Pertussis toxin (PT), produced by Bordetella pertussis
, plays a key role in the disease caused by this organism (23
). In order to damage eukaryotic cells, this toxin must cross the membrane barriers of the bacteria after it is synthesized in the cytoplasm of the cell. The individual polypeptide chains of the toxin are believed to utilize a Sec-like system to transit the inner membrane of the bacterium, after which they assemble into the holotoxin form of the molecule. The completely assembled toxin then crosses the outer membrane of the bacterium utilizing a type IV transport system (9
), and this is followed by release of the toxin from the bacterial cell into the extracellular milieu.
The PT secretion system, known as the Ptl transporter, is composed of nine distinct proteins, PtlA to PtlI (8
). The genes encoding the Ptl proteins are located on the B. pertussis
chromosome directly downstream from the ptx
genes encoding the PT subunits (Fig. ) (24
). The Ptl transporter belongs to the family containing the type IV transporters, the prototype of which is the VirB system of Agrobacterium tumefaciens
, which exhibits considerable homology with the Ptl transporter (4
). On the basis of this homology, we propose that the general architecture of the Ptl transporter likely resembles that of the VirB system. Work on the VirB system has revealed that the VirB transporter can be divided into three basic segments: the engine of the transporter, which consists of ATPases that are located in the inner membrane; the core of the transporter, which connects this energy source to the outer membrane; and the pilus, which extends outward from the outer membrane (4
). Based on the level of amino acid sequence identity between the Ptl and VirB proteins, we predict that the analogous segments of the Ptl transporter are composed of PtlC and PtlH (engine), PtlD, PtlE, PtlF, PtlG, and PtlI (core), and PtlA (pilus or piluslike structure).
FIG. 1. ptx-ptl region and comparison of PtlD to VirB6. (A) Schematic diagram of the ptx-ptl operon along with the regions deleted in BP536ΔptlD and BP536Δptxptl936-8003. The nucleotide numbering system is that described previously (18, 24). (B) (more ...)
In this study, we examined more closely one of the components of the Ptl transporter core, PtlD. The VirB homologue of PtlD, VirB6, is a hydrophobic integral membrane protein that spans the inner membrane multiple times (12
). While computer analysis of PtlD also predicted a transmembrane structure, the level of amino acid identity (20%) of PtlD and VirB6 is lower than the levels of amino acid identity of other proteins of the Ptl transporter and their VirB homologues (24
). Moreover, the sizes of PtlD and VirB6 differ dramatically (Fig. ). PtlD consists of 463 amino acids, whereas VirB6 is comprised of only 295 amino acids (22
). The extreme N-terminal and C-terminal regions of PtlD appear to lack homologous counterparts in VirB6 (24
). In this study, we investigated PtlD, including the role that it plays in PT transport and in transporter stability. In particular, we focused on the role of the C-terminal end of PtlD, which has no obvious homologous counterpart in VirB6.
In order to better understand the interaction of PtlD with other Ptl transporter proteins, we examined whether deletion of ptlD
had any effect on the stability of other Ptl proteins, specifically two of the core proteins, PtlE and PtlF, and the engine protein PtlH, since we were able to generate antibodies to these proteins. Repeated attempts to generate antibodies to other Ptl proteins did not yield antibodies that were useful in our experiments. To initiate this study, an in-frame ptlD
deletion strain was constructed using PCR and primers 1a, 1b, 2a, and 2b (Table ) to generate DNA fragments corresponding to the regions flanking ptlD
, which were then ligated and inserted into the vector pSS1129. (Table shows all of the strains and plasmids used in this study.) All PCR-amplified regions generated in this study were verified to be correct by sequence analysis. The resulting plasmid was introduced into B. pertussis
BP536 by conjugation. Deletion of ptlD
occurred upon homologous recombination followed by selection of the desired mutant with appropriate antibiotics (8
). The deletion was verified by PCR. As shown in Fig. , the lack of production of PtlD in the ptlD
deletion strain had a dramatic effect on the stability of PtlE, PtlF, and PtlH. In all cases, densitometric analysis of replicate immunoblots from independent experiments indicated that deletion of ptlD
resulted in a significant decrease in the amount of Ptl protein detected in the bacterial cell (P
< 0.05, as determined by Tukey's honestly significant difference test following analysis of variance). These results suggest that PtlD is required for stabilization of a number of Ptl proteins that comprise the Ptl transporter. Previously, other workers have shown that deletion of virB6
, a ptlD
homologue, correlates with reductions in the levels of several VirB transporter proteins, including VirB8 (PtlE homologue), VirB9 (PtlF homologue), and VirB11 (PtlH homologue), under conditions that favor VirB protein turnover (11
Strains and plasmids used in this study
FIG. 2. Effect of ptlD on the stability of Ptl proteins. B. pertussis cells were suspended in phosphate-buffered saline to an A550 of 2.0. Samples of cell extracts (50 μl) were precipitated with an equal volume of 20% trichloroacetic acid and (more ...)
