Functional limits of conformation, hydrophobicity, and steric constraints in prokaryotic signal peptide cleavage regions. Wild type transport by a simple polymeric signal sequence

Identification and characterization of N-glycosylation site on a Mucor circinelloides aspartic protease expressed in Pichia pastoris: effect on secretion, activity and thermo-stability

Methylotrophic yeasts have widely been used as model organisms for understanding cellular functions and biochemical activities in lower eukaryotes. The gene encoding an aspartic protease (MCAP) from Mucor circinelloides DSM 2183 was cloned and expressed into Pichia pastoris using both the native M. circinelloides signal peptide (mcSP) and α-factor secretion signal from Saccharomyces cerevisiae (α-MF). When expressed in P. pastoris using α-MF and mcSP, MCAP was secreted into the culture medium at a concentration 200 mg L-1 (410 MCU mL-1) and 110 mg L-1 (249 MCU mL-1), respectively. The SDS-PAGE analysis of each culture shows that the protein was secreted in the media in two forms with molecular weights of approximately 33 and 37 kDa. Upon digestion using endoglycosidase H (Endo H), only one band at 33 kDa was observed, indicating that the protein might be glycosylated. One putative N-glycosylation site was found and a site-directed mutagenesis at position Asn331-Gln of the sequence produce only one form of the protein of 33 kDa, similar to that obtained when digested with Endo H. The optimum temperature and pH activity of the expressed MCAP was found to be at 60 °C and 3.6, respectively.

Export of a hyperexpressed mammalian globular cytochrome b5 precursor in Escherichia coli is dramatically affected by the nature of the amino acid flanking the secretory signal sequence cleavage bond

A chimeric mammalian globular cytochrome b(5) fused to Escherichia coli alkaline phosphatase signal sequence (SS) was used as a model probe to investigate the influence of substituting each one of the standard 20 amino acids at its N-terminus on the Sec-dependent export of the precursor to the periplasmic space of E. coli. Substituting the native Met(+1) of the passenger protein flanking the SS with any one of the remaining 19 amino acids introduced significant changes in the export of cytochrome b(5) without jamming the Sec-dependent translocon. Acidic and hydrophilic residues proved to be the most efficient promoters of export. Small, nonbulky and basic residues yielded intermediate levels of the hemoprotein export. Replacement with a Cys(+1) residue generated significant quantities of both monomeric and disulfide-linked dimeric forms. However, bulky, aromatic and hydrophobic residues caused a significant decline in the rates of secretion. In expectation with their absences in the natural periplasmically secreted proteins, Pro and Ile-tagged cytochrome b(5) precursors failed to generate any detectable secreted recombinant products. Although Ala, amongst the native E. coli periplasmic proteins, is the preferred X(+1) residue with an occurrence of 50% frequency, it proved half as effective in promoting export when inserted proximally to the SS of cytochrome b(5). The mechanisms involved for these export variations are discussed. The findings will prove beneficial for high-level generation of recombinant proteins by secretory means for pharmaceutical and related biotechnological applications.

Bioinformatics predictions of localization and targeting

One of the major challenges in the post-genomic era with hundreds of genomes sequenced is the annotation of protein structure and function. Computational predictions of subcellular localization are an important step toward this end. The development of computational tools that predict targeting and localization has, therefore, been a very active area of research, in particular since the first release of the groundbreaking program PSORT in 1991. The most reliable means of annotating protein structure and function remains homology-based inference, i.e. the transfer of experimental annotations from one protein to its homologs. However, annotations about localization demonstrate how much can be gained from advanced machine learning: more proteins can be annotated more reliably. Contemporary computational tools for the annotation of protein targeting include automatic methods that mine the textual information from the biological literature and molecular biology databases. Some machine learning-based methods that accurately predict features of sorting signals and that use sequence-derived features to predict localization have reached remarkable levels of performance. Sustained prediction accuracy has increased by more than 30 percentage points over the last decade. Here, we review some of the most recent methods for the prediction of subcellular localization and protein targeting that contributed toward this breakthrough.

