR3-expressing and CR3-non-expressing target cells. Using B. pertussis mutants secreting such CyaA variants, we show that a selective enhancement of the cell-permeabilizing activity of CyaA does not increase the overall virulence and lethality of pneumonic B. pertussis infection of mice any further. In turn, a reduction of the cell-permeabilizing activity of CyaA did not reduce B. pertussis virulence any importantly. These results suggest that the phagocyte-paralyzing cAMP-elevating capacity of CyaA prevails over the cell-permeabilizing activity of CyaA that appears to play an auxiliary role in the biological activity of the CyaA toxin in the course of B. pertussis infections in vivo.


Introduction
The Gram-negative coccobacillus Bordetella pertussis causes a highly contagious respiratory disease called pertussis, or whooping cough. Pertussis affects children and adults of all age groups and it used to be the major cause of infant mortality before the global introduction of whole cell-based pertussis vaccines seven decades ago [1][2][3]. Unrecognized infections and/or mild disease due to B. pertussis continue to persist also in highly vaccinated populations, which makes whooping cough the least controlled vaccinepreventable infectious disease. It is estimated that more than 24 million pertussis cases and 160,000 pertussis-related deaths occur annually worldwide [4].
to cause a lethal infection even when applied at 10-times higher inoculation doses than is the LD 50 dose of the wild-type bacteria [42]. Moreover, at an equal bacterial burden in the infected lungs, the B. pertussis mutant secreting CyaA-E570Q+K860R elicited a significantly lower level of inflammatory pathology than the parental B. pertussis bacteria [42]. This led us to propose that the cell-permeabilizing (hemolytic) activity of CyaA may be required for full virulence of B. pertussis infections. Nevertheless, a caveat remained, since the CyaA-E570Q+K860R toxin exhibited a very low capacity to elevate cAMP concentrations in non-myeloid CR3 − cells. Hence, it could not be ruled out that cAMP intoxication of the CR3 − airway epithelial cells by the intact CyaA, rather than the cell-permeabilizing activity of CyaA, accounted for most of the pro-inflammatory signaling of infected epithelia and the resulting virulence of B. pertussis infection. The non-hemolytic mutant secreting the CyaA-E570Q+K860R toxin would not trigger such signaling and the reduction of its virulence could thus not be clearly attributed to the absence of cell-permeabilizing activity of CyaA.
Here, we describe CyaA variants with either selectively and strongly enhanced, or reduced cell-permeabilizing activity, that still exhibit a full capacity to elevate cAMP levels also in CR3 − cells, like intact CyaA. Using B. pertussis mutants secreting such CyaA variants, we show that the selective enhancement of the cell-permeabilizing capacity of the secreted CyaA toxin does not further increase the virulence and mouse lung colonization capacity of B. pertussis.

