JBC Papers in Press. Published on October 21, 2002 as Manuscript M205781200 Processing of Escherichia coli Alkaline Phosphatase: Sequence Requirements and Possible Conformations of the –6 to –4 Region of the Signal Peptide. Andrey V. Kajava1*, Sergey N. Zolov2, Konstantin I. Pyatkov3, Andrey E. Kalinin2# and Marina A. Nesmeyanova2** 1Center for Molecular Modeling, CIT, NIH, Bldg 12A, room 2047, Bethesda, MD 20892, USA. 2Laboratory of protein secretion in bacteria, Skryabin Institute of Biochemistry and Physiology of Microorganisms, Russian Academy of Sciences, 142290 Pushchino, D o w n lo Moscow Region, Russia. 3Institute of Cellular Biophysics, 142290 Pushchino, Russia. ad e d fro m h ttp Running title: Signal peptide cleavage ://w w w .jb Abbreviations used: c-region, cleavage region; PhoA, alkaline phosphatase; prePhoA, c.o rg b/ y alkaline phosphatase precursor; SDS-PAGE, sodium dodecyl sulfate polyacrylamide-gel g u e s t o n electrophoresis; SPase, bacterial type I signal peptidase J a n u a ry Keywords: leader peptide, processing site, alkaline phosphatase, sequence analysis, 1 3 , 2 0 1 docking 9 *To whom correspondence should be sent: CRBM-CNRS UPR 1086, 1919, route de Mende, 34293 Montpellier, Cedex 5, France, FAX: (33) 4 67523681, e-mail: [email protected] ** To whom requests for biological samples should be sent: e-mail: [email protected] # Present address: Laboratory of Skin Biology, NIAMS, NIH, Bldg. 50, room 1527, Bethesda, MD 20892, USA Summary: Analysis of the precursors of bacterial exported proteins revealed that those having bulky hydrophobic residues at position –5 have a high incidence of Pro residues at positions –6 and –4, Val at position –3 and Ser at position –4 and –2. This led to a hypothesis that the previously observed inhibition of processing by bulky residues at position –5 can be suppressed by introduction of Pro, Ser or Val in the corresponding nearby positions. Subsequent mutational analysis of E.coli alkaline phosphatase showed that, as it was predicted, Pro on either side of bulky hydrophobic –5 Leu, Ile or Tyr completely restores efficiency of the maturation. Introduction of Val at position –3 also partially suppresses D o w n lo the inhibition imposed by –5 Leu, while Ser residue at position –4 or –2 does not restore ad e d fro m processing. In addition, effective maturation of a mutant with Pro residues at positions h ttp ://w from –6 throughout –4 proved that polyproline conformation of this region is permissive w w .jb c for processing. To understand the effects of the mutations, we modeled a peptide .o rg b/ y substrate into the active site of the signal peptidase by using the known position of the β- gu e s t o n J lactam inhibitor. The inhibitory effect of –5 residues and its suppression by either Pro –6 an u a ry 1 or –4 can be explained if to assume that Pro-containing –6 to –4 regions adopt a 3 , 2 0 1 9 polyproline conformation while the one without Pro residues have a β-conformation. These results allow to specify sequence requirements at –6, –5, –4 positions for efficient processing and to improve the prediction of yet unknown cleavage sites. 2 INTRODUCTION In prokaryotes and eukaryotes, most exported proteins are synthesized as precursors with an amino-terminal extension called the leader or signal peptide. The signal peptide directs protein translocation across membranes and is removed by a membrane-bound peptidase after transition through the membrane (1). Despite their common purpose, signal peptides have very little amino acid sequence similarity, although they do share general features. Typically 15-30 amino acids long, signal peptides of prokaryotic proteins consist of three distinct regions: a 1-5-residue amino- D o w n lo terminal positively charged segment, a 10-15-residue central hydrophobic core, and a ad e d fro m more polar 5-7-residue carboxyl-terminal cleavage region (c-region) (2,3). In addition, h ttp ://w most bacterial proteins have a 14-18-residue region of the mature part immediately w w .jb c downstream of the signal sequence which has a negative or neutral net charge (4-6). As a .o rg b/ y g result of extensive research over the last two decades, the role of each region of the u e s t o n exported proteins has been mainly elucidated (for review see (7)). The export of proteins Ja n u a ry is initiated by interactions of the positively charged amino terminus with negatively 13 , 2 0 1 9 charged phospholipid headgroups of the cytoplasmic membrane (8-12) and by insertion of the hydrophobic core of the signal peptide into the apolar environment of the membrane (3,13). The insertion of the signal peptide into the lipid bilayer proceeds in association with proteins of the Sec translocation machinery (7,14,15). The positive