Differential regulation and posttranslational processing of the class II hydrophobin genes from the biocontrol fungus Hypocrea atroviridis

Hydrophobins are small extracellular proteins, unique to and ubiquitous in filamentous fungi, which mediate interactions between the fungus and environment. The mycoparasitic fungus Hypocrea atroviridis has recently been shown to possess 10 different class II hydrophobin genes, which is a much higher number than that of any other ascomycete investigated so far. In order to learn the potential advantage of this hydrophobin multiplicity for the fungus, we have investigated their expression patterns under different physiological conditions (e.g., vegetative growth), various conditions inducing sporulation (light, carbon starvation, and mechanical injury-induced stress), and confrontation with potential hosts for mycoparasitism. The results show that the 10 hydrophobins display different patterns of response to these conditions: one hydrophobin (encoded by hfb-2b) is constitutively induced under all conditions, whereas other hydrophobins were formed only under conditions of carbon starvation (encoded by hfb-1c and hfb-6c) or light plus carbon starvation (encoded by hfb-2c, hfb-6a, and hfb-6b). The hydrophobins encoded by hfb-1b and hfb-5a were primarily formed during vegetative growth and under mechanical injury-provoked stress. hfb-22a was not expressed under any conditions and is likely a pseudogene. None of the 10 genes showed a specific expression pattern during mycoparasitic interaction. Most, but not all, of the expression patterns under the three different conditions of sporulation were dependent on one or both of the two blue-light regulator proteins BLR1 and BLR2, as shown by the use of respective loss-of-function mutants. Matrix-assisted laser desorption ionization-time of flight mass spectrometry of mycelial solvent extracts provided sets of molecular ions corresponding to HFB-1b, HFB-2a, HFB-2b, and HFB-5a in their oxidized and processed forms. These in silico-deduced sequences of the hydrophobins indicate cleavages at known signal peptide sites as well as additional N- and C-terminal processing. Mass peaks observed during confrontation with plant-pathogenic fungi indicate further proteolytic attack on the hydrophobins. Our study illustrates both divergent and redundant functions of the 10 hydrophobins of H. atroviridis.

Use of synthetic signal sequences to explore the protein export machinery

The information for correct localization of newly synthesized proteins in both prokaryotes and eukaryotes resides in self-contained, often transportable targeting sequences. Of these, signal sequences specify that a protein should be secreted from a cell or incorporated into the cytoplasmic membrane. A central puzzle is presented by the lack of primary structural homology among signal sequences, although they share common features in their sequences. Synthetic signal peptides have enabled a wide range of studies of how these "zipcodes" for protein secretion are decoded and used to target proteins to the protein machinery that facilitates their translocation across and integration into membranes. We review research on how the information in signal sequences enables their passenger proteins to be correctly and efficiently localized. Synthetic signal peptides have made possible binding and crosslinking studies to explore how selectivity is achieved in recognition by the signal sequence-binding receptors, signal recognition particle, or SRP, which functions in all organisms, and SecA, which functions in prokaryotes and some organelles of prokaryotic origins. While progress has been made, the absence of atomic resolution structures for complexes of signal peptides and their receptors has definitely left many questions to be answered in the future.

Interactions that drive Sec-dependent bacterial protein transport

Understanding the transport of hydrophilic proteins across biological membranes continues to be an important undertaking. The general secretory (Sec) pathway in Escherichia coli transports the majority of E. coli proteins from their point of synthesis in the cytoplasm to their sites of final localization, associating sequentially with a number of protein components of the transport machinery. The targeting signals for these substrates must be discriminated from those of proteins transported via other pathways. While targeting signals for each route have common overall characteristics, individual signal peptides vary greatly in their amino acid sequences. How do these diverse signals interact specifically with the proteins that comprise the appropriate transport machinery and, at the same time, avoid targeting to an alternate route? The recent publication of the crystal structures of components of the Sec transport machinery now allows a more thorough consideration of the interactions of signal sequences with these components.

Expression of three different mutations in the arginine vasopressin gene suggests genotype-phenotype correlation in familial neurohypophyseal diabetes insipidus kindreds

OBJECTIVE AND STUDY DESIGN

The autosomal dominant form of familial neurohypophyseal diabetes insipidus (adFNDI) is a rare disease characterized by a severe and progressive deficiency of AVP secondary to mutations in the gene encoding the AVP precursor. Whereas a number of studies have investigated the pathogenetic mechanisms behind the disease only few studies have included detailed clinical characterization of the affected patients, thereby making genotype-phenotype correlations difficult. The aims of the present study were to investigate the cellular effects of three different adFNDI mutations (A19T, L81P and C110X) by heterologous expression in a neurogenic cell line and to correlate these findings to the corresponding clinical phenotype as determined by extensive clinical tests.

