Bacterial Type I Signal Peptidases as Antibiotic Targets

Smitha Rao CV; Jozef Anné


Future Microbiol. 2011;6(11):1279-1296. 

In This Article

Bacterial SPases: Characteristics, Structure, Mechanism & Assays

Type I SPases (EC belong to the serine protease family S26 and are classified into the evolutionary clan of serine proteases, SF.[36] SPases from different bacteria have five conserved regions denoted as boxes A–E, as revealed by sequence-alignment data (Figure 3A).[17] Box A contains the transmembrane anchor, Box B and D have the catalytic Ser and Lys residues, respectively, Box C and E contain residues important for the formation of the substrate-binding pocket and for positioning of the catalytic residues. SPases from Gram-positive and Gram-negative bacteria are functionally identical but differ in the number of genes encoding SPase within a cell, membrane topology, overall size and substrate specificity (see review[37]). Gram-negative bacteria typically have one chromosomally encoded SPase, which is essential and constitutively expressed, whereas Gram-positive bacteria often have more than one SPase, with none of the individual enzymes being essential for cell viability on their own. SPases from Gram-positive bacteria are generally smaller in size and have a smaller catalytic domain compared with those from Gram-negative bacteria.

Figure 3.

Conserved regions, structure and membrane topology of Escherichia coli signal peptidase.
(A) A line diagram showing the conserved regions (Boxes) in bacterial type I signal peptidases: the consensus sequence obtained by sequence alignment of SPases from different bacteria10,15 is shown in the boxes. The conserved residues are shown within the boxes with absolutely conserved residues in bold and those not conserved indicated as x. The residue numbers are indicated on the boxes using Escherichia coli SPase nomenclature. Box A is a part of the transmembrane anchor. Box B contains Ser (colored brown) that serves as the nucleophile in the catalytic mechanism. Box C and Box C' contain the conserved residues Leu and Val, respectively, which are a part of the substrate binding pocket (S3). Box D has Tyr, which forms a part of the substrate binding pocket (S1) followed by the general base Lys residue (colored green) and the invariant Arg which, in some bacteria plays a structural role. Box E contains Ser (residue 278, E. coli numbering) that is thought to help position the lysine general base toward the nucleophilic serine.15 (B) Schematic representation of the membrane topology of E. coli LepB: the full-length SPase with two transmembrane segments, a cytoplasmic loop and the C-terminal periplasmic domain are shown. The periplasmic region shows the orientation of the catalytic domain of the SPase apoenzyme (PDB 1kn9) in the membrane, calculated based on a computational approach.77 The latter part of the Figure has been adapted with permission from.202 (C) A ribbon diagram of the general fold of truncated E. coli LepB (Δ 2-75). The conserved Domain I β-sheet is in green and the Domain II β-sheet is in blue. The β-hairpin extension protruding from Domain I is in purple. Active site residues Ser and Lys are labeled 90 and 145 respectively.
Reproduced with permision from [43].

In vitro behavior and biochemical properties of SPases have been extensively studied in E. coli LepB, followed by the SPases from Gram-positive Bacillus subtilis, Streptomyces lividans, S. pneumoniae (see review[37]) and S. aureus.[38] These studies have shown that the bacterial SPases display several common properties in vitro, such as intermolecular self-cleavage of the enzyme; enhanced activity in the presence of some nonionic detergents and phospholipids; reduced activity of the truncated SPases (devoid of the transmembrane segment) compared with the full-length SPases; high optimum pH (ranging from pH 8 to 11) in contrast to the classical serine proteases.

