Retargeting from the CR3 to the LFA-1 receptor uncovers the adenylyl cyclase enzyme–translocating segment of Bordetella adenylate cyclase toxin

Abstract

The Bordetella adenylate cyclase toxin-hemolysin (CyaA) and the α-hemolysin (HlyA) of Escherichia coli belong to the family of cytolytic pore-forming Repeats in ToXin (RTX) cytotoxins. HlyA preferentially binds the α_Lβ₂ integrin LFA-1 (CD11a/CD18) of leukocytes and can promiscuously bind and also permeabilize many other cells. CyaA bears an N-terminal adenylyl cyclase (AC) domain linked to a pore-forming RTX cytolysin (Hly) moiety, binds the complement receptor 3 (CR3, α_Mβ₂, CD11b/CD18, or Mac-1) of myeloid phagocytes, penetrates their plasma membrane, and delivers the AC enzyme into the cytosol. We constructed a set of CyaA/HlyA chimeras and show that the CyaC-acylated segment and the CR3-binding RTX domain of CyaA can be functionally replaced by the HlyC-acylated segment and the much shorter RTX domain of HlyA. Instead of binding CR3, a CyaA_1-710/HlyA_411-1024 chimera bound the LFA-1 receptor and effectively delivered AC into Jurkat T cells. At high chimera concentrations (25 nm), the interaction with LFA-1 was not required for CyaA_1-710/HlyA_411-1024 binding to CHO cells. However, interaction with the LFA-1 receptor strongly enhanced the specific capacity of the bound CyaA_1-710/HlyA_411-1024 chimera to penetrate cells and deliver the AC enzyme into their cytosol. Hence, interaction of the acylated segment and/or the RTX domain of HlyA with LFA-1 promoted a productive membrane interaction of the chimera. These results help delimit residues 400–710 of CyaA as an “AC translocon” sufficient for translocation of the AC polypeptide across the plasma membrane of target cells.

The pore-forming Repeats in ToXin (RTX) cytotoxins are secreted by a broad range of pathogenic Gram-negative bacteria (1). The RTX adenylate cyclase toxin-hemolysin (CyaA, ACT, or Hly-AC) is secreted by Bordetellae pathogenic to mammals and acts as a “swift saboteur” of host immune cell functions (2–4). CyaA is a 1706-residue–long polypeptide (Fig. 1) that inserts directly into the plasma membrane of target cells, and without the need for endocytosis, it delivers directly into the cell cytosol an N-terminal adenylyl cyclase (AC) enzyme domain of ∼400 residues (5). Within eukaryotic cells, the AC enzyme is activated by binding of cytosolic calmodulin and rapidly catalyzes unregulated conversion of intracellular ATP to the key second messenger signaling molecule cAMP (6). This near-instantaneously annihilates the bactericidal capacities of host phagocytes (2). Due to an extremely high catalytic power of the activated AC enzyme, the CyaA toxin was found to detectably bind and promiscuously elevate cytosolic cAMP concentrations in a broad array of eukaryotic cells, most likely due to a low affinity binding to glycan moieties of gangliosides and N-linked glycans of surface glycoproteins (7–9). However, the specific efficacy of cell binding, penetration, and cAMP intoxication of target cells by the CyaA toxin is enhanced by about 2 orders of magnitude through an interaction of CyaA with the CD11b subunit of the α_Mβ₂ integrin receptor (CD11b/CD18), known as the complement receptor 3 (CR3), or Mac-1 of myeloid phagocytes (10). CyaA specifically binds the CD11b subunit of CR3 and does not recognize the other two highly homologous integrin heterodimers of the β₂ integrin family, namely CD11a/CD18 (LFA-1 or α_Lβ₂) and CD11c/CD18 (CR4, α_xβ₂ or p150,95), which share the CD18 subunit with CD11b/CD18 (10–12). The RTX hemolysin (Hly) moiety of CyaA (∼1300 carboxyl-proximal residues) is functionally independent and is itself capable to penetrate cell membranes to form small oligomeric cation-selective membrane pores in the absence of the AC domain (13–15). Hly insertion permeabilizes cell membranes for potassium efflux (16, 17) and can provoke colloid-osmotic (oncotic) lysis of cells (18–21).

Figure 1.

Schematic representation of CyaA, HlyA, and hybrid CyaA/HlyA molecules. Individual domains of CyaA and HlyA are indicated by the colored rectangles. AC, adenylate cyclase domain; LS, AC-to-Hly linker segment; PF, pore-forming domain; AS, acylated segment; RTX, calcium-binding repeats; SS, secretion signal. The numbers that follow the CyaA or HlyA in the names of the CyaA/HlyA hybrid chimera represent the number of the first and last residue of the segment of the given protein according to the sequences of full-length CyaA and HlyA, respectively. Design of the hybrid molecules: (i) in CyaA_1-501/HlyA_142-1024, the AC domain and the adjacent AC-to-Hly linker segment of CyaA (residues 1-501 of CyaA) are fused to a sequence beginning with the first putative transmembrane α-helix of HlyA (residues 142-1024 of HlyA); (ii) in CyaA_1-528/HlyA_238-1024, the AC domain, the adjacent AC-to-Hly linker segment, and the first α-helix of CyaA (residues 1-528 of CyaA) are fused to a sequence beginning with the second putative α-helix of HlyA (residues 238-1024 of HlyA); (iii) in CyaA_1-606/HlyA_368-1024, the AC domain, the adjacent AC-to-Hly linker segment, and the first three α-helices of CyaA (residues 1-606 of CyaA) are fused to a sequence beginning with the fourth putative α-helix of HlyA (residues 368-1024 of HlyA); (iv) in CyaA_1-710/HlyA_411-1024, the AC domain, the adjacent AC-to-Hly linker segment and the hydrophobic domain of CyaA (1-710 of CyaA) are fused to a sequence beginning with the acylated segment of HlyA (residues 411-1024 of HlyA); (v) in CyaA_1-800/HlyA_501-1024, the AC domain, the adjacent AC-to-Hly linker segment, the hydrophobic domain, and part of the acylated segment of CyaA (residues 1- 800 of CyaA) are fused to a sequence beginning with the truncated acylated segment of HlyA (residues 501-1024 of HlyA).

The key biological role of the Hly moiety of CyaA would consist in the delivery of the AC domain into cytosol of target cells. The Hly moiety harbors all structural information involved in the translocation of the “passive passenger” AC domain polypeptide across the lipid bilayer of the cell membrane. Indeed, the Hly can deliver instead of the AC domain also large heterologous or artificial polypeptides, providing that these bear a net overall positive charge and do not adopt a stably folded tertiary structure (22–25). Furthermore, the pore-forming and the AC translocating activities associated with the Hly polypeptide appear to be functionally fully independent and occur in parallel, being most likely associated with alternative conformers of the same Hly moiety. In fact, the oligomeric pores formed by Hly appear as too small (0.6–0.8 nm in diameter) for passage of even an unfolded polypeptide chain (13). Moreover, specific nonhemolytic CyaA mutant variants, unable to form membrane pores and permeabilize cellular membrane, do still efficiently deliver the AC polypeptide across the target membrane into the cytosol of cells (26). Hence, it appears that for AC polypeptide translocation across the lipid bilayer of the cell membrane it is the membrane-penetrating capacity of the hydrophobic (pore-forming) domain of the Hly moiety that is required and not its capacity to form membrane pores.

The Hly portion involved in AC domain translocation consists of several subdomains of CyaA (Fig. 1). Its most N-terminal portion forms an “AC-to-Hly linking segment” (residues ∼400 to 500) that participates in membrane penetration of the toxin (27–29). This segment precedes a hydrophobic pore-forming domain (residues 500 to 700) that consists of five predicted hydrophobic and amphipathic transmembrane α-helices (30–36). This is further followed by an acylated segment (residues 700 and 1000), where the protoxin is activated through covalent acylation of the ε-amino groups of internal lysine residues Lys-860 and Lys-983 by the dedicated CyaC acyltransferase (32, 37–39). Adjacent is the typical calcium-binding nonapeptide repeat domain (RTX) that harbors ∼40 calcium-binding sites and upon Ca²⁺ loading folds into five β-roll blocks (40–42).