CyaA Variants with Selectively Up-or Down-Modulated Pore-Forming (Hemolytic) Activity Retain Full cAMP-Elevating Capacity When Secreted from B. pertussis
To assess the impact of the pore-forming activity of CyaA on the virulence of B. pertussis in mice, we used two previously described recombinant CyaA variants that exhibit an intact cAMP-elevating capacity on both CR3 + and CR3 − cells, while exhibiting a selectively enhanced or reduced specific pore-forming (hemolytic) activity [44,45]. The 'hyperhemolytic' CyaA-D445N+D446N+E448Q (CyaA-NNQ) mutant produced as a recombinant CyaC-activated protein in E. coli was previously found to exhibit a several times higher specific pore-forming activity on both CR3 − and CR3 + cells than the intact CyaA [44]. In contrast, the recombinant CyaA-V695P variant showed about five times reduced specific pore-forming activity on CR3 − cells compared to intact CyaA [45]. Therefore, we introduced the corresponding mutations into the cyaA allele on the chromosome of the B. pertussis Tohama I strain by marker-less allelic exchange [46].
It was important to duly characterize the activities of the here studied toxin variants produced by B. pertussis (Bp-CyaA, Figure 1a), since the previously characterized recombinant toxin variants (rEc-CyaA) were extracted from E. coli inclusion bodies with 8 M urea and purified by ion exchange chromatography. In contrast, in B. pertussis the toxin is secreted by the type 1 secretion system and in the absence of physiological calcium concentrations (2 mM) in the Stainer-Scholte (SS) medium, Bp-CyaA aggregates on the surface of the bacterial outer membrane and can thus be extracted with 4 M urea [21,47]. Due to low yields, Bp-CyaA is purified by affinity chromatography on calmodulin-Sepharose beads [48]. More importantly perhaps, in B. pertussis, proCyaA is posttranslationally activated by CyaC-catalyzed acylation of the K860 and K983 residues by the saturated palmitoyl (C16:0) chains [17,18]. In contrast, in E. coli, the rEc-CyaA is acylated by a mixture of saturated palmitoyl (C16:0) and unsaturated palmitoleyl (cis ∆9 C16:1) chains [45,[49][50][51]. Previous work then revealed that attachment of unsaturated C16:1 chains selectively decreases the pore-forming and hemolytic capacity of rEc-CyaA in respect to the exclusively C16:0 acylated Bp-CyaA, whereas the specific membrane penetration and cAMP-elevating capacities of rEc-CyaA and Bp-CyaA are the same [18,[49][50][51][52].  antibody recognizing the C-terminal RTX domain of CyaA. (b) Sheep erythrocytes (5 × 10 8 /mL) were incubated at 37 • C with 0.2 U/mL of the CyaA variants purified on calmodulin-Sepharose beads and after 30 min, aliquots were taken for determination of the cell-associated AC enzyme activity (binding) and of the AC enzyme activity internalized into erythrocytes (invasive AC). The activity of intact CyaA was taken as 100%. Each bar represents the mean value with the standard deviation (SD) of three (binding) or four (invasive AC) independent experiments (p-value > 0.05; ANOVA). (c) Hemolytic activity was measured upon incubation of sheep erythrocytes with 0.2 U/mL of the CyaA variants as the amount of released hemoglobin by photometric determination (A 541 ). Each point represents the mean value ± SD of at least three independent experiments. Significant differences between the mean values of hemolytic activities of intact CyaA and its mutant variants are shown (* p < 0.05; **** p < 0.0001; ANOVA). (d) Binding of the CyaA variants to J774A.1 cells (1 × 10 6 ) was determined as the amount of cell-associated AC enzyme activity upon incubation of cells with 0.2 U/mL of the CyaA variants for 30 min at 4 • C. cAMP intoxication was assessed by determining the intracellular concentration of cAMP generated in cells after 30 min of incubation of J774A.1 cells (1.5 × 10 5 ) with 32, 16, and 8 mU/mL of the CyaA variants. The activity of intact CyaA was taken as 100%. Each bar represents the mean value with SD of three independent experiments (p-value > 0.05; ANOVA). (e) Cell viability of CyaA-treated THP-1 cells (1.5 × 10 5 /well) was determined as the capacity of mitochondrial dehydrogenases to reduce the tetrazolium salt WST-1 to its formazan product after 2 h at 37 • C. Each point represents the mean value ± SD of at least three independent experiments (p-value > 0.05; ANOVA).
For comparison with previous work [44,45], the toxin activities of the CyaA variants were assessed using sheep erythrocytes as model CR3 − cells and mouse J774A.1 monocytes/macrophages as model CR3 + cells. As shown in Figure 1b, both CyaA mutant variants exhibited an intact capacity to bind erythrocytes and translocate the AC domain across the erythrocyte membrane in a trypsin protection assay. Compared to intact CyaA, the CyaA-NNQ mutant showed a strongly enhanced specific hemolytic activity, whereas the CyaA-V695P mutant permeabilized erythrocytes with a significantly lower activity than the intact toxin ( Figure 1c). Under the used conditions, 0.2 U/mL of intact CyaA lysed~50% of erythrocytes within 7 h, whereas 0.2 U/mL of the CyaA-V695P protein lysed only~20% of erythrocytes within the same incubation time, and 0.2 U/mL of the CyaA-NNQ toxin reproducibly caused~50% lysis of erythrocytes already in 4.5 h of incubation ( Figure 1c). When tested on CR3 + J774A.1 cells, the CyaA-NNQ and CyaA-V695P toxin variants exhibited quite similar binding and cAMP-elevating capacity as the intact CyaA ( Figure 1d). Hence, the intact CyaA and the CyaA-NNQ and CyaA-V695P toxin variants exhibited a very similar capacity to produce the cAMP-induced loss of viability of human CR3 + THP-1 monocytes (Figure 1e).
All these data show that the CyaA-NNQ and CyaA-V695P mutants secreted by B. pertussis bound to and delivered the AC domain into both CR3 − and CR3 + cells with the same efficacy as intact CyaA, but differ greatly in their specific hemolytic activities. The CyaA-NNQ mutant exhibited a strongly enhanced capacity to permeabilize the target cell membrane, whereas the pore-forming activity of the CyaA-V695P mutant was significantly reduced. Hence, the fact that in B. pertussis the CyaA proteins were exclusively acylated by the saturated C16:0 acyl residues did not alter any importantly the relative toxin activities found previously with the E. coli-produced rEc-CyaA toxin variants mostly acylated by a mixture of C16:0 and C16:1 acyl residues (Table 1). To assess the impact of the selectively enhanced or reduced hemolytic (pore-forming) capacity of the secreted CyaA variants on the virulence of B. pertussis strains that produce them, we first determined their 50% lethal dose (LD 50 ) values using the mouse intranasal challenge model. BALB/cByJ mice were intranasally inoculated with 50 µL of serially diluted bacterial suspensions of the parental B. pertussis strain (Bp-WT), or its mutant derivatives Bp-CyaA-NNQ and Bp-CyaA-V695P, respectively, and the survival of infected animals was monitored over 10 days. The experiment was repeated twice, and representative results obtained at two bacterial challenge doses are shown in Figure 2. were taken for determination of the cell-associated AC enzyme activity (binding) and of the AC enzyme activity internalized into erythrocytes (invasive AC). The activity of intact CyaA was taken as 100%. Each bar represents the mean value with the standard deviation (SD) of three (binding) or four (invasive AC) independent experiments (p-value > 0.05; ANOVA). (c) Hemolytic activity was measured upon incubation of sheep erythrocytes with 0.2 U/mL of the CyaA variants as the amount of released hemoglobin by photometric determination (A541). Each point represents the mean value ± SD of at least three independent experiments. Significant differences between the mean values of hemolytic activities of intact CyaA and its mutant variants are shown (*, p < 0.05; ****, p < 0.0001; ANOVA). (d) Binding of the CyaA variants to J774A.1 cells (1 × 10 6 ) was determined as the amount of cell-associated AC enzyme activity upon incubation of cells with 0.2 U/mL of the CyaA variants for 30 min at 4 °C. cAMP intoxication was assessed by determining the intracellular concentration of cAMP generated in cells after 30 min of incubation of J774A.1 cells (1.5 × 10 5 ) with 32, 16, and 8 mU/mL of the CyaA variants. The activity of intact CyaA was taken as 100%. Each bar represents the mean value with SD of three independent experiments (p-value > 0.05; ANOVA). (e) Cell viability of CyaA-treated THP-1 cells (1.5 × 10 5 /well) was determined as the capacity of mitochondrial dehydrogenases to reduce the tetrazolium salt WST-1 to its formazan product after 2 h at 37 °C. Each point represents the mean value ± SD of at least three independent experiments (p-value > 0.05; ANOVA).