Structural Studies

E. coli LepB is an integral membrane protein (MW 35.9 kDa) with two transmembrane segments (residues 4–28 and 58–76), a cytoplasmic loop (residues 29–57) and a periplasmic C-terminal catalytic domain (residues 77–323) (Figure 3B). Crystallization of membrane proteins, such as the full-length SPase, is a cumbersome task, more so because of its poor in vitro stability due to self-cleavage activity and solubility issues. In order to circumvent these problems, a truncated derivative of E. coli, LepB (referred to as Δ2–75 SPase), lacking residues from 2–75, corresponding to the N-terminal transmembrane regions, was designed. The Δ2–75 SPase has been purified, characterized,[39,40] crystallized in its apoform[41,42] and in complex with different types of SPase inhibitors (β-lactam,[43] lipopeptide,[44] lipopeptide and a β-sultam inhibitor[45]), and has also been used in a nuclear magnetic resonance (NMR) study.[46] This mutant also requires the nonionic detergent Triton®-X-100 for optimal activity[40] and crystallization.[41] The addition of an SPase inhibitor greatly facilitated in obtaining high-resolution (1.95 Å) structural data.[43]

The structure of Δ2–75 SPase (Figure 3C) consists primarily of a β-sheet protein fold, consisting of two antiparallel β-sheet domains.[43] Domain I, termed the 'catalytic core', contains the conserved regions (Boxes B–E), all of which reside near to, or are a part of the active site. Domain I also contains an unusually large exposed hydrophobic surface that is consistent with a membrane association surface and possibly the detergent/lipid requirement of the truncated SPase.[43] Structure-based multiple sequence alignment data of representative type I SPases reveal that Domain I is conserved throughout evolution. By contrast, Domain II is relatively smaller and the extended β-ribbon is lacking in Gram-positive bacterial SPases, which is attributed to the generally smaller size of these SPases.[10]

The co-crystal structures of Δ2–75 SPase-inhibitor complexes and the structure of the apoenzyme confirmed the residues involved in catalysis, oxyanion stabilization, those forming the substrate-binding sites and the ones important for optimum activity of the enzyme. The structural data also provides plausible reasons for the high pKa of the general base Lys,[43] high pH requirement and the requirement for detergent.[45]

Nuclear magnetic resonance spectroscopy has been partially performed on the full-length SPase (LepB) in detergent micelles.[47] This may be refined in the future to be a useful tool in drug function studies as well as to obtain the 3D structure of this enzyme, which could ultimately shed some light on the structural relationship between the catalytic and the transmembrane domains.

The Substrate-binding Pockets & Substrate Specificity

Signal peptides in preproteins are, on average, 20 amino acids long (generally longer in case of Gram-positive bacteria) and have very little sequence identity except in the SPase-cleavage site region. However, three distinct regions can be recognized, namely, a positively charged n-region, a hydrophobic core (h-region) and a polar c-terminal region (Figure 4).[48] Based on a common pattern in the c-region of the SP at positions -1 and -3, the region preceding the cleavage site was proposed as the SPase-recognition site.[49] Site-directed mutagenesis confirmed the importance of these residues. The frequent presence of Ala residues at these positions led to the formulation of (-3, -1) rule or Ala-X-Ala of cleavage site specificity (see review[15]). Consistent with these data, SPase has a substrate specificity for small uncharged residues at -1 (P1) and small or larger aliphatic residues at -3 (P3).

Figure 4.

Schematic representation of a typical signal peptide.
Bacterial signal peptides have a tripartite structure that consists of an amino-terminal positively charged region (n-region), a central hydrophobic core (h-region) and a neutral but polar carboxy terminal region (c-region) containing the SPase recognition sequence (AXA motif). Helix-breaking residues (Pro or Gly) are found in the middle of the h-region and at the boundary between the h- and the c-regions (-6 position relative to the cleavage site).