In the absence of structural data, the transmembrane topology and organization of the CyaA translocon that delivers the AC polypeptide across the cell membrane remains elusive. Translocation of the AC enzyme across the cell membrane is a rapid process with a half-time of dozen of seconds (43). It occurs from cholesterol-enriched membrane lipid microdomains (lipid rafts), into which the membrane inserted CyaA laterally separates together with its receptor CD11b/CD18, once the membrane-inserted toxin has opened a transient conduit for influx of extracellular calcium ions into cells (44, 45). Mutational studies have identified a set of residues within the hydrophobic domain of Hly that appear to play a functional or structural role in the AC translocation process (20, 27–30, 32, 33, 36, 46). Recently, it was claimed that CyaA harbors an intrinsic phospholipase A activity (PLA) that would be involved in AC domain translocation (47). However, the association of PLA activity with CyaA has been disproved by two independent studies (48–50).

Among the other best-characterized pore-forming RTX toxins is the α-hemolysin (HlyA), secreted by uropathogenic and some commensal fecal isolates of Escherichia coli (51–53). The N-terminal hydrophobic domain of HlyA is predicted to contain amphipathic α-helices and it is assumed that this region mediates irreversible anchoring of HlyA to the plasma membrane (54, 55). The post-translational activation of HlyA involves the covalent acylation of internal Lys-564 and Lys-690 residues by the acyltransferase HlyC (56–58). The RTX domain of HlyA (residues ∼724 to 852) comprises 11 to 17 calcium-binding repeats and is much shorter than the RTX domain of CyaA. HlyA was proposed to bind two different protein receptors, the glycophorin on the membrane of erythrocytes (59) and the α_Lβ₂ integrin LFA-1 expressed on leukocytes (60, 61). However, none of these two receptors is expressed on some other cells that are effectively permeabilized by HlyA, such as the epithelial cells, indicating that HlyA, like some other RTX toxins, is to some extent competent for a promiscuous and protein receptor-independent association with cell membrane (62).

Given the unique AC enzyme domain of CyaA, it remained unclear if its C-terminal Hly moiety has evolved to specifically participate in the translocation of the N-terminal AC domain into target cells, or whether a quite different RTX cytolysin would also support the delivery of the AC enzyme domain into cells. Therefore, we constructed a set of CyaA/HlyA hybrid molecules acylated by either CyaC or HlyC acyltransferases and show that upon swapping the acylated and RTX domains of CyaA for that of HlyA, the AC toxin can be retargeted from its CR3 receptor to bind LFA-1. The AC-delivering capacity of the acylated CyaA_1-710/Hly_411-1024 chimera then defines the “AC translocon” as confined within the residues 400 to 710 of CyaA.

Results

Selection of different fatty acyl chains and acylation sites by the CyaC and HlyC acyltransferases differentially determines the activity of the CyaA/HlyA chimeras

To test the capacity of HlyA to support delivery of the AC domain of CyaA into cells, we have generated a set of CyaA/HlyA molecular hybrids (Fig. 1). Each of the five CyaA/HlyA chimeras was produced in E. coli cells in the presence of either the CyaA-activating CyaC acyltransferase, or the HlyA-activating HlyC enzyme. The 10 acylated toxin hybrids were then purified by affinity chromatography on calmodulin-Sepharose, taking advantage of the affinity of the N-terminal AC domain of CyaA for calmodulin (Fig. S1).

Because activation by posttranslational fatty acylation is crucial for activities of RTX toxins on cells (1), we first characterized the acylation state of the individual hybrid toxins by analyzing their tryptic digests by liquid chromatography coupled to ultra high-resolution Fourier transform ion cyclotron resonance MS (LC FT-ICR MS). As shown in Table 1, the CyaC acyltransferase modified only the Lys-690 residue within the HlyA segment of the CyaA/HlyA hybrids. The Lys-564 residue of HlyA was not recognized and remained largely unacylated by CyaC. In contrast, the cognate HlyC acyltransferase modified both the Lys-564 and Lys-690 lysine residues of HlyA in all tested CyaA/HlyA chimeras that contained the entire acylated segment of HlyA (Fig. 1 and Table 1).

Table 1.

Acylation status of the CyaA, HlyA and hybrid proteins

Proteina	CyaC+			HlyC+
	Modification	Lys-860b	Lys-983b	Modification	Lys-860b	Lys-983b
CyaA	Nonmodified	31	1	Nonmodified	99	20
	C16:0	32	45	C14:0	0	70
	C16:1	34	43	C14:0-OH	1	7
	C18:1	3	8
	Modification	Lys-564b	Lys-690b	Modification	Lys-564b	Lys-690b
HlyA	Nonmodified	93	0	Nonmodified	10	0
	C16:0	1	23	C14:0	13	58
	C16:1	6	67	C14:0-OH	71	35
	C18:1	0	10
CyaA1-501/HlyA142-1024	Nonmodified	97	0	Nonmodified	2	0
	C16:0	1	44	C14:0	57	84
	C16:1	2	33	C14:0-OH	35	9
	C18:1	0	23
CyaA1-528/HlyA238-1024	Nonmodified	98	0	Nonmodified	2	0
	C16:0	1	39	C14:0	67	83
	C16:1	1	32	C14:0-OH	26	9
	C18:1	0	27
CyaA1-606/HlyA368-1024	Nonmodified	96	0	Nonmodified	1	0
	C16:0	1	40	C14:0	64	84
	C16:1	2	35	C14:0-OH	30	8
	C18:1	1	25
CyaA1-710/HlyA411-1024	Nonmodified	99	0	Nonmodified	0	0
	C16:0	0	40	C14:0	72	85
	C16:1	1	32	C14:0-OH	23	8
	C18:1	0	28
CyaA1-800/HlyA501-1024	Nonmodified	99	0	Nonmodified	0	0
	C16:0	0	39	C14:0	67	82
	C16:1	1	42	C14:0-OH	27	8
	C18:1	0	17

^aProteins were produced in the E. coli strain BL21/pMM100 and purified close to homogeneity.

^bPercentage distribution of fatty acid modification of the ε-amino groups of the Lys-860, Lys-983, Lys-564, and Lys-690 residues. Average values are derived from determinations performed with two different toxin preparations (one with CyaC-modified CyaA_1-710/HlyA_411-1024). The numbering of the acylated lysine residues is according to the sequence of the full-length CyaA or HlyA proteins. The remaining lysine residues to 100% are acylated by C12:0, C12:0-OH, C14:0, C14:1; C16:0, C16:1, and/or C16:1-OH.

Intriguingly, as documented in Table 1, the HlyC acyltransferase selected from the E. coli acyl-ACP pool near-exclusively the 14-carbon myristoyl (C14:0) and hydroxymyristoyl (C14:0-OH) chains for acylation of the Lys-564 and Lys-690 residues of HlyA. In contrast, the CyaC acyltransferase selected predominantly the 16-carbon palmitoyl (C16:0) and palmitoleyl (C16:1) acyl chains, with some octadecenoyl (C18:1) acyls also being used for modification of the Lys-690 residue of HlyA (Table 1). Hence, the choice of the acyltransferase determined the site at which the CyaA/HlyA hybrids were acylated and also the nature (length) of the acyl chains (14-carbon versus 16/18-carbon) attached to the modified lysine residues.