Selective Increase or Down-Modulation of the Specific Pore-Forming Activity of CyaA Does Not Affect the Lung Colonization Capacity and Virulence of B. pertussis
To assess the impact of the selectively enhanced or reduced hemolytic (pore-forming) capacity of the secreted CyaA variants on the virulence of B. pertussis strains that produce them, we first determined their 50% lethal dose (LD50) values using the mouse intranasal challenge model. BALB/cByJ mice were intranasally inoculated with 50 µ L of serially diluted bacterial suspensions of the parental B. pertussis strain (Bp-WT), or its mutant derivatives Bp-CyaA-NNQ and Bp-CyaA-V695P, respectively, and the survival of infected animals was monitored over 10 days. The experiment was repeated twice, and representative results obtained at two bacterial challenge doses are shown in Figure 2.  At the challenge dose of 3 × 10 7 colony forming units (CFU), approximately 20% of the infected mice died within 3 (Bp-WT and Bp-CyaA-NNQ) or 4 (Bp-CyaA-V695P) days from intranasal infection (Figure 2a). An additional 10% of mice challenged with Bp-WT died on day 7. When an increased challenge dose of 6.5 × 10 7 CFU was administered, all infected mice died within 2 (Bp-WT) or 4 (Bp-CyaA-NNQ and Bp-CyaA-V695P) days ( Figure 2b). The calculated LD 50 values for Bp-WT (~3 × 10 7 CFU/mouse), Bp-CyaA-NNQ (~1.8 × 10 7 CFU/mouse), and Bp-CyaA-V695P (~2.6 × 10 7 CFU/mouse) were not statistically significantly different. Hence, the selective increase or decrease of the specific pore-forming activity of the secreted CyaA variants did not affect the overall virulence of B. pertussis in the mouse pneumonic infection model any importantly, and the mutants secreting the CyaA-NNQ and CyaA-V695P variants caused a comparably lethal course of infection in mice.
To corroborate this observation, the capacity of the mutants to proliferate in mouse lungs was compared to that of the parental B. pertussis strain using a sublethal infection dose. BALB/cByJ mice were intranasally inoculated with 50 µL of suspensions containing 1 × 10 5 CFU/mouse of Bp-WT or its mutant derivatives Bp-CyaA-NNQ and Bp-CyaA-V695P and lung colonization course was followed for 21 days. Mice were sacrificed by cervical dislocation at 2 h and 5, 8, 12, and 21 days after infection, the lungs were aseptically removed, homogenized in physiological salt solution, and serial dilutions of lung homogenates were plated on Bordet-Gengou agar plates for B. pertussis CFU determination after 72 h of incubation at 37 • C. As shown in Figure 3, the Bp-CyaA-NNQ and Bp-CyaA-V695P mutant strains colonized the lungs of mice as efficiently as the parental Bp-WT strain, exhibiting an increase in CFU counts in the lungs by almost two orders of magnitude at the peak of infection 5 days after challenge (~7 × 10 6 CFU/mouse). The CFU counts of all three B. pertussis strains then gradually decreased with a similar time course to a value of~1 × 10 3 -4 × 10 4 CFU/mouse on day 21. These results demonstrated that despite a substantial difference in the specific pore-forming activity of the secreted CyaA toxin, the CyaA-NNQ and CyaA-V695P variants supported the capacity of B. pertussis to multiply and persist in mouse lungs to the same extent as intact CyaA. × 10 7 CFU/mouse), and Bp-CyaA-V695P (~2.6 × 10 7 CFU/mouse) were not statistically significantly different. Hence, the selective increase or decrease of the specific pore-forming activity of the secreted CyaA variants did not affect the overall virulence of B. pertussis in the mouse pneumonic infection model any importantly, and the mutants secreting the CyaA-NNQ and CyaA-V695P variants caused a comparably lethal course of infection in mice.
To corroborate this observation, the capacity of the mutants to proliferate in mouse lungs was compared to that of the parental B. pertussis strain using a sublethal infection dose. BALB/cByJ mice were intranasally inoculated with 50 µ L of suspensions containing ~1 × 10 5 CFU/mouse of Bp-WT or its mutant derivatives Bp-CyaA-NNQ and Bp-CyaA-V695P and lung colonization course was followed for 21 days. Mice were sacrificed by cervical dislocation at 2 h and 5, 8, 12, and 21 days after infection, the lungs were aseptically removed, homogenized in physiological salt solution, and serial dilutions of lung homogenates were plated on Bordet-Gengou agar plates for B. pertussis CFU determination after 72 h of incubation at 37 °C. As shown in Figure 3, the Bp-CyaA-NNQ and Bp-CyaA-V695P mutant strains colonized the lungs of mice as efficiently as the parental Bp-WT strain, exhibiting an increase in CFU counts in the lungs by almost two orders of magnitude at the peak of infection 5 days after