|Home | About | Journals | Submit | Contact Us | Français|
Prolidases hydrolyze Xaa-Pro dipeptides and can also cleave the P-F and P-O bonds found in organophosphorus (OP) compounds, including the nerve agents soman and sarin. Ph1prol (PH0974) has previously been isolated and characterized from Pyrococcus horikoshii and was shown to have higher catalytic activity over a broader pH range, higher affinity for metal, and increased thermostability compared to P. furiosus prolidase, Pfprol (PF1343). To obtain a better enzyme for OP nerve agent decontamination and to investigate the structural factors that may influence protein thermostability and thermoactivity, randomly mutated Ph1prol enzymes were prepared. Four Ph1prol mutants (A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol) were isolated which had greater thermostability and improved activity over a broader range of temperatures against Xaa-Pro dipeptides and OP nerve agents compared to wild type Pyrococcus prolidases.
Pyrococcus horikoshii and Pyrococcus furiosus are both hyperthermophilic archaea, growing optimally at 98–100°C that were isolated from a deep hydrothermal vent in the Okinawa Trough in the northeastern Pacific Ocean and from a shallow marine solfatara at Vulcano Island off the coast of Italy,respectively [1, 2]. Pyrococcus spp. are some of the most studied hyperthermophilic archaea to date owing in part to their utility for a variety of biotechnological applications [3–7]. For example, recombinant prolidases from Pyrococcus spp. are being studied for their potential use in bio-decontamination applications .
Prolidases function in vivo to hydrolyze dipeptides with proline in the C-terminus, Xaa-Pro, and a non-polar amino acid in the N-terminus . However, studies have demonstrated that prolidases can also hydrolyze and detoxify organophosphate (OP) compounds such as chemical warfare agents (CWA) . Two enzymes that have been characterized for potential field detoxification of nerve agents are organophosphorus acid anhydrolase (OPAA) and phosphotriesterase (PTE) [10, 11]. Recently, the crystal structure of OPAA from Alteromonas sp. JD6.5 strain has been solved, and it has been determined to be a prolidase . While OPAA does have the capability to degrade OP nerve agents, its activity can be limited by exposure to high temperatures and solvents during use in field situations [13, 14].
In 2008, the Defense Threat Reduction Agency (DTRA), under the auspices of the Department of Defense recognized the importance of developing enzyme-based OP nerve agent detoxification systems and created an initiative calling for new enzymes and biocatalysts that are stable over a broad temperature and pH range, in the presence of salts and surfactants, and that do not pose an environmental hazard . In response to the need to develop stable OP nerve agent degrading enzyme systems, thermostable prolidases from Pyrococcus spp. were studied [16–18].
Specifically, Pyrococcus prolidases from P. furiosus (PF1343 or Pfprol) and P. horikoshii (PH1149 or Phprol and PH0974 or Ph1prol) were characterized both structurally and enzymatically [16, 18, 19]. The Pyrococcus prolidases were determined to be very similar, with Pfprol showing 88% amino acid similarity to Phprol and 55% similarity to Ph1prol. Although they have high similarity to each other, the kinetic properties of Ph1prol appeared to be more favorable for application purposes than those of Pfprol and Phprol. Ph1prol has demonstrated higher activity at lower temperatures and over a broader pH range. It has long-term stability, higher affinity for substrates, and a lower metal requirement for catalysis . Therefore, it was deemed to be advantageous to use Ph1prol in mutagenesis studies in order to develop a better enzyme for OP nerve agent detoxification and to further investigate factors that influence the catalysis of thermophilic metalloenzymes. To this end, a random mutagenesis Ph1prol gene library was constructed and screened for production of mutants that showed increased prolidase activity at 30°C compared to wild type Ph1prol. Four Ph1prol mutants were isolated and subsequently characterized to determine their substrate catalysis over a broad range of temperatures and their performance against OP nerve agent analogs in comparison to Ph1prol and the previously characterized Pfprol.
The E. coli K-12 derivative NK5525 (proA::Tn10) was used to construct the selection strain JD1(λDE3) as described in  for screening of cold-adapted P. horikoshii prolidase variants. The P. horikoshii prolidase expression plasmid pET-Ph1prol was previously constructed as described in . The E. coli strains were cultured either in Luria-Bertani (LB) broth or M9 selective minimal medium supplemented with 0.2% glucose, 1mM MgSO4, 0.05% VitB1, 20μM IPTG, 20μM Leu-Pro. Ampicillin (100μg/mL), kanamycin (50μg/mL), chloramphenicol (34μg/mL), and tetracycline (6μg/mL) were added into the medium when required.