In order to confirm that PtlD is required for transporter stability, we complemented the ptlD deletion strain with a plasmid capable of expressing ptlD. We constructed this plasmid (pAMC249) by amplifying the ptlD region using PCR and primers 3a and 3b. The resulting fragment was inserted into pUFR047. As shown in Fig. , when we complemented the ΔptlD strain with this plasmid capable of expressing the full-length ptlD gene under the control of the lac promoter, we found that the levels of PtlE, PtlF, and PtlH increased significantly (P < 0.05). The levels of PtlE, PtlF, and PtlH in the complemented strain approached those in wild-type strain BP536.
Because of the lack of homology between the C-terminal regions of PtlD and VirB6 of A. tumefaciens
and therefore the paucity of information available about the probable biological role of this region of PtlD, we were particularly interested in determining the role of this portion of the protein in transporter structure and function. A clue that this region of the protein might contribute to transporter stability came from a comparison of the stabilities of PtlF in two different strains, BP536Δptxptl
(936-8003), a strain that lacks all but the extreme 3′ end of ptlD
(Fig. ), and a ΔptlD
deletion strain, in which the entire ptlD
gene, including its 3′ end, has been deleted. We noted that BP536Δptxptl
(936-8003), which would be expected to produce a protein comprised of the first 143 amino acids of the precursor form of the S1 subunit of PT fused to the terminal 62 amino acids of PtlD, produced significant quantities of PtlF (3
), an observation that was in contrast to the significant decrease in the PtlF level that we observed in this study with the strain in which the entire ptlD
gene was deleted (Fig. ). These results suggested that the C-terminal end of PtlD might significantly contribute to stabilization of the Ptl transporter. In order to confirm this, we examined the role of the terminal 72 amino acids of PtlD in stabilization of other Ptl proteins. In order to express only the 3′ end of the ptlD
gene, we fused nucleotides 507 to 935 of the ptx-ptl
region (containing the first 429 nucleotides of the ptxS1
gene, amplified using primers 4a and 4b) to nucleotides 7977 to 8195 of the region (encoding the terminal 72 amino acids of PtlD, derived from the NotI-HindIII fragment of pAMC249). We then cloned this fragment behind the lac
promoter of pUFR047 and introduced this plasmid into the ΔptlD
strain. As a control, we also introduced a plasmid (pAMC267) that contained only nucleotides 507 to 935 of the ptx-ptl
region into the same strain. We found that the presence of the 3′ end of ptlD
was able to restore the PtlE, PtlF, and PtlH protein levels in the ΔptlD
strain to levels that were not significantly different from those seen when the ΔptlD
strain was complemented with full-length ptlD
(Fig. ), as determined by performing a densitometric analysis of replicate immunoblots and using Tukey's honestly significant difference test following analysis of variance, but were significantly different than the negative control levels (P
FIG. 3. Effect of the 3′ end of ptlD on the stability of Ptl proteins. Samples of cell extracts were prepared and subjected to immunoblot analysis using antibodies specific to the Ptl proteins indicated as described in the legend to Fig. (more ...)
Previously, other workers have shown that PtlD is critical for PT secretion (6
) since mutations in the the ptlD
gene eliminated secretion. When we examined secretion, we found that introduction of a plasmid that carried the entire ptlD
gene under the control of the lac
promoter into the ΔptlD
strain restored secretion in this strain; however, complementation of the same strain with the 3′ end of ptlD
did not restore secretion (data not shown). These data suggest that while the C-terminal 72 amino acids of the 463-amino-acid PtlD protein are sufficient for stabilization of transporter proteins, a larger region of PtlD is required for secretion of the toxin.
In order to further localize the region of PtlD critical for transporter protein stability, we constructed a series of plasmids by using PCR and the primers indicated in Table , which produced increasingly smaller regions of the C-terminal end of PtlD (Fig. ). Each of these limited regions of ptlD was fused to DNA containing nucleotides 507 to 935 of ptxS1 and inserted into pUFR047. The fusion constructs were introduced into BP536ΔptlD, and production of PtlF was examined. As shown in Fig. (lanes 4 to 6), fusion constructs that encoded PtlD amino acids 405 to 463 (pAMC265.1), 415 to 463 (pAMC265.2), and 425 to 463 (pAMC265.3) were each capable of complementing the ΔptlD deletion since wild-type levels of PtlF were produced. The results demonstrated that a portion of PtlD limited to its 39 terminal amino acids can stabilize PtlF. Similar results were obtained when we examined the stability of PtlE and PtlH (data not shown).
