Membrane Topology of the Streptomyces lividans Type I Signal Peptidases

Abstract

Most bacterial membranes contain one or two type I signal peptidases (SPases) for the removal of signal peptides from export proteins. For Streptomyces lividans, four different type I SPases (denoted SipW, SipX, SipY, and SipZ) were previously described. In this communication, we report the experimental determination of the membrane topology of these SPases. A protease protection assay of SPase tendamistat fusions confirmed the presence of the N- as well as the C-terminal transmembrane anchor for SipY. SipX and SipZ have a predicted topology similar to that of SipY. These three S. lividans SPases are currently the only known prokaryotic-type type I SPases of gram-positive bacteria with a C-terminal transmembrane anchor, thereby establishing a new subclass of type I SPases. In contrast, S. lividans SipW contains only the N-terminal transmembrane segment, similar to most type I SPases of gram-positive bacteria. Functional analysis showed that the C-terminal transmembrane anchor of SipY is important to enhance the processing activity, both in vitro as well as in vivo. Moreover, for the S. lividans SPases, a relation seems to exist between the presence or absence of the C-terminal anchor and the relative contributions to the total SPase processing activity in the cell. SipY and SipZ, two SPases with a C-terminal anchor, were shown to be of major importance to the cell. Accordingly, for SipW, missing the C-terminal anchor, a minor role in preprotein processing was found.

Most bacterial proteins exported by the general secretion pathway (Sec pathway) are synthesized in the cytoplasm as preproteins containing a signal peptide. The signal peptide is required for targeting the precursor proteins to the translocase (13, 35). Upon translocation, a type I signal peptidase (SPase) (10, 11) removes the signal peptide. This is required for the release of the mature protein from the plasma membrane (9). Type I SPases are divided in two subfamilies, prokaryotic (P)-type and eukaryotic endoplasmic reticulum (ER)-type SPases (40). P-type SPases contain conserved serine and lysine residues which function as a catalytic serine/lysine dyad (41, 42). In ER-type SPases, the lysine residue is replaced by a histidine residue.

Based on computer predictions for topological characterization, SPases can be divided into four groups. The first group comprises SPases with one N-terminal anchor. This group contains most type I SPases from gram-positive bacteria, the SPase from the gram-negative species Bradyrhizobium japonicum (SipS [28]), an SPase from the mitochondria (Imp1p [2]), and finally, the catalytic subunits of the ER-type SPases (26). SPases having two N-terminal anchors are classified into the second group. They were identified only in gram-negative bacteria. Haemophilus influenzae SPase (16) is the only known member of the third group, harboring SPases with three N-terminal anchors. In the last group, SPases with both an N-terminal and a C-terminal anchor are classified, e.g., Imp2p from the yeast mitochondria (29), the SPase of the gram-negative bacterium Rhodobacter capsulatus (22), and SipW of Bacillus subtilis (40), the only ER-type SPase identified in bacteria.

Streptomyces bacteria are gram-positive soil bacteria characterized by their mycelial growth and the production of vegetative exospores. Members of this genus produce over 60% of the naturally occurring antibiotics, as well as several industrially important hydrolytic enzymes. Thanks to its excellent secretion capacity, Streptomyces, and Streptomyces lividans in particular, has been intensively investigated for the secretory production of heterologous proteins (1, 3, 17, 25, 45). Recently, four different type I SPases in S. lividans were identified (31, 32, 37). Transcriptional analysis revealed that three genes (sipW, sipX, sipY) constitute an operon, while the fourth (sipZ) is the first gene of another operon together with three unrelated genes (32). Only B. subtilis was shown to have more (five chromosome-encoded and two plasmid-encoded) type I SPases (40).

In this report, the membrane topology of the S. lividans SPases was experimentally analyzed. Therefore, the monomeric tendamistat protein (TM) (47) was used as a novel reporter protein in protease protection experiments. Interestingly, TM fusion analyses confirmed the computer-based topology with both an N- and C-terminal transmembrane anchor for SipY, a unique feature for P-type type I SPases of gram-positive bacteria. Functional analysis demonstrated the importance of this domain in the processing efficiency. In addition, SipY and SipZ, both containing the C-terminal anchor, were found to be determining factors in the total processing capacity of the cell. The membrane topology of S. lividans SipW was experimentally proven to be similar to the predicted membrane topology of most type I SPases of gram-positive bacteria.

MATERIALS AND METHODS

Bacterial strains and growth conditions

For cloning purposes, Escherichia coli TG1 cells were used and grown at 37°C (300 rpm) in Luria broth (LB) (27) in the presence of ampicillin (50 μg/ml). For purification purposes, sip genes were expressed in E. coli BL21(DE3)pLysS (38) grown in LB in the presence of chloramphenicol (25 μg/ml) and ampicillin (50 μg/ml). For leader peptidase (Lep) complementation assays, E. coli IT89 cells (20), which were grown in LB supplemented with 50 μg of ampicillin/ml and 5 μg of tetracycline/ml, were used. For solid media, 15 g of agar was added per liter.