challenge (~7 × 10 6 CFU/mouse). The CFU counts of all three B. pertussis strains then gradually decreased with a similar time course to a value of ~1 × 10 3 -4 × 10 4 CFU/mouse on day 21. These results demonstrated that despite a substantial difference in the specific pore-forming activity of the secreted CyaA toxin, the CyaA-NNQ and CyaA-V695P variants supported the capacity of B. pertussis to multiply and persist in mouse lungs to the same extent as intact CyaA. The in vitro cytotoxicity of CyaA action on CR3 + phagocytic cells was found to be mainly due to the toxin-catalyzed conversion of cellular ATP to cAMP, which causes BimEL stabilization, Bax activation, and mitochondrial outer membrane permeabilization already at CyaA concentrations as low as 10 ng/mL [38]. This cytotoxic action of CyaA would prevail over the cytotoxic effects of the reduced or enhanced pore-forming activity of the CyaA variants that come in play at orders of magnitude higher CyaA concentrations above ~1 μg/mL [25]. To test this hypothesis, we constructed a set of B. pertussis mutants The in vitro cytotoxicity of CyaA action on CR3 + phagocytic cells was found to be mainly due to the toxin-catalyzed conversion of cellular ATP to cAMP, which causes BimEL stabilization, Bax activation, and mitochondrial outer membrane permeabilization already at CyaA concentrations as low as 10 ng/mL [38]. This cytotoxic action of CyaA would prevail over the cytotoxic effects of the reduced or enhanced pore-forming activity of the CyaA variants that come in play at orders of magnitude higher CyaA concentrations abovẽ 1 µg/mL [25]. To test this hypothesis, we constructed a set of B. pertussis mutants secreting the CyaA-AC − toxoids that were unable to convert ATP to cAMP due to a dipeptide insert between residues 188 and 189, which disrupts the catalytic activity of the AC enzyme [42]. This allowed us to assess the net contribution of the altered pore-forming activities to lung colonization capacities of the B. pertussis mutants. As shown in Figure 3, such B. pertussis mutants secreting the CyaA-AC − toxoids proliferated in the infected lungs by only one order of magnitude over the inoculation dose. This likely reflected their deficiency in CyaA-mediated disarming of immune cells infiltrating into the infected tissue, such as the inhibition of oxidative burst and of complement-mediated opsonophagocytic killing of bacteria by host neutrophils and macrophages, respectively [30,32,36,37,53]. Moreover, the course of lung infection by the Bp-CyaA-AC − -NNQ and Bp-CyaA-AC − -V695P mutant strains secreting the toxoid variants was very similar (Figure 3). Hence, no important impact on the capacity of B. pertussis to multiply and persist in mouse lungs was observed with the mutants secreting CyaA-AC − toxoids that exhibited a selectively and importantly enhanced or reduced specific pore-forming capacity. This implies that the pore-forming hemolysin capacity and host cell permeabilization resulting from CyaA toxin action play only an auxiliary role in the contribution of CyaA to the in vivo virulence of B. pertussis, which appears to depend exclusively on the capacity of CyaA to intoxicate host cells by elevated cAMP concentrations.
Intriguingly, the B. pertussis mutants secreting the CyaA-AC − toxoids exhibited steadỹ 10 6 CFU/lung levels on days 5, 8, and 12 post challenge and persisted at higher CFU levels for a longer time than the wild-type B. pertussis, reaching a level of~5 × 10 4 -3 × 10 5 CFU/lung on day 21 ( Figure 3). This surprising phenotype of a persistent infection by the Bp-CyaA-AC − bacteria was likely due to a lower level of harnessing of innate immune defense mechanisms by the lower numbers of the Bp-CyaA-AC − bacteria producing also lower levels of TLR ligands, making the infection more 'complaisant' and chronic. Moreover, the cAMP-intoxicating capacity of the CyaA toxin has previously been shown to elicit the production of interleukin 6 by tracheal epithelial cells, which would activate the cytotoxic action of neutrophils at the site of infection. Furthermore, CyaA-produced cAMP signaling was shown to induce cyclooxygenase 2 expression in macrophages, leading to release of prostaglandins that serve in chemoattraction of neutrophils [7,41,54,55]. It is thus plausible to assume that collateral activation of the innate defenses by the effects of CyaA-produced cAMP signaling would yield high neutrophil recruitment and rapid killing of invading bacteria in the infected lung tissue, once the peak of infection was reached on day 8, resulting in rapid decline of the CFU counts of B. pertussis strains producing the active CyaA toxin variants, as documented in Figure 3.