Construction of pAMC265 subclones
FIG. 4. Complementation of BP536ΔptlD with plasmids encoding specific regions of the C-terminal end of PtlD. (A) Schematic diagram of fusion constructs encoding specific regions of the C-terminal end of PtlD. The designation of each fusion construct is (more ...)
We extended our deletion analysis and constructed plasmids containing regions encoding amino acids 392 to 454 (pAMC265.4), 392 to 444 (pAMC265.5), and 392 to 434 (pAMC265.6) of PtlD, each fused at the 5′ end to nucleotides 507 to 935 of ptxS1. Again, we found that each of these constructs was capable of complementing the ΔptlD deletion since wild-type levels of PtlF (Fig. , lanes 7 to 9), PtlE, and PtlH (data not shown) were observed.
The region common to the deletion constructs that we examined was limited to nucleotides encoding amino acids 425 to 434 of PtlD (Fig. ). In order to confirm that this 10-amino-acid stretch of PtlD is critical for stabilization of transporter proteins by the C terminus of PtlD, we constructed two more plasmids, pAMC265.9 and pAMC265.10 (Fig. ). Plasmid pAMC265.9 encoded amino acids 356 to 434 of PtlD, and plasmid pAMC265.10 encoded amino acids 356 to 424 of PtlD (both fused to nucleotides 507 to 935 of ptxS1
). Therefore, the latter construct did not produce amino acids 425 to 434 of the protein. Whereas the fusion construct encoding amino acids 356 to 434 of PtlD could complement the ΔptlD
deletion, the fusion construct encoding amino acids 356 to 424 of PtlD could not. These results indicate that a 10-amino-acid stretch of amino acids, amino acids 425 to 434 of PtlD, contributes to transporter stability. When we conducted a secondary structure analysis of this 10-amino-acid region using either GOR secondary structural prediction (5
) or PSIPRED secondary structural analysis (1
), we found that the analyses predicted that the first seven of these amino acids assume a helical structure. This helix might participate directly in protein-protein interactions with other transporter proteins.
To examine whether these 10 amino acids are critical for transporter stability when they are considered in the context of full-length PtlD, we created a mutant with an in-frame deletion of amino acids 425 to 434 of PtlD in wild-type strain BP536. We constructed this deletion mutant using PCR and primers 14, 14a, 15, and 15a to generate fragments flanking the region to be deleted. The fragments were ligated and introduced into pSS1129. The resulting plasmid was introduced into BP536 by conjugation. Deletion of the region occurred upon homologous recombination followed by selection with appropriate antibiotics (8
) and was verified by sequence analysis. When the levels of PtlF and PtlH in this deletion strain, BP536ptlD
Δ425-434, were analyzed, they were found to be the same as those in wild-type strain BP536 (data not shown). Taken together, our results indicate that while amino acids 425 to 434 contribute to the stability of the transporter, other portions of PtlD can compensate for the loss of these 10 amino acids.
A limited amount of information concerning the potential role of the C-terminal portion of PtlD homologues of the type IV transporter family is available. Jakubowski et al. (12
) proposed that the PtlD homologue of the VirB transporter of A. tumefaciens
, VirB6, along with VirB8 forms an inner membrane channel through which substrates traverse. The data of these workers suggest that the C-terminal 30 amino acids of VirB6 are necessary for substrate transfer to periplasmic and outer membrane-associated subunits of the transporter. Das and colleagues have demonstrated that the C-terminal 20 amino acids of VirB6 are essential for targeting of the protein to the cell pole (16
). They found that five proteins, VirB7 to VirB11, are required for polar localization of VirB6, suggesting that VirB6 likely interacts with a complex of other VirB proteins. Thus, they postulated that the C-terminal domain of VirB6 may participate in protein-protein interactions with other subunits of the VirB transporter. These results, along with the results reported here demonstrating that the C-terminal end of PtlD stabilizes a number of other transporter proteins, suggest that the extreme C-terminal ends of PtlD and its homologues may be involved in directly interacting with other proteins of the transport apparatus. This similarity in function occurs despite a lack of amino acid homology in this region. The extreme C-terminal end of PtlD is one region that displays little homology to VirB6. Homology analysis of PtlD and other VirB6 homologues, including those from Brucella
spp. and Bartonella
spp., indicated that the C-terminal regions of these proteins display limited amino acid homology when they are analyzed by using ClustalW analysis (data not shown). Nonetheless, our findings demonstrating that the C-terminal end of PtlD has a role in Ptl transporter stability, along with the findings of other workers suggesting that the C-terminal end of VirB6 has a role in VirB transporter interactions, suggest that the C-terminal ends of members of this family of proteins may have a general role in transporter protein interactions.