charge of the amino terminus can also govern the N -C orientation of the signal in out peptide within the membrane (16). In this orientation, the c-region of the signal peptide is exposed on the periplasmic side where it can be recognized and cleaved by the signal peptidase (SPase) between positions –1 and +1 (1,17,18). A sequence motif with small 3 residues at positions –3 and –1 defines the cleavage site (2,3,19). The conformational characteristics of the signal peptide are also mainly established. There is a consensus view based on several in vitro experimental studies (20,21) that the region of the signal peptide, inserted into the membrane, adopts an α-helical conformation. Now, it is known that the –3 to –1 region has an extended β-structural conformation in order to be recognized by SPase (19,22). Despite this progress, the critical physical and structural characteristics of residues –6, –5 and –4 that delimit the hydrophobic core and peptidase recognition site of the signal peptide are still poorly understood. Bacterial signal peptides frequently have α-helix-breaking residues such as proline and glycine at –6 to –4 Do w n lo a d positions (16) and this suggested that the disruption of the helical conformation in this e d fro m region is an important requirement for efficient processing. A number of experimental h ttp ://w w data supported this conclusion (23,24). Based on the analysis of natural sequences (16) w .jb c .o and experimental evidence, it was also proposed that the hydrophilicity of this region rg b/ y g u rather than its conformation may be important for the maturation (25). However, none of es t o n J a these rules have an absolute support from the recent collection of natural sequences: there n u a ry 1 3 are exported proteins with c-regions consisting of only apolar or helix-fostering residues. , 2 0 1 9 The conformation of the –4 to –6 region is also unknown. The Pro, Gly and Ser residues that frequently occupy –4, –5 and –6 positions (2,16) are typical for β-turns of globular proteins (26). This observation resulted in a widely accepted opinion that this region has a β-turn conformation (2,27). Furthermore, some mutagenesis studies of exported proteins showed that a decrease of the processing efficiency in mutant proteins correlates with a low probability of β-turn formation (24,28). However, it was shown that when Pro residues are simultaneously present at both the –5 and –4 positions of alkaline phosphatase from E.coli, this protein is processed properly (29). The sterical constraints 4 of this Pro-Pro tandem allow only a β-conformation of the –5 residue and, as a consequence, this result cast doubt on the presence of the β-turn in the –6 to –4 region. Instead, it was suggested that the c-region has an extended β-conformation (29). In this conformation, the –5 residue may have a contact with SPase and this can explain why the processing is sensitive to the size of the –5 residue (29). The determination of the three- dimensional structure of the bacterial type I SPase co-crystallized with its inhibitor (19) allows a final rejection of the β-turn hypothesis and favors the extended conformation of the c-region. However, despite the knowledge of the active site of the SPase and docking D of the peptide substrate into its binding pocket, the exact conformation of the –6 to –4 ow n lo a d region remains unknown. This could be considered as a minor academic problem if it was ed fro m not known that amino acid substitutions within this region can significantly diminish or http ://w w even block the maturation of exported proteins (23,29-31). w .jb c .o The goal of this work was to define the sequence requirements and conformation rg b/ y g u of the –6 to –4 region and its interactions with SPase during the processing. We es t o n J a approached this problem by using sequence analysis of exported proteins, mutational nu a ry 1 3 analysis and molecular modeling. , 2 0 1 9 EXPERIMENTAL PROCEDURES Sequence analysis -- Sequences from gram-negative bacteria were taken from SwissProt using Sequence Retrieval System software (http://www.ebi.ac.uk/srs/) and then checked manually. They were: 110 proteins of E. coli (68 with known and 42 with well- predicted cleavage sites) and 81 proteins of other gram-negative bacteria with known cleavage sites. The collection did not include highly homologous sequences with more than 80% of identity. Anomalous signal sequences (those whose lengths of the 5 hydrophobic core did not fall into the range between 7 and 17 residues) and proteins, secreted by other or modified secretion machineries (hydrogenases, having RR**F*K pattern within the signal sequence (32), pili (15) and lipoproteins) were also excluded. The collection of the 191 sequences