S. lividans TK24 cells were grown at 27°C with continuous shaking at 300 rpm in phage medium (23) or Nutrient Broth 2 (NB) (Lab M, Bury, United Kingdom) buffered at pH 7.2 with 0.05 M MOPS (3-[N-morpholino]propanesulfonic acid) (44). When necessary, kanamycin (50 μg/ml) was added. Spore isolation, protoplast formation, and transformation of S. lividans were carried out as described by Hopwood et al. (19).

DNA techniques

All DNA manipulations used in this work were performed by standard procedures (36). Restriction endonucleases and other DNA-modifying enzymes were purchased from Roche Diagnostics (Mannheim, Germany), Eurogentec (Seraing, Belgium), and Life Technologies (Gaithersburg, Md.). Oligonucleotides (Table 1) were obtained from Amersham Pharmacia Biotech (Rainham, United Kingdom). All plasmids used are listed in Table 2. PCR was carried out with Pfu polymerase (Stratagene Europe, Amsterdam, The Netherlands). DNA sequence analysis was performed according to the dideoxy chain termination method with a Thermo sequenase fluorescent labeled primer cycle sequencing kit with 7-deaza-dGTP (Amersham Pharmacia Biotech). Random primed DNA labeling with digoxigenin-11-dUTP (DIG) was performed as previously described (18).

TABLE 1.

Oligonucleotides used in this study

Oligonucleotide	Sequencea	Restriction site(s)
MaTM5′	5′-GGGCCGGCCTCCGTCGACACGACGACC-3′	HincII
TM3′	5′-TAGAATTCGGGCTAAAGGCAGCGAG-3′	EcoRI
Sipy5′	5′-TACATATGGGGGATGTGGCGGTTGGGGC-3′	NdeI
SipyHIS5′	5′-TACATATGCACCATCATCACCATCACGGGGATGTGGCGGTTGGGGC-3′	NdeI
Sipy260	5′-TACCCGGGCAGCTCCTTCCAGAA-3′	SmaI
Sipy900	5′-ATGCCGGCGGCCGAGGATCGGGCCTGC-3′	NaeI
Sipy3′	5′-TAAGGCCTGGTCTCGGCCGGGGCGG-3′	StuI
Sipw5′	5′-TACATATGGACACCGAAACACAGCACACG-3′	NdeI
Sipw3′	5′-TAATGCATCCATCGGCGTCCCCGTCGGG-3′	NsiI
Sipy01	5′-ATGGATCCTCAGCGGGTCTCGGCCGGG-3′	BamHI
Sipy02	5′-AGGATCCTCAGGCGGCCGAGGATCGGGCC-3′	BamHI
DR-1	5′-AGTACCATGGGCGCATATGGCTGTTTCCTGTGTGAAATTG-3′	NcoI, NdeI
DR-2	5′-AGTCCCATGGCGCGGATCCGGTGTGAAATACCGCACAGATGC-3′	NcoI, BamHI

^aRestriction cleavage sites are underlined. 

TABLE 2.

Plasmids used in this work

Plasmid	Relevant properties	Reference or source
E. coli
pGEM-T Easy	Used for cloning of PCR amplified DNA fragments, Apr	Promega
pDR540	Contains the galK gene under control of the tac promoter, Apr	Accession no. U13847
pSN40	pRM1 derivative (30) containing an S. lividans TK21 chromosomal BamHI fragment encoding the four SPases, Apr	32
pEX50	pDR540 derivative, tac promoter, Apr	This work
pET3a	T7 promoter, Apr	Novagen
pEX20	pET3a derivative encoding hexahistidine-tagged SipY336	This work
pEX23	pET3a derivative encoding hexahistidine-tagged SipY304	This work
pTM	pGEM-T Easy derivative containing the gene encoding mature tendamistat	This work
pSipTM	pTM derivative, contains different SPase-TM fusion genes	This work
pEXSipY336	pEX50 derivative encoding wild-type SipY	This work
pEXSipY304	pEX50 derivative encoding SipY derivative lacking the C-terminal anchor	This work
Streptomyces
pAX5a	E. coli-Streptomyces shuttle vector encoding pre-tendamistat, Apr Tsr	15
pIJ6021	Contains a multiple cloning site after a tipA promoter, Kmr	39
pIJSipTM	pIJ6021 derivative encoding the different SPase-TM fusion proteins	This work

Colony hybridization

S. lividans colonies were grown on nylon filters (Hybond N+; Amersham Pharmacia Biotech) placed on agar plates containing kanamycin (50 μg/ml) at 27°C for 2 days. The filters (colony side up) were first transferred onto five stacked sheets of Whatman 3MM filter paper soaked for 15 min in a solution containing 4 mg of lysozyme/ml in Tris-EDTA buffer (pH 7.8), then transferred onto five sheets of paper soaked in 1.5 M NaCl–0.5 M NaOH for 20 min, and finally transferred onto five sheets of paper soaked in neutralization solution (0.5 M Tris-HCl [pH 7.2], 1.5 M NaCl, 1 mM EDTA) for 5 min. This step was repeated three times. After washing with a 2× SSC solution (0.3 M NaCl, 0.03 M Na₃citrate), the filters were dried and the DNA was cross-linked to the membrane by UV radiation. Subsequently, the fixed DNA was hybridized with a tendamistat fragment labeled with DIG according to a method described by Engler-Blum et al. (14). The chemiluminescent detection procedure was performed as described by Hoeltke et al. (18).

Construction of expression plasmids

To express the SPase-TM fusions in S. lividans, expression vectors were constructed by inserting the coding sequence for the desired fusion protein in the pIJ6021 plasmid downstream from the tipA promoter. Using pSN40 (Table 2) as a template, DNA fragments coding for the SPases or N-terminal fragments thereof were generated by PCR. Primers were designed for cloning these fragments upstream of the tendamistat reporter gene in pTM (Fig. 1) in such a way that a fusion protein was obtained with tendamistat. The appropriate PCR products were treated with NdeI-SmaI for SipY86, with NdeI-NaeI for SipY304, with NdeI-StuI for SipY336, or with NdeI-NsiI (blunted with T4 DNA polymerase) for SipW259 and were cloned into the NdeI-HincII sites of pTM, resulting in pY86TM, pY304TM, pY336TM, and pW259TM, respectively. Finally, the NdeI-EcoRI fragments of pY86TM, pY304TM, pY336TM, and pW259TM were cloned into the corresponding sites of the S. lividans pIJ6021 plasmid, resulting in pIJY86TM, pIJY304TM, pIJY336TM, and pIJW259TM, respectively, in such a way that the fusion proteins were under the control of the thiostrepton inducible promoter tipA (39).

FIG. 1.

Schematic representation of pEX50 and pTM. (A) pEX50 was constructed with the primers DR-1 and DR-2 by using plasmid pDR540 DNA. After cleavage with NcoI, the amplified fragment containing the introduced restriction sites for NdeI, NcoI, and BamHI was self-ligated, resulting in the removal of the galK gene. (B) pTM was constructed by inserting the gene encoding the mature part of tendamistat, which was amplified by PCR with the primers MaTM5′ and TM3′ by using pAX5a as a template, into the pGEM-T Easy plasmid. The resulting vector was used for the construction of SPase-tendamistat gene fusions (SipTM).

To construct pEX20, encoding the N-terminally hexahistidine-tagged SipY336 protein, a PCR with the primers SipyHIS5′ and Sipy01 was performed using the plasmid pSN40 as a template. The amplified fragment was subsequently cut with NdeI-BamHI and ligated into the corresponding sites of pET3a. pEX23, encoding the N-terminally hexahistidine-tagged SipY304 protein, was similarly constructed using the primers SipyHIS5′ and Sipy02.

pEXSipY336 and pEXSipY304, used for Lep complementation assays, were constructed by inserting into the corresponding sites of pEX50 (Fig. 1) an NdeI-BamHI PCR amplified fragment encoding the respective SPase. This fragment was amplified using the primers Sipy5′-Sipy01 for SipY336 and Sipy5′-Sipy02 for SipY304 with pSN40 as a template.

Expression of SPase-TM fusion proteins in S. lividans.

Five milliliters of liquid phage medium was inoculated with 4 × 10⁶ S. lividans spores. The mycelium from this preculture, grown for 3 days at 27°C, was fragmented in a glass homogenizer, and 1 ml of homogenate was used to inoculate 50 ml of NB medium. After 12 h of growth, the expression of the fusion proteins was induced by the addition of 5 μg of thiostrepton (Calbiochem, San Diego, Calif.)/ml (39) and incubation was continued for 24 h. A 1-ml sample was centrifuged (5 min, 2,000 × g) to harvest the mycelium. The cell pellet was subsequently resuspended in 1 ml of ice-cold 50 mM Tris-HCl, pH 7.5, and cell lysates were prepared by sonication on ice (three times for 30 s each time; 100 W Ultrasonic disintegrator [Measuring & Scientific Equipment, Ltd., London, United Kingdom]).

Expression and purification of SipY304 and SipY336

A 500-ml culture of E. coli BL21(DE3)pLysS cells harboring pEX20 or pEX23 was grown until it reached an optical density at 600 nm (OD₆₀₀) of 0.6. Expression of the Sip proteins was then induced by the addition of IPTG (isopropyl-β-d-thiogalactopyranoside) to a final concentration of 1 mM. Growth of the culture was continued for 3 h, after which the cells were harvested by centrifugation (10 min, 5,000 × g). The collected cells were resuspended in 25 ml of lysis buffer (50 mM Tris-HCl, pH 8, 20% sucrose, 10% glycerol). Cells were lysed by passing them three times through a French pressure cell at 15,000 lb/in², and the cell debris was removed by centrifugation (20 min, 12,000 × g). To isolate the membrane fraction, the cell lysate was centrifuged for 2 h at 100,000 × g. Subsequently, the pellet was resuspended in 5 ml of solubilization buffer (50 mM NaH₂PO₄, pH 8, 300 mM NaCl, 10% glycerol, 0.5% Triton X-100). After 1 h on ice, the sample was recentrifuged (1 h, 100,000 × g) and the extracted membrane proteins were loaded onto a 5-ml Ni-nitrilotriacetic acid column equilibrated with solubilization buffer. The column was washed with 5 ml of solubilization buffer plus 10 and 20 mM imidazol, successively. Fractions of 1 ml were collected and analyzed for purity on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels.

Isolation of S. lividans membranes.

After homogenization of the mycelium of a 5-ml S. lividans preculture in phage medium, 1 ml was used to inoculate 50 ml of NB medium which was incubated for an additional 24 h. The mycelium was harvested by centrifugation (5 min, 1,000 × g), and the cells were lysed in a French pressure cell (three times, 15,000 lb/in²). After removal of the cell debris by centrifugation (20 min, 12,000 × g), the cell lysate was centrifuged for 2 h at 100,000 × g. The cell pellet was resuspended in 50 mM Tris-HCl (pH 8)–5 mM MgSO₄ containing complete Mini EDTA-free protease inhibitor cocktail tablets (Roche Diagnostics) and normalized for protein content (10 mg/ml) (Bio-Rad protein assay).

Preparation of anti-tendamistat and anti-SipY antibodies.

Purified tendamistat (250 μg in 500 μl), kindly provided by J. W. Engels, was mixed with 500 μl of complete Freund's adjuvant and intramuscularly injected (two 1-ml injections at a 3-week interval) in a Hollander rabbit. Two weeks after the last injection, blood was sampled and the serum was collected by centrifugation (5 min, 150 × g). Because of the high background observed, the anti-tendamistat antibodies were further purified on a 1-ml HiTrap N-hydroxysuccinimide-activated column with the antigen tendamistat covalently bound to the matrix by using the ÄKTA prime purification system (Amersham Pharmacia Biotech). Coupling of tendamistat to the matrix was performed according to the manufacturer's recommendations. After dialysis of the serum against binding buffer (75 mM Tris-HCl, pH 8; overnight at 4°C), the soluble fraction was transferred to the prepared column. Following this, the column was washed with 10 column volumes of binding buffer. Bound anti-tendamistat antibodies were eluted with a solution of 100 mM glycine-HCl–500 mM NaCl (pH 2.7) and collected in 1-ml fractions. Elution fractions were neutralized with 1 N NaOH and analyzed by SDS-PAGE. Specific anti-SipY antibodies were prepared analogously to the described protocol.

Western blotting and protein detection

Proteins were separated by SDS-PAGE as described by Laemmli (24). Transfer of proteins onto a nitrocellulose Porablot membrane (Macherey Nagel, Düren, Germany) was performed with the aid of a Transblot semidry transfer cell (Bio-Rad) according to the manufacturer's recommendations. The fusion proteins were detected colorimetrically with polyclonal rabbit anti-tendamistat antiserum and anti-rabbit immunoglobulin G conjugated to alkaline phosphatase (AP).

Protease protection assay of SPase-TM fusion proteins

To 15 μl of a protoplast suspension (OD₆₀₀ = 1.0), 2 μl of water or 2 μl of 10% Triton X-100 was added, followed by 3 μl of a solution of trypsin (10 mg/ml in 50 mM Tris-HCl, pH 8.3). After 15 to 30 min of incubation at room temperature, 4 μl of 6× SDS-loading buffer (0.35 M Tris-HCl, pH 6.8, 0.35 M SDS, 30% [vol/vol] glycerol, 0.6 M dithiothreitol, 0.175 mM bromophenol blue) was added and the samples were placed at 100°C for 2 min. After separation by SDS-PAGE, the proteins were transferred to a nitrocellulose membrane and detected following Western blotting as described above.

In vitro activity assay.

Quantitative analysis of the in vitro activities of purified SipY304 and SipY336 was performed using an assay previously described for Lep (49). An internally quenched fluorescent peptide derived from the E. coli maltose binding protein signal peptide (Californian Peptide Research) was used as a substrate. After a 3-min preincubation period of the peptide (final concentration, 20 μM) at 42°C in assay buffer (50 mM Tris-HCl, pH 9, 1% Triton X-100), the reaction was initiated by the addition of the SPase (final concentration, 2.5 μM). Cleavage by the SPase leads to the removal of the quenching group, which in turn results in an increase of fluorescence. The fluorescent emission signal at 400 nm was measured as a function of time on a Spex 1860 Double Spectrophotometer (SPEX Instruments S.A., Longjumeau, France) using an excitation wavelength of 340 nm. The data shown are the averages of three independent measurements.

The reaction velocity was calculated by converting the units of the observed velocities from fluorescence units per second to nanomoles per second by using the following equation: v_nM/s = v_FI/s/[(FI_t=∞ − FI_{t = 0})/S₀], where v_nM/s is the reaction velocity (in nanomolar per second), v_FI/s is the reaction velocity (in fluorescence units per second), FI_t=∞ is the fluorescence intensity at the end point reading, FI_t=0 is the fluorescence intensity at zero time, and S₀ is the initial substrate nanomolar concentration. Because the reactions were too slow to enable monitoring the reaction to completion, the end point was determined by adding 20 μl of an alkaline protease solution (Promega, Madison, Wis.) in order to completely hydrolyze the substrate.