The crystal structures have helped to explain the Ala-X-Ala substrate specificity by revealing two shallow hydrophobic pockets (S1 and S3) adjacent to the catalytic residues.[42,43] The model of SP from E. coli outer membrane protein A (OmpA) bound to the E. coli SPase suggests possible substrate–enzyme contacts all the way up to -8 (P8) of the SP and that the P2, P4 and P5 side chains are most likely exposed to solvent, which is consistent with the lack of specificity at these positions. Karla et al. examined the importance of residues that make up the S1 and S3 binding pockets using site-directed mutagenesis, mass spectrometric analysis, and in vivo and in vitro activity assays, as well as molecular modeling.[50] This led to the identification of a residue contributing to the high fidelity of cleavage by SPase, isoleucine 144 (E. coli numbering). Mutation of Ile144 to Cys results in cleavage at multiple sites, in contrast to wild-type SPase, which cleaves exclusively at the correct site. In addition, the residues contributing to substrate specificity were identified to be Ile 86 and Ile 144.[50] In a structure-based model of the SPase, the two residues are seen at the borderline between the S1 and S3 binding pockets, contributing atoms to both the pockets.[50] Furthermore, analysis of substrate specificity of various SPase Ile144 single mutants and Ile144/Ile86 double mutants by a fluorescence resonance energy transfer-based peptide library approach showed that the specificity is relaxed at the -3 position.[51] This work established the two Ile residues as key P3 substrate specificity determinants for SPases.

Interestingly, an NMR study on the Δ2–75 SPase revealed that a small subset of 18 residues undergo a conformational change upon SP binding, while the rest of the enzyme structure (94% of the residues) remains substantially unchanged.[46] The residues perturbed include the Ser–Lys dyad and those contributing to the S1 and S3 substrate-binding pockets, demonstrating that the SP binding is specific and involves only local changes in Δ2–75 SPase structure. The residues of S1 and S3 reorient in order to achieve close complementation with the docked -1 and -3 SP residues, consistent with an induced fit hypothesis. The NMR data supports the overall picture of SPase–SP interaction inferred from the co-crystal structure of SPase with inhibitors[43,44] and by molecular modeling.

Recent data favors the contact points between the SPase and the SP extending slightly beyond the S1 and S3 binding sites. For example, the NMR study by Musial-Siwek et al. also identified three residues (located away from the S1 and S3 pocket), which suggests interaction of the SPase with the C-terminus of the hydrophobic core region of the SP.[46] A theoretical model of a preprotein/SPase complex, based on the structures of SPase/inhibitor complex defined subsites on both sides of the active site; both S and the S/subsites. In the co-crystal structure of the Δ2–75 SPase-inhibitor complex, Luo et al. observed electron density for a Triton X-100 fragment within a crevice along the surface of SPase just beyond the active site (opposite side from the S1 and S3 subsites) suggesting that this might provide a clue for the location of a possible S/subsite.[45]

Proposed Catalytic Mechanism

SPase-inhibition studies provided the first clue on the unique catalytic mechanism[10] and the si-face attack of the substrate.[52] The initial evidences for the SPases utilizing an unconventional Ser/Lys dyad mechanism came from site-directed mutagenesis and chemical modification studies on E. coli LepB, B. subtilis SipS (see review[17]) and S. pneumoniae Spi,[53] which was finally confirmed by the crystal structure of LepB. A mechanism of catalysis for preprotein cleavage by E. coli LepB was proposed[10,15] and is in agreement with data accumulated thus far. In short, the SP binds to the SPase with the P1–P4 residues in an extended β-conformation, forming hydrogen bonds with the β-strand containing Lys145. Upon binding of the preprotein, the neutral Lys145 ε-amino group, orientated by the Oγ Ser278, abstracts a proton from the side-chain hydroxyl of Ser90, activating the Ser90 Oγ for nucleophilic attack on the carbonyl (si-face) of the scissile peptide bond (i.e., the SPase cleavage site). A tetrahedral intermediate is formed, which is electrostatically stabilized by hydrogen bonds to the oxyanion hole, formed by the Ser90 main-chain amide and the Ser88 side-chain hydroxyl group. The ammonium group from the side chain of Lys145 would then donate a proton to the leaving group amide (i.e., the N-terminus of the mature protein). This releases the mature protein from the SPase and a SP-bound acyl-enzyme intermediate is formed. The deacylating water then loses its proton to the Lys145 general base and the hydroxyl group attacks the ester carbonyl to produce another tetrahedral intermediate. This then breaks down to release the SP, regenerating the SPase active site.