RESULTS

The clinical studies showed a later age of onset in the family carrying the A19T mutation (3.4 years, range 2-9 years) compared with families with the L81P and C110X mutations [0.75 year, range 0.5-1 year and 1.0 year (n = 1), respectively]. No other differences could be demonstrated in the clinical phenotype between families. Expression studies showed that each of the three mutant genes caused significant reduction of the amount of immunoreactive AVP in the cell culture medium and severe impairment of the intracellular trafficking and processing of the AVP prohormone, supporting the disease causing nature of all three mutations. However, the A19T mutation was associated with some capacity for processing and trafficking consistent with the clinical observations. Immunoflourescence studies provided evidence of reticular accumulation of protein within the ER in the A19T and C110X mutants but a unique accumulation of much larger aggregates in the L81P, which were localized both within and immediately outside the ER.

CONCLUSION

The study suggests a genotype-phenotype correlation with regard to age of onset of diabetes insipidus symptoms and provides support by expression studies.

A directed evolution strategy for optimized export of recombinant proteins reveals critical determinants for preprotein discharge

A directed evolutionary approach is described that searches short, random peptide sequences for appendage at the secretory signal peptide-mature protein junction to seek ideal algorithms for both efficient and hyper export of recombinant proteins to the periplasm of Escherichia coli. The strategy employs simple, visual detection of positive clones using a PINK expression system that faithfully reports on export status of a mammalian hemoprotein in E. coli. With-in "sequence spaces" ranging from 1 to 13 residues, a significant but highly variable secretory fitness was scored such that the rate of secretion reciprocally correlated with the membrane-associated precursor pool of the evolved exportable hemoproteins. Three clusters of hyper, median, and hypo exporters were isolated. These had corresponding net charges of -1, 0, and +1 within the evolved sequence space, which in turn clearly correlated with the prevailing magnitude and polarity of the membrane energization states. The findings suggest that both the nature of the charged residue and the proximal sequence in the early mature region are the crucial determinants of the protonophore-dependent electrophoretic discharge of the precursor across the inner membrane of E. coli. We conclude that the directed evolutionary approach will find ready application in engineering recombinant proteins for their efficient secretion via the sec export pathway in E. coli.

Structure and function of SecA, the preprotein translocase nanomotor

Most secretory proteins that are destined for the periplasm or the outer membrane are exported through the bacterial plasma membrane by the Sec translocase. Translocase is a complex nanomachine that moves processively along its aminoacyl polymeric substrates effectively pumping them to the periplasmic space. The salient features of this process are: (a) a membrane-embedded "clamp" formed by the trimeric SecYEG protein, (b) a "motor" provided by the dimeric SecA ATPase, (c) regulatory subunits that optimize catalysis and (d) both chemical and electrochemical metabolic energy. Significant recent strides have allowed structural, biochemical and biophysical dissection of the export reaction. A model incorporating stepwise strokes of the translocase nanomachine at work is discussed.

Processing of Escherichia coli alkaline phosphatase. Sequence requirements and possible conformations of the -6 to -4 region of the signal peptide

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 positions -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 Escherichia 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 the inhibition imposed by -5 Leu, while a Ser residue at position -4 or -2 does not restore processing. In addition, effective maturation of a mutant with Pro residues at positions from -6 throughout -4 proved that polyproline conformation of this region is permissive for processing. To understand the effects of the mutations, we modeled a peptide substrate into the active site of the signal peptidase using the known position of the beta-lactam inhibitor. The inhibitory effect of the -5 residue and its suppression by either Pro -6 or Pro -4 can be explained if we assume that Pro-containing -6 to -4 regions adopt a polyproline conformation whereas the region without Pro residues has a beta-conformation. These results permit us to specify sequence requirements at -6, -5, and -4 positions for efficient processing and to improve the prediction of yet unknown cleavage sites.

Cloning, expression, and functional characterization of the Mycobacterium tuberculosis secA gene

To better understand the protein secretion mechanisms involved in the growth and pathogenesis of Mycobacterium tuberculosis, we examined the secA gene from M. tuberculosis (tbsecA; cosmid sequence accession No. z95121.gb_ba). We generated plasmids containing the full-length tbsecA gene or a fusion containing the 5' sequence from the M. tuberculosis secA gene and the remainder from the Escherichia coli secA gene and evaluated the ability of each construct to complement the defective SecA protein in E. coli MM52ts when grown at the non-permissive temperature. The full-length tbsecA gene was unable to compensate for the temperature-sensitive defect, whereas E. coli MM52ts that has been transformed with plasmid pMF8TB226 containing a chimeric secA gene was able to grow at 42 degrees C. This work confirms that the topography of SecA and its ATP binding sites are highly conserved, whereas its membrane insertion domains are species specific.