Selective Increase or Decrease of the Specific Pore-Forming Activity of CyaA Does Not Alter the Inflammatory Pathology of B. pertussis-Infected Lungs
Previously, we observed that infection with a B. pertussis strain secreting a CyaA-E570Q+K860R toxin variant capable to intoxicate by cAMP only CR3 + myeloid cells and not CR3 − cells caused a significantly lower inflammatory pathology in infected mouse lungs than wild-type B. pertussis, despite an equal capacity to proliferate in the lungs [42]. Therefore, we investigated the extent of lung pathology and inflammation provoked by infections with the B. pertussis strains that secreted CyaA toxin variants exhibiting a selectively enhanced or reduced pore-forming capacity, but capable to intoxicate by cAMP elevation both CR3 + and CR3 − cell types. BALB/cByJ mice were intranasally infected with~2 × 10 5 CFU of bacterial suspensions and sacrificed on day 7. Formalin-fixed lungs of infected animals were processed for histological examination, and six representative 2-µm thick sections of the left lobe of each sample set were cut every 150 µm and stained with hematoxylin and eosin (H&E). Pathological changes in the H&E stained lung sections were assessed (Figure 4a) and quantified using an inflammation score ranging from 0 to 4 ( Figure 4b) by an experienced pathologist in a blinded manner.  Using this scoring, the lungs of control mice inoculated with SS medium yielded an inflammation score of less than 0.5 ( Figure 4b) and exhibited an unaltered microscopic structure with continuous alveolar septa, intact bronchi and bronchioles, and well-preserved respiratory epithelium (Figure 4a). In contrast, the lungs of mice inoculated with Bp-WT and its mutant derivatives Bp-CyaA-NNQ and Bp-CyaA-V695P exhibited a high level of inflammation with conspicuous bronchopneumonic lesions that spread peribronchially with involvement of the peribronchial parenchyma (Figure 4a). The inflammatory cellular infiltrates consisted predominantly of polymorphonuclear granulocytes, substantially higher numbers of monocytes/macrophages, and swollen alveolar cells. The bronchial respiratory epithelium showed a characteristic focal to discrete hypertrophy and marked congestion in the swollen pulmonary regions. Quantification of the inflammation score revealed that the infection with the mutant strains Bp-CyaA-NNQ (score 3.3 ± 0.7) and Bp-CyaA-V695P (score 3.6 ± 0.6) elicited a comparable level of inflammatory pathology in mouse lungs as the parental Bp-WT infections (score 3.1 ± 0.8) (Figure 4b).
In contrast, as also shown in Figure 4a, the infections of mouse lungs with the Bp-CyaA-AC − , Bp-CyaA-AC − -NNQ, and Bp-CyaA-AC − -V695P mutant strains secreting enzymatically inactive toxoids elicited a distinctly milder lung pathology than the infections with the B. pertussis strains secreting the CyaA toxin variants that proliferated in the lungs to about an order of magnitude higher levels ( Figure 3). The alveoli in lungs infected with the toxoid-secreting strains remained well preserved, and only occasional focal alveolar pneumonic foci were observed. The peribronchial parenchyma was usually uniform and without signs of hypertrophy. Occasionally, a peribronchial inflammatory infiltrate was present, extending into the vicinity of the bronchi, but rarely affecting the adjacent alveoli. Mild lung congestion was also observed, but it was less pronounced than in the lungs infected with the B. pertussis strains secreting the enzymatically active CyaA toxin variants (Figure 4a). Quantification of the inflammation score of the infected lungs revealed that the Bp-CyaA-AC − -NNQ (score 2.4 ± 0.7) and Bp-CyaA-AC − -V695P (score 2.4 ± 0.8) mutant strains caused lower inflammatory damage of infected mouse lungs, alike the Bp-CyaA-AC − infection (score 2.1 ± 1.0) (Figure 4b).
To determine whether the up-or down-modulated specific pore-forming activity of CyaA produced by infecting bacteria affects macrophage recruitment into infected lungs, we assessed the infiltration of macrophages on lung sections by immunostaining for the F4/80 macrophage marker. As shown in Figure 5a, a strongly increased signal corresponding to macrophages was observed in the inflamed parenchyma of lungs infected with Bp-WT and its mutant derivatives Bp-CyaA-NNQ and Bp-CyaA-V695P, compared to the lungs of control mice inoculated with sterile SS medium. Nevertheless, quantification of F4/80-stained macrophages using the QuPath software revealed that the Bp-CyaA-NNQ and Bp-CyaA-V695P mutant infections elicited recruitment of similar numbers of macrophages into the inflamed lungs as the parental Bp-WT strain (Figure 5b). In contrast, the Bp-CyaA-AC − , Bp-CyaA-AC − -NNQ, and Bp-CyaA-AC − -V695P mutant strains, secreting toxoid forms of CyaA and proliferating to about an order of magnitude lower bacterial counts in the lungs, provoked only low level of macrophage infiltration into lung tissue (Figure 5a). Infections by all three toxoid-secreting strains then attracted comparable numbers of macrophages into the lungs (Figure 5b). toxoid forms of CyaA and proliferating to about an order of magnitude lower bacterial counts in the lungs, provoked only low level of macrophage infiltration into lung tissue (Figure 5a). Infections by all three toxoid-secreting strains then attracted comparable numbers of macrophages into the lungs (Figure 5b). (b) Quantification of F4/80-stained macrophages was performed by an automated evaluation for positive pixel count for DAB using the QuPath software on 6 consecutive lung sections for 3 infected animals per group (18 sections analyzed in total; ** p < 0.01; *** p < 0.001; **** p < 0.0001; ns, not significant; ANOVA).
In summary, all these results show that a substantial difference in pore-forming activity of the secreted CyaA toxin variants harboring the NNQ or V695P substitutions had no clear impact on the pathology and inflammation elicited by B. pertussis infection of mouse lungs. The inflammatory damage was driven by the level of bacterial proliferation in the lungs and depended on the enzymatic capacity of CyaA to intoxicate cells by conversion of ATP to cAMP and not on the cell-permeabilizing activity of the toxin.

Bacterial Strains and Growth Conditions
The Escherichia coli strain XL1-Blue was used throughout this work for plasmid construction and the E. coli strain SM10 λ pir was used for plasmid transfer into B. pertussis by bacterial conjugation. E. coli strains were cultivated at 37 • C on Luria-Bertani (LB) agar medium or in LB broth. When appropriate, LB culture media were supplemented with ampicillin (pSS4245 plasmid-carrying transformants, 100 µg/mL). The B. pertussis Tohama I strain was obtained from the Institute Pasteur Collection of Cultures (Paris, France) under the catalogue No. CIP 81.32. The parental B. pertussis and derived mutant strains were grown on Bordet-Gengou (BG) agar plates (Difco, Franklin Lakes, NJ, USA) supplemented with 1% glycerol, 15% defibrinated sheep blood (LabMediaServis, Jaromer, Czech Republic) at 37 • C in a 5% CO 2 atmosphere for 72 h to visualize hemolysis. Liquid cultures were obtained by growing bacteria overnight in modified Stainer-Scholte (SS) medium [56] supplemented with 3 g/L of Casamino Acids (Difco, Franklin Lakes, NJ, USA) and 1 g/L of heptakis (2.6-di-O-dimethyl) β-cyclodextrin (Sigma-Aldrich, St. Louis, MO, USA) to a mid-exponential phase (B. pertussis OD 600 = 1.0) at 37 • C.