is available over the World Wide Web (http://cmm.cit.nih.gov/kajava/gram-negat.dat). The data sets of 114 exported proteins from gram-positive bacteria and 1011 human exported proteins have been taken from SIGNALP database http://www.cbs.dtu.dk/services/SignalP/sp_matrices.html (33). Bacterial strains and plasmids -- E. coli strain E15 (Hfr ∆phoA8 fadL701 tonA22 garB10 ompF627 relA1 pit-10spoT1T2) (34) was used as a host strain for the expression D o w n lo of wild-type and mutant phoA genes cloned in plasmids. E. coli strain Z85 (thi ∆ (lac- ade d fro m proAB) ∆ (srl-recA) hsdR::Tn10 (F' traD proAB lacIq∆ZM15)) (35) was used to construct h ttp ://w w mutant phoA genes. w .jb c .o Wild-type alkaline phosphatase gene (phoA) was cloned into HindIII/BamHI sites rg b/ y g u of vector p15SK(-) containing multiple cloning site identical to pBluescript SK es t o n J a (Stratagene, USA), p15A ori of replication and chloramphenicol-acetyltransferase gene n u a ry 1 3 (Fischer and Hengstenberg, unpublished). Resulting phagemid was used to construct and , 2 0 1 9 express mutant phoA genes. Helper phage R408 was used to isolate single-strand recombinant phagemids. Plasmid harboring the gene of amber suppressor tRNAAla2 of E. coli in the vector pGFIB (36) was provided by Dr. J. Miller. Media and culture conditions -- Bacteria for cloning and oligonucleotide-directed mutagenesis were grown on the LB or 2YT media at 37oC. All media were supplemented with 25 µg/ml of chloramphenicol to either select for or maintain phoA-containing plasmids. To screen for colonies expressing active alkaline phosphatase, E. coli cells were grown on agar plates made of LB medium free of inorganic phosphate and 6 containing 40 µg/ml of 5-bromo-4-chloro-3-indolyl-phosphate (37). For the alkaline phosphatase expression, cells were grown on minimal medium (38) with 1 mM K HPO 2 4 and 0.1% peptone to the mid-log phase and transferred to the medium without orthophosphate and peptone. Oligonucleotide-directed mutagenesis -- To generate mutant forms of phoA, we used a new two step method which allowed us to omit hybridization with labeled nucleotides during selection of clones containing mutant genes (6,39). Isolation of single-strand phagemid DNA and plasmid DNA, electrophoresis of DNA fragments in agar gels, phosphorylation of oligonucleotides, and transformation of E. coli cells were D o w n lo performed by standard procedures (40). Mutations (Table 1) were confirmed by DNA ad e d fro m sequencing (41). h ttp ://w Alkaline phosphatase maturation -- Pulse-chase experiments were used to analyze w w .jb c the alkaline phosphatase maturation. E. coli cells grown to the mid-log phase in the .o rg b/ y g minimal medium with 1 mM K HPO were harvested, washed, and incubated for 10 min u 2 4 e s t o n in the same medium without orthophosphate to induce alkaline phosphatase synthesis. Ja n u a ry The cells were labeled with 50 µCi/ml [35S]methionine for 60 seconds and chased for 0.1, 13 , 2 0 1 9 1.0, 5.0, or 60.0 minutes by addition of unlabeled methionine to a final concentration of 0.05%. Proteins were precipitated with 10% TCA. Alkaline phosphatase and its precursor were immunoprecipitated with rabbit antibodies, and separated by 10% SDS- PAGE followed by autoradiography. Proteins were quantified using a LKB UltroScan laser densitometer. The relative quantity of mature alkaline phosphatase and its precursor was calculated with adjustment for the difference in number of methionine residues between the precursor and mature form. 7 Alkaline phosphatase isoforms and activity -- Cells expressing alkaline phosphatase were harvested and converted to spheroplasts in 20 mM Tris-HCl, pH 7.5, 10 mM EDTA, 50 mM sucrose and 1 mg/ml lysozyme for 15 minutes on ice. Periplasmic fraction was separated from the cell debris by centrifugation at 12,000 g for 5 minutes. The samples were analyzed by non-denaturing electrophoresis in 7.5% PAGE (42). Staining of the alkaline phosphatase isoforms was performed by incubation of the gel with α-naphthyl phosphate (Sigma, N-7255) and Fast Red Dye TR (Chemapol, Czech Republic) (43). The alkaline phosphatase activity was determined by measuring the rate of p-nitrophenylphosphate hydrolysis, taking the activity of hydrolysis of 1µmol of Do w n lo a substrate per 1 minute at 37oC as a unit of enzymatic activity (U). Total cell protein was de d fro m assayed by the Lowry method (44). h ttp ://w Molecular modeling -- Initial docking of a peptide corresponding to the –3 to +1 w w .jb c .o region of the alkaline phosphatase into the active site of SPase was made manually based rg b/ y g on the known position of the β-lactam inhibitor and using Insight II program (45). ue s t o n J a Possible conformations of the region –6 to –4 were selected based on two constraints: n u a ry 1 3 first, the absence of sterical clashes within the peptide chain and between the peptide and , 2 0 1 9 SPase; second, direction of signal peptide α-helix (residues –21 to –7) into the cytoplasmic membrane. Then the complexes between SPase and alkaline phosphatase precursor (-21 to +2) were subjected to energy minimization using DISCOVER module of Insight II (300 steps of minimization based on the steepest descent algorithm and the next 500 steps using conjugate gradients algorithm). The CHARMM force field (46) and the distance-dependent dielectric constant were used for the energy calculations. During the minimization, (i) the backbone atoms of SPase were tethered to their positions in the crystal structure, (ii) a carbonyl carbon atom of –1 residue was covalently linked to the 8 oxygen atom of the Ser90 side chain forming a tetrahedral intermediate and (iii) several hydrogen bonds (between oxygen of the peptide group of –1 residue and hydroxyl group of Ser88, between backbone oxygen of –2 residue and NH-group of Ile144, between backbone nitrogen of –2 residue and backbone oxygen of Asp142) were enforced by setting the distance constraints with moderate force (K=50), in order to improve their geometry. In addition, when the region –6 to –4 in β-conformation was energy minimized the distance constraints were imposed on hydrogen bonds between backbone CO-group of Gln85 and NH-group of –3 residue; backbone NH-group of Gln85 and CO-group of –4 residue; CO-group of Pro83 and NH-group of –5 residue. To allay the concern that these D o w n lo a constraints generated significant tension in the minimized structure, the last calculation d e d fro m was performed without any restrictions to an RMS derivative of 0.4 kcal/(mol*Å). A h ttp ://w module “Struct_Check” of Insight II program (45) was used to check the quality of the w w .jb c modeled complexes. Figures were generated with Molscript (47). .org b/ y g u e s t o n J RESULTS an u a ry 1 Rationale of the selected amino acid substitutions in the –6 to –4 region -- Our 3 , 2 0 1 9 previous study showed that the introduction of bulky residues at position -5 of E. coli alkaline phosphatase causes a decrease in the efficiency of its maturation (29). In agreement with this result, the occurrence of large hydrophobic residues Trp, Ile, Phe, Leu, Met, and Tyr (written in the order of decreasing hydrophobicity (48)) at position –5 for the c-regions of eukaryotic and gram-positive bacterial proteins is lower than at positions –6 and –4 (17% vs. 45% and 27%, and 7% vs. 20% and 18% correspondingly). Surprisingly, the exported proteins of gram-negative bacteria have an opposite distribution: 21% of large residues at position –5 against 14% and 17% at positions –6 9 and –4. We analyzed a subset of bacterial proteins with bulky hydrophobic residues at position –5 and found that they have a higher incidence of Pro residue at positions –6 and –4, Val residue at position –3, and Ser residue at position –4 and –2 as compared with the complete collection of these proteins (Fig. 1). This observation led to the hypothesis that the inhibitory effect of bulky hydrophobic residue in position –5 can be suppressed by introducing Pro, Ser or Val in the corresponding nearby positions. In accordance with this, a series of mutant E.coli alkaline phosphatases was obtained (Table 1) to test the hypothesis. In addition, a mutant having Pro residues in all –6, –5 and –4 positions was also obtained. The –6 to –4 region of such protein is sterically constrained in the D o w n lo polyproline conformation and it was interesting to examine whether it is processed or not. ad e d fro m The fact that such tandem of three Pro was not found in the natural sequences provided h ttp ://w an additional motivation to study this mutant. w w .jb c .o rg b/ y g Effect of the mutations on translocation and processing of the alkaline u e s t o n phosphatase -- All mutant proteins were enzymatically active in the cells (data not Ja n u a ry shown). This result implies that the mutants were translocated across the cytoplasmic 13 , 2 0 1 9 membrane because it is known that alkaline phosphatase becomes active only after translocation into the periplasm, where disulfide bond formation and enzyme dimerization take place (49). The effect of the amino acid substitutions on alkaline phosphatase maturation was assessed by the rate of conversion of a pulse-labeled mutant protein precursor into the mature form in vivo using the standard pulse-chase method. As shown in Fig. 2, presence of bulky Leu, Ile or Tyr residue at position –5 (proteins L(-5), I(-5) and Y(-5)) notably impaired the maturation of the precursor in comparison with wild-type protein. Even 10