To characterize how this selection of the acylated site and length of the used acyl chain impacted on the activities of the hybrid toxins, we first determined the capacity of the CyaA/HlyA hybrids to bind and permeabilize sheep erythrocytes. The red blood cells were used as a convenient model of target cells that express neither the CR3 receptor of CyaA, nor the LFA-1 receptor of HlyA. Further advantage was taken of the highly active N-terminal AC enzyme domain contained in the hybrid proteins, which allowed to quantify the specific capacity of the CyaA/Hly hybrids to tightly associate with the erythrocyte membrane (binding) and enabled to follow the delivery of the AC domain into the cytosol of erythrocytes (invasive AC). In parallel, we measured the specific capacity of the CyaA/HlyA hybrids to permeabilize erythrocytes by hemolytic pores and release hemoglobin through provoking oncotic erythrocyte lysis over time. As shown in Fig. 2A, the CyaC-activated hybrid molecules (monoacylated by predominantly the 16-carbon acyls on Lys-690 of HlyA) bound erythrocytes with an efficacy ranging from ∼50 to ∼75% of intact doubly acylated CyaA (100%). The HlyC-activated CyaA/HlyA hybrids, doubly acylated by 14-carbon acyls on the Lys-564 and Lys-690 residues of the HlyA moiety, bound erythrocytes more efficiently than their CyaC-monoacylated counterparts. This was particularly apparent when a portion of the pore-forming domain of HlyA was preserved in the CyaA/HlyA hybrid, such as in the CyaA_1-501/HlyA_142-1024 and CyaA_1-528/HlyA_238-1024 constructs, which bound to cells in several times higher amounts than the intact CyaC-activated CyaA (Fig. 2A). However, compared with intact CyaA, which under the conditions used provoked lysis of ∼50% red blood cells in 4 h, the hybrids containing the pore-forming domain of CyaA, such as the CyaA_1-606/HlyA_368-1024, CyaA_1-710/HlyA_411-1024, and CyaA_1-800/HlyA_501-1024 constructs, did not exhibit any hemolytic activity. This was true irrespective of whether these proteins were monoacylated by CyaC (Fig. 2B), or doubly acylated by HlyC (Fig. 2C). Only the CyaA_1-501/HlyA_142-1024 chimera, with a large portion of the pore-forming domain of HlyA preserved, exhibited a rapidly manifesting HlyA-like hemolytic activity even when monoacylated by CyaC (Fig. 2B). Hence, the 16-carbon–monoacylated HlyA moiety was also able to form pores in erythrocyte membrane. Furthermore, its hemolytic activity, as well as the hemolytic activity of the CyaA_1-528/HlyA_238-1024 hybrid, was much higher when these hybrids were acylated by the HlyC-attached 14-carbon acyls on the Lys-564 and Lys-690 residues of the HlyA moiety (Fig. 2C). These results show that replacement of up to 238 N-terminal residues of HlyA by the first 528 residues of CyaA (comprising the AC domain, the AC-to-Hly linking segment, and the first amphipathic α-helix of the pore-forming domain of CyaA) still preserved the hemolysin function of the HlyC-activated HlyA moiety. In contrast, the other HlyC- or CyaC-acylated CyaA/HlyA hybrids were unable to properly insert into erythrocyte membrane and form hemolytic pores.

Figure 2.

16-Carbon mono- and 14-carbon doubly-acylated (CyaC- and HlyC-activated) hybrid molecules bind and lyse erythrocytes with importantly differing efficacies. A, sheep erythrocytes (5 × 10⁸/ml) were incubated in the presence of 75 mm sucrose as osmoprotectant with 5 nm purified proteins at 37 °C and after 30 min, aliquots were taken for determinations of the cell-associated AC activity (Binding). Activities are expressed as percentages of the activity of the intact 16-carbon doubly-acylated, CyaC-activated CyaA and represent average mean ± S.D. from at least three independent determinations performed in duplicate with two different toxin preparations. Sheep erythrocytes (5 × 10⁸/ml) were incubated at 37 °C in the presence of CyaC-activated (B) or HlyC-activated (C) proteins (25 nm). Hemolytic activity was measured as the amount of released hemoglobin by photometric determination (A541) (n = 3).

The AC-to-Hly linking segment and the intact hydrophobic domain of CyaA are necessary and sufficient for AC domain translocation into cells by a mechanism that does not provoke permeabilization of the membrane by CyaA pores

Importantly, as shown in Fig. 3A, whereas single acylation by CyaC was insufficient, the double modification of the Lys-564 and Lys-690 residues of the HlyA moiety with the 14-carbon acyls conferred on the nonhemolytic CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 hybrids the capacity to translocate the AC domain into erythrocytes. These two chimeras bound erythrocytes comparably well as the CyaC-acylated CyaA (Fig. 2A). Their specific capacity to penetrate the erythrocyte membrane and translocate the N-terminal AC enzyme domain into erythrocyte cytosol corresponded to ∼35 to ∼40% of the specific capacity of CyaA (Fig. 3A). The doubly acylated CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 hybrid proteins not only inserted the AC enzyme across the erythrocyte membrane, but the AC also became protected from degradation by externally added trypsin. The translocated AC enzyme was truly delivered into erythrocyte cytosol containing ATP and calmodulin and it catalyzed an increase of cellular cAMP concentrations to roughly ∼40% of the levels produced by the CyaC-activated CyaA (Fig. 3B). Moreover, as for the intact CyaA, the erythrocyte binding and cell-invasive activity of the CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 chimeras depended fully on the presence of free (2 mm) calcium ions (Fig. 3C).

Figure 3.

Nonhemolytic HlyC-activated chimeras CyaA_1-710/HlyA_411-1024, and CyaA_1-800/HlyA_501-1024 translocate the AC domain across the plasma membrane of sheep erythrocytes in a calcium-dependent manner. A, sheep erythrocytes (5 × 10⁸/ml) were incubated in the presence of 75 mm sucrose as osmoprotectant with 5 nm purified proteins at 37 °C and after 30 min, aliquots were taken for determinations of the AC activity internalized into erythrocytes and protected against digestion by externally added trypsin (Invasive AC). Activities are expressed as the activity of the intact, 16-carbon doubly-acylated, CyaC-activated CyaA and represent average mean ± S.D. from at least three independent determinations performed in duplicate with two different toxin preparations. B, cAMP intoxication was assessed by determining the intracellular concentration of cAMP generated in sheep erythrocytes (5 × 10⁸/ml) after 30 min of incubation of cells with 25 nm CyaA or the hybrid proteins. Activities are expressed as percentages of intact CyaC-activated CyaA activity and represent average mean ± S.D. from four independent determinations performed in duplicate. C, sheep erythrocytes (5 × 10⁸/ml) were incubated in the presence of 2 mm calcium (+Ca²⁺) or in absence of calcium and presence of 5 mm EDTA (+EDTA) at 37 °C with 5 nm proteins. After 30 min, aliquots were taken for determinations of the cell-associated AC activity (Binding) and AC activity internalized into erythrocytes and protected against digestion by externally added trypsin (Invasive AC). Activities are expressed as percentages of intact CyaC-activated CyaA activity (n = 3-7).

Since the AC-translocating CyaA_1-710/HlyA_411-1024 hybrid was not hemolytic on erythrocytes, we examined whether the hybrid proteins exhibited a capacity to permeabilize naked planar asolectin lipid bilayers with applied negative voltage. Whereas 250 pm CyaA elicited a steep increase of conductance across the planar lipid bilayer over time, the same amounts of the CyaC- or HlyC-activated CyaA_1-710/HlyA_411-1024 hybrid proteins produced negligible, if any, conductance across the black lipid membrane (Fig. 4A). Nevertheless, with a much reduced propensity, the hybrid protein occasionally still formed single pore conductance units that exhibited similar characteristics as the pores formed at much higher frequency by the intact CyaC-activated CyaA (Fig. 4B and Fig. S2). Hence, the HlyC-acylated CyaA_1-710/HlyA_411-1024 hybrid was capable to efficiently translocate the AC enzyme across the erythrocyte membrane despite a close to nil hemolytic and membrane-permeabilizing (pore-forming) capacity. This result corroborates our earlier observation that translocation of the AC enzyme polypeptide across the lipid bilayer of cell membrane does not involve the formation of oligomeric membrane-permeabilizing CyaA pores and that the translocating AC polypeptide is presumably delivered by membrane-inserted monomers of CyaA through a sealed proteolipidic translocon (26).

Figure 4.

The cell-invasive CyaA_1-710/HlyA_411-1024 hybrid molecules exhibit a very low membrane-permeabilizing activity. A, overall membrane activities of intact CyaC-acylated CyaA and CyaC- or HlyC-activated CyaA_1-710/HlyA_411-1024 hybrid molecules on asolectin/decane:butanol (9:1) membranes. Conditions of measurement were: 150 mm KCl, 10 mm Tris-HCl (pH 7.4), 2 mm CaCl₂; the applied voltage was 50 mV; temperature was 25 °C, and the recording was filtered at 10 Hz; the protein concentration was 250 pm. B, single-pore recordings of asolectin membranes in the presence of 10 pm purified protein variants under otherwise identical conditions as in A.

