Transport of Preproteins by the Accessory Sec System Requires a Specific Domain Adjacent to the Signal Peptide

Abstract

The accessory Sec (SecA2/Y2) systems of streptococci and staphylococci are dedicated to the transport of large serine-rich repeat (SRR) glycoproteins to the bacterial cell surface. The means by which the glycosylated preproteins are selectively recognized by the accessory Sec system have not been fully characterized. In Streptococcus gordonii, the SRR glycoprotein GspB has a 90-residue amino-terminal signal sequence that is essential for transport by SecA2/Y2 but is not sufficient to mediate the transport of heterologous proteins by this specialized transporter. We now report that a preprotein must remain at least partially unfolded prior to transport by the accessory Sec system. In addition, a region of approximately 20 residues from the amino-terminal end of mature GspB (the accessory Sec transport or AST domain) is essential for SecA2/Y2-dependent transport. The replacement of several AST domain residues with glycine strongly interferes with export, which suggests that a helical conformation may be important. Analysis of GspB variants with alterations in the AST domain, in combination with the results with a SecY2 variant, indicates that the AST domain is essential both for targeting to the SecA2/Y2 translocase and for initiating translocation through the SecY2 channel. The combined results suggest a unique mechanism that ensures the transport of a single substrate by the SecA2/Y2 system.

The serine-rich repeat (SRR) glycoproteins of streptococci and staphylococci are a unique family of bacterial cell surface proteins. They are highly similar in domain organization, with an unusually long amino-terminal signal peptide followed by a short serine-rich region, an acidic or basic region, an extremely long serine-rich region, and a carboxy-terminal cell wall-anchoring motif (3, 18, 25, 31, 35, 38, 50). Despite these similarities, the SRR glycoproteins are a highly diverse family of adhesins, binding to a broad range of human tissues and proteins, including endothelial cells, epithelial cells, erythrocytes, platelet membrane proteins, salivary glycoproteins, and keratins (21, 28, 33, 46, 48, 52). A subset of the SRR proteins bind to distinct carbohydrate structures on their respective ligands and, thus, display lectin-like activity (1, 25, 39, 41, 42, 53). For several of the SRR proteins, the ligand-binding domain has been mapped to the region between the two serine-rich regions (28, 33, 41, 53). Many of the SRR glycoproteins have a significant impact on virulence (21, 23, 26, 31, 35, 40, 48, 51), and evidence suggests they are also important for biofilm formation (14, 18).

The SRR proteins are typically encoded in a chromosomal locus that also encodes components that mediate their posttranslational modification and transport to the bacterial cell surface. The addition of carbohydrate moieties to the two serine-rich regions occurs very rapidly, through the combined action of two or more enzymes (7, 21, 43, 54). The glycosylated preprotein is then transported to the cell surface via a dedicated transporter known as the accessory Sec (or SecA2/Y2) system (9, 21, 34, 44). This specialized transporter invariably includes SecY2 (the putative transmembrane channel) and the SecA2 ATPase (4), along with Asp1, Asp2, and Asp3. In some cases, the accessory Sec system also includes one or two short transmembrane proteins (Asp4 or Asp5) that are required for transport (45).

The reasons why a dedicated transporter is necessary for transporting these glycoproteins and the key differences between the SecA2/Y2 and canonical Sec (SecA/Y) systems are only partially understood. While the SecA2/Y2 systems may have evolved as more efficient transporters of glycosylated proteins, it is apparent that the SecA/Y systems of some organisms can transport SRR glycoproteins fairly readily. For example, the Staphylococcus aureus SRR SraP can be transported via SecA/Y during the late log phase of growth, and the accessory Sec system may be more important when cells are dividing rapidly (34). In most streptococci, however, the glycoproteins are not readily accommodated by the SecA/Y system, and transport is therefore more strictly dependent on SecA2/Y2. Whether this discrepancy is due to the extent of modification of the SRR glycoproteins or to the nature of the translocases is not yet known.

The means by which the SRR glycoproteins are so efficiently trafficked through SecA2/Y2, while other cell surface proteins are excluded from this pathway, are not entirely understood. To address this issue, we have previously used a truncated version of the SRR glycoprotein GspB (Fig. 1) as a model substrate for analyzing accessory Sec versus canonical Sec transport in Streptococcus gordonii. Use of this GspB variant has revealed some key requirements for transport by the SecA2/Y2 versus the SecA/Y system (2, 5). First, the carbohydrate moieties on GspB influence trafficking but do so only indirectly, in that they strongly impede SecA/Y transport. However, the carbohydrate is not necessary for recognition by the accessory Sec system, since nonglycosylated variants of GspB are readily transported by SecA2/Y2. Second, the 90-residue N-terminal signal peptide of GspB has a direct effect on trafficking. It is essential for SecA2/Y2 transport but is inefficient at facilitating transport via SecA/Y. Somewhat surprisingly, the extended N region of the signal peptide does not have an influence on the export route. Instead, three glycine residues in the hydrophobic core of the signal sequence interfere with export via the canonical pathway and concomitantly facilitate SecA2/Y2 transport. This is likely to be a common mode of trafficking to the SecA2/Y2 systems, as all of the SRR glycoproteins have two or three glycines in the hydrophobic core of the signal peptide. Reducing the number of glycines can reroute the preprotein so that it is trafficked more efficiently to SecA/Y. Whether the effect of the glycine residues on trafficking is due to a preferential targeting to SecA2 versus SecA, to preferential interaction with SecY2 versus SecY, or to both is not yet known.