The processing rates of membrane fractions isolated from S. lividans wild-type and SPase knock-out strains were measured according to an identical procedure, but at 27°C. As a negative control, wild-type S. lividans membranes, which were preincubated for 5 min with 0.2 mM penem SB-214357 (an SPase inhibitor) (4, 34), were added. All membranes were tested at a final protein concentration of 0.1 mg/ml.

RESULTS

Computer-aided hydrophobicity analysis

The positions of the putative membrane-spanning segments in the amino acid sequences of the four SPases were identified by the TopPred 2 program (http://www.biokemi.su.se/∼server/toppred2 /toppredServer.cgi) (8, 48), a program which generates hydrophobicity profiles. S. lividans SipY was predicted to have two hydrophobic domains which could act as possible transmembrane anchors, amino acids (aa) 87 to 100 [A(nchor)I] and aa 305 to 325 [A(nchor)II]. SipX and SipZ have a predicted topology similar to that of SipY. For SipW, only one hydrophobic region (aa 48 to 63) was predicted to be a transmembrane segment (Fig. 2).

FIG. 2.

Predicted membrane topology of the S. lividans SPases and gene fusion approach. The predicted topologies of the S. lividans SPases are shown here for SipW and SipY. Regions conserved in all type I SPases are indicated. Almost all amino acids important for activity and structure of the bacterial SPases are situated in these regions, denoted boxes A, B, C, D, and E. Box A comprises the hydrophobic residues of the predicted transmembrane segments. Box B contains the catalytic serine, while the lysine is situated in the D region. N, amino terminus; C, carboxy terminus; in, the cytoplasmic side of the membrane; out, the cell wall-exposed side of the membrane. To experimentally confirm the membrane topology, the tendamistat reporter was fused to SipW and SipY at the indicated sites. The number in the protein notation indicates the total number of amino acids of the SPase or its truncated form present in the fusion protein. Three Sip-TM fusions were constructed with SipY fragments. One fusion contained the SipY fragment N-terminal of the first putative transmembrane anchor (aa 1 to 86, designated SipY86TM), and the second included the SipY region in front of the second predicted transmembrane segment (aa 1 to 304, designated SipY304TM). The complete SipY used as a fusion partner (aa 1 to 336) was designated SipY336TM. In addition, one TM fusion was constructed with complete SipW (aa 1 to 259, named SipW259TM).

Expression of fusion proteins in S. lividans and determination of membrane topology by using a trypsin protection assay

The predicted topology was investigated experimentally by constructing a series of fusion proteins with SipY and SipW, the former predicted to have one C-terminal anchor and the latter to have none (Fig. 2). As a reporter protein, tendamistat (the α-amylase inhibitor of Streptomyces tendae) (47) was chosen. All fusions were constructed at the C terminus of a hydrophilic region to avoid the disruption of the two known classes of topogenic signals (membrane-spanning segments and short hydrophilic regions with a net positive charge) (5).

S. lividans TK24 was transformed with the pIJSipTM constructs, and isolates containing the desired plasmid were identified following colony hybridization with a DIG-labeled tendamistat-specific probe. From each of the different S. lividans TK24 [pIJSipTM] strains, a spore suspension was prepared which was used as inoculum to grow cultures for overexpression of the chimeric proteins. Cultures were grown for 48 h in the presence of thiostrepton to induce expression of the fusion proteins, after which the mycelium was harvested by centrifugation. To identify the fusion proteins that were encoded by the pIJSipTM constructs, proteins in the cell lysates were separated by SDS-PAGE and transferred to a nitrocellulose membrane. Using immunodetection with anti-tendamistat antibodies, the presence of all constructed fusion proteins was confirmed (data not shown).

To determine whether the tendamistat reporter protein was located cytoplasmically or extracellularly, protoplasts were prepared from the different S. lividans TK24 strains expressing the fusion proteins. Next, trypsin sensitivity of the different fusions was analyzed after SDS-PAGE by immunodetection with anti-TM specific antibodies (Fig. 3). Cytoplasmically localized tendamistat is protected against the proteolytic activity of trypsin because of the presence of the membrane barrier, while tendamistat at the outer side of the membrane becomes hydrolyzed. When the membrane barrier is disrupted due to the addition of Triton X-100, all proteins are accessible for trypsin.

FIG. 3.

Trypsin protection assay of the different fusion proteins. After protoplasting the S. lividans cells expressing the chimeric proteins, protease protection was analyzed by SDS-PAGE, Western blotting, and immunodetection with anti-tendamistat specific antibodies. Fusion proteins remain intact or become digested with trypsin depending on their intracellular or extracellular localization. In addition, Triton X-100 (1%)-disrupted protoplasts were subjected to the same protocol. Western blots are shown for SipY86TM (A), SipY304TM (B), SipY336TM (C), and SipW259TM (D). (A) Trypsin sensitivity of the SipY86TM fusion. S. lividans protoplasts expressing the SipY86TM fusion treated with trypsin in the absence of Triton X-100 left the fusion protein intact. Two different incubation times were tested: 15 min (lane 4) and 30 min (lane 5) at room temperature. No detectable amount of fusion protein was present when Triton X-100 was added prior to the trypsin digestion (lane 1). (B) Trypsin sensitivity of the SipY304TM fusion. Addition of trypsin to S. lividans protoplasts expressing SipY304TM led to complete degradation of the fusion protein, both in the presence (lane 5) and absence (lane 3, 15-min incubation at room temperature, and lane 4, 30-min incubation at room temperature) of Triton X-100. (C) Trypsin sensitivity of the SipY336TM fusion. S. lividans protoplasts expressing the SipY336TM fusion treated with trypsin resulted in the accumulation of a 12-kDa fragment (lane 3, 15-min incubation, and lane 4, 30-min incubation) which could be degraded when Triton X-100 was added (lane 5). The size of the resulting fragment can be explained as the sum of the TM reporter protein (10 kDa on gel) and the part of SipY located at the inner side of the plasma membrane. Degradation of this fragment in the presence of Triton X-100 ruled out that the protected fragment was a result of trypsin insensitivity. (D) Trypsin sensitivity of SipW259TM fusion. After trypsin treatment of protoplasts expressing the SipW259TM fusion, a digestion pattern comparable to that of the SipY304TM fusion was detected. No detectable amount of fusion protein was present after addition of trypsin in all conditions tested, with (lane 5) or without (lane 3, 15-min incubation, and lane 4, 30-min incubation) Triton X-100. MW, molecular mass (in kilodaltons).

In the case of the SipY86TM (Fig. 3A) and SipY336TM (Fig. 3C) fusions, the reporter protein tendamistat could be degraded by trypsin only when Triton X-100 was added prior to the digestion (lanes 1 and 5, respectively). This implies a cytoplasmic localization of the reporter protein for these fusions. In contrast, degradation of tendamistat was observed for the SipY304TM (Fig. 3B) and SipW259TM (Fig. 3D) fusions both in the presence and absence of Triton X-100 (lanes 3, 4, and 5). It can be concluded that in these cases, the reporter protein was localized extracellularly.

These experimental data are consistent with a topological model for SipY having a C-terminal transmembrane anchor in addition to the N-terminally located one. SipW contains only one N-terminal transmembrane segment.

A considerable level of nonspecific reaction was observed when trypsin-treated protoplasts were probed with the anti-SipY antibodies. This was probably due to the close taxonomic relation between Streptomyces and Mycobacterium, the latter present as adjuvant in the complete Freund's solution used for immunization. As a consequence, anti-SPase antibodies could not be used for immunodetection of proteolytical fragments after trypsin treatment.

In addition to fusions of the SPase fragments with tendamistat, we constructed fusions with E. coli AP, a more commonly used reporter protein in topological studies. In turn, SipY86, SipY304, SipY336, and SipW259 were fused to AP. However, when expressed from the tipA promoter in S. lividans, the fusions of SipW and SipY fragments to AP could not be detected with anti-AP antibodies. In contrast, when expressed from the tac promoter in E. coli, only the fusion proteins of which AP was located periplasmically could be detected with anti-AP antibodies (not shown).

Functional analysis of the presence of the C-terminal anchor

The importance of the C-terminal anchor in the SPase processing activity was studied in vitro and in vivo by comparing the activity of a SipY derivative lacking the C-terminal anchor with the wild-type SipY.

The in vivo activity of the wild-type SipY and the C-truncated SipY derivative was assessed by their capacity to complement Lep. Therefore, E. coli IT89, which possesses a temperature-sensitive defect in Lep activity, was transformed with the plasmids pEXSipY304 and pEXSipY336. At nonpermissive temperature, the growth curves were monitored (Fig. 4A). Both SipY336 and the C-truncated SipY304 mutant were able to functionally replace the E. coli Lep. Deletion of the C-transmembrane anchor was, as a consequence, not inhibitory for the SPase activity by itself. However, we observed a slower growth rate for E. coli IT89 (pEXSipY304) cells compared to that of the cells containing pEXSipY336. It was considered that the difference could be due to a lower stability or activity of the mutant protein relative to that of the authentic SipY. Figure 4B shows the result of the immunoblot analysis of total protein extract from E. coli IT89 cells expressing SipY304 or SipY336, grown at 42°C until reaching an OD₆₀₀ of 0.4, probed with rabbit polyclonal antiserum to SipY. Similar immunodetectable concentrations of both enzymes were present, thereby demonstrating the stable accumulation of the SipY304 mutant. As a consequence, the lower growth rate could not be attributed to partial degradation or a decrease in the concentration of the C-truncated SipY derivative relative to that of wild-type SipY.

FIG. 4.