These results further demonstrate that the AC-to-Hly linking segment and the intact hydrophobic (pore-forming) domain, comprised between residues 400 and 710 of CyaA, are necessary and sufficient for accomplishing the calcium-dependent AC domain translocation across the cell membrane. The C-terminal portion of CyaA, containing the acylated segment and the RTX domain, does not appear to be involved in the process of AC translocation, as it could be functionally replaced with the acylated segment and the much shorter RTX domain of E. coli HlyA that did not evolve to deliver the AC enzyme.

Swapping of the acylated segment and the RTX domain of HlyA into CyaA retargets the adenylyl cyclase toxin from its CR3 receptor to LFA-1–expressing cells

CyaA specifically binds phagocytic cells through their α_Mβ₂ integrin CR3 (Fig. S3A) and thanks to this the toxin swiftly exerts a complex array of cytotoxic and immunosubversive activities on host myeloid phagocytic cells (10, 11, 63–68). The segment of CyaA involved in the interaction of the toxin with CR3 was previously mapped within residues 1166 and 1287 of the RTX domain and is located at the interface of the RTX blocks II and III (11, 69–71). In the CyaA_1-710/HlyA_411-1024 hybrid the CR3-binding RTX domain of CyaA was replaced by the much shorter RTX domain of HlyA. It was demonstrated that HlyA preferentially targets another β₂ integrin, LFA-1 (60), but the integrin-interacting segment of HlyA is currently unknown. Therefore, we examined if the domain swap retargeted the hybrid protein from CR3 to α_Lβ₂ LFA-1. As shown in Fig. 5A, indeed, the HlyC-activated CyaA_1-710/HlyA_411-1024 hybrid bound approximately 10 times less to the CR3-expressing murine J774A.1 cells than the intact CyaC-activated CyaA. The residual binding of the hybrid to J774A.1 cells did not involve an interaction with CR3, as it was not inhibited by the mAb M1/70 that competes with CyaA for binding to the same segment of the CD11b subunit of CR3 (11). Accordingly, at the low concentrations needed for assays with CyaA (0.1–0.5 nm), the HlyC-activated CyaA_1-710/HlyA_411-1024 hybrid did not elevate any detectably the cAMP concentrations in the cytosol of J774A.1 cells, showing that the residual CR3-independent binding interaction to J774A.1 cells was rather unproductive (Fig. 5B).

Figure 5.

The CyaA_1-710/HlyA_411-1024 hybrid preferentially binds LFA-1–positive cells and monoacylation by CyaC is sufficient for AC translocating activity. A, binding of the intact CyaC-activated CyaA or HlyC-activated CyaA_1-710/HlyA_411-1024 hybrid variant to J774A.1 mouse macrophages (1 × 10⁶) was determined as the amount of total cell-associated AC enzyme activity upon incubation of cells with 5 nm protein for 30 min at 4 °C. To block the binding site of CyaA on the CR3 receptor, the J774A.1 cells (1 × 10⁶) were preincubated for 30 min on ice with 5 μg/ml of the CD11b-specific mAb M1/70 prior to addition of 5 nm CyaA (CyaC⁺) or CyaA_1-710/HlyA_411-1024 (HlyC⁺). Activities are expressed as percentages of intact CyaA activity in the absence of M1/70 mAb and represent average mean ± S.D. from two independent determinations performed in duplicate with two different toxin preparations. B, cAMP intoxication was assessed by determining the intracellular concentration of cAMP after 30 min of incubation of J774A.1 cells (2 × 10⁵) with toxins (n = 3). C, binding of CyaA or CyaA_1-710/HlyA_411-1024 (HlyC⁺) to Jurkat T-cells (1 × 10⁶) was determined as the amount of total cell-associated AC enzyme activity upon incubation of cells with 5 and 50 nm toxin for 30 min at 4 °C. Activities are expressed as percentages of intact CyaA activity and represent average mean ± S.D. from two independent determinations performed in duplicate with two different toxin preparations. D, cAMP intoxication was assessed by determining the intracellular concentration of cAMP generated in Jurkat T-cells (3 × 10⁵) after 30 min of incubation at different toxin concentrations (n = 5). E, binding of CyaA_1-710/HlyA_411-1024 (HlyC⁺) or intact CyaA to CHO cells expressing CD11a/CD18 or mock-transfected CHO cells (1 × 10⁶) was determined as the amount of total cell-associated AC enzyme activity upon incubation of cells with indicated toxin concentrations for 30 min at 4 °C. Activities represent average mean ± S.D. from three independent determinations performed in duplicate with two different toxin preparations. F, cAMP intoxication was assessed by determining the intracellular concentration of cAMP generated in CHO cells expressing CD11a/CD18 or mock-transfected CHO cells after 30 min of incubation of cells (1 × 10⁵) with different concentrations of CyaA_1-710/HlyA_411-1024 (HlyC⁺) or intact CyaA (n = 4). *, statistically significant differences (p < 0.05); **, statistically significant differences (p < 0.01); ***, statistically significant differences (p < 0.001); ****, statistically significant differences (p < 0.0001).

We therefore assessed the cell-penetration capacity of the HlyC-acylated CyaA_1-710/HlyA_411-1024 hybrid using Jurkat lymphoblastoma T-cells that do not express CR3 but produce modest amounts of LFA-1 (Fig. S3B) (72). Higher toxin concentrations had thus to be used (5 or 50 nm) and at equal molar concentrations of the toxins, the LFA-1-expressing Jurkat cells bound about two times lower amounts of the CyaA toxin than of the CyaA_1-710/HlyA_411-1024 hybrid toxin (Fig. 5C). Furthermore, compared with intact CyaA, the hybrid produced comparable cAMP levels in Jurkat cells over a range of toxin concentrations (Fig. 5D). This goes well with the results of the assays on erythrocytes, where the efficacy of translocation of the HlyC-acylated cell-bound CyaA_1-710/HlyA_411-1024 hybrid toxin ranged between ∼35 and 40% of that of cell-bound intact CyaA (cf. Fig. 3, A and B).

To corroborate that the swapping of the acylated segment and the RTX domain with those of HlyA retargeted AC toxin from CR3 to LFA-1, we further used transfected CHO cells that expressed high levels of the LFA-1 receptor (Fig. S3B). Compared to mock-transfected cells, the transfected CHO cells expressing LFA-1 on the cell surface bound much higher amounts of the HlyC-acylated CyaA_1-710/HlyA_411-1024 hybrid, when this was used at concentrations below or equal to 5 nm (Fig. 5E). At a high concentration of 25 nm the hybrid protein bound at similar levels to both mock-transfected and LFA-1–producing CHO cells (Fig. 5E). However, the specific interaction with LFA-1 still enabled an importantly more efficient penetration of the bound hybrid toxin across the cell membrane (Fig. 5F). Despite comparable amounts of the protein bound to cells, the hybrid toxin produced about 10-fold higher cAMP levels in the LFA-1-expressing cells than it produced in the mock CHO cells (Fig. 5F). Hence, the specific interaction with LFA-1 presumably imposed on the bound hybrid toxin an oriented and more productive topology that enhanced its capacity to penetrate cell membrane and translocate the AC domain into cytosol of LFA-1-expressing cells.

Moreover, interaction with the β₂ integrin rescued the AC translocation defect observed on erythrocytes (cf. Fig. 3, A and B) for the monoacylated CyaC-activated CyaA_1-710/HlyA_411-1024 hybrid protein (acylated only on the Lys-690 residue by 16-carbon acyls). The monoacylated protein intoxicated the LFA-1–expressing cells to the same levels as the doubly acylated HlyC-activated hybrid (c.f. Fig. 5D and Fig. S4). Hence, the interaction with LFA-1 enabled a productive interaction of the CyaA_1-710/HlyA_411-1024 hybrid toxin with the plasma membrane of the target cells and thereby promoted the translocation of its AC domain across the membrane even in the absence of the second acylation of the Lys-564 residue in the HlyA moiety.