Functional signal peptides bind a soluble N-terminal fragment of SecA and inhibit its ATPase activity

The selective recognition of pre-secretory proteins by SecA is essential to the process of protein export from Escherichia coli, yet very little is known about the requirements for recognition and the mode of binding of precursors to SecA. The major reason for this is the lack of a soluble system suitable for biophysical study of the SecA-precursor complex. Complicating the development of such a system is the likelihood that SecA interacts with the precursor in a high affinity, productive manner only when it is activated by binding to membrane and SecYEG. A critical aspect of the precursor/SecA interaction is that it is regulated by various SecA ligands (nucleotide, lipid, SecYEG) to facilitate the release of the precursor, most likely in a stepwise fashion, for translocation. Several recent reports show that functions of SecA can be studied using separated domains. Using this approach, we have isolated a proteolytically generated N-terminal fragment of SecA, which is stably folded, has high ATPase activity, and represents an activated version of SecA. We report here that this fragment, termed SecA64, binds signal peptides with significantly higher affinity than does SecA. Moreover, the ATPase activity of SecA64 is inhibited by signal peptides to an extent that correlates with the ability of these signal peptides to inhibit either SecA translocation ATPase or in vitro protein translocation, arguing that the interaction with SecA64 is functionally significant. Thus, SecA64 offers a soluble, well defined system to study the mode of recognition of signal peptides by SecA and the regulation of signal peptide release.

The role of the membrane-spanning domain of type I signal peptidases in substrate cleavage site selection

Type I signal peptidase (SPase I) catalyzes the cleavage of the amino-terminal signal sequences from preproteins destined for cell export. Preproteins contain a signal sequence with a positively charged n-region, a hydrophobic h-region, and a neutral but polar c-region. Despite having no distinct consensus sequence other than a commonly found c-region "Ala-X-Ala" motif preceding the cleavage site, signal sequences are recognized by SPase I with high fidelity. Remarkably, other potential Ala-X-Ala sites are not cleaved within the preprotein. One hypothesis is that the source of this fidelity is due to the anchoring of both the SPase I enzyme (by way of its transmembrane segment) and the preprotein substrate (by the h-region in the signal sequence) in the membrane. This limits the enzyme-substrate interactions such that cleavage occurs at only one site. In this work we have, for the first time, successfully isolated Bacillus subtilis type I signal peptidase (SipS) and a truncated version lacking the transmembrane domain (SipS-P2). With purified full-length as well as truncated constructs of both B. subtilis and Escherichia coli (Lep) SPase I, in vitro specificity studies indicate that the transmembrane domains of either enzyme are not important determinants of in vitro cleavage fidelity, since enzyme constructs lacking them reveal no alternate site processing of pro-OmpA nuclease A substrate. In addition, experiments with mutant pro-OmpA nuclease A substrate constructs indicate that the h-region of the signal peptide is also not critical for substrate specificity. In contrast, certain mutants in the c-region of the signal peptide result in alternate site cleavage by both Lep and SipS enzymes.

How many potentially secreted proteins are contained in a bacterial genome?

Artificial neural networks were trained on the prediction of the subcellular location of bacterial proteins. A cross-validated average prediction accuracy of 93% was reached for distinction between cytoplasmic and non-cytoplasmic proteins, based on the analysis of protein amino-acid composition. Principal component analysis and self-organizing maps were used to create graphical representations of amino-acid sequence space. A clear separation of cytoplasmic, periplasmic, and extracellular proteins was observed. The neural network system was applied to predicting potentially secreted proteins in 15 complete genomes. For mesophile bacteria the predicted fractions of non-cytoplasmic proteins agree with previously published estimates, ranging between 15% and 30%. Characteristics of thermophile genomes might lead to an under-estimation of the fraction of secreted proteins by presently available prediction systems. A self-organizing map was constructed from all 15 bacterial genomes. This technique can reveal additional sequence features independent from exhaustive pair-wise sequence alignment. The Treponema pallidum and Mycobacterium tuberculosis data formed separate clusters indicating unusual characteristics of these genomes.

Processing of Escherichia coli alkaline phosphatase: role of the primary structure of the signal peptide cleavage region

A wide range (69) of mutant Escherichia coli alkaline phosphatases with single amino acid substitutions at positions from -5 to +1 of the signal peptide were obtained for studying protein processing as a function of the primary structure of the cleavage region. Amber suppressor mutagenesis, used to create mutant proteins, included: (i) introduction of amber mutations into respective positions of the phoA gene; and (ii) expression of each mutant phoA allele in E. coli strains producing amber suppressor tRNAs specific to Ala, Cys, Gln, Glu, Gly, His, Leu, Lys, Phe, Pro, Ser and Tyr. Most amino acid substitutions at positions -3 and -1 resulted in a complete block of protein processing. These data give new experimental support for the "-3, -1 rule". Only Ala, Gly and Ser at position -1 allowed protein processing, and Ala provided the highest rate of processing. The results revealed the more conservative nature of the amino acids at the -1 position of signal peptides of Gram-negative bacteria as compared with those of eukaryotic organisms. Position -3 was less regular, since not only Ala, Ser and Gly, but also Leu and Cys at this position, allowed the processing. Mutations at position -4 had an insignificant effect on the processing. Surprisingly, efficient processing was provided mainly by large amino acid residues at position -2 and by middle-sized residues at position -5, indicating that the processing rate is affected by the size of amino acid residues not only at positions -1 and -3. Conformation analysis of the cleavage site taken together with the mutation and statistical data suggests an extended beta-conformation of the -5 to -1 region in the signal peptidase binding pocket.

Peptide design aided by neural networks: biological activity of artificial signal peptidase I cleavage sites

De novo designed signal peptidase I cleavage sites were tested for their biological activity in vivo in an Escherichia coli expression and secretion system. The artificial cleavage site sequences were generated by two different computer-based design techniques, a simple statistical method, and a neural network approach. In previous experiments, a neural network was used for feature extraction from a set of known signal peptidase I cleavage sites and served as the fitness function in an evolutionary design cycle leading to idealized cleavage site sequences. The cleavage sites proposed by the two algorithms were active in vivo as predicted. There seems to be an interdependence between several cleavage site features for the constitution of sequences recognized by signal peptidase. It is concluded that neural networks are useful tools for sequence-oriented peptide design.

Identification of a sequence motif that confers SecB dependence on a SecB-independent secretory protein in vivo

SecB is a cytosolic chaperone which facilitates the transport of a subset of proteins, including membrane proteins such as PhoE and LamB and some periplasmic proteins such as maltose-binding protein, in Escherichia coli. However, not all proteins require SecB for transport, and proteins such as ribose-binding protein are exported efficiently even in SecB-null strains. The characteristics which confer SecB dependence on some proteins but not others have not been defined. To determine the sequence characteristics that are responsible for the SecB requirement, we have inserted a systematic series of short, polymeric sequences into the SecB-independent protein alkaline phosphatase (PhoA). The extent to which these simple sequences convert alkaline phosphatase into a SecB-requiring protein was evaluated in vivo. Using this approach we have examined the roles of the polarity and charge of the sequence, as well as its location within the mature region, in conferring SecB dependence. We find that an insert with as few as 10 residues, of which 3 are basic, confers SecB dependence and that the mutant protein is efficiently exported in the presence of SecB. Remarkably, the basic motifs caused the protein to be translocated in a strict membrane potential-dependent fashion, indicating that the membrane potential is not a barrier to, but rather a requirement for, translocation of the motif. The alkaline phosphatase mutants most sensitive to the loss of SecB are those most sensitive to inhibition of SecA via azide treatment, consistent with the necessity for formation of a preprotein-SecB-SecA complex. Furthermore, the impact of the basic motif depends on location within the mature protein and parallels the accessibility of the location to the secretion apparatus.

Definition of Mycobacterium tuberculosis culture filtrate proteins by two-dimensional polyacrylamide gel electrophoresis, N-terminal amino acid sequencing, and electrospray mass spectrometry

A number of the culture filtrate proteins secreted by Mycobacterium tuberculosis are known to contribute to the immunology of tuberculosis and to possess enzymatic activities associated with pathogenicity. However, a complete analysis of the protein composition of this fraction has been lacking. By using two-dimensional polyacrylamide gel electrophoresis, detailed maps of the culture filtrate proteins of M. tuberculosis H37Rv were generated. In total, 205 protein spots were observed. The coupling of this electrophoretic technique with Western blot analysis allowed the identification and mapping of 32 proteins. Further molecular characterization of abundant proteins within this fraction was achieved by N-terminal amino acid sequencing and liquid chromatography-mass spectrometry. Eighteen proteins were subjected to N-group analysis; of these, only 10 could be sequenced by Edman degradation. Among the most interesting were a novel 52-kDa protein demonstrating significant homology to an alpha-hydroxysteroid dehydrogenase of Eubacterium sp. strain VPI 12708, a 25-kDa protein corresponding to open reading frame 28 of the M. tuberculosis cosmid MTCY1A11, and a 31-kDa protein exhibiting an amino acid sequence identical to that of antigen 85A and 85B. This latter product migrated with an isoelectric point between those of antigen 85A and 85C but did not react with the antibody specific for this complex, suggesting that there is a fourth member of the antigen 85 complex. Novel N-terminal amino acid sequences were obtained for three additional culture filtrate proteins; however, these did not yield significant homology to known protein sequences. A protein cluster of 85 to 88 kDa, recognized by the monoclonal antibodies IT-57 and IT-42 and known to react with sera from a large proportion of tuberculosis patients, was refractory to N-group analysis. Nevertheless, mass spectrometry of peptides obtained from one member of this complex identified it as the M. tuberculosis KatG catalase/peroxidase. Thus, the detailed mapping of M. tuberculosis proteins, combined with state-of-the-art analytical techniques such as mass spectrometry, provides a basis for further analysis and rapid identification of biologically relevant molecules.

Displaying human interleukin-2 on the surface of bacteriophage

Previous attempts to produce active human Interleukin-2 (hIL-2) in E. coli have failed, due to its aggregation in the form of cytoplasmic inclusion bodies, and the inability of the protein to enter the periplasmic export pathway, when fused to bacterial signal sequences. We have reasoned that these limitations could be overcome by introducing changes in the signal sequence and/or in some hIL-2 residues, not critical for its biological activity; and proceeded to test this hypothesis using a phagemid vector carrying the pelB secretion signal sequence, and the filamentous phage display system. Deletion of the Pro +2 in hIL-2 led to the export of a correct size (processed) molecule to the bacterial periplasm of Su- cells by the phagemid vector. However, this was achieved under growth conditions that would not favor phage assembly in Su+ strains. Changing the hydrophobic core of the leader peptide reversed this situation and allowed phage assembly and display of a pIII/hIL-2 hybrid protein in TG1 cells. The phage-displayed hIL-2 is correctly folded, as judged by its ability to interact with a conformation-specific anti-hIL-2 monoclonal antibody, and maintains its biological activity when tested in a CTLL-2 cell proliferation assay. The changes introduced in hIL-2 and the signal sequence will make possible to use the powerful phage display technology for the selection of high-affinity variants from libraries of hIL-2 mutants.

Competition between functional signal peptides demonstrates variation in affinity for the secretion pathway

We have developed a system for examining the relative affinity of two different signal peptides for the protein secretion pathway in Escherichia coli. This system involves the expression of a modified alkaline phosphatase which possesses two signal peptides arranged in tandem. When both signal peptides have the wild-type sequence, cleavage after the first and cleavage after the second occur with nearly equal frequency. In both cases the remainder of the protein is transported to the periplasm. Thus both signal peptides effectively compete with each other for entrance to the secretion pathway. When the hydrophobicity of the second signal peptide is altered by small increments, we find that the more hydrophobic signal peptide is preferentially utilized. Thus, a more hydrophobic signal peptide can outcompete even an efficient wild-type signal sequence. The crossover point, for utilization of the second to the first signal peptide, is marked and occurs over a very small change in hydrophobicity. Our results suggest that the small differences in the hydrophobicity of wild-type signal peptides may have critical consequences: preproteins with the more hydrophobic signals could dominate one pathway, leaving those with only slightly less hydrophobic signals to require additional factors such as chaperonins, SecB, and other binding proteins.

The amino-terminal charge and core region hydrophobicity interdependently contribute to the function of signal sequences

We have constructed a series of signal sequence mutants that contain negatively charged amino termini and simplified core regions of varying hydrophobicity levels. This series provides a means of exploring the relative roles of the amino terminus and the hydrophobic core region during transport. The signal peptides with highly hydrophobic core regions support a rapid rate of transport in the presence of a negatively charged amino terminus. We have found that these negatively charged mutants are secreted in a manner similar to the wild-type signal sequence; sodium azide and carbonyl cyanide 3-chlorophenylhydrazone treatments indicate that the negatively charged mutants depend on SecA and the protonmotive force, respectively. These same mutants also demonstrate reduced competition with coexpressed beta-lactamase, reflecting the lower overall affinity for the transport pathway due to the net negative charge at the amino terminus. In addition, the pronounced effects of introducing three negative charges support the conclusion that the two regions function in a concerted manner.

Protein transport via amino-terminal targeting sequences: common themes in diverse systems

Many proteins that are synthesized in the cytoplasm of cells are ultimately found in non-cytoplasmic locations. The correct targeting and transport of proteins must occur across bacterial cell membranes, the endoplasmic reticulum membrane, and those of mitochondria and chloroplasts. One unifying feature among transported proteins in these systems is the requirement for an amino-terminal targeting signal. Although the primary sequence of targeting signals varies substantially, many patterns involving overall properties are shared. A recent surge in the identification of components of the transport apparatus from many different systems has revealed that these are also closely related. In this review we describe some of the key components of different transport systems and highlight these common features.

Physical and conformational properties of synthetic idealized signal sequences parallel their biological function

Transported proteins often contain an extension sequence called the signal peptide. The alkaline phosphatase (PhoA) signal sequence represents a typical signal peptide for comparison to idealized sequences both in vivo and in vitro. We have designed a series of idealized signal sequences which vary in amino terminal charge and core region hydrophobicity with minimal variation in amino acid composition. The idealized core regions contain different proportions of leucine and alanine residues, effectively producing hydrophobicities above and below the threshold level required for efficient secretion. The flanking amino and carboxyl termini were designed to maintain the general features and relative hydrophobicity of their counterparts in the wild-type PhoA signal sequence. Using the phoA gene, the signal peptide region was modified to generate mutants corresponding to the model sequences. Transport studies in Escherichia coli confirmed that completely idealized signal sequences, which lack a helix-breaking proline or glycine residue, can be functional if the core region is sufficiently hydrophobic and that one positively charged residue in the amino terminus is adequate for efficient transport. The corresponding peptides were chemically synthesized and exhibited HPLC retention times that reflect the relative hydrophobicities of the sequences. Structural analyses of the isolated peptides by circular dichroism demonstrate solvent dependence and exceptionally stable alpha-helix formation by the functional signal peptides in trifluoroethanol. Although leucine and alanine residues are often predicted to have similar propensities for forming an alpha-helix, considerably higher alpha-helical content is observed in the signal peptides which contain predominantly polyleucine core regions.(ABSTRACT TRUNCATED AT 250 WORDS)

Signal peptides: exquisitely designed transport promoters

Prokaryotic proteins destined for transport out of the cytoplasm typically contain an N-terminal extension sequence, called the signal peptide, which is required for export. It is evident that many secretory proteins utilize a common export system, yet the signal sequences themselves display very little primary sequence homology. In attempting to understand how different signal peptides are able to promote protein secretion through the same pathway, the physical features of natural signal sequences have been extensively examined for similarities that might play a part in function. Experimental data have confirmed statistical analyses which highlighted dominant features of natural signal sequences in Escherichia coli: a net positive charge in the N-terminus increases efficiency of transport; the core region must maintain a threshold level of hydrophobicity within a range of length limitations; the central portion adopts an alpha-helical conformation in hydrophobic environments; and the signal cleavage region is ideally six residues long, with small side-chain amino acids in the -1 and -3 positions. This review focuses on the parallels between signal peptide physical features and their functions, which emerge when the results of a variety of experimental approaches are combined. The requirement for each property may be ascribed to a potential interaction that is critical for efficient protein export. The summation of the key physical features produces signal peptides with the flexibility to function in multiple roles in order to expedite secretion. In this way, nature has indeed evolved exquisitely tuned signal sequences.

Selection of functional signal peptide cleavage sites from a library of random sequences

The export of proteins to the periplasmic compartment of bacterial cells is mediated by an amino-terminal signal peptide. After transport, the signal peptide is cleaved by a processing enzyme, signal peptidase I. A comparison of the cleavage sites of many exported proteins has identified a conserved feature of small, uncharged amino acids at positions -1 and -3 relative to the cleavage site. To determine experimentally the sequences required for efficient signal peptide cleavage, we simultaneously randomized the amino acid residues from positions -4 to +2 of the TEM-1 beta-lactamase enzyme to form a library of random sequences. Mutants that provide wild-type levels of ampicillin resistance were then selected from the random-sequence library. The sequences of 15 mutants indicated a bias towards small amino acids. The N-terminal amino acid sequence of the mature enzyme was determined for nine of the mutants to assign the new -1 and -3 residues. Alanine was present in the -1 position for all nine of these mutants, strongly supporting the importance of alanine at the -1 position. The amino acids at the -3 position were much less conserved but were consistent with the -3 rules derived from sequence comparisons. Compared with the wild type, two of the nine mutants have an altered cleavage position, suggesting that sequence is more important than position for processing of the signal peptide.