In vivo activity of SipY304 and SipY336. (A) E. coli IT89 cells containing pEXSipY304 or pEXSipY336 were grown overnight in LB medium at the permissive temperature (27°C). After dilution to an OD₅₄₀ of 0.02 with LB medium at 42°C (nonpermissive temperature), the cultures were further incubated and the OD₅₄₀ was plotted as a function of time. All points are an average of two experiments. (B) Relative concentrations of the wild-type SipY336 and the C-truncated SipY304 mutant were investigated using SDS-PAGE, Western blotting, and immunodetection with anti-SipY specific antibodies. Total-cell lysates were prepared from E. coli IT89 cells harboring the plasmids pEXSipY304 (lane 1), pEX50 (lane 2), or pEXSipY336 (lane 3). Samples were taken at an OD₅₄₀ of 0.4. No significant difference in concentration nor degradation of SipY304 could be demonstrated.

To quantitatively analyze the effect of the deletion of the C-terminal anchor on the processing activity, in vitro activity tests of purified SipY304 and SipY336, overexpressed in E. coli cells, were performed. The resulting fluorescence spectra (Fig. 5) showed that cleavage of the peptide was linear as a function of time, comparable with the processing of this substrate by Lep (49). The processing rates were calculated to be 0.392 nM/s for SipY336 and 0.227 nM/s for SipY304. Removal of the C-terminal anchor thus resulted in a ∼1.7-fold decrease of the in vitro processing activity of SipY, corresponding to a relative rate of 0.58 for SipY304 compared to the processing rate of the wild-type enzyme (1.0).

FIG. 5.

In vitro activity of SipY304 and SipY336. Progressive cleavage of Y(NO₂)FSASALAKIK(Abz) was by SipY304 and SipY336. The assay was performed in 50 mM Tris-HCl (pH 9)–1% Triton X-100. After a preincubation of the peptide at 42°C for 3 min, the reaction was initiated by the addition of the SPase (t = 0), and the change in fluorescence was monitored. Final concentrations of the SPases were 2.5 μM, and the substrate concentration was 20 μM.

In order to get an idea of the relative contributions of the SPases containing the C-terminal anchor to the total processing activity of the cell compared to that of SipW, several S. lividans SPase knock-out strains were constructed by marker replacement (V. Parro et al., unpublished data). To date, three single mutants (SipX⁻, SipY⁻, SipZ⁻), one double mutant (SipW⁻X⁻), and one triple mutant (SipW⁻X⁻Y⁻) were obtained. All these S. lividans SPase knock-out strains were viable, indicating that none of the SPases is essential by itself for cell viability, similar to the five B. subtilis SPases, and that the SPases can at least partly complement each other. A mutant strain lacking both SipY and SipZ could not be obtained so far. This might indicate that such a double knock-out mutant would not be viable. In a next step, the in vitro SPase activity of membrane fractions isolated from these strains was compared with that of membranes isolated from the wild-type strain. In this test, the peptide substrate described above was used. The calculated processing rates are listed in Table 3. As a blank, wild-type S. lividans membranes incubated with penem SB-214357 were used. This made it possible to detect background processing from proteases which were not inhibited. In practice, no significant processing rates were detected, showing that the assay was specific for SPase activity. The highest decrease in processing activity was observed for S. lividans membrane fractions lacking SipW⁻X⁻Y⁻. Furthermore, only deletion of SipY or SipZ led to a significant reduction in activity. Deleting SipX or SipX plus SipW did not significantly affect the processing activity of the membranes. In first instance, it seems that in S. lividans cells, SipZ is necessary but not sufficient to obtain a high processing activity. In contrast, compared to the SipW⁻X⁻Y⁻ membranes, in which only SipZ is active, the addition of SipY alone (in the SipW⁻X⁻ membranes) was sufficient to exhibit almost wild-type processing activities. This indicates that both SipY and SipZ are important contributors to the SPase processing capacity of the cell. Interestingly, almost no negative influence was observed on the processing activity of the membranes only when SipY and SipZ were both active.

TABLE 3.

Initial processing rates of peptide substrate by membranes isolated from S. lividans SPase knock-out strains

S. lividans strain	Active SPase(s)	Processing rate (nM/s)a	Relative rate
Wild type	SipW, SipX, SipY, SipZ	6.565	1.00
SipX−	SipW, SipY, SipZ	6.285	0.96
SipY−	SipW, SipX, SipZ	4.554	0.69
SipZ−	SipW, SipX, SipY	3.022	0.46
SipW−X−	SipY, SipZ	6.186	0.94
SipW−X−Y−	SipZ	2.680	0.41
Blank		0.164	0.02

^aProcessing rates were calculated as described in Materials and Methods. 

DISCUSSION

The most important result of this work is the experimental confirmation of the C-terminal transmembrane anchor of SipY, which is presently unique to P-type type I SPases of gram-positive bacteria. According to the TopPred 2 program, S. lividans SipX and SipZ have a similar membrane topology and belong to the same exceptional group of P-type SPases. Both N- and C-terminal anchors are also present in the SPase of gram-negative R. capsulatus (22), Imp2p of Saccharomyces cerevisiae mitochondria (29), and two ER-type SPases: SipW of B. subtilis (40) and SCP21 of Archaeoglobus fulgidus (21).

