ABSTRACT
Pernisine is a subtilisin-like protease that was originally identified in the hyperthermophilic archaeon Aeropyrum pernix, which lives in extreme marine environments. Pernisine shows exceptional stability and activity due to the high-temperature conditions experienced by A. pernix. Pernisine is of interest for industrial purposes, as it is one of the few proteases that has demonstrated prion-degrading activity. Like other extracellular subtilisins, pernisine is synthesized in its inactive pro-form (pro-pernisine), which needs to undergo maturation to become proteolytically active. The maturation processes of mesophilic subtilisins have been investigated in detail; however, less is known about the maturation of their thermophilic homologs, such as pernisine. Here, we show that the structure of pro-pernisine is disordered in the absence of Ca2+ ions. In contrast to the mesophilic subtilisins, pro-pernisine requires Ca2+ ions to adopt the conformation suitable for its subsequent maturation. In addition to several Ca2+-binding sites that have been conserved from the thermostable Tk-subtilisin, pernisine has an additional insertion sequence with a Ca2+-binding motif. We demonstrate the importance of this insertion for efficient folding and stabilization of pernisine during its maturation. Moreover, analysis of the pernisine propeptide explains the high-temperature requirement for pro-pernisine maturation. Of note, the propeptide inhibits the pernisine catalytic domain more potently at high temperatures. After dissociation, the propeptide is destabilized at high temperatures only, which leads to its degradation and finally to pernisine activation. Our data provide new insights into and understanding of the thermostable subtilisin autoactivation mechanism.
IMPORTANCE Enzymes from thermophilic organisms are of particular importance for use in industrial applications, due to their exceptional stability and activity. Pernisine, from the hyperthermophilic archaeon Aeropyrum pernix, is a proteolytic enzyme that can degrade infective prion proteins and thus has a potential use for disinfection of prion-contaminated surfaces. Like other subtilisin-like proteases, pernisine needs to mature through an autocatalytic process to become an active protease. In the present study, we address the maturation of pernisine and show that the process is regulated specifically at high temperatures by the propeptide. Furthermore, we demonstrate the importance of a unique Ca2+-binding insertion for stabilization of mature pernisine. Our results provide a novel understanding of thermostable subtilisin autoactivation, which might advance the development of these enzymes for commercial use.
INTRODUCTION
Proteolytic enzymes from the subtilisin family are widely distributed among all domains of life, owing to their diverse functions (1). Extracellular subtilisins are synthesized as inactive precursors that are composed of the signal sequence, the propeptide, and the protease catalytic domain. The signal sequence directs the protease through the cell envelope into the extracellular matrix, where the pro-subtilisin matures into active subtilisin (2). This process starts with folding of the catalytic domain and autocatalytic cleavage of the scissile peptide bond, which generates the noncovalent autoprocessed complex of propeptide-subtilisin. Maturation is completed by dissociation and degradation of the propeptide by subtilisin, which subsequently becomes available for degradation of other substrates (3).
A common feature of subtilisin-like proteases is stabilization of their catalytic domain by Ca2+ ions (4–6). In general, subtilisins from mesophilic Bacillus species have a high-affinity Ca2+-binding site and a low-affinity Ca2+-binding site, as was determined from the crystal structures of various Bacillus subtilisins (4, 7–11). Ca2+ ions stabilize only mature subtilisin, and they are not required for the folding and autocatalytic maturation of the unprocessed pro-subtilisin (12). Instead, the folding of the subtilisin catalytic domain is under the control of the propeptide, which acts as an intramolecular chaperone (13–15). In contrast, pro-Tk-subtilisin from the hyperthermophilic archaeon Thermococcus kodakarensis requires Ca2+ to adopt a suitable conformation for autocatalytic processing into its mature form (16, 17).
The crystal structures of Tk-subtilisin have revealed that it binds seven Ca2+ ions (17, 18), which is the highest number known to date for a subtilisin-like protease. These multiple Ca2+-binding sites in Tk-subtilisin result from two insertion sequences (IS) within the catalytic domain (18). Both of them can promote folding and subsequent maturation of pro-Tk-subtilisin only upon binding Ca2+, while the propeptide acts mainly as an inhibitor of the catalytic domain (16–19). Apparently, the stimuli that trigger this maturation differ between the mesophilic and thermophilic subtilisins, which might be related to the different environmental conditions under which their host organisms are found.
To further address the dependence of thermophilic activation of subtilisins on their Ca2+ binding, we investigated the maturation of pernisine as a Tk-subtilisin homolog from the hyperthermophilic archaeon Aeropyrum pernix. This archaeon was isolated from a coastal solfataric vent and grows at 98°C to 103°C in situ (20). Pernisine is of interest for industrial purposes due to its extraordinary stability and its potential use for degradation of prion proteins (21–23). In this study, we show that maturation of pro-pernisine (pro-Per) is coupled to the Ca2+-induced transition from a disordered Ca2+-free state into the folded Ca2+-bound state. Both of the Tk-subtilisin insertions and its seven Ca2+-binding sites are conserved in pernisine, which implies similar modes of maturation for these proteases. However, an additional insertion sequence was identified in pernisine, with a consensus motif that defines an additional Ca2+-binding site. We show that this unique insertion is crucial for the efficient folding and maturation of pernisine through stabilization of the catalytic domain. Moreover, based on functional and spectroscopic analyses of the isolated pernisine propeptide, we propose a novel model for activation of the archaeal thermophilic subtilisins that explains the high-temperature requirement for their maturation.
RESULTS
Ca2+-binding sites and insertions in the pernisine sequence.Alignment of the pernisine and Tk-subtilisin amino acid sequences revealed that all seven of the Tk-subtilisin Ca2+-binding sites are conserved in pernisine (Fig. 1, Ca1 to Ca7). Of note, some of the Asn and Gln residues in the Ca1 and Ca5 Ca2+-binding sites of Tk-subtilisin are replaced by charged Asp and Glu residues in pernisine. Also, both of the insertion sequences from Tk-subtilisin that are absent in the mesophilic subtilisins (e.g., in subtilisin E) are conserved in pernisine (Fig. 1, IS1 and IS2). However, the pernisine catalytic domain contains an additional insertion of eight amino acids (Fig. 1, Val129 to Asn136) that is not seen for Tk-subtilisin. Together with downstream Asp138 and Gly139, Asp134 and Asn136 from this insertion form a consensus Ca2+-binding motif, DX(D/N)XDG (24). According to the three-dimensional model of pernisine, this unique insertion leads to the formation of a surface loop that contains this further Ca2+-binding motif, which is, again, not seen in Tk-subtilisin (see Fig. S1 in the supplemental material).
Schematic representation of pro-pernisine primary structure and predicted Ca2+ binding sites. (Top) The numbers above the diagram define the amino acid numbering. The signal sequence (SIG.) starts with Met1, the propeptide starts with Ala27, and the catalytic domain starts with Ala95. The unique insertion in pernisine (violet bar) is compared to the corresponding regions of Tk-subtilisin and subtilisin E. The consensus Ca2+-binding motif in the insertion is underlined, and the predicted Ca2+-binding residues are highlighted in violet. (Bottom) Defined amino acid sequences of the insertions (IS1 and IS2) and the predicted Ca2+-binding sites (Ca1 to Ca7) inferred by the alignment of pernisine (UniProt accession no. Q9YFI3), Tk-subtilisin (UniProt accession no. P58502), and subtilisin E (UniProt accession no. P04189). The amino acids shown to interact with Ca2+ in Tk-subtilisin (18) and the corresponding residues in pernisine and subtilisin E are highlighted in violet. Black shading, identical amino acids; gray shading, similar amino acids. Clustal Omega (41) and BoxShade tools were used for the alignments.
Preparation of pro-pernisine variants for maturation assays.Pro-pernisine was overexpressed in Escherichia coli and isolated as described in Materials and Methods. After isolation from the E. coli periplasm, several pro-Per intermediates were observed on SDS-PAGE gels (Fig. 2A). The unprocessed pro-Per migrated as an ∼55-kDa SDS-PAGE band. The N terminus of this unprocessed form was identified as Ala27. This confirmed that pro-Per consisted of the propeptide covalently linked to the catalytic domain and that the N-terminal signal peptide was missing. In addition to the unprocessed form, the autoprocessed pro-Per intermediate (pro-Per*) was present. This autoprocessed intermediate was a noncovalent complex of the propeptide-catalytic domain, which migrated as two bands on SDS-PAGE: an ∼40-kDa band (the catalytic domain) and a faint ∼10-kDa band (the propeptide) (Fig. 2A). The N terminus of the ∼40-kDa SDS-PAGE band was identified as Ala95. This confirmed that the Met94-Ala95 scissile peptide bond between the propeptide and the catalytic domain was cleaved during E. coli cultivation or isolation. As no proteolytic activity was observed without further thermal activation (as described further below), this ∼40-kDa band was not regarded as the mature pernisine form. Additionally, several intermediates that migrated between ∼30 kDa and ∼37 kDa on SDS-PAGE were observed (Fig. 2A). These possibly belonged to degraded forms of pro-Per that resulted from autolysis or degradation by E. coli endogenous proteases.
Pro-pernisine variants for maturation assays. (A) SDS-PAGE analysis of pro-Per, pro-PerΔCa, and pro-PerΔIS isolated by affinity chromatography. The positions of the unprocessed pro-pernisine, pro-pernisineΔCa, and pro-pernisineΔIS; the autoprocessed forms of pro-pernisine and pro-pernisineΔIS (pro-Per* and pro-PerΔIS*, respectively); and the propeptide are indicated with arrows. Note that the autoprocessed form of pro-pernisineΔCa was not present. Degraded forms of these proteins are also marked. The N termini of the SDS-PAGE bands corresponding to pro-pernisine and pro-pernisine* were determined to be Ala27 and Ala95, respectively. Lanes M, protein weight markers, with molecular masses shown next to the gels. (B) Schematic representation of the unprocessed pernisine variants (pro-Per, pro-PerΔCa, and pro-PerΔIS) and their autoprocessed forms (pro-Per* and pro-PerΔIS*). The unique insertion is represented by the violet bars. For pro-pernisineΔCa, the mutations in the predicted Ca2+-binding motif within the insertion (D134A/N136A/D138A) are indicated in red. Amino acid numbering is as in Fig. 1. For the autoprocessed forms, the cleaved scissile bond between the propeptide and the catalytic domain is indicated by a dashed line. Note that the signal sequence of the OmpA protein and the native pernisine signal sequence were not present after isolation from the periplasm, as confirmed by N-terminal sequencing of pro-Per. The 10× histidine tags (His10) are also indicated.
To investigate the role of the unique insertion in pro-Per maturation, two variants of pro-Per were also prepared, designated pro-pernisineΔIS (pro-PerΔIS) and pro-pernisineΔCa (pro-PerΔCa). Pro-PerΔIS had the insertion (V129→N136) deleted, whereas pro-PerΔCa had the Ca2+-binding motif within this insertion mutated, as described in Materials and Methods. As was observed for pro-Per, the isolated pro-PerΔIS and pro-PerΔCa also consisted of several intermediates (Fig. 2A). According to the intermediates that were identified for the isolated pro-Per, the ∼55-kDa bands observed for pro-PerΔIS and pro-PerΔCa corresponded to their unprocessed forms (i.e., pro-PerΔIS and pro-PerΔCa, respectively). Interestingly, the autoprocessed intermediate of ∼40 kDa was observed for pro-ProΔIS (pro-PerΔIS*), but not for pro-PerΔCa. Both of these variants also contained truncated forms that migrated between ∼30 kDa and ∼37 kDa on SDS-PAGE.
All of the pro-Per variants that were prepared for maturation assays and their intermediates are schematically presented in Fig. 2B.
Maturation of pro-pernisine.The maturation of pro-Per was initially investigated at different CaCl2 concentrations over 12-h incubations at 90°C (Fig. 3). In the absence of Ca2+, the unprocessed pro-Per was rapidly converted into the autoprocessed pro-Per*. However, the maturation was not complete without CaCl2, as no proteolytic activity was detected in the azocasein assays over 12 h of incubation, and autolysis occurred. In the presence of CaCl2, the ratio between pro-Per and pro-Per* did not change initially. With further incubation, the autoprocessed pro-Per* started to be converted into the mature form via degradation of the propeptide. This was accompanied by an increase in proteolytic activity. Note that the ∼40-kDa SDS-PAGE band was regarded as the true mature form of pernisine only when proteolytic activity was detected using the azocasein assay and the propeptide was degraded. This was because the SDS-PAGE band that corresponded to the mature form (i.e., the catalytic domain alone) could not be distinguished from the band that corresponded to the autoprocessed form (i.e., the noncovalent complex propeptide-catalytic domain that dissociated during SDS-PAGE), as both forms migrated at ∼40 kDa (Fig. 3). Multiple N termini were identified for the mature form (e.g., Ser106, Asn104, and Gly102), which indicated that the N terminus of the catalytic domain was truncated by ∼10 residues upon complete maturation. Of note, only the autoprocessed intermediate pro-Per* was converted into the mature form, which then degraded the unprocessed pro-Per. This was evident, as the intensity of the ∼40-kDa SDS-PAGE band did not increase with the disappearance of the ∼55-kDa band (Fig. 3). Apparently, most of the unprocessed pro-Per was degraded by the mature pernisine before it was subjected to autoprocessing and further maturation. In addition to the unprocessed pro-Per, the different degraded pro-pernisine forms that migrated between ∼30 kDa and ∼40 kDa on SDS-PAGE were also hydrolyzed by the mature pernisine (Fig. 3). Thus, upon complete maturation, only the mature pernisine was observed on SDS-PAGE gels (Fig. 3).
Maturation of pro-pernisine and its variants with the unique Ca2+ site mutated (pro-pernisineΔCa) and the insertion deleted (pro-pernisineΔIS) at different CaCl2 concentrations. Pro-pernisine, pro-pernisineΔCa, and pro-pernisineΔIS (40 μg/ml) were incubated in the absence of CaCl2 and with 100 μM, 1 mM, and 10 mM CaCl2 at 90°C for the indicated times. Their maturation was analyzed using SDS-PAGE and azocasein assays, as described in Materials and Methods. Lanes M, protein weight markers, with molecular masses shown next to the gels. The arrows indicate the unprocessed form (gray), the autoprocessed/mature form (black), and the propeptide (dashed). The proteolytic activities were measured at the indicated times at 90°C in triplicate and are plotted below the gels, in alignment with their corresponding lanes. The error bars indicate standard deviations. The CaCl2 concentration in the azocasein assays was 1 mM.
The time at which the maturation of pro-Per was completed depended on the CaCl2 concentration. At 100 μM CaCl2, full proteolytic activity was achieved after 6 h at 90°C. When CaCl2 was increased to 1 mM and 10 mM, maturation was completed after 5 h and 2 h, respectively (Fig. 3). In all cases, the proteolytic activity did not decrease significantly even several hours after maturation was completed. Thus, the Ca2+ ions facilitated the maturation process and also stabilized the mature form of pernisine.
The concentrations of CaCl2 used in the maturation assays described above corresponded to molar ratios of CaCl2 to protein of approximately 100:1 at 100 μM CaCl2 to 10,000:1 at 10 mM CaCl2. Next, we examined the minimal CaCl2-to-protein ratio required for complete maturation of pro-Per at 90°C (Fig. 4). A CaCl2-to-protein ratio of 25:1 was sufficient to stabilize the unprocessed pro-Per and the autoprocessed pro-Per*, but no significant proteolytic activity was seen at that ratio. At ratios of 50:1 and 75:1, the proteolytic activity gradually increased. Complete maturation and full activity of pernisine were reached at ratios of at least 100:1, when all pro-Per* molecules were converted into the mature form and the unprocessed pro-Per was degraded.
Maturation of pro-pernisine (A) and its variants with the unique Ca2+ site mutated (B) and with the insertion deleted (C) at different CaCl2-to-protein molar ratios. The proteins were incubated at 90°C for 6 h at the indicated molar ratios of CaCl2 to protein and analyzed using SDS-PAGE and azocasein assays, as described in Materials and Methods. The molar concentration of the target protein (1 μM) was estimated from the mass concentration (40 μg/ml). Lanes M, molecular masses of the protein markers, as shown next to the gels. The arrows indicate the unprocessed form (gray), the autoprocessed/mature form (black), and the propeptide (dashed). The proteolytic activities at the indicated ratios were recorded at 90°C in triplicate and are plotted below the gels, in alignment with their corresponding lanes. The error bars indicate standard deviations. The CaCl2 concentration in azocasein assays was 1 mM.
The maturation of pro-Per was also investigated at lower temperatures in the presence of 10 mM CaCl2 (Fig. 5). At 60°C and 70°C, no maturation was observed, as the unprocessed and autoprocessed forms persisted throughout the 12 h of incubation. Accordingly, no proteolytic activity was detected. The mature pernisine with full proteolytic activity was observed only after 12 h at 80°C.
Maturation of pro-pernisine and its variants with the unique Ca2+ site mutated (pro-pernisineΔCa) and with the insertion deleted (pro-pernisineΔIS) at different temperatures, as indicated. Pro-pernisine, pro-pernisineΔCa, and pro-pernisineΔIS (40 μg/ml) were incubated at 60°C, 70°C, and 80°C for the indicated times in the presence of 10 mM CaCl2. Their maturation was analyzed using SDS-PAGE and azocasein assays, as described in Materials and Methods. Lanes M, molecular masses of the protein markers, as shown next to the gels. The arrows indicate the unprocessed form (gray), the autoprocessed/mature form (black), and the propeptide (dashed). The proteolytic activities at the indicated times were measured at 90°C in triplicate, with the data plotted below the gels, in alignment with their corresponding lanes. The error bars indicate standard deviations. The CaCl2 concentration in the azocasein assays was 1 mM.
Maturation of pro-pernisineΔIS and pro-pernisineΔCa.Deletion of the insertion sequence or mutation of the Ca2+-binding site in the insertion resulted in inefficient maturation of pro-Per. At 0 μM and 100 μM CaCl2, the unprocessed pro-PerΔCa and pro-PerΔIS were converted into their autoprocessed forms but did not further mature over 12 h of incubation at 90°C (Fig. 3). At 1 mM and 10 mM CaCl2, weak proteolytic activity was detected for pro-PerΔCa after 3 h and 4 h. When this proteolytic activity occurred, only the ∼37-kDa SDS-PAGE band was present, along with the unprocessed pro-PerΔCa. This ∼37-kDa band appeared to correspond to the truncated mature pernisineΔCa, which retained proteolytic activity. After 5 h, the ∼37-kDa form was further truncated to the ∼36-kDa form, which was then autolyzed, and the unprocessed pro-PerΔCa was degraded (Fig. 3). Of note, the ∼40-kDa autoprocessed pro-PerΔCa* was not observed at any point during maturation; instead, the autoprocessed pro-PerΔCa* migrated at ∼37 kDa. This might be because the predicted surface loop that contained the mutated Ca2+-binding motif was destabilized. Presumably, the ∼40-kDa autoprocessed form was rapidly converted into the ∼37-kDa autoprocessed form due to proteolytic cleavage in the region. This ∼37-kDa form of pernisineΔCa apparently still associated with the propeptide, as the proteolytic activity increased when the propeptide was degraded (Fig. 3).
For pro-PerΔIS, the mature form was observed only after 4 h with 10 mM CaCl2 (Fig. 3). However, the yield of the mature pernisineΔIS was significantly lower than that for pernisine with an intact insertion, so that no proteolytic activity was detected using the azocasein assay. The mature pernisineΔIS was also rapidly autolyzed during further incubation (Fig. 3). This autoproteolysis of pernisineΔCa and pernisineΔIS occurred over various ranges of CaCl2-to-protein ratios (Fig. 4B and C). This again implied that both of the variants mature in an excess of Ca2+ but are not responsive to stabilization by Ca2+ and are rapidly degraded.
The maturation of pro-PerΔCa and pro-PerΔIS was too slow at 60°C and 70°C to detect their mature forms after 12 h (Fig. 5). At 80°C, weak proteolytic activity was observed for pro-PerΔCa after 12 h, which implied that maturation was initiated at that temperature. The mature form of pernisineΔIS was also observed only after 12 h at 80°C. Again, the yield of mature pernisineΔIS was insufficient to detect its proteolytic activity using the azocasein assay.
Preparation of unprocessed pro-pernisine variants for folding analyses.To understand the Ca2+-facilitated maturation of pro-Per, we investigated the impact of Ca2+ ions on the unprocessed pro-Per conformation. To obtain the unprocessed pro-Per variants, the Ser355 in the active site of the catalytic domain was replaced with Ala. In this way, the autoprocessing and subsequent maturation of pro-Per were prevented. Three pro-Per variants with the active site mutated were prepared: pro-pernisineS355A (pro-PerS355A), pro-pernisineΔISS355A (pro-PerΔISS355A), and pro-pernisineΔCaS355A (pro-PerΔCaS355A). Pro-PerΔISS355A had the unique insertion deleted, whereas pro-PerΔCaS355A had the Ca2+-binding motif within the insertion mutated, as described in Materials and Methods.
After affinity chromatography, the isolated pro-PerS355A contained two N-terminally truncated forms (Asn136 through Ser430 [Asn136–Ser430] and Ala146–Ser430), as well as the target unprocessed pro-PerS355A (Ala27–Ser430), as confirmed by Western blotting and N-terminal sequencing (Fig. 6A). The N terminus of the unprocessed pro-PerS355A (Ala27) was determined after the final purification step, as well. After the final purification step, the N-terminally truncated form of Ala146–Ser430 was still present, representing ∼20% of the total protein (Fig. 6A). Attempts to separate this truncated form from the unprocessed pro-PerS355A by ion-exchange or size exclusion chromatography were not successful, due to the low resolution and significant loss of protein. This truncated form was also observed for pro-PerΔCaS355A after purification (Fig. 7H). In contrast, pro-PerΔISS355A was purified to a single ∼55-kDa band on SDS-PAGE (Fig. 7I) using the same purification strategy as for pro-PerS355A and pro-PerΔCaS355A. The unprocessed pro-Per variants with the active site mutated are schematically presented in Fig. 6B.
Active-site mutants of the pro-pernisine variants. (A) Analysis of pro-pernisineS355A (pro-PerS355A) purity after IMAC and after the final purification step by SDS-PAGE and Western blotting. Lanes M, molecular masses of the protein markers, as shown next to the gels. The arrows (right) indicate the bands that correspond to the unprocessed pro-pernisineS355A and the two N-terminally truncated variants of pro-pernisineS355A (Asn136–Ser430 and Ala146–Ser430). These bands were subjected to N-terminal sequencing after IMAC. The band corresponding to the target pro-pernisineS355A was also subjected to N-terminal sequencing after the final purification step. (B) Schematic representations of the isolated unprocessed pro-pernisine variants that were used for folding analyses: pro-pernisineS355A, the unprocessed pro-pernisineS355A; pro-pernisineΔCaS355A (pro-PerΔCaS355A), the unprocessed pro-pernisineΔCaS355A; and pro-pernisineΔISS355A (pro-PerΔISS355A), the unprocessed pro-pernisineΔISS355A. The unique insertion and the mutations in the Ca2+-binding motif are indicated as in Fig. 2. The Ser355→Ala355 substitution in the active site is indicated by the asterisk (S355A). Amino acid numbering is as in Fig. 1. Note that the signal sequence of the OmpA protein and the native pernisine sequence were not present after isolation from the periplasm.
Folding analyses of the pro-pernisine active-site mutant (pro-pernisineS355A) and its variants with the unique Ca2+-binding site mutated (pro-pernisineΔCaS355A) or insertion deleted (pro-pernisineΔISS355A). (A to C) Far-UV CD (A), near-UV CD (B), and tryptophan fluorescence (C) spectra. Red lines, pro-pernisineS355A; black lines, pro-pernisineΔCaS355A; gray lines, pro-pernisineΔISS355A; dashed lines, without CaCl2; solid lines, with CaCl2 (CaCl2-to-protein molar ratio, 200:1). (C) The fluorescence spectra of pro-pernisineS355A, pro-pernisineΔCaS355A, and pro-pernisineΔISS355A were normalized to their respective spectra measured in the absence of CaCl2. (D to F) The molar ellipticities at 200 nm (D) and 280 nm (E) and the ratios of the fluorescence intensities at 330 nm and 350 nm (F) were plotted as a function of the CaCl2-to-protein ratio. Red, pro-pernisineS355A; black, pro-pernisineΔCaS355A; gray, pro-pernisineΔISS355A. (G to I) SDS-PAGE analysis of the Ca2+-induced folding of pro-pernisineS355A (G), pro-pernisineΔCaS355A (H), and pro-pernisineΔIS S355A (I). The ratios of CaCl2 to protein are as indicated above the lanes. Gray arrows, Ca2+-free form of pro-pernisineS355A, pro-pernisineΔCaS355A, or pro-pernisineΔISS355A; black arrows, Ca2+-bound form of pro-pernisineS355A, pro-pernisineΔCaS355A, or pro-pernisineΔISS355A; dashed arrows, Ca2+-free form of N-terminally truncated pro-pernisineS355A or pro-pernisineΔCaS355A. Lanes where the sample was precipitated with trichloroacetic acid prior to SDS-PAGE are indicated with black dots.
Spectroscopic analysis of pro-pernisine folding.The pro-PerS355A in Ca2+-free medium was predominantly unstructured, with strong molar ellipticity at 200 nm (Fig. 7A). Upon addition of CaCl2, there was a substantial increase in the secondary structure, as the far-UV circular dichroism (CD) spectrum of Ca2+-bound pro-PerS355A was characteristic for α-helical folding. This conformational transition of pro-PerS355A occurred at molar ratios of CaCl2 to pro-PerS355A from 60:1 to 80:1 (Fig. 7D). The near-UV CD spectra of pro-PerS355A showed a stronger band at ∼280 nm upon addition of CaCl2 (Fig. 7B), which implied an increase in the tertiary structure. Similar to the far-UV CD, this transition occurred at molar ratios of CaCl2 to pro-PerS355A from 60:1 to 90:1 (Fig. 7E). Ca2+-induced folding of pro-PerS355A was also evident from the intrinsic tryptophan fluorescence. With the addition of CaCl2, the fluorescence intensity increased and the maximum fluorescence wavelength shifted from 346 nm to 334 nm, which implied a transition of the tryptophan residues into a more hydrophobic environment (Fig. 7C). This transition was observed at molar ratios of CaCl2 to pro-PerS355A from 60:1 to 100:1 (Fig. 7F).
Pro-PerΔCaS355A and pro-PerΔISS355A were also unstructured in the absence of Ca2+, as was seen from the far-UV CD (Fig. 7A), near-UV CD (Fig. 7B), and fluorescence (Fig. 7C) spectra. The conformational transition of pro-PerΔCaS355A upon addition of CaCl2 was nearly identical to that of pro-PerS355A according to the far-UV CD and intrinsic tryptophan fluorescence (Fig. 7A and C). However, the near-UV CD signal of the Ca2+-bound pro-PerΔCaS355A was significantly weaker than that of the Ca2+-bound pro-PerS355A (Fig. 7B). This implied that disruption of the Ca2+-binding motif in the unique insertion perturbed the Ca2+-induced folding of pro-PerΔCaS355A into the tertiary structure. This folding was even more affected when the whole insertion was deleted; the far-UV CD spectrum of pro-PerΔISS355A showed a higher degree of random coils after addition of CaCl2 than that of pro-PerS355A (Fig. 7A). Also, the increase in the pro-PerΔISS355A fluorescence intensity upon addition of CaCl2 was not as prominent as for pro-PerS355A (Fig. 7C). As was observed for pro-PerΔCaS355A, pro-PerΔISS355A also showed a weaker near-UV CD signal than pro-PerS355A both in the absence of CaCl2 and with excess CaCl2.
Conformational transitions of pro-PerΔCaS355A and pro-PerΔISS355A occurred at higher CaCl2-to-protein molar ratios than those of pro-PerS355A (Fig. 7D to F). The folding midpoints of these transitions are given in Table 1, as estimated from the folding curves (Fig. 7D to F).
Molar ratios of CaCl2 to protein for pro-PerS355A and pro-PerΔISS355A at which the folding midpoint was reacheda
Electrophoretic analysis of pro-pernisine folding.The impact of CaCl2 on the pro-Per conformation was further analyzed using SDS-PAGE, where it was possible to distinguish pro-Per in the Ca2+-free form and the Ca2+-bound form (Fig. 7G to I). In the absence of Ca2+ ions, pro-PerS355A migrated as a 55-kDa band. When pro-PerS355A was incubated with CaCl2 prior to SDS-PAGE, the 55-kDa band appeared fainter and was accompanied by a band at 30 kDa (Fig. 7G). If the pro-PerS355A preincubated with CaCl2 was precipitated with trichloroacetic acid prior to SDS-PAGE, it again migrated as the 55-kDa band. Therefore, the 30-kDa band was a result of the Ca2+-induced conformational change of pro-PerS355A rather than any truncation caused by the binding of Ca2+. Apparently, the Ca2+-bound conformation was stable enough not to denature under SDS-PAGE conditions, and consequently, the Ca2+-bound form migrated faster than the Ca2+-free form. Thus, the 55-kDa and 30-kDa bands corresponded to the Ca2+-free and Ca2+-bound forms, respectively. The Ca2+-bound form of pro-PerS355A appeared at molar ratios of CaCl2 to protein of 80:1 and higher, which was concomitant with the decreased intensity of the 55-kDa band (Fig. 7G). The Ca2+-bound forms of pro-PerΔCaS355A and pro-PerΔISS355A appeared at CaCl2-to-protein ratios higher than those of pro-PerS355A, as they were detected at ratios of at least 200:1 (for pro-PerΔCaS355A) (Fig. 7H) and 120:1 (for pro-PerΔISS355A) (Fig. 7I).
Thermal stability of pro-pernisine.The effects of temperature on the conformational stability of pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A were investigated by CD in the presence of CaCl2. The secondary structure of pro-PerS355A retained the α-helical fold at 90°C, although the signal recorded at 90°C was a little weaker than that at 20°C (Fig. 8A). However, the tertiary structure of pro-PerS355A appeared to remain intact upon heating to 90°C (Fig. 8B). Also, for pro-PerΔCaS355A, the far-UV and near-UV CD spectra were not considerably changed at 90°C (Fig. 8A and B). Interestingly, pro-PerΔISS355A underwent a small conformational change at 90°C that resulted in a secondary structure that had a lower degree of random coiling than at 20°C (Fig. 8A). The tertiary structure of pro-PerΔISS355A was not considerably changed at 90°C (Fig. 8B).
Conformational stability of the pro-pernisine active-site mutant (pro-pernisineS355A) and its variants with the unique Ca2+-binding site mutated (pro-pernisineΔCaS355A) or with the insertion deleted (pro-pernisineΔISS355A). The far-UV CD (A) and near-UV CD (B) spectra were measured at a CaCl2-to-protein molar ratio of 200:1 at 20°C (solid lines) and at 90°C (dashed lines). Red lines, pro-pernisineS355A; black lines, pro-pernisineΔCaS355A; gray lines, pro-pernisineΔISS355A.
Propeptide conformation and inhibitory properties.The propeptide of pernisine was overexpressed in E. coli and purified to homogeneity (see Fig. S2 in the supplemental material), as described in Materials and Methods. The purified propeptide was folded at 20°C, as confirmed by far-UV and near-UV CD spectroscopy (Fig. 9A and B). After heating to 90°C, the far-UV signal was slightly decreased, implying minor disruption to the secondary structure (Fig. 9A). A greater change occurred in the near-UV region at 90°C, which indicated significant loss of tertiary structure of the propeptide upon heating (Fig. 9B).
Characterization of propeptide conformation and inhibitory activity. (A and B) The far-UV CD (A) and near-UV CD (B) spectra of the propeptide were recorded at 20°C (black lines) and 90°C (red lines). (C to F) The activities of the mature pernisine (2 nM) at the indicated concentrations of the propeptide were recorded at 20°C (C and D) and 90°C (E and F) using Suc-AAPF-pNA (0.3 mM) as a substrate, as described in Materials and Methods. The propeptide was either mixed with Suc-AAPF-pNA prior to addition of the mature pernisine (C and E) or incubated with the mature pernisine prior to addition of Suc-AAPF-pNA (D and F).
Inhibition of the mature pernisine by the propeptide was assayed using the synthetic substrate Suc-AAPF-pNA (N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide) at 20°C and 90°C (Fig. 9C to F). The propeptide was either premixed with the substrate before addition of the mature pernisine (Fig. 9C and E) or preincubated with the mature pernisine before addition of the substrate (Fig. 9D and F). When the propeptide was premixed with the substrate, the progress lines were curved at 20°C, showing slow binding inhibition (Fig. 9C). At 20°C, ∼50% inhibition was achieved with 10 nM propeptide (Fig. 9C). When the propeptide was preincubated with the mature pernisine, the progress lines were straight (Fig. 9D). This was because the binding of the propeptide to the mature pernisine had already reached equilibrium during the preincubation. Consequently, ∼50% inhibition had already been achieved with 2 nM propeptide at 20°C (Fig. 9D). The inhibition was stronger at 90°C than at 20°C. When the propeptide was premixed with the substrate, ∼50% inhibition was achieved with 1 nM propeptide, and 5 nM propeptide completely inhibited the mature pernisine at 90°C (Fig. 9E). Interestingly, preincubation of the propeptide with the mature pernisine resulted in weaker inhibition at 90°C; however, complete inhibition was again achieved with 5 nM propeptide (Fig. 9F). Of note, 5 nM propeptide did not completely inhibit the mature pernisine at 20°C, as the corresponding progress lines were still ascending (Fig. 9C and D).
DISCUSSION
The slow maturation of pro-pernisine.In this study, we demonstrate that, in contrast to bacterial pro-subtilisins, the autocatalytic maturation of pro-Per into its active form is mediated by Ca2+ ions. This dependence of pro-Per maturation on Ca2+ is similar to that reported for the pernisine homolog Tk-subtilisin (16). However, pro-Per matured at considerably lower rates than pro-Tk-subtilisin. At 90°C, the maturation of pro-Per was complete after 120 min in the presence of excess Ca2+, whereas pro-Tk-subtilisin was already fully matured after 10 min at 80°C (25). Furthermore, pro-Per maturation was inefficient at temperatures lower than 80°C, whereas pro-Tk-subtilisin completed maturation even at 60°C (25). In contrast to maturation, the autoprocessing step of pro-Per (i.e., cleavage of the scissile peptide bond) is not dependent on high temperatures.
The rate-limiting steps of pro-subtilisin maturation are propeptide release from the cleaved propeptide-subtilisin complex and degradation of the released propeptide (3). Thus, these two steps might contribute to slower maturation of pro-Per than of pro-Tk-subtilisin. Following isolation, the propeptides from subtilisins of mesophilic bacteria are disordered and adopt their folded conformations only when in complex with their cognate catalytic domains (15, 26, 27). Upon dissociation from the catalytic domain, the propeptide is destabilized and degraded, which leads to pro-subtilisin maturation at low temperatures (28). In contrast, we show here that the isolated propeptide of pernisine adopts the folded conformation and acts as a slow-binding inhibitor of the mature pernisine at 20°C, similar to the Tk-subtilisin propeptide (29, 30). However, direct comparisons of stabilities and inhibitory activities between the pernisine and Tk-subtilisin propeptides would be required to definitively explain the slower maturation of pro-Per.
Thermally induced pro-pernisine maturation is linked to the propeptide conformation.Marie-Claire et al. (31) reported denaturation of the thermophilic aqualysin I propeptide at 80°C and suggested that it had lower inhibitory activity at the high temperatures that are optimal for aqualysin I function. In contrast, here, we show that inhibition of pernisine by the propeptide was stronger at higher temperatures. This implies that the propeptide is not unfolded at high temperatures when complexed with the pernisine catalytic domain and that the interactions between the two are stronger under such conditions. However, upon dissociation from the catalytic domain, the pernisine propeptide might unfold at 90°C, but not at 20°C, as indicated by the near-UV CD. Consequently, degradation of the pernisine propeptide by the catalytic domain is effective only at temperatures that are high enough for destabilization of the dissociated propeptide. This might explain the high-temperature requirements for pro-Per maturation. To our knowledge, no previous studies have reported functionality of the thermophilic propeptide at the temperature that is optimal for its cognate subtilisin activity. Together with mutational studies of Tk-subtilisin propeptide stability at 20°C (29, 30), the data presented here support the hypothesis that thermostable subtilisins and their cognate propeptides have coadapted to the high-temperature environment.
The role of Ca2+ in pro-pernisine maturation.As discussed above, the maturation of pro-Per requires Ca2+ and high temperatures. Our results show that Ca2+ ions induce folding of the unprocessed pro-Per. The molar ratio of CaCl2 to pro-PerS355A that was required for complete folding was in line with the ratio at which pro-Per* completed maturation and reached its full proteolytic activity. This demonstrates that Ca2+-induced folding of pro-Per is a prerequisite for its activation. Nevertheless, the maturation process of pro-Per was unusual in one key point: the unprocessed pro-Per was rapidly autoprocessed into pro-Per* only in the absence of CaCl2. When CaCl2 was added, this conversion (pro-Per→pro-Per*) was not observed; instead, pro-Per was degraded by the mature pernisine. The mature pernisine was produced from the preexisting pro-Per*, which was copurified with the unprocessed pro-Per from E. coli. Importantly, pro-Per* was converted into mature pernisine only when CaCl2 was present. Based on these observations, Ca2+ ions (i) induce folding of unprocessed pro-Per, (ii) considerably lower the rate of autoprocessing (i.e., pro-Per→pro-Per* conversion), (iii) enable complete maturation of pro-Per* (i.e., pro-Per*→mature pernisine conversion), and (iv) stabilize the mature form. This Ca2+- and temperature-dependent in vitro maturation of pro-Per is summarized in Fig. 10.
Schematic model for pro-pernisine maturation. The propeptide (violet) is connected to the catalytic domain (blue) via the peptide linker (black line), which contains the scissile bond. The linker is composed of the C-terminal region of the propeptide and the N-terminal region of the catalytic domain. The active site on the catalytic domain is represented by an orange asterisk. Disordered/destabilized domains are bordered by dotted lines, whereas folded domains are bordered by solid lines. (A) In the absence of Ca2+, the disordered pro-pernisine undergoes fast autoprocessing (i.e., cleavage of the scissile bond) into the noncovalent complex of the propeptide-catalytic domain that dissociates reversibly. At 20°C, the dissociated propeptide is resistant to degradation by the catalytic domain. At 90°C, the autoprocessed complex is subjected to autolysis. (B) In the presence of Ca2+, the unprocessed pro-pernisine adopts a folded conformation. Autoprocessing of this Ca2+-bound pro-pernisine is slow. At 20°C, the noncovalent complex of the propeptide-catalytic domain dissociates more readily than at 90°C, as shown by the arrows. The dissociated propeptide is destabilized and degraded by the Ca2+-bound catalytic domain only at 90°C, which leads to formation of the mature pernisine.
The disordered, Ca2+-free pro-Per is therefore more amenable to autoprocessing than the folded, Ca2+-bound pro-Per. When produced in E. coli, the unprocessed pro-Per had apparently been exposed to low Ca2+ concentrations. Consequently, a proportion of the pro-Per molecules underwent autoprocessing into pro-Per*, as was observed after purification. There is a lack of evidence for the intracellular levels of free Ca2+ in A. pernix and other marine hyperthermophilic archaea. However, Ca2+ is abundant in marine water, which is the extracellular environment of these microorganisms. As pro-Per is secreted by A. pernix, it might be unlikely that it is exposed to a Ca2+-free environment in vivo (i.e., the A. pernix cytoplasm), at least not long enough to undergo fast autoprocessing. If the secretion rate of the newly synthesized pro-Per is high, the secreted pro-Per would first bind Ca2+ and fold into the ordered conformation. Then, the Ca2+-bound pro-Per would undergo slow autoprocessing and maturation. A similar process of in vivo maturation has also been discussed for Tk-subtilisin (29) and thermolysin-like proteases (32). Nonetheless, other routes of pro-Per in vivo maturation are possible, as well. For example, the disordered pro-Per might already be autoprocessed in the A. pernix cytoplasm. After secretion, the autoprocessed pro-Per* would bind Ca2+ and adopt the ordered conformation, which would then enable the degradation of the propeptide to complete maturation.
Role of the unique pernisine insertion in pro-pernisine maturation.The resemblance of Ca2+-modulated folding and maturation of pro-Per to that of pro-Tk-subtilisin is not surprising, as all seven of the Ca2+-binding sites of Tk-subtilisin are conserved in pernisine. Interestingly, the glutamine and asparagine residues that form the Ca1 and Ca5 sites in Tk-subtilisin are replaced by glutamate and aspartate in pernisine. This introduction of negatively charged side chains might increase the affinity for Ca2+ at these sites, as was also reported for thermitase (33). Pro-Tk-subtilisin already binds Ca2+ in its unprocessed form and undergoes subsequent folding and autocatalytic activation, mainly through the two insertion sequences (18, 19, 34, 35). As these two insertions are conserved in pernisine, similar modes of Ca2+ binding and maturation are expected in the two proteases. However, pernisine contains an additional insertion that forms a further Ca2+-binding motif. It appears that this insertion contributes to the formation of a surface loop just upstream of the central β-strand that builds the core αβα structure of the catalytic domain. The two N-terminally truncated forms of pro-PerS355A that were observed during the isolation from E. coli resulted from cleavage in the region that corresponds to the predicted surface loop. The N-terminally truncated form was also observed for pro-PerΔCaS355A, but not for pro-PerΔISS355A, which suggested that this insertion indeed contributes to the formation of an unstable region. Such a region might correspond to a surface loop that is flexible and prone to cleavage by endogenous proteases in the absence of Ca2+.
Our results show that this unique insertion with the Ca2+-binding motif is crucial for efficient folding of the unprocessed pro-Per and for proteolytic stability of the catalytic domain during maturation, although it is not required for conformational stability of pro-Per at temperatures up to 90°C. Such prominent stabilizing effects of the insertion during maturation were not expected, considering that the homologous pro-Tk-subtilisin is efficiently matured even though it is devoid of the insertion (16). This investigated insertion and the corresponding Ca2+-binding motif appear to compensate for other structural dissimilarities that have emerged in pernisine compared to its homolog, Tk-subtilisin. It has been reported that pyrolysin also contains an insertion within the catalytic domain that contributes to its activation and thermostability through the binding of Ca2+ (36). Therefore, acquisition of insertion sequences as Ca2+-binding motifs appears to be a common strategy for thermal adaptation of thermostable subtilisin-like proteases.
MATERIALS AND METHODS
Construction of expression plasmids.The DNA primers used in this study are listed in Table 2. Phusion high-fidelity DNA polymerase (Thermo Scientific) was used for all of the DNA amplification steps. Codon-optimized nucleotide sequences that encode full-length pernisine, including its propeptide and signal sequence (pro-Per) and its active-site mutant (pro-PerS355A), were amplified from the pMCSG7 expression plasmids used in our previous study (22) by PCR using primers P1 and P2. The sequence that encodes pro-PerS355A had codon GCC in place of TCG, to generate the Ser355→Ala355 mutation. The PCR products were purified using GeneJET purification kits (Thermo Scientific) and treated with the XhoI and EcoRI restriction enzymes. The resulting DNA fragments were ligated between XhoI and EcoRI sites in plasmid pMD204 (37) to generate pMD204_proPer and pMD204_proPerS355A. These vector constructs encoded pro-Per and pro-PerS355A from Gly2 to Ser430 (UniProt accession number Q9YFI3) with the OmpA signal from E. coli at the N terminus and a His10 tag at the C terminus.
Oligonucleotide sequences used in this study
The insertion sequence encoded residues Val129 to Asn136, and it was deleted from the pro-Per sequence by overlap extension PCR (38). The pMCSG7 expression plasmids from our previous study (22) were used as templates for the amplification steps. The region upstream of the IS in the genes for pro-Per and pro-PerS355A was amplified using the primer pair P1 and P3 in a molar ratio of 10:1, and the region downstream of the IS was amplified using the primer pair P4 and P2 in a molar ratio of 1:10. The amplified single strands that corresponded to the upstream and downstream regions were combined and hybridized using a temperature gradient (90°C to 4°C over 60 min). The complementary strands were elongated and further amplified using five PCR amplification cycles. The resulting DNA fragments that encoded the full-length pernisine and its active-site mutant (with deleted IS; pro-PerΔIS and pro-PerΔISS355A, respectively) were cloned into the pMD204 plasmid between the XhoI and EcoRI sites, as described above, to obtain pMD204_proPerΔIS and pMD204_proPerΔISS355A. These plasmid constructs encoded pro-PerΔIS and pro-PerΔISS355A with an OmpA signal at the N terminus and a His10 tag at the C terminus.
The Ca2+-binding motif within the IS was mutated by replacing Asp134, Asn136, and Asp138 with alanines (i.e., Asp134→Ala134, Asn136→Ala136, and Asp138→Ala138). These mutations were introduced by overlap extension PCR as described above for deletion of the IS. For this, primer pair P1 and P5 was used in a molar ratio of 10:1, and primer pair P6 and P2 was used in a molar ratio of 1:10. The amplified single strands were hybridized, elongated, amplified, and cloned into the pMD204 plasmid as described above. The plasmids pMD204_proPerΔCa and pMD204_proPerΔCaS355A obtained encoded pro-PerΔCa and pro-PerΔCaS355A, respectively.
To obtain the tag-free propeptide of pernisine, the empty plasmid pMCSG7 was first digested with the NdeI and XhoI restriction enzymes to remove the His6 tag downstream of the T7 promoter. The DNA fragment that encoded the propeptide (Ala27–Met94) was amplified from the pMCSG7 plasmid that encoded pro-Per (22) using the primers P7 and P8 and then purified as described above. The PCR product was digested with the NdeI and XhoI restriction enzymes and ligated into vector pMCSG7 between the NdeI and XhoI restriction sites to obtain pMB_Pro.
All of the plasmids constructed in this study were propagated in competent E. coli JM107 cells and purified using GeneJET plasmid miniprep kits (Thermo Scientific). The nucleotide sequences of the target DNAs were determined by sequencing (Macrogen).
Overexpression and purification of recombinant proteins.For biosynthesis of pro-Per, pro-PerS355A, pro-PerΔCa, pro-PerΔCaS355A, pro-PerΔIS, and pro-PerΔISS355A, competent E. coli BL21(DE3) cells were transformed with the pMD204 expression constructs as described above. The cells were grown at 37°C in 2× yeast extract tryptone broth (Carl Roth) supplemented with chloramphenicol (34 μg/ml). When an optical density at 600 nm (OD600) of ∼1.5 was reached, protein expression was induced with 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG), followed by incubation at 23°C with agitation at 225 rpm for 16 h. For biosynthesis of the propeptide, competent E. coli BL21(DE3) cells were transformed with pMB_Pro and grown at 37°C in Luria–Bertani broth (Carl Roth) supplemented with ampicillin (110 μg/ml). Propeptide expression was induced at an OD600 of ∼0.8, using 1 mM IPTG. After addition of the IPTG, the cells were incubated at 23°C with agitation at 225 rpm for 16 h.
All of the chromatographic steps were carried out using an NGC chromatography system (Bio-Rad). The His10-tagged pernisine variants (i.e., pro-Per, pro-PerS355A, pro-PerΔCa, pro-PerΔCaS355A, pro-PerΔIS, and pro-PerΔISS355A) were isolated from the E. coli periplasm using immobilized metal affinity chromatography (IMAC). The periplasmic fractions were obtained as described previously (39) with some modifications. Briefly, the cells were harvested by centrifugation at 3,000 × g for 20 min at 4°C and resuspended in TSE buffer (100 mM Tris-HCl, pH 8.0, 20% sucrose, 0.5 mM EDTA). The cell suspensions were incubated on ice for 15 min and centrifuged at 4,000 × g for 15 min at 4°C. The pellets were resuspended in 10 mM Tris (pH 8.0), incubated on ice for 60 min, and centrifuged at 8,000 × g for 20 min at 4°C. The collected supernatants were regarded as the periplasmic fractions. For purification of pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A, the periplasmic fractions were supplemented with 300 mM NaCl and 10 mM imidazole and passed through HisTrap HP columns (GE Healthcare), which were preequilibrated with IMAC buffer (50 mM NaH2PO4, 300 mM NaCl, pH 8.0) containing 10 mM imidazole. The columns were then washed with 25 column volumes of the same buffer. Finally, pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A were eluted with IMAC buffer containing 250 mM imidazole and dialyzed against 10 mM Tris (pH 8.0) overnight. After dialysis, the proteins were incubated with 5 mM CaCl2 at 90°C for 30 min and then centrifuged at 18,000 × g for 15 min at 4°C to remove heat-labile protein contaminants. The remaining protein impurities in the supernatants were hydrolyzed by addition of activated pernisine (2 μg/ml), with incubation at 90°C for 30 min. After centrifugation at 18,000 × g for 15 min at 4°C, the supernatants containing the purified pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A were treated with 10 mM EDTA and dialyzed against 10 mM Tris-HCl (pH 8.0) for 30 h, with two buffer changes, to obtain the proteins in their Ca2+-free forms. After the final dialysis, the purified proteins were concentrated using 10-kDa-cutoff centrifuge filters (Ultra-15; Amicon) and then stored at −80°C. The purification procedures for pro-Per, pro-PerΔCa, and pro-PerΔIS were the same as for pro-PerS355A, pro-PerΔCaS355A, and pro-Per-ΔISS355A, except that the proteins eluted from the HisTrap HP columns were directly stored at −80°C after the first dialysis against 10 mM Tris (pH 8.0). The concentrations of the purified proteins were determined with a combination of bicinchoninic acid assays and SDS-PAGE using 12% polyacrylamide gels.
The propeptide was isolated from the E. coli cytoplasm. The cells were harvested with centrifugation at 4,000 × g for 25 min at 4°C, resuspended in 50 mM Tris (pH 7.0), and sonicated on ice. Cellular debris was removed by centrifugation at 15,000 × g for 15 min at 4°C. To remove the heat-labile proteins, the clarified lysate was incubated at 80°C for 20 min and then cooled on ice and centrifuged at 18,000 × g for 15 min at 4°C. The supernatants were collected and combined with 1.5 M (NH4)2SO4 (final concentration). After centrifugation at 18,000 × g for 15 min at 4°C, the supernatants were loaded onto HiTrap phenyl (LS) columns (GE Healthcare) that were preequilibrated with 1.5 M (NH4)2SO4 in 50 mM Tris (pH 7.0). The columns were then washed with 30 column volumes of the same buffer. The propeptide was eluted with 0.6 M (NH4)2SO4 in 50 mM Tris (pH 7.0) and dialyzed against 10 mM Tris (pH 8.0) overnight. After dialysis, the propeptide was loaded onto a HiTrap Q HP column (GE Healthcare) and eluted with a linear gradient from 0 mM to 500 mM NaCl in 20 mM Tris-HCl (pH 8.0) as a single peak at ∼180 mM NaCl. The purified propeptide was dialyzed against 10 mM Tris (pH 8.0) and stored at −80°C. The concentrations of the propeptide were calculated using absorbance at 280 nm and an extinction coefficient of 1,490 M−1 cm−1. This extinction coefficient was obtained using the online ProtParam tool (ExPASy). The purity of the propeptide was analyzed using SDS-PAGE with 15% polyacrylamide gels.
Western blotting and N-terminal sequencing.Proteins were separated using SDS-PAGE and then electrotransferred from the gels onto polyvinylidene difluoride (PVDF) membranes. For Western blotting, the membranes were first blocked in Tris-buffered saline supplemented with 5% (wt/vol) dry milk at room temperature for 30 min. Afterward, the membranes were incubated with a rabbit polyclonal anti-histidine antibody (dilution, 1:1,500; Abcam), for 2 h at room temperature. After washing with phosphate-buffered saline, the membranes were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG antibodies (dilution, 1:2,000; Abcam) overnight at 4°C. After a final wash with phosphate-buffered saline, the bound peroxidase-conjugated antibodies were detected using ECL Plus reagent (GE Healthcare) according to the manufacturer’s instructions.
For N-terminal sequencing, the PVDF membranes were stained with Coomassie R-250. The bands of interest were cut out, and the N termini were sequenced using the Edman degradation method (Procise protein-sequencing system 492A; PE Applied Biosystems). The resulting phenylthiohydantoin-amino acid derivatives were identified using a Spheri-5 RP-18 (C18) column (Brownlee) by high-performance liquid chromatography (HPLC) (140C system; PE Applied Biosystems).
Preparation of mature pernisine.To produce the active, mature form of pernisine, the pro-Per isolated by IMAC was diluted to a final concentration of 40 μg protein/ml in 50 mM HEPES (pH 8.9) supplemented with 1 mM CaCl2. To produce the mature pernisine, this reaction mixture was incubated at 90°C for 6 h. After heat activation, the mature pernisine was stored at −80°C until it was used. The concentrations of mature pernisine were estimated using SDS-PAGE with 15% gels, with purified pro-PerS355A as the standard.
Maturation assays.The isolated pro-Per, pro-PerΔCa, and pro-PerΔIS were diluted in 50 mM HEPES (pH 8.9) to 40 μg protein/ml and incubated at defined temperatures in 250-μl aliquots with or without different concentrations of CaCl2. At the appropriate time intervals, 50 μl was withdrawn from each aliquot and immediately frozen at −20°C for later proteolytic activity analysis. The maturation in the remaining 200 μl was stopped by addition of 22.4 μl 100% trichloroacetic acid. The precipitated proteins were pelleted by centrifugation at 18,000 × g for 10 min at 4°C, washed with 100% acetone, and centrifuged again, as before. The pellets were dissolved in 4× sample buffer for SDS-PAGE, supplemented with 4 mM phenylmethylsulfonyl fluoride and 4 mM EDTA, boiled at 100°C for 5 min, and analyzed using SDS-PAGE with 15% polyacrylamide gels.
Proteolytic activity.Proteolytic activity was determined using azocasein or Suc-AAPF-pNA as a substrate. The azocasein protease assays were conducted as described previously (16) with some modifications. Briefly, 10 μl sample was mixed with 90 μl azocasein (Sigma) in 50 mM HEPES (pH 8.9) for a final azocasein concentration of 1.5% (wt/vol), with a final CaCl2 concentration of 1 mM. The reaction mixtures were incubated for 20 min at 90°C, and proteolysis was terminated by the addition of 30 μl 15% trichloroacetic acid. After centrifugation at 18,000 × g for 10 min, 100 μl of the supernatants was collected and combined with 30 μl 5 M NaOH, with the absorbance then determined at 440 nm.
The enzymatic activities were determined using the synthetic substrate Suc-AAPF-pNA in 600-μl reaction mixtures that contained 2 nM mature pernisine, 0.3 mM Suc-AAPF-pNA (Sigma), 1 mM CaCl2, and different concentrations of the propeptide in 50 mM HEPES (pH 8.9). The mature pernisine was either added to the mixture of the propeptide and the substrate or was incubated with the propeptide (30 min; 20°C) prior to addition of the substrate. The reactions were performed at 20°C and 90°C in a spectrophotometer (Cary 100 Bio UV-visible; Varian), and the p-nitroaniline released was followed by its absorbance at 410 nm, as described previously (40).
Tryptophan spectrofluorimetry.The intrinsic tryptophan fluorescence spectra were recorded using a fluorescence spectrophotometer (Cary Eclipse; Varian). The active-site mutants pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A were diluted in 10 mM Tris (pH 8.0) to a final concentration of 2 μM and then excited at 295 nm under various CaCl2 concentrations. The emission spectra were scanned from 300 nm to 600 nm at 20°C. The excitation and emission slits were set at 5 nm, with a scanning rate of 120 nm/min and a signal averaging time of 0.5 s. Each spectrum was scanned three times and averaged.
Circular-dichroism spectroscopy.The far-UV and near-UV CD spectra were collected with a CD spectrometer (J-1500; Jasco). The bandwidth was set at 1.0 nm and the scanning speed at 20 nm/min. All of the proteins were dissolved in 10 mM Tris (pH 8.0). For the far-UV spectra, 0.1 mg/ml pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A was used, and 0.2 mg/ml propeptide. The spectra were scanned from 250 nm to 195 nm at 20°C and 90°C under the various CaCl2 concentrations. For the near-UV spectra, 1.0 mg/ml pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A was used, and 2.0 mg/ml propeptide. The spectra were scanned from 320 nm to 250 nm at 20°C and 90°C at the various CaCl2 concentrations. Cuvettes with optical paths of 1 mm and 10 mm were used for the far-UV CD and near-UV CD, respectively. Mean residue weights of 102.37, 102.10, 102.25, and 100.60 were used to calculate the molar ellipticities of the pro-PerS355A, pro-PerΔCaS355A, pro-PerΔISS355A, and propeptide, respectively.
SDS-PAGE folding analysis.The active-site mutants pro-PerS355A, pro-PerΔCaS355A, and pro-PerΔISS355A (2 μM) were incubated in 10 mM Tris (pH 8.0) with the different concentrations of CaCl2 at room temperature in 12-μl aliquots. After 15 min, 4 μl of 4× sample buffer for SDS-PAGE was added to each aliquot. These mixtures were then analyzed using SDS-PAGE on 12% gels.
Data availability.No new nucleotide or amino acid sequences, strains, protein structures, etc., are reported in this study. UniProt and PDB accession numbers for all proteins and structures mentioned are given in the text and supplemental material.
ACKNOWLEDGMENTS
We thank Ajda Taler Verčič and Andreja Sekirnik for technical advice and assistance with Western blotting, Adrijana Leonardi for the N-terminal sequencing, and Marko Dolinar for providing the pMD204 plasmid.
This work was supported by the Slovenian Research Agency through postgraduate research funding (to M.B.), grant L7-8277 (to N.P.U.), and infrastructural grant I0-0048 (to D.T.). We are also thankful to Borer Chemie AG for financial support through the L7-8277 grant.
M.B. and N.P.U. conceived and designed the study. M.B. and M.Š. conducted the research. M.B. and N.P.U. analyzed and interpreted the data. M.B. wrote the manuscript. D.T. and N.P.U. provided materials/analysis tools. We all read the manuscript and approved it for final submission.
FOOTNOTES
- Received 25 April 2020.
- Accepted 16 June 2020.
- Accepted manuscript posted online 19 June 2020.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.