FIG. 1.

Diagram of GspB736flag and the GspB::MalE fusions. GspB736flag is a C-terminally truncated and FLAG-tagged version of GspB that has been used as a model substrate for SecA2/Y2 in S. gordonii. MalE31 is a variant of MalE that has two amino acid substitutions which together affect the protein-folding kinetics (30). SP, signal peptide; AST, accessory Sec transport domain; SRR1 and SRR2, serine-rich repeat regions 1 and 2, respectively; BR, basic region.

Although the GspB signal peptide is essential for transport, it is not sufficient to mediate the export of other substrates by SecA2/Y2 in S. gordonii. In this report, we provide evidence that preprotein folding can prevent transport via SecA2/Y2 and that a heterologous protein can indeed be transported by SecA2/Y2 if it remains in a sufficiently unfolded state. In addition, we describe a domain of GspB adjacent to the signal peptidase cleavage site that is absolutely required for SecA2/Y2 transport. The region is apparently required both for targeting to the SecA2/Y2 translocase and for initiating translocation through the SecY2 channel. The same region is not necessary for transport by SecA/Y, which suggests that engagement of the SecA2/Y2 translocase is a more stringent process.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and plasmids used in this study are listed in Table 1. Oligonucleotide primer sequences are listed in Table 2. S. gordonii strains were grown in Todd-Hewitt broth (THB; Becton, Dickinson and Company) at 37°C in a 5% CO₂ environment. Transformed strains were initially selected by plating on sheep blood agar containing 15 μg/ml erythromycin or 100 μg/ml spectinomycin as required. Subsequent retention of the chromosomally integrated plasmids or DNA fragments did not require antibiotic selection.

TABLE 1.

Strains and plasmids used in the present study

Strain or plasmid	Relevant characteristicsa	Reference or source
Strains
M99	Streptococcus gordonii parental strain	37
PS516	M99 ΔsecA2; in-frame deletion	3
PS727	PS516 ΔgtfA::spec	2
PS846	M99 ΔgspB::pEVP3	5
PS998	PS516 ΔgspB::pEVP3	2
PS999	PS727 ΔgspB::pEVP3	2
Plasmids
pMAL-c2x	malE lacking the native signal sequence	New England Biolabs
pBluescript KS-	Amprori (E. coli)	Stratagene
pVA891	erm (Gram positive), cat ori (E. coli)	20
pS326	spec (Gram positive and E. coli), ori (E. coli)	44
pB736flagR	gspB736flag in pVA891	5
pB736flagRTn72	pB736flagR with TN::EZΔNotI at codon 72	5
pB736flagRTn80	pB736flagR with TN::EZΔNotI at codon 80	5
pB736flagRΔ8-68	pB736flagR with a deletion of the indicated codons	5
pB736flagRΔ9-79	pB736flagR with a deletion of the indicated codons	5
pB736flagRΔ81-191	pB736flagR with a deletion of the indicated codons	5
pBMalE	pB736flagR with codons 146 to 560 replaced by malE	This study
pBMalE31	pB736flagR with codons 146 to 560 replaced by malE31	This study
pB97M	pB736flagR with codons 98 to 736 replaced by malE31	This study
pB110M	pB736flagR with codons 111 to 736 replaced by malE31	This study
pB111M	pB736flagR with codons 112 to 736 replaced by malE31	This study
pB117M	pB736flagR with codons 118 to 736 replaced by malE31	This study
pB129M	pB736flagR with codons 130 to 736 replaced by malE31	This study
pB736flag/96X	pB736flagR/H96L	This study
pB736flag/96X118B	pB736flagR/H96L/A118G/A119S	This study
pB736flagRΔ94-117	pB736flagR with a deletion of the indicated codons	This study
pB736flagRΔ118-131	pB736flagR with a deletion of the indicated codons	This study
pB736flagRΔ132-223	pB736flagR with a deletion of the indicated codons	This study
pB736flagRΔ98-102	pB736flag/96X with a deletion of the indicated codons	This study
pB736flagRΔ103-109	pB736flag/96X118B with a deletion of the indicated codons	This study
pB736flagRΔ112-117	pB736flag/96X118B with a deletion of the indicated codons	This study
pB736flagRΔ116-117	pB736flag/96X118B with a deletion of the indicated codons	This study
pB736flagR/E97A and others	pB736flag/96X118B with an alteration of the indicated codon	This study
pY2KO	3′ nss and 5′ asp1 in pS326	This study
pY2+flanks	3′ nss, secY2 and 5′ asp1 in pBluescript KS-	This study
pY2I382N	3′ nss, secY2(I382N) and 5′ asp1 in pBluescript KS-	This study

^acat, chloramphenicol resistance; spec, spectinomycin resistance; erm, erythromycin resistance; Amp^r, ampicillin resistance.

TABLE 2.