Error-prone PCR mutagenesis using the Genemorph II Random mutagenesis kit (Stratagene, La Jolla, Calif) was used to amplify and insert mutations into the P. horikoshii prolidase gene (PH0974). PCR amplification was carried out for 30 cycles: (60sec at 95°C, 60sec at 55°C, 120sec at 72°C), with a 10-minute final extension at 72°C. Reactions contained Mutazyme II reaction buffer, 125ng/μL of each primer, 40mM dNTP mix, and 2.5U of Mutazyme II DNA polymerase. Initial DNA template amounts used were 250ng and 750ng in order to select for medium-to-low mutation rates, respectively. The Genemorph II EZClone (Stratagene, La Jolla, Calif) reaction was employed to clone the mutated prolidase gene into the expression vector pET-21b.
The EZClone reaction included EZClone enzyme mix, 50ng of template plasmid (pET-prol), 500ng of megaprimer (mutated prolidase PCR product), and EZClone solution. The reactions were amplified for 25 cycles: (50sec at 95°C, 50sec at 60°C, 14min at 68°C). Amplified products were digested with Dpn I for 2 hours at 37°C to remove template plasmid. XL1-Gold super competent E. coli cells were transformed with the mutant plasmid mixture.
pET-Ph1prol plasmids from the mutant P. horikoshii library were transformed into the selective strain JD1(λDE3) and were plated on M9 selective agar plates. Colonies that grew after being incubated for 3–7d at 20°C were isolated on LB plates and then grown in 10mL LB medium at 37°C with shaking (200rpm) until an optical density of 0.6–0.8 was reached. IPTG was then added to the cell culture to a final concentration of 1mM. The induced culture was shaken at 37°C for 3h before harvesting the cells. These cell pellets were lysed using 300μL of B-per buffer (Thermo Scientific, Rockford, Ill), and the resulting cell extracts were used for enzyme activity assays conducted at 30°C and at 100°C. Heat-treated soluble protein samples were heated at 80°C for 20min. Four mutant colonies that exhibited at least 2-3-fold higher activities compared to the cells expressing the wild type P. horikoshii prolidase were selected for characterization, and their plasmids were isolated. The prolidase genes present in those isolated plasmids were sequenced using the T7 promoter and T7 terminator primers (MWG Biotech, Huntsville, AL).
Production of P. horikoshii prolidase variants (A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol) was carried out in transformed E. coli BL21 (λDE3) cells with the appropriate pET-Ph1prol plasmid and pRIL vector. The transformants were grown in 1L cultures of autoinduction media  incubated at 37°C with shaking (200 RPM) overnight.
Cell pellets from the four Ph1prol variants (A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol) were suspended in 50mM Tris-HCl, pH 8.0 (3mL Tris per 1 gram of cell paste), with 1mM benzamidine and 1mM DTT. For each variant, diluted cell slurry was passed through a French pressure cell (20,000lb/in2) three times. Cell lysates were centrifuged at 38,720xg for 30min at 4°C, and then the supernatants were heated to 80°C for 30min anaerobically to denature any proteins not stable at that temperature. Heat-treated supernatants were centrifuged at 38,720xg to remove the denatured proteins. Supernatants were sampled both before and after heat treatment for activity analysis. (NH4)2SO4 was added gradually to the supernatants to make a final concentration of 1.5M prior to loading onto a 5mL Phenyl-Sepharose hydrophobic interaction chromatography column (Hi-Trap Phenyl HP Column, GE Healthcare Life Sciences, Piscataway, NJ). Fractions containing the prolidase mutants were pooled and dialyzed overnight into 4L of 50mM Tris HCl, pH 8.0 at 4°C, and were further purified on a 5mL Q Sepharose anion exchange chromatography column (Hi-Trap Q FF Column, GE Healthcare Life Sciences, Piscataway, NJ). Buffers for both purification steps have been described in . Fractions from both purification steps were further visualized using SDS-PAGE (12.5% SDS-polyacrylamide gels) and were tested for enzyme activity. Fractions were then pooled based on gel images, and enzyme stocks were stored at −80°C. The molecular weights of Ph1prol and mutants are approximately 40.04kDa. The purity of each Ph1prol mutant was estimated to be greater than 95% using both visual inspection of SDS-polyacrylamide gels and electrophoretic microchip analysis.