Swapping of the acylated segment and the RTX domain retargets the CyaA/HlyA hybrid onto a spectrum of mouse spleen cells

To analyze which cell types could be targeted by the CyaA_1-710/HlyA_411-1024 hybrid molecule following the acylated segment and the RTX domain swap, we tested in vitro the binding of the fluorescently labeled (Dy495) and genetically detoxified (AC⁻) HlyC-acylated toxoid (CyaA-AC^–_1-710/HlyA_411-1024–Dy495) to mouse splenocytes comprising a broad mixture of cells of various types. A fluorescently labeled (Dy650) and CyaC-acylated CyaA toxoid (CyaA-AC⁻–Dy650) was used as a control. At first, the CyaA-AC^–_1-710/HlyA_411-1024–Dy495 and CyaA-AC⁻–Dy650 proteins were incubated with spleen cells separately at concentrations ranging up to 140 nm (Fig. 6A and Fig. S5). Although the CyaA-AC⁻–Dy650 toxoid used at the highest concentration stained ∼5% of all cells, at the same concentration the chimera stained almost 35% of splenocytes in the sample (Fig. 6A and Fig. S5). Because the SSC and FSC characteristics suggested binding of both proteins to different cell types (Fig. S5), a multicolor antibody panel was introduced to identify the target cell populations (Fig. S6A). Incubation of an equimolar mixture of the two proteins (56 nm) with cells yielded a positive detection of the CyaA-AC⁻–Dy650 stain on 5.3% of all cells. 25.6% of the cells were stained by CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 when the corresponding nonfluorescent proteins were used to set the positive gates for fluorescent proteins (Fig. S6B). As shown in Fig. 6B, compared with binding of CyaA-AC⁻–Dy650, the CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 toxoid exhibited importantly higher binding to nonleukocytic cells devoid of LFA-1 and CR3 (cf. Fig. 6C). Similarly, CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 bound better to B and T lymphocytes (Fig. 6B) expressing LFA-1, but lacking CR3 (Fig. 6C). Both proteins then bound with high affinity to neutrophils and myeloid cells (dendritic cells, monocytes, and macrophages) that express high levels of both LFA-1 and CR3 in parallel (Fig. 6, B and C). Adding the CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 and CyaA-AC⁻–Dy650 proteins together or separately to spleen cells did not significantly influence their binding patterns, except for the higher binding of the individually added chimera to LFA-1–negative cells and to LFA-1–epressing B lymphocytes (Table S1). This ruled out a possible interaction of the two proteins. All these data confirm that CR3 is the bona fide receptor of CyaA, whereas the acylated segment and the RTX domain of HlyA swapped into the CyaA-AC⁻_1-710/HlyA_411-1024 protein relocated a substantial portion of the hybrid molecules to bind cell populations expressing only LFA-1, or even to nonleukocytic cells devoid of both LFA-1 and CR3 integrins.

Figure 6.

Binding of fluorescently labeled CyaA-AC⁻_1-710/HlyA_411-1024 and CyaA-AC⁻ to spleen cells in vitro. A, spleen single cell suspensions (1 × 10⁶) from 7-week–old Balb/c mice were incubated in vitro with a range of equimolar concentrations of separately added Dy495-labeled CyaA-AC⁻_1-710/HlyA_411-1024 or Dy650-labeled CyaA-AC⁻ and were analyzed by flow cytometry (for gating strategy see Fig. S5). B, spleen single cell suspensions (1 × 10⁶) from 7-week-old Balb/c mice were incubated with an equimolar mixture of 56 nm fluorescently-labeled or nonfluorescent CyaA-AC⁻_1-710/HlyA_411-1024 and CyaA-AC⁻ for 30 min at 4 °C. Cells were then stained with a mixture of fluorescently-labeled antibodies against cell-surface antigens. Binding of CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 or CyaA-AC⁻–Dy650 to nonleukocytic cells (CD11a⁻), B lymphocytes (CD19⁺), T lymphocytes (CD3⁺), neutrophils (Ly-6G⁺CD11b⁺), CD11b^high myeloid cells (Lineage^neg CD11b^high) and CD11b^int myeloid cells (Lineage^neg CD11b^int) was determined by flow cytometry and depicted as histograms of Dy650 or Dy495 signal in each population. Gray histograms represent nonfluorescent CyaA-AC⁻_1-710/HlyA_411-1024 or CyaA-AC⁻ controls. Dy650- or Dy495-positive cells in each population are shown in percentages. Values in top right corners indicate the mean fluorescence intensity (MFI) for each histogram with subtracted MFI of nonfluorescent controls. C, spleen single cell suspensions from 7-week-old Balb/c mice were stained with a mixture of fluorescently labeled antibodies against cell-surface antigens. Expression of CD11a and CD11b on corresponding cell populations was determined by flow cytometry. Gray histograms represent corresponding isotype control. Values in top right corners indicate mean fluorescent intensity for each histogram with subtracted mean fluorescent intensity of isotype controls. A result from one representative experiment is shown in each panel. In vitro experiment was repeated two times independently with similar results.

Discussion

We show here that swapping of the acylated segment and the RTX domain of HlyA into the CyaA toxin molecule retargets the AC toxin from the CR3 β₂ integrin for binding to the LFA-1 β₂ integrin. Moreover, the cell-invasive capacity of the doubly acylated CyaA_1-710/HlyA_411-1024 hybrid revealed that the CyaA translocon unit that delivers the AC enzyme polypeptide across the target cell membrane is formed by the structure consisting of the AC-to-Hly linking segment and the hydrophobic domain segment of CyaA comprised within residues 400 to 710 of the CyaA toxin molecule (Fig. 7A). This CyaA segment was necessary and sufficient to accomplish AC enzyme translocation across the lipid bilayer of target cell membrane. The results with the CyaA_1-710/HlyA_411-1024 hybrid also demonstrated that swapping of the acylated segment and the RTX domain retargeted the AC toxin from CR3-expressing myeloid phagocytic cells to LFA-1–expressing B and T leukocytes and broadened, at higher toxin concentrations, the target cell spectrum of the hybrid toxin also to nonleukocyte types of cells targeted by HlyA. More importantly, the presented data reveal that interaction with the β₂ integrin receptor is itself not crucial for membrane penetration of the CyaA translocon moiety and the delivery of the AC polypeptide across cellular membrane. As suggested by the results reported here, interaction with the β₂ integrin would only facilitate and accelerate the productive interaction of cell surface-associated toxin molecules with the lipid bilayer of the target cell membrane. This is most likely due to imposing an oriented topology on the receptor-bound toxin molecules, thus increasing the probability of their productive insertion into the lipid bilayer and the formation of a proper AC enzyme translocon unit for delivery of the AC polypeptide across the lipid bilayer of the membrane.

Figure 7.

Schematic representation of CyaA and HlyA domains sufficient for AC translocation and LFA-1 binding. A, the AC domain of Bordetella pertussis CyaA has to be fused to the adjacent AC-to-Hly linker segment and to the five predicted transmembrane α-helices for an efficient translocation across the cell membrane, whereas the acylated segment and the RTX domain segment of CyaA can be replaced with the corresponding segments of the E. coli HlyA toxin. B, the doubly acylated HlyC-modified 14-carbon-acylated CyaA_1-710/HlyA_411-1024 hybrid, bearing C14:0 or C14:0-OH on the two Lys-564 and Lys-690 residues translocated the AC domain into cell cytosol of both, LFA-1–positive and LFA-1–negative cells. C, the monoacylated CyaA_1-710/Hly_A411-1024 hybrid, bearing C16:0, C16:1, and/or C18:1 on the Lys-690 residue intoxicated the LFA-1–positive cells with the same efficacy as the double acylated hybrid but was poorly active on erythrocytes lacking the β₂ integrin LFA-1. Because CyaA does not bind the LFA-1 integrin (Fig. S3), the LFA-1–binding segment of the HlyC-acylated CyaA_1-710/HlyA_411-1024 molecule is confined in the acylated segment and/or the RTX domain of HlyA (residues 411 to 1024).