Assays for the SPases

In vivo activity of SPase is commonly monitored in whole cells by using temperature-sensitive mutants (E. coli IT41/IT89, described earlier) in which the SPase activity is inactivated at the nonpermissive temperature by controlling the enzymatic activity by constructing a strain that regulates the expression of the SPase or by controlling activity with specific SPase inhibitors.[15] The latter approach is also used to establish the potency of inhibitors against the enzyme in vivo. For instance, an E. coli strain underexpressing the lepB gene was utilized to set up a cellular assay for testing compounds that inhibit SPase activity in vitro.[54] Assays of this type may be useful in preliminary investigation of target inhibition in whole cells. Another cell-based assay, the β-lactamase secretion assay, was employed for testing the potency of inhibitors against the SPase of S. aureus.[55] In this assay, a chromogenic β-lactamase substrate, nitrocefin was used to determine the secretion of β-lactamase (a native substrate of SPase) by bacterial cells, incubated in the presence or absence of an inhibitor. The secretion of β-lactamase is blocked in the presence of a SPase inhibitor, most likely owing to the uncleaved SP anchoring the β-lactamase to the cytoplasmic membrane. Recently, an immunofluorescence assay was employed to demonstrate the cytoplasmic accumulation and thereby lack of secretion of a native SPase substrate owing to inhibition of enzyme activity by a known SPase inhibitor.[33]

In vitro SPase assays are performed using preprotein substrates or short synthetic peptides. The preprotein processing assay involves SDS-PAGE. Previously, this assay was used to measure the kinetic constants of SPase cleavage using different substrates[15] and to demonstrate in vitro inhibition of the enzyme.[38] Although useful, this assay is not suitable for rapid screening of inhibitors as it is time consuming and labor intensive. Assays involving short synthetic peptide substrates are preferred, as kinetic constants of SPase cleavage can be easily measured and high-throughput screening (HTS) of inhibitors is possible.

Peptide substrates are designed by incorporating the SPase recognition sequence and the residues flanking it, based on the sequence information in the SP region of preprotein substrates. Although, the minimum substrate sequence recognized by SPase is a pentapeptide,[56] more efficient and common substrates are at least nonapeptides or 9-mers. Previously, enormous disparities in catalytic efficiencies were noted between preprotein and peptide substrates.[10] However, the quality of substrates has improved significantly in recent years, resulting from a better understanding of the substrate requirements for efficient hydrolysis by SPases, in vitro. Table 1 summarizes the activities of SPases in assays involving different peptide substrates with a protein substrate, pro-OmpA-nucleaseA, included as a 5. The latter is a hybrid protein of the S. aureus nuclease A attached to the SP of the E. coli outer membrane protein A, and gives the best catalytic constants reported so far against E. coli LepB. Substrate B (Table 1) is based on the SP region of maltose-binding protein in E. coli and is the first fluorogenic peptide reported for SPase. In order to provide SP-like sequence to substrate B, Stein et al. appended positively charged residues (K5) and a hydrophobic core (L10) to its N-terminus (substrate C; Table 1).[57] This resulted in a dramatic increase in activity, albeit in presence of the detergents. Substrates D–H, designed specifically for S. aureus SpsB or S. pneumoniae Spi, are based on the preprotein sequence data from Staphylococcus or Streptococcus spp., respectively. The difference in activities of LepB and SpsB in processing a common substrate suggests different substrate requirements of the SPases. Substrate G is the best reported for SpsB. Using this peptide, Bruton et al. demonstrated that apart from the AXA cassette, a decanoyl moiety (mimicking a membrane anchor) and a turn-inducing motif (proline at P5 position) are important for efficient processing of substrates by the SPase.[58] As seen in Table 1, most assays designed in the recent past are fluorescence resonance energy transfer-based, allowing SPase–substrate reactions to be monitored continuously. Many of these assays have been in use for high-throughput screening of SPase inhibitors.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: