Secondary structure as a conformational determinant in processing of peptide precursors at dibasic amino acids

The aim of this work was to establish the functional role of selected secondary structure motifs of peptide hormone precursors in their selective recognition by the corresponding converting endoproteases. The strategy was based on the use of synthetic peptides either reproducing or mimicking the sequences of the cleavage regions of two distinct models; i.e. proocytocin–neurophysin and prosomatostatin. Both prohormones were capable to release their biologically active sequences either by cleavage at a dibasic stretch or by proteolysis at a monobasic site. Both kinetic and thermodynamic parameters of peptide cleavage by various convertases were measured. They were examined in light of structural data on preferred conformations adopted by these substrates, which were obtained by a combination of spectroscopical techniques including CD, FT-IR and proton NMR. In the case of prosomatostatin, these approaches were in addition paralleled by site-directed mutagenesis experiments. The wealth of collected data point toward the conclusion that β-turns and/or loops, favored by sequences bearing basic residues, constitute a key feature in the specification of those peptide loci which are proteolytically processed in vivo. They will be discussed with respect of other processing mechanisms where these mechanisms were also shown.


Introduction
Proteolytic processing is an important regulatory mechanism used by cells to control gene expression at the post-translational level.It is generally associated to mechanisms of activation or inactivation of many proteins and the regulation of their cellular localization [1][2][3][4][5][6].Peptide hormones and peptide transmitters are generated from polypeptide precursors by specific cleavage reactions which take place principally at sites formed by single or paired basic residues [6][7][8][9].Not all the possible cleavage sites are utilised [6,7], however, and the degree of processing of many propeptides has been found to vary according to the tissue of origin [1][2][3][4][5].The restricted nature of processing reactions could point to the existence of a series of enzymes with stringent specificities [2,4,6], recognising regions of structure in addition to the single or paired basic residues.Alternatively the action of processing enzymes may be directed by conformation of the pro-peptide which could focus the action of a protease onto or away from a particular site [2,7].
Earlier analysis of amino acid sequences around proteolytic cleavage sites of several biosynthetic precursors by secondary structure predictive methods [6], indicated that the dibasic residues constituting these processing sites which are cleaved in vivo are situated in, or immediately adjacent to, β-turns [6] or loops [10].In contrast, the potential sites which are not cleaved in vivo are associated with ordered structures as α-helices or β-sheets.Turns or loops, often associated to exposed regions in proteins [11,12], possess a greater degree of flexibility and local mobility, compared to more rigid structures such as α-helices and β-sheets.Therefore, the presence of turns or loops, in the vicinity of dibasic cleavage sites, is not surprising since they might favor both accessibility and adaptability of substrates to proteolytic enzymes.
The purpose of this review is to summarize the strategies and experimental approaches that were designed to analyze the role of precursor structure in the recognition of dibasic cleavage sites by their corresponding processing proteases (for reviews see [13,14]).To address this issue, the proocytocin/neurophysin (pro-OT/Np) and the prosomatostatin (ProSom) precursors were used as models.
These conclusions, recently confirmed by studying the OT/Np(1-20) peptide as a model [20], support the concept that the β-turn structure might constitute a recognition signal for processing enzymes.

The β-turn is an interchangeable structural motif
It was previous shown that dibasic site recognition by processing endoproteases was not correlated with the existence of a consensus primary sequence but rather with the presence of β-turns or loops around these proteolytic loci [7,10].Therefore, replacement of the P 7 -L-G-G 10 sequence by nonhomologous peptide stretches, known to organize as β-turn structures in proteins [21], could not abolish the endoproteolytic cleavage of peptide OT/Np(7-15) derivatives.
In view of the general role played by this type of secondary structures in a large number of biological processes [24,25,27], β-turn structures appear to be interchangeable motifs [26] that provide readily accessible regions for selective proteases.

Role of β-turn in the endoproteolytic cleavage of substrates
Since the β-turn structure appeared to be essential for the cleavage of substrates by the processing proteases, one would expect that its replacement by ordered structures might affect the kinetics of the enzymatic reaction.The peptides [I 7 -L 10 ]-OT/Np(7-15), [I 7 -L 10 ]-OT/Np(7-20) and [A 3 -V 10 ]-OT/Np(1-20) (Table 1), designed to promote formation of β-sheet or α-helix in the vicinity of the dibasic cleavage site [21], were used as models to test this hypothesis, Although those substrate analogs, as well as the reference peptides OT/Np(7-15), OT/Np(7-20) and OT/Np(1-20), were cleaved by the pro-OT/Np convertase, data in Table 1 reveal drastic changes in their kinetic parameters.Indeed, pairwise comparisons of peptides with the same size (OT/Np(7-15) and [I 7 -L 10 ]-OT/Np(7-15); OT/Np(7-20) and [I 7 -L 10 ]-OT/Np(7-20); OT/Np(1-20) and [A 3 -V 10 ]-OT/Np(1-20)) indicated an increase in K m of 43-, 20-and 47-fold, whereas the corresponding V max values were modified only by factors 6, 7 and 2, respectively (Table 1).Structural support for these in vitro data was provided by analyzing the solution conformation of those peptides (Fig. 5).In H 2 O, the far-UV CD spectra of all the peptides were characterized by a negative band centered at 198 nm, indicative of peptides in a random-coil conformation [22,23].In TFE, their spectra were strongly modified.Indeed, the CD spectrum exhibited by peptide [I 7 -L 10 ]-OT/Np(7-15) (Fig. 5B) is typical of a β-sheet conformation with a negative band near 216 nm and a positive one between 195 and 200 nm [22,23].For peptides [I 7 -L 10 ]-OT/Np(7-20) and [A 3 -V 10 ]-OT/Np(1-20), the shape of their CD spectra (Fig. 5C,D) become typical of α-helical folding with two minima at 208 and 222 nm and a maximum at 193 nm.The helix content of these two peptides, estimated from the ellipticity at 222 nm, increased when compared to reference peptides OT/Np(7-20) (Fig. 2, curve 4) and OT/Np(1-20) (Fig. 5A), respectively.These data indicate that the type of secondary structure at the cleavage loci is the major determinant of the K m value.Consequently, it can be concluded that β-turn structures, or loops, constitute recognition signals which allowed for dibasic specific convertases to discriminate between in vivo cleaved and uncleaved sites [7,8,27].

The prosomatostatin model
Compared to the latter precursor, prosomatostatin (proSom) constitutes a useful model in the study of the formation of multiple hormone products by differential processing of a single polyfunctional precursor [3].Indeed, proSom undergoes both monobasic (R −15 ) and dibasic (R −2 -K −1 ) cleavages to release somatostatin-28 (S-28) and somatostatin-14 (S-14), respectively (Fig. 6).Both basic loci are separated by a dodecapeptide segment S-28(1-12) that corresponds to the NH 2 -terminal sequence of S-28 (Fig. 5).Secondary structure prediction on this connecting region reveals the presence of several β-turn structures [7,14].In addition, the S-28(1-12) sequence includes proline residues which, by both their particular structural properties and their multiplicity in certain protein domains, play a special role in protein structure and function [28][29][30][31].On the basis of theses observations, a combined use of site-directed mutagenesis performed on prosomatostatin substrate, with synthetic peptides selectively modified in their primary structures was adopted in order to evaluate the functional role(s) of the S-28(1-12) domain in the control of cleavage at both loci [14].

Evidence for β-turn in the vicinity of the Arg-Lys doublet
The existence of β-turn in the vicinity of the Arg-Lys doublet was provided by analyzing the conformation of peptides Som(−5-+5) and Som(−9-+5) (Table 2) which share the sequence involved in β-turn formation; i.e. the tetrapeptide P As shown in Fig. 7 (curve 2), the profile of CD spectrum obtained for the peptide Som(−9-+5) is typical of an equilibrium between α-helix, aperiodic structures, and another component of the β-turn type [22,23].The peptide Som(−5-+5), derived from peptide Som(−9-+5) by NH 2 -terminal deletion, exhibits the same spectral pattern (Fig. 7, curve 1) but the negative band around 220 nm decreased whereas the negative one around 203 nm shifted to 198 nm with a concomitant increase in magnitude.
These spectroscopical data, corroborated by IR spectra of both peptides (Fig. 8, panels A and B), argue in favor of the presence of β-turn in the vicinity of the basic doublet [32,33].

Role of β-turn in in vivo processing at the Arg-Lys doublet
To test for the importance of β-turn in in vivo processing, different mutants of human prosomatostatin were constructed in which the sequence P −5 -R-E-R −2 was partially or totally substituted.
These data support the concept that in vivo the scissible bond in the Arg-Lys doublet is situated in the vicinity of a β-turn structure.However, the proportions of S-14 and S-28, recovered from cell extracts of [G −5 ], [S −5 -N −3 ] and [Y −5 -G −3 ] mutants, were not similar to that found in the case of the wild type (Table 3).These observations underline the importance of certain structural features other than the sole β-turn structure in specifying correct processing at both cleavage sites.

Role of P-(X) 3 -P motif in the conformation of S-28(1-12) domain
The presence of proline residues, arranged as P-(X) 3 -P pattern in S-28(1-12) sequence, might be responsible for the differences observed precedently.In order to evaluate the potential role(s) of this motif in prosomatostatin processing, different mutants were constructed in which the P-(X) 3 -P motif was either deleted or mutated.
Results presented in Table 3 underline large differences in the processing of mutants in which the P −5 -A-M-A-P −9 pattern was partially deleted.Indeed, the ratio of S-28 to S-14 produced, which was 1 in Fig. 9. Secondary structure prediction of the human prosomatostatin sequence Som-28 (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12).Helicity per residue calculated for peptides in which (A) -the P −9 -(X)3-P −5 motif was partially deleted and (B) -P −5 or P −9 was mutated by Ala (from [36]).cells expressing the non-mutated precursor [34], was reduced for the ∆[AMA] mutant and raised for the ∆[PP] mutant.Moreover, while increase in cleavage at either the monobasic (∆[PP] mutant) or the dibasic (∆[AMA] mutant) sites resulted in a decrease in cleavage at the other site, deletion of proline residues induced additionally a decrease in the processing efficiency of the ∆[PP] mutant.Data, obtained by the AGADIR method [35], reveal also large variations in helical propensity of the sequences corresponding to these mutants (Fig. 9A).Indeed, deletion of the A-M-A tripeptide (PXXXP motif) decreased exclusively the helicity values per residue of the dibasic site containing domain whereas deletion of proline residues (XAMAX motif) increased both the helicity values and the size of the helical sequence.Together, these results emphasize the functional role of the P −5 -A-M-A-P −9 motif in the generation of equal amounts of S-28 and S-14 from their common precursor.It is noteworthy that statistical studies have revealed that, amongst the P-(X)n-P patterns analyzed in proteins, the P-(X) 3 -P motif is the most frequent [30].
Final demonstration of the importance of the P-(X) 3 -P pattern was provided by examination of prosomatostatin mutants in which proline residues were mutated.As shown in Table 3, replacement of P −9 or P −5 by Ala (α-helix promoting residue) almost abolished selectively the cleavage of the precursor at the monobasic ([A −9 ] mutant) and the dibasic ([A −5 ] mutant) sites.Structural effects induced by these substitutions were firstly investigated by the AGADIR method.Analysis of helicity profiles, obtained for the S-28(1-12) sequences bearing these mutations (Fig. 9B), indicated that substitution of proline residues by Ala favored the extension of an α-helix towards the dibasic site (from N −10 to G +2 for the A −5 mutation: PAMAA motif) and the monobasic one (from N −12 to P −5 for the A −9 mutation: AAMAP motif), respectively.Secondly, conformational analysis of synthetic peptides, corresponding to these mutations, was performed by CD and IR spectroscopies [36].Table 4 summarizes the percentage of α-helix The percentage values of α-helix were calculated from CD spectra using the Fasman method [22,23].b The percentage values of α-helix determined from FT-IR spectra using curve-fitting procedures [37][38][39][40][41]. estimated from their CD and IR spectra [36].Altogether, CD and FTIR results, in agreement with the predictions from the AGADIR method, underline the respective role of each proline residue in both the stability and the precise location of the helical structure adopted by the A-M-A tripeptide motif.
The major emerging concept from this study is that the P −5 -A-M-A-P −9 motif is an helicalpromoting seed whose integrity is essential for alternative prosomatostatin processing at both basic cleavage sites.

Conclusion
In this work was illustrated the usefulness of combined spectroscopical techniques and enzymatic studies in vitro, paralleled by in vivo/in vitro site-directed mutagenesis analysis of the selective effects of various mutations, in establishing the functional role of secondary structure in peptide hormone precursors.These observations provide a new example about the key participation of exposed structural motifs in protein recognition by converting endoproteolytic enzymes called also prohormones convertases [4,6].Similar conclusions were reached in studies of signal peptide excision [42], peptide hormone active forms [43], protein phosphorylation [44], glycosylation of precursors [45] and signal receptor-mediated internalization of lipoproteins [46].
Albeit no crystal structure of precursors so far has emerged, this type of approach should encourage future work on the 3D-structure of particular protein models in structure-function relationships investigations combining high performance spectroscopy and in vivo experiments.It can be envisioned that, in the near future, from stimulating research on pro-proteins and pro-peptides will emerge an heuristic and accurate view on processing mechanisms at the molecular level.This should provide a solid basis for the design of compounds capable to interfere selectively with these essential cellular processes.They might prove useful in the atriogenic control of certain dysregulations resulting from inappropriate protein post-translational modifications which play a central role in a wide range of cellular events.

Fig. 6 .
Fig. 6.A schematic representation of the 92-amino acid human prosomatostatin.Sites I and II correspond to the dibasic (RK) and monobasic (R) cleavage loci for the production of S-14 and S-28, respectively.

Table 1 Kinetic
parameters for the cleavage of OT/Np-related peptides adopting β-turn, β-sheet or α-helix conformations

Table 2
Amino acid sequences of prosomatostatin-related peptide substrates

Table 3
Effects of various mutations on prosomatostatin processing in transfected Neuro2A cells