Previously, Westrop and co-workers (73) constructed hybrid molecules of leukotoxin (LktA) from Mannheimia hemolytica and the CyaA toxin. The first hybrid molecule, containing N-terminal residues 1 to 687 of CyaA and C-terminal residues 379 to 953 of LktA, showed no toxic activities (73). The second hybrid molecule, containing N-terminal residues 1 to 687 of CyaA, residues 379 to 616 of LktA, and C-terminal residues 919 to 1706 of CyaA, exhibited a hemolytic activity, but its capacity to deliver the AC domain to the cell cytosol was ∼20 times lower than the activity of intact CyaA (73). Our CyaA_1-710/HlyA_411-1024 hybrid molecule was able to translocate the AC domain across the membrane of cells devoid of CR3 with a much higher efficacy of ∼40% of the intact CyaA. This indicates that the presence of all five transmembrane α-helices, predicted between residues 502 and 698 of CyaA (33), may be required for translocation of the AC enzyme across the plasma membrane. However, because the CyaA_1-710/HlyA_411-1024 chimera translocated the AC domain with lower efficacy than intact CyaA, it is difficult to conclusively discriminate if the reduced translocation efficacy was due to structural constraints imposed on the AC translocon unit in the artificial hybrid molecule (assembled without any structural data guidance), or whether the additional CyaA segments participate indirectly and facilitate by structural input the optimal function of the CyaA translocon. Indeed, the calcium-binding RTX domain of CyaA_1-710/HlyA_411-1024 comes from HlyA and contains a significantly lower number of RTX repeats than CyaA (1). Thus, it can be hypothesized that the shorter RTX domain of HlyA may yield upon binding the hybrid molecule to the cell surface, a less appropriate positioning of the AC-to-Hly linking segment and/or of the hydrophobic domain of CyaA in respect to lipid bilayer of the membrane. This might reduce the efficacy of the subsequent formation and operation of the AC translocon.

Also, little is known about the role of the acylated segment of the RTX toxins in their mechanism of penetration into target membrane. In contrast to the activity on liposomes lacking membrane potential, the capacity of CyaA and HlyA to penetrate cellular membranes strictly depends on activation of their protoxins by covalent posttranslational fatty acylation (37, 58, 74–76). It was shown that acylation of CyaA alters and dictates the folding path and the overall conformation of the acylated segment of the toxin molecule (77, 78). By dictating the overall conformation of the protein, the acylation also determines the immunogenicity and the capacity of the CyaA antigen to induce toxin/neutralizing antibodies and confer protective immunity against Bordetella infection (79). Here we show that the acylation status of the toxin molecule also impacts on the translocation efficacy of its AC domain across the cell membrane in the absence of the integrin receptor. The doubly acylated HlyC-modified CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 variants, bearing C14:0 or C14:0-OH on the two Lys-564 and Lys-690 residues, translocated the AC domain into the erythrocyte cytosol with quite high efficacy of ∼35 to 40% of intact CyaA. In contrast, the AC domain translocation capacity of the monoacylated CyaC-modified variants, acylated with a mixture of C16:0, C16:1, and C18:1 exclusively at the Lys-690, was very low when assayed on erythrocytes. Because the CyaC- and HlyC-acylated CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 variants bound erythrocytes with similar efficacy, it indicates that in the absence of interaction with the β₂ integrin receptor, which could orient the protein in respect to membrane lipid bilayer, the type and/or site of acylation of the toxin molecule directly affects its capacity to form the translocon and initiate AC domain translocation. On cells devoid of integrin receptor, the presence of two acyl chains per toxin molecule may thus be critical for the ability of the hybrid molecules to adopt a conformation that allows proper membrane insertion and subsequent AC domain translocation (Fig. 7, B and C). Indeed, interaction with the LFA-1 receptor also enabled the monoacylated CyaC-activated hybrid protein to penetrate cells and deliver its AC domain, showing that acylation of the first lysine residue (Lys-564 in HlyA and Lys-860 in CyaA) was required for toxin activity only on cells that do not express the cognate β₂ integrin receptor (e.g. LFA-1 or CR3, respectively) on their surface. A precedent of such a rescuing effect of a β₂ integrin receptor interaction was previously observed with the monoacylated CyaC-activated CyaA-K860R toxin variant (37, 80).

Both the CyaC-acylated and HlyC-acylated CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 variants exhibited an almost nil hemolytic capacity on erythrocytes and a very low membrane-permeabilizing activity on asolectin bilayer membranes with imposed voltage. On the other hand, the type and/or site of acylation of the CyaA_1-710/HlyA_411-1024 hybrid molecule did not affect characteristics of the rarely formed single pores made by these proteins. This goes well with our previous reports showing that irrespectively of the differences in acylation status, the truncated CyaA variants lacking the RTX domain, or even the nonacylated proCyaA protein can form pores that exhibit identical conductance and cation selectivity as intact CyaA and are just formed at a very much reduced frequency (13, 80). More importantly, the observations reported here show that the HlyC-acylated CyaA_1-710/HlyA_411-1024 and CyaA_1-800/HlyA_501-1024 variants can efficiently translocate the AC domain, despite being devoid of any detectable membrane-permeabilizing capacity. This corroborates our previous report that the pore-forming activity of CyaA is not involved in translocation of the AC domain across the target cell membrane (26, 32, 36). The data clearly show that AC translocation does not occur through the oligomeric CyaA pore, which is bypassed by the translocating AC enzyme polypeptide. The AC is most likely delivered across the lipid bilayer of the membrane by monomers of CyaA along the electrical gradient and through a tightly sealed proteolipid translocon formed by residues 400 to 710 of CyaA, as found here.

HlyA exhibits a limited target cell and species specificity, exhibiting a well-detectable cytotoxic activity on a broad spectrum of cell types from various species, including erythrocytes, monocytes, granulocytes, or endothelial cells from ruminants, mice, and primates (81–87). Despite the rather low cell specificity, HlyA binding and action on erythrocytes was shown to be dependent on the abundant glycophorin protein proposed to serve as a high-affinity receptor (59, 88). Indeed, the HlyC-activated CyaA_1-501/HlyA_1-1024 hybrid molecule, carrying the entire HlyA, bound sheep erythrocytes with a ∼7 times higher efficacy (Fig. S7) than intact CyaA that does not bind glycophorin of erythrocytes (59, 89). The glycophorin-binding region was proposed to be located between residues 914 and 936 of HlyA (88), thus overlapping with the RTX folding center of the toxin molecule (41). This opens the possibility that mutations in this region prevented proper folding and ablated HlyA activity independently of loss of glycophorin interaction. Interestingly, sequential shortening of HlyA from its N terminus in the HlyC-acylated CyaA/HlyA hybrids, still harboring an intact putative glycophorin-binding region, resulted in a substantial reduction of their cell binding capacity, as compared with CyaA_1-501/HlyA_1-1024. Although the binding of the CyaA/HlyA hybrids was assessed on sheep erythrocytes, whereas identification of the glycophorin-binding region was performed on human and horse erythrocytes (59, 88), it is plausible to hypothesize that the glycophorin-dependent binding of HlyA to erythrocytes may require the structural integrity of several regions of the toxin. Another explanation may be that upon a primary interaction of residues 914 to 936 of HlyA with glycophorin, the toxin irreversibly inserts into the membrane through the N-terminal segments of the molecule. This assumption is supported by previous reports showing that HlyA is anchored to the plasma membrane through its N-terminal structures (54), which also modulate pore-forming activity (90). Moreover, the sequential deletion of the N-terminal structures of HlyA within the CyaA/HlyA hybrids, including a hydrophobic region between residues 238 and 410, also resulted in a substantial reduction of their hemolytic capacity compared with that of CyaA_1-501/HlyA_1-1024, most likely due to their reduced cell-association capacity.

HlyA was found to preferentially bind and target leukocytes expressing the β₂ integrin LFA-1 (60). The evidence was based on observations showing that (i) HlyA-mediated cytolysis was inhibited by mAbs recognizing the CD11a and CD18 subunits of LFA-1, (ii) immobilized HlyA bound LFA-1, and (iii) the K562 cells transfected with LFA-1 became sensitive to HlyA (60). Another study (91), however, indicated that HlyA bound to erythrocytes and granulocytes in a nonsaturable manner and competition experiments showed that the binding of HlyA to the cells was receptor-independent. Here we show that Jurkat T-cells expressing LFA-1 but not CR3 bound approximately two times more of the HlyC-acylated CyaA_1-710/HlyA_411-1024 than of the intact CyaA. In addition, at low protein concentrations the hybrid bound, and intoxicated by cAMP, preferentially to the CHO cells expressing LFA-1, suggesting that interaction of the acylated segment and the RTX domain of HlyA in the hybrid facilitated its productive insertion into cell membrane and fostered its capacity to deliver the AC enzyme across the cell membrane. Our results further show that compared with CyaA toxoid, the binding of CyaA-AC⁻_1-710/HlyA_411-1024 was higher on B and T lymphocytes expressing LFA-1, but not CR3. However, due to the very high toxoid concentrations used, in the same experiment the binding of CyaA-AC⁻_1-710/HlyA_411-1024 was also substantially higher on nonleukocytic cells lacking LFA-1 and CR3. These results further emphasize some binding “promiscuity” of HlyA and demonstrate that at low concentrations the HlyA would preferentially bind to cells expressing LFA-1, whereas at higher concentrations it exhibits promiscuous binding and low target cell specificity.

Experimental Procedures

Bacterial strains

The E. coli strain XL1-Blue (Stratagene, La Jolla, CA) was used throughout this work for DNA manipulations and was grown in Luria-Bertani medium at 37 °C. The E. coli strain BL21 (Novagen, Madison, WI) carrying the plasmid pMM100 (encoding LacI and tetracycline resistance) was used for expression of the CyaA and HlyA protein variants.

Cell lines

CHO-K1 Chinese hamster ovary cells (ATCC CCL-61) transfected with human CD11a/CD18 or mock transfected were established earlier (11). Murine monocytes/macrophages J774A.1 (ATCC, number TIB-67), human THP-1 monocytes (ATCC number TIB-202), and Jurkat human T lymphocyte cells (TIB-152) were obtained from the American Type Culture Collection (ATCC, Manassas, VA). CHO cells were grown in F12K medium (GIBCO Invitrogen) supplemented with 10% fetal calf serum (FCS) (GIBCO Invitrogen) and antibiotic antimycotic solution (ATB, 0.1 mg/ml streptomycin, 100 units/ml of penicillin, and 0.25 µg/ml of amphotericin; Sigma-Aldrich). J774A.1, THP-1, and Jurkat cells were cultured in RPMI 1640 (Sigma-Aldrich) supplemented with 10% fetal calf serum and antibiotic antimycotic solution. Prior to assays, RPMI or F12K were replaced with DMEM (1.9 mm Ca²⁺) without FCS and the cells were allowed to rest in DMEM for 1 h at 37 °C in a humidified 5% CO₂ atmosphere.

Mice

7-Week-old female Balb/c mice were bred locally at the Institute of Microbiology of the CAS, v.v.i. All animal experiments were approved by the Animal Welfare Committee at the Institute of Microbiology of the CAS, v.v.i. in accordance with the Guidelines for the Care and Use of Laboratory Animals, the Act of the Czech National Assembly, Collection of Laws number 246/1992.

Plasmid construction

The pT7CACT1 plasmid (40), harboring the cyaC and cyaA genes under control of the isopropyl β-d-thiogalactopyranoside–inducible lacZp promoter, was used as a parental plasmid to generate constructs for expression of the CyaA/HlyA hybrid molecules activated by the acyltransferase CyaC or HlyC. For this purpose, the plasmid pTZHly11 (9) was used as a source of the hlyC and hlyA. First, pT7CACT1-derived plasmid expressing HlyC was constructed by replacement of the cyaC gene from its start to stop codon in pT7CACT1 by hlyC. Second, an overlapping PCR mutagenesis procedure was used to amplify DNA segments encoding required CyaA/HlyA hybrid molecules and the PCR products were cloned instead of cyaA into the appropriate pT7CACT1-derived plasmid (harboring cyaC or hlyC). For production and subsequent purification of HlyC-acylated HlyA, the cyaA gene in the pT7CACT1-derived plasmid harboring hlyC was replaced from its start to stop codon by the entire hlyA gene, which was fused in-frame at the 3′ terminus to a sequence encoding a double hexahistidine purification tag (92). The detoxified constructs CyaA-AC⁻_1-710/HlyA_411-1024 and CyaA-AC⁻ were obtained by ablation of their catalytic activity as previously described (40).

Protein production, purification, and labeling

The intact CyaA, intact HlyA, and CyaA/HlyA hybrid molecules were produced in E. coli BL21/pMM100 cells transformed with the appropriate plasmids. 500-ml cultures were grown with shaking at 37 °C in MDO medium (yeast extract, 20 g/liter; glycerol, 20 g/liter; KH₂PO₄, 1 g/liter; K₂HPO₄, 3 g/liter; NH₄Cl, 2 g/liter; Na₂SO₄, 0.5 g/liter; thiamine hydrochloride, 0.01 g/liter) containing 150 μg/ml of ampicillin and 12.5 μg/ml of tetracycline. When cultures reached OD₆₀₀ = 0.8, they were induced with 1 mm isopropyl β-d-thiogalactopyranoside and grown for an additional 4 h. For protein purification, the cells were harvested by centrifugation, washed twice with 50 mm Tris-HCl (pH 8.0), disrupted by sonication at 4 °C, and the homogenate was centrifuged at 20,000 × g for 30 min at 4 °C. The inclusion bodies collected in the pellet were washed with 50 mm Tris-HCl (pH 8.0) containing 4 m urea, then solubilized with 50 mm Tris-HCl (pH 8.0) containing 8 m urea, and the urea extract was cleared at 20,000 × g for 30 min at 4 °C.

The urea extracts containing intact CyaA or the CyaA/HlyA hybrid proteins were diluted 4 times in ice-cold TNC buffer (50 mm Tris-HCl (pH 8.0), 500 mm NaCl, and 2 mm CaCl₂) and loaded at 4 °C on a calmodulin-Sepharose 4B column (GE Healthcare) equilibrated with the same buffer. The column was washed with TNC buffer and the proteins were eluted at room temperature with TUE buffer (50 mm Tris-HCl (pH 8.0), 8 m urea, and 2 mm EDTA).

To prepare labeled CyaA-AC⁻_1-710/HlyA_411-1024 and CyaA-AC⁻ with low content of lipopolysaccharide for mouse experiments, the protein samples purified on calmodulin-Sepharose were diluted 4 times in ice-cold 50 mm Tris-HCl (pH 8.0) containing 1 m NaCl and loaded on phenyl-Sepharose CL-4B beads (Sigma–Aldrich). Endotoxin was removed by repeated washes of the protein-bound resin with 60% isopropyl alcohol (93). The phenyl-Sepharose column was then washed with 50 mm sodium bicarbonate (pH 8.3) containing 1 m NaCl and the beads were resuspended in the same buffer containing Dy495-NHS ester (Dyomics, Jena, Germany) for CyaA-AC⁻_1-710/HlyA_411-1024 or Dy650-NHS ester for CyaA-AC⁻ in a concentration to reach a Dyomics:protein molar ratio of ∼6:1. Labeling was performed at 25 °C for 2 h and the column was subsequently washed with 50 mm Tris-HCl (pH 8.0). The labeled proteins were eluted in TU buffer (50 mm Tris-HCl (pH 8.0), 8 m urea) and concentrated on 30-kDa Ultracel membrane filters (Merck, Darmstadt, Germany). The preparations used in mouse experiments thus contained less than 0.1 EU/1 µg of the protein as determined by Limulus Amebocyte Lysate assay (QCL-1000, Lonza, Walkersville, MD).

The urea extract containing intact HlyA was loaded on a nickel-nitrilotriacetic acid-agarose column (Qiagen, Germantown, MD) equilibrated with TNU buffer (50 mm Tris-HCl (pH 8.0), 200 mm NaCl, and 8 m urea). The column was washed with TNU buffer containing 20 mm imidazole and HlyA was eluted with TNU buffer containing 600 mm imidazole. The eluted fractions of HlyA were diluted 4 times in ice-cold 50 mm Tris-HCl (pH 8.0) containing 1 m NaCl and loaded on a phenyl-Sepharose CL-4B column equilibrated with the same buffer. The column was then washed with 50 mm Tris-HCl (pH 8.0) and HlyA was eluted with TUE buffer.

Cell binding, cell invasive and hemolytic activities on sheep erythrocytes

AC enzymatic activities were measured in the presence of 1 μm calmodulin as previously described (94). One unit of AC activity corresponds to 1 µmol of cAMP formed per min at 30 °C (pH 8.0). Hemolytic activity was measured by determining the hemoglobin release in time upon toxin incubation with washed sheep erythrocytes (LabMediaServis, Jaromer, Czech Republic) (5 × 10⁸/ml), as previously described (18). Cell invasive AC was measured by determining the AC protected against externally added trypsin upon internalization into sheep erythrocytes as previously described (27). Erythrocyte binding was measured by determining the membrane-associated AC activity as previously described (27). Activity of intact CyaA was taken as 100%.

Binding and cAMP elevation of CyaA on nucleated cells

Cells were incubated in DMEM with intact CyaA or CyaA/HlyA hybrid proteins for 30 min at 4 °C prior to removal of unbound toxin by three washes in DMEM. After transfer to a fresh tube, the cells were lysed with 0.1% Triton X-100 for determination of cell-bound AC enzyme activity. Activity of WT CyaA was taken as 100%. For intracellular cAMP assays, cells were incubated at 37 °C with CyaA or hybrid toxin for 30 min in DMEM, the reaction was stopped by addition of 0.2% Tween 20 in 100 mm HCl, samples were boiled for 15 min at 100 °C, neutralized by addition of 150 mm unbuffered imidazole, and cAMP was measured by a competitive immunoassay as previously described (24).

Detection of CD11a/CD18 on cell surface

2 × 10⁵ Jurkat, THP-1, CHO-CD11a/CD18, and mock-transfected CHO cells were incubated for 30 min at 4 °C in 100 µl of HEPES-buffered salt solution (HBSS buffer; 10 mm HEPES, pH 7.4, 140 mm NaCl, 5 mm KCl) supplemented with 2 mm CaCl₂, 2 mm MgCl₂, 1% (w/v) glucose, and 1% (v/v) FCS (cHBSS) in 96-well culture plates (Nunc, Roskilde, Denmark) containing the anti-human CD11a-specific MEM-25 mAb (mAb) conjugated with allophycocyanin (Exbio, Vestec, Czech Republic) diluted according to the manufacturer's instructions. After washing, cells were resuspended in HBSS and analyzed by flow cytometry in the presence of 1 µg/ml of Hoechst 33258. Data were analyzed using the FlowJo software (Tree Star, Ashland, OR) and appropriate gatings were used to exclude cell aggregates and dead cells.

Planar lipid bilayers

Measurements on planar lipid bilayers (black lipid membranes) (95) were performed in Teflon cells separated by a diaphragm with a circular hole (diameter 0.5 mm) bearing the membrane. The intact CyaA and the CyaC- or HlyC-activated CyaA_1-710/HlyA_411-1024 hybrid molecules were diluted in TU buffer and added into the grounded cis compartment with a positive potential. The membrane was formed by the painting method using soybean lecithin in n-decane–butanol (9:1, v/v). Both compartments contained 150 mm KCl, 10 mm Tris-HCl (pH 7.4), and 2 mm CaCl₂, the temperature was 25 °C. The membrane current was registered by Ag/AgCl electrodes (Theta) with salt bridges (applied voltage, 50 mV), amplified by LCA-200-100G amplifiers (Femto), and digitized by use of a KPCI-3108 card (Keithly). For lifetime determination, ∼400 of individual pore openings were recorded and the dwell times were determined using QuB software with 10-Hz low-pass filter. The kernel density estimation was fitted with a double-exponential function using Gnuplot software. The relevant model was selected by the χ2 value. The error estimates of lifetimes were obtained by the bootstrap analysis.

LC–MS analysis

The proteins were dissolved in 50 mm ammonium bicarbonate buffer (pH 8.2) to reach 4 m concentration of urea and digested with trypsin (Promega, Madison, WI, modified sequencing grade) at a trypsin:protein ratio of 1:50 for 6 h at 30 °C. The 2nd portion of trypsin was added to a final ratio of trypsin:protein of 1:25 and the reaction was carried out for another 6 h at 30 °C. When the reaction was complete, the concentration of the resulting peptides was adjusted by 0.1% TFA (TFA) to 0.1 mg/ml and 5 µl of the sample were injected into the LC–MS system. The LC separation was performed using a desalting column (ZORBAX C18 SB-300, 0.1 × 2 mm) at a flow rate of 40 µl/min (Shimadzu, Kyoto, Japan) of 0.1% formic acid (FA) and a separation column (ZORBAX C18 SB-300, 0.2 × 150 mm) at a flow rate of 10 µl/min (Agilent 1200, Santa Clara, CA) of water/acetonitrile (MeCN) (Merck, Darmstadt, Germany) gradient: 0-1 min 0.2% FA, 5% MeCN; 5 min 0.2% FA, 10% MeCN; 35 min 0.2% FA, 50% MeCN; 40 min 0.2% FA, 95% MeCN; 40–45 min 0.2% FA, 95% MeCN. A capillary column was directly connected to a mass analyzer. The MS analysis was performed on a commercial solariX XR FTMS instrument equipped with a 15 Tesla superconducting magnet and a Dual II ESI/MALDI ion source (Bruker Daltonics, Bremen, Germany). Mass spectra of the samples were obtained in the positive ion mode within an m/z range of 150-2000. The accumulation time was set at 0.2 s, LC acquisition was 45 min with a 5-min delay, and one spectrum consisted of accumulation of four experiments. The instrument was operating in survey LC–MS mode and calibrated online using Agilent tuning mix, which results in mass accuracy below 2 ppm.

Data processing and interpretation

MS data were processed by the SNAP version 2.0 algorithm of the DataAnalysis 4.4 software package (Bruker Daltonics, Billerica, MA, USA) generating a list of monoisotopic masses from deconvoluted spectra. The parameters were set as follows: export m/z range of 150-2000, maximum charge state of 8, S/N threshold of 0.75, and absolute intensity threshold 5 × 10⁵. The extracted experimental data were searched against the FASTA of single corresponding toxin molecule (CyaA, UniProtKB, P0DKX7; HlyA, UniProtKB, P08715, chimeric proteins: assembled from CyaA and HlyA according to the highlighted residue positions of individual molecules) using the home-built Linx software (RRID:SCR_018657). The Linx algorithm was set for fully tryptic restriction with a maximum of 3 missed cleavages and variable modification for methionine oxidation along with lysine acylation ranging from C12 to C18, including monosaturated and hydroxylated variants. The mass error threshold was set to ±2 ppm and all assigned peptides used for quantification were verified manually. The acylation status of lysine residues was determined by comparison of relative intensity ratio between acylated peptide ions and its unmodified counterparts. Only lysine residues at positions 564, 690, and 860, and 983 according to the sequence of the full-length HlyA or CyaA proteins, respectively, were investigated. All peptide sequences including post-translational modifications along with corresponding FASTA formats used within the search algorithm are listed in Supporting Information.

Tissue homogenization and single cell preparation

Mice were sacrificed by cervical dislocation and spleens were collected. Spleen single cell suspensions were obtained by homogenization in a gentleMACS™ Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany) and by enzymatic digestion with collagenase D (1 mg/ml; Roche Applied Science) in Hanks' balanced salt solution for 30 min at 37 °C. The reaction was stopped by 14 mm EDTA for 5 min, and cells were subsequently filtered through a 70-μm nylon cell strainer (Corning Costar, New York, NY). After lysing of erythrocytes by ACK lysing buffer (Life Technologies), cells were resuspended in FACS buffer (PBS, 2 mm EDTA, 2% heat-inactivated fetal bovine serum), and filtered through a 30-μm CellTrics strainer (Sysmex Partec GmbH, Görlitz, Germany).

Binding of labeled CyaA variants to spleen cells in vitro

Prior to staining, spleen cell suspensions were blocked by 10% Balb/c mouse serum and anti-mouse CD16/CD32 mAb (0.25 μg/reaction; Bioscience) in FACS buffer for 30 min at 4 °C. Aliquots of spleen cells (1 × 10⁶) were subsequently incubated with or without CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 and CyaA-AC⁻–Dy650 (or corresponding nonfluorescent proteins), added separately or in an equimolar mixture, for 30 min at 4 °C. Cells were washed once with FACS buffer and stained directly with labeled mAbs (Table S2) for 30 min in the dark at 4 °C and washed twice with FACS buffer. A Fixable Viability Dye eFluor™ 780 live/dead stain (eBioscience) was used to assess cell viability and the samples were analyzed using a BD LSRII flow cytometer. 2 × 10⁵ events were run for each sample and data were analyzed using FlowJo software. Gates positive for CyaA-AC⁻_1-710/HlyA_411-1024–Dy495 and CyaA-AC⁻–Dy650 were set using a nonfluorescent CyaA-AC⁻_1-710/HlyA_411-1024 and CyaA-AC⁻ as a background.

Statistical analysis

Significance of differences in values was assessed by Student's t test.

Data availability

The MS data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD018789 (96).