In contrast, S. lividans SipW was found not to have a C-terminal membrane anchor. It possesses only one N-terminally located transmembrane anchor, similar to most SPases of gram-positive bacteria described so far. Because all known SPases from gram-positive bacteria have a single N-terminal transmembrane segment and because its presence was already confirmed for one S. lividans SPase (SipY), no additional TM fusion was constructed for the SipW part N-terminal of the transmembrane segment.

In this experimental approach, two different reporter proteins were used, the small monomeric Streptomyces TM protein and the large multimeric E. coli AP protein. The latter is (together with LacZ) one of the most frequently used reporter proteins for topological studies. In our experiments, two major problems with the constructed SPase-AP fusions were encountered: (i) instability of the cytoplasmically located SPase-AP fusions proteins in E. coli (12, 43) and (ii) no expression of the SPase-AP fusions in S. lividans or no detectable production for unknown reasons. Thus, in those cases where the commonly used reporter proteins (which are all large, multimeric complex proteins) fail, the small monomeric polypeptide tendamistat could be a valuable alternative as a reporter protein to determine subcellular localization.

Little is known so far about the function of the SPase transmembrane segments. The most frequently allocated functions, based on studies for N-terminal transmembrane anchors, are the insertion of the SPase in the membrane and the prevention of intermolecular self-cleavage (46). Very recently, Carlos et al. (7) revealed that the N-terminal transmembrane anchors of E. coli Lep and the B. subtilis SipS are not important for determining the in vitro cleavage fidelity, because soluble derivatives of these enzymes retain the same substrate specificity.

Here, we showed both in vitro as well as in vivo that the C-terminal anchor of SipY is important to enhance the processing activity. We favor the idea that the C-terminal anchor, in addition to its role to position the SPase in the lipid bilayer (in vivo) or in detergent micelles (in vitro), plays a role in the overall structure of the SPase, thereby positioning the catalytic residues relative to each other. Consequently, removal of this anchor could lead to a less efficient membrane insertion and a change in conformation, resulting in a weaker interaction between the catalytically important residues, which in turn can cause the observed decrease in processing efficiency.

These topological analyses revealed that three of the four S. lividans SPases have a C-terminal anchor. In contrast, in B. subtilis, which contains five chromosomally encoded SPases, only SipW, which has a minor role in pre-protein processing, contains a C-terminal anchor. Considering this, the question was raised of whether a relationship in S. lividans exists between the presence of the C-terminal anchor and the relative importance of that respective SPase for the total processing capacity of the cell. This was analyzed by comparing the SPase activity of membranes isolated from several S. lividans SPase knock-out mutants, i.e., S. lividans SipX⁻, SipY⁻, SipZ⁻, SipW⁻X⁻, and a triple mutant, SipW⁻X⁻Y⁻. The quantitative in vitro activity measurements of isolated membranes of these strains showed the important contribution of SipY and SipZ to the total processing capacity of the cell. These findings, together with the fact that, up to now, it was not possible to isolate an S. lividans strain lacking both SipY and SipZ, strongly suggest that SipY and SipZ (SPases containing a C-terminal anchor) are of major importance for the cell. Accordingly, SipW, the only SPase without a C-terminal anchor, seems to be of minor importance.

Finally, a question was raised on whether a link exists between the unexpected finding that the majority of SPases found in S. lividans have two anchoring domains and the characteristics of the substrates processed by these SPases. When the processing activities of SPases from gram-positive bacteria were tested on known substrates of the E. coli Lep, none of them could process these substrates as efficiently as Lep (33). These findings indicate that prominent differences may exist in substrate specificity between the SPases of gram-negative and gram-positive bacteria. Moreover, even within the latter group, remarkable differences do appear, i.e., when we look more closely at signal peptide characteristics of proteins secreted by Streptomyces, its signal peptides are significantly different from those of all other bacteria, both in the total lengths of the N and C regions and in the positive charge of the N region, as well as in the number of hydrophobic residues present in the H region. Conceivably, the presence of the C-terminal anchor reflects an evolutionary adaptation of the Streptomyces SPases to this particular phenomenon.

In addition, the observed differences in membrane topology together with the existing differences between the conserved regions containing the catalytically important amino acids may not only affect the enzymatic activity itself but can also influence the substrate specificity of the different Streptomyces SPases. Interestingly, the distance in the primary structure between boxes C and D is about 30 amino acids longer in SipY (and SipX) than in SipW (and SipZ). As a consequence, the distance between the catalytic serine and lysine is about 30 amino acids longer than in all other SPases described so far. Differences in substrate specificity between the multiple SPases of B. subtilis have already been reported (6). Because processing by the SPases was identified as a bottleneck for the secretion of certain proteins in S. lividans (V. Parro, unpublished data) and B. subtilis (40), this observation could be promising for the engineering of the secretion machinery for improved secretion of both native and heterologous proteins by S. lividans.