Primers used in the present study

Primer	Sequence (5′ to 3′)a	Restriction enzyme
MalEF1	GGGAGCTCAATCGAAGAAGGTAAACTGG	SacI
MalER	GGGAGCTCTCTGCGCGTCTTTCAGGGC	SacI
MalEsfF	GAGAAAGATACGGATCCTAAAGTCACCGTTb	BamHI
MalEsfR	AACGGTGACTTTAGGATCCGTATCTTTCTCb	BamHI
MalEF2	GGACTAGTAATCGAAGAAGGTAAACTGG	SpeI
B93M	AAACTAGTTCCTCTTCAGCATAAACAG	SpeI
B110M	AAACTAGTCCACGAGTTGCTAAAACATCCC	SpeI
B111M	AAACTAGTTCTCCACGAGTTGCTAAAACATCCC	SpeI
B117M	AAACTAGTTCTTCGCTTAAAACCGCTTCTCC	SpeI
B129M	AAACTAGTGGATTGGCTTCTGTGGATGAC	SpeI
d3F	GCTGAAGAGGAAGCAGCAACTACTTTGTCATCCAC
d3R	AGTAGTTGCTGCTTCCTCTTCAGCATAAACAGAA
d4F	AGCGAAGAAGGATCCTTGTCAGATACTTTGTCTb	BamHI
d4R	TGACAAGGATCCTTCTTCGCTTAAAACCGCTTCb	BamHI
d5F	CCAGTGGAAGGATCCGAAAGTAGCCAACAGTCGb	BamHI
d5R	ACTTTCGGATCCTTCCACTGGATTGGCTTCTGTb	BamHI
96XF	GGAACAAGCACTCGAGAAAGTGATTGATACGAGGGb,c	XhoI
96XR	CCCTCGTATCAATCACTTTCTCGAGTGCTTGTTCCb,c	XhoI
118BF	GGGGATCCACTACTTTGTCATCCACAGAAGCC	BamHI
d6F	AGCACTCGAGAGGGATGTTTTAGCAACTCGTGG	XhoI
d7F	GGGGCTCGAGAAAGTGATTGATACGGGAGAAGCGGTT	XhoI
d7R	TTGGATCCTTCTTCGCTTAAAACCGCTTCTCCCGTATCAATCAC	BamHI
d8R	AAGGATCCTTCTCCACGAGTTGCTAAAACATC	BamHI
d9R	AAGGATCCGCTTAAAACCGCTTCTCCACGAGTTGC	BamHI
E97AF	GGAACAAGCACATGCAAAAGTGATTGATACGAGGG
E97AR	CCCTCGTATCAATCACTTTTGCATGTGCTTGTTCC
E97srF	GGAACAAGCACATGNNAAAGTGATTGATACGAGGG
E97srR	CCCTCGTATCAATCACTTTNNCATGTGCTTGTTCC
K98AF	AGCACTCGAGGCAGTGATTGATACGAGGG	XhoI
D101AF	AGCACTCGAGAAAGTGATTGCTACGAGGGAT	XhoI
R103AF	AGCACTCGAGAAAGTGATTGATACGGCGGATGTTTTAGCA	XhoI
R103AR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACGAGTTGCTAAAACATCCGCCGT	BamHI
D104AF	AGCACTCGAGAAAGTGATTGATACGAGGGCTGTTTTAGCA	XhoI
D104AR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACGAGTTGCTAAAACAGCCCTCGT	BamHI
L106GF	AGCACTCGAGAAAGTGATTGATACGAGGGATGTTGGAGCA	XhoI
L106GR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACGAGTTGCTCCAACATCCCTCGT	BamHI
A107GF	AGCACTCGAGAAAGTGATTGATACGAGGGATGTTTTAGGAACTCGTGGAGAAGCGG	XhoI
A107GR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACGAGTTCCT	BamHI
T108AR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACGAGCTGCTAAAA	BamHI
T108GF	AGCACTCGAGAAAGTGATTGATACGAGGGATGTTTTAGCAGGTCGTGGAGAAGCGG	XhoI
T108GR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACGACCTGCT	BamHI
R109AR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCAGCAGTTGCTAAAA	BamHI
R109GF	AGCACTCGAGAAAGTGATTGATACGAGGGATGTTTTAGCAACTGGTGGAGAAGCGG	XhoI
R109GR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTCCACCAGTTGCT	BamHI
G110AR	AAGGATCCTTCTTCGCTTAAAACCGCTTCTGCACGAGTTGC	BamHI
E111AR	AAGGATCCTTCTTCGCTTAAAACCGCTGCTCCACGAGTTGC	BamHI
E111GR	AAGGATCCTTCTTCGCTTAAAACCGCTCCTCCACGAGTTGC	BamHI
A112GR	AAGGATCCTTCTTCGCTTAAAACCCCTTCTCCACGAGTTGC	BamHI
V113GR	AAGGATCCTTCTTCGCTTAAACCCGCTTCTCCACGAGTTGC	BamHI
L114GR	AAGGATCCTTCTTCGCTTCCAACCGCTTCTCCACGAGTTGC	BamHI
L114XR	AAGGATCCTTCTTCGCTNNNAACCGCTTCTCCACGAGTTGC	BamHI
Y2I382NF	TGGTATGTTCATGAATTTTATTGGAATGGTb,c	BspHI
Y2I382NR	ACCATTCCAATAAAATTCATGAACATACCAb,c	BspHI

^aRestriction sites are underlined.

^bThe deletion or codon change introduces a restriction site.

^cIncludes a silent mutation.

Construction of MalE and MalE31 fusions with GspB.