Development of artificial neural filters for pattern recognition in protein sequences

Four different artificial neural network architectures have been tested for their suitability to extract and predict sequence features. For optimization of the network weights an evolutionary computing method has been applied. The networks have feedforward architecture and provide adaptive neural filter systems for pattern recognition in primary structures and sequence classification. The recognition and prediction of signal peptidase cleavage sites of E. coli periplasmic protein precursors serves as an example for filter development. The primary structures are represented by seven physicochemical residue properties. This amino acid description provides the feature space for network optimization. The properties hydrophobicity, hydrophilicity, side-chain volume, and polarity allowed an accurate classification of the data. A three-layer network architecture reached a learning success of 100%; the highest prediction accuracy in an independent test set of sequences was 97%. This network architecture appears to be most suited for the analysis of E. coli signal peptidase cleavage sites. Further suggestions about the design and future applications of artificial neural networks for protein sequence analysis are made.

Titration of protein transport activity by incremental changes in signal peptide hydrophobicity

A systematic series of mutants has been generated which provides a means for titrating the dependence of protein transport activity on signal peptide hydrophobicity. These mutants involve replacement of the hydrophobic core segment of the Escherichia coli alkaline phosphatase signal peptide while maintaining the natural amino- and carboxyl-terminal segments and the overall length. The new core regions vary in composition from 10:0 to 0:10 in the ratio of alanine to leucine residues. Thus, a nonfunctional polyalanine-containing signal peptide is titrated with the more hydrophobic residue, leucine. Using precursor processing to quantify transport activity, we observe a clear, nonlinear dependence on hydrophobicity. At ratios of alanine to leucine of less than or equal to 8:2, the signal peptide is essentially nonfunctional; at ratios greater than or equal to 3:7, the signal peptide functions efficiently. The midpoint is between alanine to leucine ratios of 6:4 and 5:5. Signal peptides with hydrophobicity just below the midpoint show substantial, additional precursor processing over time while the others do not. The data are consistent with a simple model involving a two-state equilibrium between the untransported and transported species and a change in the delta G of -0.85 kcal/mol for every alanine to leucine conversion.

Signal sequences containing multiple aromatic residues

Using homopolymeric units of either phenylalanine or tryptophan to replace the natural core segment of the Escherichia coli alkaline phosphatase signal peptide, the hydrophobicity requirements for protein export and processing were further explored. The mutant signal peptide containing polyphenylalanine functioned at least as efficiently as the wild-type, while the signal incorporating polytryptophan was dysfunctional. The transport properties of these mutants confirm our work with sequences rich in aliphatic residues; namely that a high mean hydrophobicity per residue is critical for complete and rapid precursor processing and for translocation of the protein. The efficient transport properties of the polyphenylalanine-containing signal peptide demonstrate that neither the bulky, aromatic nature of phenylalanine nor the unusually high hydrophobicity of this mutant peptide adversely alters function. This study also suggests that the low occurrence of phenylalanine in natural signal sequences is not of functional consequence but probably reflects the low number of DNA codons for this residue. The polytryptophan-containing precursor was membrane inserted but not translocated. This type of transport defect suggests that this is a weakly hydrophobic signal peptide, consistent with hydropathy scales, which indicate that tryptophan is comparable to alanine. This application of polymeric sequences provides a function-based assay for the evaluation of amino acid hydrophobicity.

Proteolysis in protein import and export: signal peptide processing in eu- and prokaryotes

Numerous proteins in pro- and eukaryotes must cross cellular membranes in order to reach their site of function. Many of these proteins carry signal sequences that are removed by specific signal peptidases during, or shortly after, membrane transport. Signal peptidases have been identified in the rough endoplasmic reticulum, the matrix and inner membrane of mitochondria, the stroma and thylakoid membrane of chloroplasts, the bacterial plasma membrane and the thylakoid membrane of cyanobacteria. The composition of these peptidases varies between one and several subunits. No site-specific inhibitors are known for the majority of these enzymes. Accordingly, signal peptidases recognize structural motifs rather than linear amino acid sequences. Such motifs have become evident by employing extensive site-directed mutagenesis to investigate the anatomy of signal sequences. Analysis of the reaction specificities and the primary sequences of several signal peptidases suggests that the enzymes of the endoplasmic reticulum, the inner mitochondrial membrane and the thylakoid membrane of chloroplasts all have evolved from bacterial progenitors.