Mutagenesis of the cyaA Gene on B. pertussis Chromosome
Construction of the alleles for production of the CyaA-D445N+D446N+E448Q (CyaA-NNQ) and CyaA-V695P mutant variants, as well as their CyaA-AC − variants (188-GlySer-189) in E. coli was described in detail previously [44,45]. The respective mutated cyaA gene segments were cloned into the exchange vector plasmid pSS4245 (oriV, AmpR, StrR, KmR, BleR, TetR, and an I-SceI cleavage site for counterselection) kindly provided by Dr. Scott Stibitz (U.S. CBER, FDA, MD, USA) and introduced by allelic exchange into the B. pertussis chromosome as described previously [42]. For each construct, several individual colonies were characterized for phenotypic change and the presence of introduced mutations and the absence of undesired mutations were verified by sequencing of relevant portions of PCR-amplified segments of the cyaA gene.

Assay of AC Activity
AC enzymatic activity was determined in the presence of 1 µM calmodulin, as previously described [58]. One unit (U) of adenylate cyclase enzyme (toxin) activity corresponds to 1 µmol of cAMP formed in 1 min at pH 8.0 at 30 • C.

Cell Binding, Cell-Invasive and Hemolytic Activities in Sheep Erythrocytes
Sheep erythrocytes (LabMediaServis, Jaromer, Czech Republic) were washed in TNC buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 2 mM CaCl 2 ), adjusted to 5 × 10 8 cells/mL, and incubated with the purified CyaA variants at 37 • C in TNC buffer. After 30 min, cells were washed in TNC buffer to remove unbound CyaA and divided into two aliquots. One aliquot was after extensive washing in cold TNE buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA) directly used to determine the amount of cell-associated AC activity (binding, membrane-bound CyaA). The other aliquot was treated with 20 µg/mL of trypsin for 15 min at 37 • C to inactivate the extracellular AC enzyme that did not translocate across the cellular membrane. Soybean trypsin inhibitor (40 µg/mL) was then added to the mixture to stop the reaction before the samples were washed twice in cold TNE buffer and used to determine the amount of cell-invasive AC enzyme activity. The activity of intact CyaA was taken as 100%. Hemolytic activity was measured by determining the hemoglobin release (A 541 ) upon incubation of the CyaA variants with 5 × 10 8 /mL washed erythrocytes in TNC buffer. were lysed with 0.1% Triton X-100 and the membrane-associated AC enzyme activity was determined as described above.

cAMP Determination
Mouse J774A.1 cells (1 × 10 5 /well) were incubated at 37 • C with the CyaA variants for 30 min in D-MEM. The reaction was stopped by the addition of 0.2% Tween-20 in 100 mM HCl, samples were boiled for 15 min at 100 • C, neutralized by the addition of 150 mM unbuffered imidazole, and cAMP was measured by a competitive immunoassay as previously described [59]. The activity of intact CyaA was taken as 100%.

Analysis of the Acylation Pattern of CyaA Variants
The acylation pattern of the purified CyaC-acylated CyaA variants was determined by liquid chromatography-mass spectrometry (LC-MS), as previously described [52,61].

Ethics Statement
All animal experiments were approved (permission No. 41/2019) by the Animal Welfare Committee of the Institute of Molecular Genetics of the Czech Academy of Sciences, v. v. i., Prague, Czech Republic. Handling of animals was performed according to the Guidelines for the Care and Use of Laboratory Animals, the Act of the Czech National Assembly, Collection of Laws no. 246/1992.

Animal Infection Experiments
Five-week-old female BALB/cByJ mice (Charles River, France) were used in this study. Mice were anesthetized by intraperitoneal (i.p.) injection of ketamine (80 mg/kg) and xylazine (8 mg/kg) in saline and were inoculated intranasally with 1 × 10 5 CFU of B. pertussis cell suspensions delivered in 50 µL volume. To determine viable CFU, the suspensions were diluted in PBS and plated on BG agar plates.
To determine B. pertussis lung colonization, infected BALB/cByJ mice (3 to 6 mice per time point) were sacrificed by cervical dislocation 2 h after exposure to challenge suspension (day 0) and on the indicated days thereafter (day 5, 8, 12, and 21). Lungs were aseptically removed and homogenized in physiological solution with tissue grinders (Merck, Darmstadt, Germany). Dilutions of lung homogenates were plated on BG agar plates and CFU were counted after 72 h of incubation at 37 • C.
For determination of the LD 50 values, groups of 6 or 10 mice were challenged intranasally with serially diluted bacterial suspensions 1.0 × 10 7 , 3.0 × 10 7 , or 6.5 × 10 7 CFU/mouse, and their survival was monitored over 10 days. The results are expressed as the percentage of surviving mice and are representative of two independent experiments. The LD 50 values were calculated by the probit analysis method of Finney using the freeware LD 50 /LC 50 calculator downloaded at https://goo.gl/9QYcNk, as described at https: //probitanalysis.wordpress.com/2016/07/07/first-blog-post/, (accessed on 12 November 2019).

Histological Studies
Lung morphology was examined 7 days upon the intranasal challenge of mice with a dose of 2 × 10 5 CFU of B. pertussis strains. Control mice were treated with 50 µL of SS medium only. The animals were anesthetized as described above and sacrificed by cervical dislocation. Lungs were removed, fixed in 4% phosphate-buffered formaldehyde for 48 h at 4 • C, and transferred into 70% ethanol for 24 h at 4 • C. The tissue samples were processed in a Leica ASP6025 Tissue Processor (Leica Biosystems, Nussloch, Germany) with an automated protocol and embedded in paraffin blocks. Six representative 2-µm thick sections of the left lobe were cut every 150 µm from each sample using a Leica RM 2255 Microtome (Leica Biosystems, Nussloch, Germany) and placed onto SuperFrost+ slides (Epredia, Kalamazoo, MI, USA).
Hematoxylin and eosin (H&E) staining was performed using an automated Leica ST5020 Staining Machine (Leica Biosystems, Nussloch, Germany) and coverslipping was carried out using a Leica CV5030 Coverslipper (Leica Biosystems, Nussloch, Germany). Quantification of H&E stained slides was performed manually by inflammation score 0-4 (0-no inflammation; 1-1-25% inflammation; 2-26-50% inflammation; 3-51-75% inflammation; 4-76-100% inflammation). It was calculated as the ratio of the total inflammatory lesion area to the total area of each lung section and values calculated for six consecutive lung sections for 3 infected mice per group were averaged and expressed as means with SD (18 sections analyzed in total).
For immunohistochemistry staining for the macrophage marker F4/80, the tissue sections were dried for 1 h at 65 • C, deparaffinized using an automated LEICA ST5020 machine (Leica Biosystems, Nussloch, Germany), and rehydrated. Antigen retrieval was performed with citrate buffer HIER, pH 6.0 (Zytomed, Berlin, Germany) in a pressure cooker for 15 min at 110 • C. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide in methanol (v/v) for 15 min at room temperature (RT) and 2% BSA (Carl Roth, Karlsruhe, Germany) was used for 20 min at RT to prevent unspecific binding. The tissue sections were then stained with the rabbit anti-F4/80 mAb (Cell Signaling Technology, Danvers, MA, USA) at a 1:800 dilution in Antibody Diluent (Zytomed, Berlin, Germany) overnight at 4 • C, followed by incubation with anti-rabbit HRP-conjugated polymer for 30 min at RT. The immune reaction was developed using 3,3'-diaminobenzidine (DAB) solution (Dako, Glostrup, Denmark) for 10 min at RT in the dark, and counterstaining was performed with hematoxylin H (Biognost, Zagreb, Croatia) for 2 min at RT. Slides were mounted with Aquatex mounting medium (Merck-Millipore, Darmstadt, Germany) and visualized using an AxioScan.Z1 automated slide scanner (Carl Zeiss, Göttingen, Germany) or a light microscope Zeiss Scope A1 (Carl Zeiss, Göttingen, Germany). Quantification of F4/80-stained macrophages was performed by an automated evaluation for positive pixel count for DAB on six consecutive lung sections for 3 infected mice per group (18 sections analyzed in total) using the QuPath Software, v0.2.3 (University of Edinburgh, Edinburgh, UK).

Statistical Analysis
The results are expressed as the arithmetic means ± SD of the mean. Statistical analysis was performed by one-way ANOVA with Tukey multiple comparisons test, by two-way ANOVA with Sidak's multiple comparisons test, or Log-rank (Mantel-Cox) test using GraphPad Prism 9.1.0 (GraphPad Software, La Jolla, CA, USA).

Conclusions
We previously constructed a B. pertussis strain secreting the CyaA-E570Q+K860R mutant toxin that was almost completely inactive on non-myeloid CR3 − cells, such as erythrocytes or airway epithelial cells [43]. However, this toxin variant retained a full capacity to elevate cAMP levels in the cytosol of CR3 + cells, despite a selective reduction of its pore-forming activity on CR3 + cells by about an order of magnitude, as compared to intact CyaA [42,43]. In the mouse pneumonic infection model, the Bp-CyaA-E570Q+K860R mutant strain exhibited an approximately tenfold reduction in virulence, as judged by an increase of LD 50 from~4.1 × 10 7 CFU/mouse of the parental Bp-WT strain to LD 50 of 3.5 × 10 8 CFU/mouse, with infected mice dying over a week later than upon infection with the WT strain [42]. This indicated that the pore-forming activity of CyaA and/or its capacity to elevate cAMP in CR3 − cells was required for full virulence of B. pertussis [42]. However, the used mutant did not allow to exclude that it was rather the contribution of the cAMP elevation in CR3 − cells that was missing for full virulence, than the pore-forming activity. Here, we found that B. pertussis mutants secreting CyaA capable to intoxicate both CR3 + and CR3 − cells and exhibiting either a selectively and strongly enhanced or a selectively reduced hemolytic (pore-forming) activity, exhibited equal virulence and mouse lung colonization capacity as the parental Bp-WT strain. Since the modulation of the pore-forming activity did not affect the lethality of B. pertussis infection, it appears that the full virulence of B. pertussis depends predominantly on the capacity of CyaA to elevate cAMP in both myeloid CR3 + and non-myeloid CR3 − cells, such as airway epithelial cells colonized by B. pertussis bacteria.