The plasmid pB736flagR contains the entire gspB736flag gene sequence (GspB736flag is a variant of GspB that is truncated at residue 736 of 3,072 residues and has a C-terminal 3×FLAG tag), along with 1.2 kb of DNA from the chromosomal region upstream of gspB to facilitate integration into the normal chromosomal locus (5). For construction of the gspB736flag::malE fusion, malE was amplified from pMAL-c2x using the primers MalEF1 and MalER. The PCR product was digested with SacI and then ligated to pB736flagR that had been digested similarly. The gspB736flag::malE31 fusion was generated by a two-stage PCR procedure. In the first stage, primers 32F (5) and MalEsfR or primers MalEsfF and MalER were used to amplify the upstream or downstream segments, respectively. The two PCR products were combined for the second stage and then amplified using primers 32F and MalER. The PCR product was digested with SacI and then ligated to pB736flagR that had been digested similarly.

For fusion of MalE31 to the N-terminal end of GspB, the malE31 coding sequence was generated by two-stage PCR as described above except using the SpeI-linked upstream primer MalEF2 along with the M13 −40 primer. The fragment was digested with SpeI and XbaI and then ligated to pB736flagR that had been digested with SpeI. This removes codons 295 to 736 of gspB736flag but leaves the carboxy-terminal 3×FLAG coding sequence. The upstream NsiI-SpeI fragment, containing the 5′ end of gspB736flag through codon 294, was then replaced with shorter fragments that had been generated by PCR using primer 32F along with SpeI-linked reverse primer B93M, B110M, B111M, B117M, or B129M.

The plasmids were propagated in Escherichia coli, and the presence of only the intended changes was confirmed by DNA sequence analysis (Sequetech). Plasmids were then used to transform S. gordonii strain PS846 or PS998 as indicated.

Targeted deletions and substitutions in GspB736flag.

Three deletions in gspB736flag (codons 94 to 117 [Δ94-117], Δ118-131, and Δ132-223) were generated by a two-stage PCR procedure as described above, except using primer 32F along with d3R, d4R, or d5R for the first-stage upstream reactions and d3F, d4F, or d5F along with 24R (5) for the first-stage downstream reactions. The PCR products were digested with NsiI and HpaI and then used to replace the corresponding region of pB736flagR.

Two derivatives of pB736flagR were subsequently generated in order to facilitate the exchange of short regions of gspB736flag. Plasmid pB736flagR/96X was generated by a two-stage PCR procedure as described above, except using primers 32F and 96XR or 96XF and 24R for the first-stage upstream and downstream reactions, respectively. The PCR product was digested with NsiI and HpaI and then used to replace the corresponding region of pB736flagR. Plasmid pB736flagR/96X118B was generated by replacing the XhoI-HpaI fragment of pB736flagR/96X with an XhoI-BamHI and a BamHI-HpaI fragment which had been amplified by PCR using primers 96XF and d4R or 118BF and 24R, respectively, and then digested accordingly.

To construct pB736flagRΔ98-102, a DNA fragment was generated by PCR using the XhoI-linked primer d6F along with primer 24R. The PCR product was digested with XhoI and HpaI and then used to replace the XhoI-HpaI fragment of pB736flagR/96X. Three other deletions in gspB736flag (Δ103-109, Δ112-117, and Δ116-117) were generated by replacing the XhoI-BamHI fragment of pB736flagR/96X118B with one that had been amplified by PCR using primers d7F and d7R, 96XF and d8R, or 96XF and d9R, respectively.

The E97 codon in gspB736flag was replaced by a two-stage PCR procedure. In the first stage, primers 32F and E97R or primers E97AF and d4R were used to amplify the upstream or downstream segments, respectively. The two PCR products were combined for the second stage and then amplified using primers 32F and d4R. The PCR product was digested with NsiI and BamHI and then ligated to pB736flagR/96X118B that had been digested similarly. All other codon changes were made by a single PCR, using either an XhoI-linked forward primer (codons 98 and 101) along with d4R, an XhoI-linked forward primer along with a BamHI-linked reverse primer (codons 103 to 109), or primer 96XF along with a BamHI-linked reverse primer (codons 108 to 114). The PCR products were digested with XhoI and BamHI and then ligated to pB736flagR/96X118B that had been digested similarly.

The plasmids were propagated in E. coli, and the incorporation of only the intended changes was verified by DNA sequence analysis. Plasmids were then used to transform S. gordonii strain PS846, PS998, or PS999 as indicated.

Deletion of secY2 and construction of SecY2(I382N)-expressing strains.

M99 derivative strains that coexpress SecY2(I382N) along with a variant of GspB736flag were derived in two steps, with the end result being a markerless replacement of secY2 in the native chromosomal locus. For each GspB736flag variant strain, the wild-type secY2 gene was first replaced with a spectinomycin resistance cassette by transformation with pY2KO. Next, pY2I382N [which carries secY2(I382N) along with upstream nss and downstream asp1 flanks] was used to replace the spectinomycin resistance cassette. Transformants were plated on sheep blood agar without antibiotic selection and then scored for loss of spectinomycin resistance by replica plating on medium with or without spectinomycin. Spectinomycin-sensitive colonies [which had incorporated secY2(I382)] were typically obtained at a frequency of approximately 1 in 20 (5 × 10⁻²).

The ΔsecY2::spec cassette (pY2KO) was constructed by cloning the regions that normally flank secY2 in the chromosome (the 3′ half of nss or the 5′ half of asp1) upstream or downstream, respectively, of the spec gene in pS326. The secY2(I382N) replacement vector (pY2I382N) was constructed by two-stage PCR mutagenesis of pY2+flanks (pBluescript carrying the 3′ half of nss followed by secY2 and the 5′ half of asp1), using primers Y2I382NF and Y2I382NR along with standard M13 forward and reverse primers. A silent mutation that incorporates a BspHI restriction site was included in the primer sequence in order to facilitate tracking of the mutation (Table 2).

Analysis of secreted and nonsecreted proteins.

Proteins from S. gordonii culture supernatants or protoplasted cells were prepared and analyzed by Western blotting as described previously (2), with minor modifications. Strains were grown for 18 h in THB (the final cell density of all cultures was approximately 8 × 10⁸ CFU per ml). For analysis of secreted products, cells were removed by centrifugation at 16,000 × g for 5 min, and the spent medium was combined directly with SDS-PAGE sample buffer and boiled for 5 min. For analysis of nonsecreted proteins, protoplasts (generated by digestion of the cell wall with mutanolysin) were lysed by suspension in SDS-PAGE sample buffer, followed by boiling for 5 min. The proteins present in 20 μl of spent culture medium or in protoplasts recovered from 20 μl of the culture were loaded onto gels and compared by Western blotting for the relative amounts of the secreted versus nonsecreted proteins. Anti-FLAG monoclonal antibody (Sigma) was used at 1 μg/ml, along with peroxidase-conjugated anti-mouse antibodies (Sigma) at a 1:10,000 dilution. A 50-kDa cytoplasmic protein that cross-reacts with the anti-FLAG antibody served as a loading control (not shown in some cases).

Secondary structure predictions.

Secondary structure predictions were made using the Consensus Data Mining (combined GOR V and FDM) and Jpred algorithms (10, 32). Helical wheel diagrams and helical properties were determined by using HeliQuest (15).

RESULTS

Preprotein folding interferes with SecA2/Y2 transport.

In order to identify regions of GspB that were both necessary and sufficient for transport by the accessory Sec system, we had previously tested fusions of GspB (129 or more amino-terminal residues) with heterologous proteins for their ability to undergo transport by this pathway (5). However, none of the fusion proteins were successfully exported by the SecA2/Y2 system, indicating that the heterologous protein had inhibited GspB transport. One possible explanation for the inability of the SecA2/Y2 system to transport heterologous proteins is that specific interactions between the SecA2/Y2 translocase and the mature region of the preprotein are required for transport (12). Alternatively, preprotein folding prior to transport might block passage through the channel (6, 22).

To test the latter possibility, we employed a well-characterized slow-folding variant (MalE31) of the maltose binding protein MalE (30), which has been used in analyses of the AIDA-I autotransporter in E. coli (27). We used the MalE or MalE31 coding sequence to replace codons 146 through 560 of gspB736flag (Fig. 1) and then looked for transport via SecA2/Y2. Whereas no transport of native MalE was evident upon fusion to GspBflag (Fig. 2, lanes 1 and 4), the MalE31 slow-folding variant was readily transported (lanes 2 and 5). No transport of GspBflag::MalE31 was seen in the absence of SecA2 (lanes 3 and 6), indicating that transport of the slow-folding fusion protein occurs strictly via the accessory Sec system. The GspBflag::MalE31 preprotein, when not exported from the S. gordonii cytoplasm, appeared to be more degraded than GspB736flag::MalE (compare the protoplast fractions in lane 6 versus lane 4), which is consistent with the possibility that MalE31 folds more slowly in the cytoplasm of S. gordonii. Thus, the results indicate that a heterologous protein can be transported by the accessory Sec system, as long as it does not fold too rapidly.

FIG. 2.

SecA2-dependent transport of GspBflag::MalE31. Lanes contain proteins present in 20 μl of cultures grown for 18 h. The GspBflag::MalE and GspBflag::MalE31 fusion proteins were detected using anti-FLAG antibodies. The fusion proteins undergo glycosylation of the serine-rich regions, resulting in an apparent molecular mass of 110 to 120 kDa. Small amounts of the unglycosylated fusion proteins, which migrate at the predicted molecular mass of 80 kDa, are evident. Lanes 1 and 4, PS846 gspB736flag::pBMalE; lanes 2 and 5, PS846 gspB736flag::pBMalE31; lanes 3 and 6, PS998 gspB736flag::pBMalE31. wt (wild type), GspB::MalE; MalE31, GspB::MalE31.

GspB_1-117 linked to MalE31 is necessary and sufficient for SecA2/Y2 transport.

We next used MalE31 to determine whether the GspB signal peptide was sufficient to promote transport of the slow-folding heterologous protein via SecA2/Y2. A MalE31 fusion protein that included the GspB signal peptide along with three adjacent acidic residues (residues 1 to 93) (Fig. 3 A) showed no evidence of SecA2-dependent transport (Fig. 3B, lanes 1 and 2). Likewise, the signal peptide along with 20 residues from the mature region of GspB (residues 1 to 110) did not facilitate transport of MalE31 by SecA2/Y2 (lanes 3 and 4). However, upon the inclusion of 21 residues along with the 90-residue signal peptide (residues 1 to 111), some SecA2-dependent export was evident (lanes 5 and 6). SecA2/Y2 transport was most readily apparent upon the fusion of 117 or more residues from the amino terminus of GspB to MalE31 (lanes 7 to 10). The results indicate that transport of a heterologous protein by SecA2/Y2 requires a portion of the mature region of GspB in addition to the N-terminal signal peptide.

FIG. 3.

GspB_1-117 is sufficient to promote SecA2-dependent transport of MalE31. (A) N-terminal sequence of GspB. GspB has a tripartite signal peptide with an atypically long amino-terminal region (N), a hydrophobic core (H), and a signal peptidase cleavage region (C). Cleavage normally occurs between residues 90 and 91 (5). −, acidic residue; +, basic residue. (B) The C-terminally FLAG-tagged MalE31 coding sequence was fused in-frame to gspB at the indicated codon, and the resulting plasmids (pB93M-pB129M) were used to transform the S. gordonii strain PS846 (SecA2 positive) or PS998 (SecA2 negative). Lanes contain proteins present in 20 μl of culture medium after growth of the strains for 18 h. Note that the fusion proteins do not undergo glycosylation, since they lack the SRR1 and SRR2 regions. As a result, some transport of the fusion proteins via the canonical SecA/Y system can occur (5) and SecA2/Y2 transport is detected as an increase in transport in the presence of SecA2.

GspB_94-117 is essential for SecA2/Y2 transport and does not inhibit export via SecA/Y.

To verify the importance of the amino terminus of mature GspB in SecA2-dependent transport, we examined the effect of altering this region of GspB736flag. The deletion of 24 residues (residues 94 to 117) that are adjacent to the signal peptidase cleavage site abolished transport (Fig. 4 A, lane 2), whereas deleting 14 residues (residues 118 to 131) preceding the SRR region 1 (SRR1) domain or deleting the SRR1 domain (residues 132 to 223) had a negligible effect (lanes 3 and 4, respectively).

FIG. 4.

The AST domain is essential for transport by SecA2/Y2 but not by SecA/Y. Lanes contain proteins present in 20 μl of cultures grown for 18 h. (A) SecA2-dependent transport. Plasmids carrying the indicated gspB736flag variations were used to transform the ΔgspB strain PS846. M, medium fraction; P, protoplast fraction. (B) SecA2-independent transport. In the absence of GtfA and SecA2, nonglycosylated gspB736flag is transported via SecA/Y (5). Plasmids carrying the indicated gspB736flag variants were used to transform the ΔgspB ΔsecA2 ΔgtfA strain PS999.

We next asked whether the 24 residues (94 to 117) were required specifically for SecA2/Y2 (versus SecA/Y) transport and whether they might interfere with SecA/Y transport (thereby indirectly facilitating SecA2/Y2 transport). As mentioned above, nonglycosylated GspB736flag (produced by expressing the protein in a ΔgtfA strain) can be exported inefficiently by SecA/Y (2, 5). We therefore examined whether the export of selected GspB736flag variants was increased or decreased compared with the export of wild-type GspB736flag when expressed in a ΔgtfA ΔsecA2 background. Deletion of residues 94 to 117, which completely abolished SecA2-dependent transport, neither abolished nor significantly increased transport by SecA/Y (Fig. 4B, compare lanes 2 and 6 with lanes 1 and 5). A larger deletion spanning the same region (residues 81 to 191 [Δ81-191]), which also abolished SecA2-dependent export (5), had no apparent effect on SecA2-independent transport (Fig. 4B, lanes 3 and 7). In addition, an insertion of 19 amino acids adjacent to the hydrophobic core of the signal peptide (Tn80) which severely reduced SecA2-dependent export (5) (see below) had little effect on SecA/Y transport (Fig. 4B, lanes 4 and 8). The combined data therefore confirm that a region of GspB adjacent to the signal peptide is required specifically for SecA2/Y2 transport and does not simply interfere with SecA/Y transport. Thus, the region will be referred to as the accessory Sec transport or “AST” domain.

Characteristics of the AST domain.

A series of smaller deletions and amino acid substitutions was made in GspB736flag to further define the essential features of the AST domain. To facilitate exchange within this short region, restriction sites were incorporated at codons 96 and 118. The resulting GspB736flag variant (96X118B), despite having three codon alterations (H96L/A118G/A119S), showed export comparable to that of GspB736flag (Fig. 4A, lane 5 versus lane 1). The deletion of residues 98 to 102 led to a severe reduction in transport (lane 6), and the deletion of residues 103 to 109 (lane 7) or 112 to 117 (lane 8) completely abolished transport. Alanine-scanning mutagenesis of selected charged and polar residues within this region indicated that E97 was important for export, since the replacement of this residue resulted in a substantial decrease in export, with a simultaneous accumulation of the preprotein in the protoplasts (Fig. 5 A, lane 2). Replacement of the other charged or polar residues (Fig. 5A, lanes 3 through 10) or the deletion of two acidic residues (Fig. 4A, lane 9) had little or no effect on export. Thus, although the AST domain has a high net negative charge, this property does not appear to be critical for transport.

FIG. 5.

Identification of key residues in the AST domain. (A) Results of alanine replacement of selected residues. (B) Results of glycine-scanning mutagenesis of selected residues. (C) Results of replacement of E97 and L114. Plasmids carrying the indicated gspB736flag codon substitution were used to transform the ΔgspB strain PS846. Lanes contain proteins present in 20 μl of cultures grown for 18 h. M, medium fraction; P, protoplast fraction.

The secondary structure predictions of the GspB preprotein indicated that the AST domain might form an α-helix (data not shown). Alanine substitutions within this region did not drastically alter the predicted structure. We therefore generated a series of substitutions by using the helix-destabilizing residue glycine in order to determine whether the structure of the AST domain was likely to be important. In this case, the replacement of several residues led to a substantial (L106 or R109) or severe (L114) decrease in export (Fig. 5B, lanes 2, 5, and 9, respectively). The results support the possibility that a helical conformation of the AST domain is essential for transport by SecA2/Y2.

To further characterize the essential features of the AST domain, additional amino acids were substituted for either E97 or L114. For E97, this included either a conservative substitution or replacement with a helix-destabilizing residue. For L114, the substitutions included an assortment of polar and nonpolar amino acids. As seen by the results in Fig. 5C, aspartic acid could only partially substitute for the glutamic acid residue (E97), since the substitution resulted in a decrease in the amount transported into the culture medium along with an accumulation of the preprotein in the protoplasts (lane 2). Replacement of the same residue with glycine had a stronger negative impact on transport (lane 3). The replacement of L114 with highly hydrophobic residues, such as phenylalanine, isoleucine, or valine, had no adverse effect (lanes 5, 7, and 10, respectively), whereas substitution with alanine led to a partial reduction in transport (lane 4). The replacement of L114 with histidine, proline, or serine resulted in a severe reduction in transport (lanes 6, 8, and 9, respectively). These results indicate that a glutamic acid residue at position 97 and a highly hydrophobic residue at position 114 are essential requirements of the AST domain.

SecY2(I382N) can transport GspB variants with defective signal peptides.

Studies of the Sec system in E. coli have indicated that certain variants of SecY are capable of transporting preproteins that have defective signal peptides. One of the most extensively studied SecY mutants is the prlA4 (protein localization) mutant (13), which has an asparagine substitution for an isoleucine in the 10th transmembrane helix (I408N). The isoleucine residue is one of five that form a constriction known as the “pore ring,” which seals the channel when it is not engaged in protein transport (17, 29, 47). The I408N substitution destabilizes the closed state of the channel and is thought to override a gating mechanism that is usually facilitated by signal peptide intercalation between the second and seventh transmembrane segments of SecY (19, 36).

We reasoned that an analogous variant of SecY2 [SecY2(I382N)] could be used to distinguish gating defects [as evidenced by increased export of the variant preprotein from a strain expressing SecY2(I382N) compared with the export by wild-type SecY2] from defects in targeting (i.e., delivery to the translocase, as evidenced by no increase in export). We first examined whether secY2(I382N) resulted in a phenotype equivalent to that of the prlA4 mutation of SecY. As seen by the results in Fig. 6 A, SecY2(I382N) (lane 2) transports GspB736flag comparably to the wild-type SecY2 (lane 1). Moreover, it can more readily transport GspB736flag variants with at least some types of signal peptide defects (lanes 3 to 10). That is, GspB736flag variants with a large deletion in the N region (Δ8-68), an insertion of 19 residues between the N and H regions (Tn72), or an insertion of 19 residues between the H and C regions (Tn80), were all exported more readily by the SecY2(I382N)-expressing strain. However, a variant of GspB736flag that has a deletion of nearly the entire signal peptide (Δ9-79) was not exported, which suggests that this variant was not targeted to the SecA2/Y2(I382N) translocase. Thus, SecY2(I382N), like E. coli SecY(I408N), results in a prl phenotype and can be used to detect gating defects.

FIG. 6.

Export of GspB736flag signal peptide (A) or AST domain (B) variants by SecY2(I382N) (IN) or SecY2 (WT). Plasmids carrying the indicated gspB736flag alterations were used to transform the ΔgspB strain PS846. The chromosomal secY2 gene in each of these strains was subsequently replaced with secY2(I382N) as described in Materials and Methods. Increased transport of any GspB variants by SecY2(I382N) in comparison to transport by SecY2 indicates that the variant is properly targeted to the translocon but is unable to initiate translocation. A lack of increased export by SecY2(I382N) in comparison to the transport by SecY2 is suggestive of a targeting defect. Lanes contain proteins present in 20 μl of cultures grown for 18 h. M, medium fraction; P, protoplast fraction.

SecY2(I382N) can transport a subset of the AST domain variants.

We next sought to assess the extent to which SecY2(I382N) could transport GspB variants with alterations in the AST domain. As seen by the results in Fig. 6B, two of the variants (E97G and Δ98-102), although inefficiently transported by wild-type SecY2, were transported relatively well by SecY2(I382N) (lanes 1 to 4). These variants were therefore not defective at targeting to SecA2/Y2 but, rather, were defective in gating. Two other variants, Δ81-191 and Δ112-117, showed no transport by SecY2(I382N) (data not shown) and were therefore apparently not targeted to SecA2/Y2. All other AST domain variants that were examined (including Δ103-109, L106G, and L114G) (Fig. 6B, lanes 5 to 10) showed only a very slight increase in transport by SecY2(I382N), indicating that these regions may be involved in targeting, as well as gating. Thus, the combined results indicate that the more distal portion of the AST domain (residues 103 to 115) and an intact signal peptide are required for targeting to SecA2/Y2, whereas the more proximal portion of the AST domain (residues 91 to 102) and the distance between the AST region and the hydrophobic core of the signal peptide (Tn80) are essential for gating.

DISCUSSION

The studies presented here describe a region of the SRR glycoprotein GspB that is specifically required for transport by SecA2/Y2. The essential part of the preprotein, the AST domain, corresponds to approximately 20 residues from the amino-terminal end of the mature protein. When combined with a functional signal peptide, the AST region is necessary and sufficient to promote the transport of an unfolded heterologous protein through the specialized SecA2/Y2 transporter. The AST domain does not appear simply to interfere with the canonical Sec system components and, thus, has a positive role in SecA2/Y2-dependent transport.

The precise requirements for the structure and composition of the AST region of GspB have not been fully defined. The replacement of E97 with alanine resulted in decreased transport of GspB736flag, but the replacement of other charged residues with alanine had little or no effect on the export of GspB736flag. Likewise, the deletion of two of the acidic residues (E116 and E117) had no apparent impact on transport. This suggests that, although the AST domain has a high net negative charge, this property is not critically important for the transport process. Conversely, the replacement of several residues (E97, L106, and L114) with glycine or proline had a strong impact on GspB736flag export, which suggests that a helical conformation of the AST domain is important. L114 could be replaced with a highly hydrophobic residue (phenylalanine, isoleucine, or valine) with no negative impact on transport, but its replacement with less hydrophobic residues (alanine, histidine, or serine) resulted in a severe reduction in transport (Fig. 5C). These findings suggest that the AST domain may form an amphipathic helix, with a glutamic acid residue at the seventh position from the signal peptidase cleavage site.

An apparent discrepancy exists between the results of the MalE fusion and GspB736flag deletion analyses regarding the essentiality of residues 112 to 114. GspB736flag missing these residues was not transported at all, whereas GspB111::MalE31 (which lacks these residues) showed some SecA2-dependent transport. However, inspection of the GspB111::MalE31 fusion protein sequence indicates that the N-terminal end of MalE is similar to residues 112 to 114 (LVI versus AVL), so the distal portion of the putative amphipathic helix would be maintained.

The function of the AST domain in trafficking appears to be 2-fold. First, it has a profound impact on targeting of the preprotein to the translocase. Many of the AST region variants were at best only slightly more efficiently transported by SecY2(I382N) than by wild-type SecY2 (Fig. 6B and data not shown), which suggests that they were inefficiently targeted to the translocon. Whether the AST domain directly interacts with SecA2 or some other targeting factor (such as the Asps) remains to be determined. Second, the AST domain is also essential for gating of the translocon. Two alterations, E97G and Δ98-102, had a detrimental impact on transport but did not strongly interfere with targeting, since transport of these variants by SecY2(I382N) was relatively efficient (Fig. 6B, lanes 2 and 4). The combined results indicate that there may be specific contacts between the proximal portion of the AST domain (residues 97 to 102) and the translocase (SecA2, SecY2, or both) that are essential for the initiation of translocation.

The finding that a heterologous protein can be exported if coupled to the GspB signal peptide and the AST domain also indicates that there are not likely to be specific contacts between the translocase and the remainder of the mature region of the preprotein during transport. As long as the preprotein is sufficiently unfolded and has the AST domain, it can be transported. Thus, the main factor affecting the exclusivity of the accessory Sec system may be that other preproteins lack the AST domain. In addition, the SecA2/Y2 translocase may be less able than the SecA/Y translocases to unfold partially folded substrates (22).

It is not yet known whether the AST domain is an essential feature of the SecA2/Y2 substrates in general. An alignment of the SRR glycoprotein sequences indicates that the AST domain (or putative N-terminal end of the mature protein) is less conserved than the signal peptide. Despite the lack of sequence similarity, at least some of these putative AST domains are predicted to form amphipathic helices (data not shown). Moreover, although the essential domains have not been fully defined, portions of the mature region of the SRR glycoprotein Fap1 are required for SecA2-dependent transport in Streptococcus parasanguinis (8). This supports the possibility that an amino-terminal AST domain may be required for SecA2/Y2 transport in other organisms.

It would be ideal to compare the mechanisms of trafficking through the accessory Sec system (targeting and gating) with trafficking through the canonical Sec system. However, little is known about posttranslational trafficking of preproteins through the general Sec system in Gram-positive bacteria. In the Gram-negative bacterium E. coli, the SecB chaperone binds to the mature region of preproteins and targets them to SecA (reviewed in reference 11). Preproteins are also targeted to the Sec system through the direct binding of the signal peptide domain to SecA. The preprotein is then delivered to SecY, where the signal peptide appears to trigger channel opening by intercalating between two transmembrane segments of SecY (24, 49). For the accessory Sec system of S. gordonii, the signal peptide and the AST domain are both important for targeting to the translocase and for gating of the channel. Separation of the hydrophobic core of the signal peptide and the AST domain by 19 amino acids (Tn80) resulted in a gating defect (Fig. 6A, lanes 7 and 8), which suggests that the signal peptide and AST domain need to be adjacent for proper gating but can be separated without a detrimental effect on targeting. This is consistent with the recent report that the signal peptide and mature region of preproteins can bind separately to E. coli SecA but contrasts with the finding that the signal peptide can act in trans to trigger channel opening and the transport of SecA-bound mature regions (16). For SecA2/Y2, the signal peptide and a specific portion of the mature region must be contiguous for gating. It may be that the GspB signal peptide and AST region act as a two-pronged wedge and that both are necessary to open the SecY2 channel. Future experiments will address which components of the accessory Sec system directly contact the AST domain, why this region is essential for targeting, and how it facilitates gating of the SecY2 translocon.