The enzyme activity assay protocol is based on a method previously described by [17, 20]. The reaction mixture (500μL) contained 50mM MOPS buffer (3-[N-morpholino] propanesulfonic acid) pH 7.0, 200mM NaCl, water, 5% (vol/vol) glycerol, 100μg/mL BSA (bovine serum albumin) protein, 0.2mM CoCl2, and the enzyme. The reaction mixture was heated at 100°C for 5min allowing time for the metal and enzyme to interact. The reaction was initiated by the addition of substrate (Xaa-Pro, 4mM final concentration) and allowed to proceed for 10min at 100°C. The reaction was stopped with 500μL glacial acetic acid and 500μL ninhydrin reagent (3% (wt. vol)) and heated again for 10min at 100°C. The reaction was then cooled to 23°C. Prolidase samples were assayed in triplicate and specific activity was calculated using the absorbance value at 515nm and an extinction coefficient of 4,570M-1cm−1 for the ninhydrin-proline complex.
For assays evaluating the temperature profile, WT-Ph1prol and the four prolidase mutants were assayed in triplicate for activity with 4mM Met-Pro at 10°C, 20°C, 35°C, 50°C, 70°C, and 100°C. Experiments were performed in duplicate.
In order to study substrate specificity, the following Xaa-Pro dipeptides were used as substrates (4mM final concentration) in the enzyme activity assays for WT-Ph1prol and the four prolidase mutants: Val-Pro, Met-Pro, Phe-Pro, Leu-Pro, Ala-Pro, and Gly-Pro. Prolidase samples were assayed with two additional substrates, Pro-Ala and Val-Leu-Pro, to further illustrate prolidase preference of Xaa-Pro dipeptides . Kinetic parameters of the Ph1prol mutants were determined at 70°C using a range of Leu-Pro concentrations (0.25–12mM).
Thermostability experiments were performed in duplicate on WT-Ph1prol and the four mutants. Each enzyme was diluted to a concentration of 0.04mg/mL in 50mM MOPS, pH 7.0, and 200mM NaCl and incubated in an anaerobic sealed vial at 90°C. An initial sample was taken to represent Time = 0h, and additional samples were taken at Time = 24, 48, and 72h. Samples were diluted to 0.4μg/mL in 50mM MOPS, pH 7.0, and 200mM NaCl and were then assayed in triplicate in accordance with the enzyme activity assay protocol described in Section 2.6. In all cases, the substrate used in the activity assay was 4mM Met-Pro.
Pot-life experiments were also performed in duplicate. Each enzyme was diluted to a concentration of 0.04mg/mL in 50mM MOPS, pH 7.0, and 200mM NaCl and incubated anaerobically in a sealed vial at 70°C. An initial sample was taken to represent Time = 0 days, and additional samples were taken at Time = 1, 7, 14, 16, and 21 days. Samples were diluted to 0.4μg/mL in 50mM MOPS, pH 7.0, and 200mM NaCl and were then assayed in triplicate to determine specific activity.
Differential scanning calorimetry was performed using a MicroCal VP-DSC scanning calorimeter. The calorimetric samples contained ~1mg/mL protein in 50mM MOPS, 200mM NaCl, 0.2mM CoCl2, pH 7.0. Protein samples were dialyzed 15h against this buffer before the experiment. Samples were degassed before loading into the chamber cell. The calorimetric experiment was performed by heating the samples at a scan rate of 100°C/hr.
The hydrolysis of DFP by prolidases was measured by monitoring fluoride release with a fluoride-specific electrode as previously described . Assays were performed at 35°C and 50°C, with continuous stirring in 2.5mL of buffer (50mM MOPS, 200mM NaCl, pH 7.0), 0.2mM CoCl2 and 3mM DFP. The enzyme and metal were incubated at the reaction temperature 5min prior to the start of the reaction. The background of DFP hydrolysis was measured by running a reaction without enzyme present at 35°C and 50°C. The background hydrolysis of DFP was subtracted from enzymatic hydrolysis to determine specific activity of the enzyme.
Prolidase hydrolysis of p-nitrophenyl soman was monitored by accumulation of p-nitrophenolate [10, 22]. The p-Nitrophenyl Soman was synthesized at Edgewood Chemical Biological Center in Aberdeen Proving Ground, Md. and contained a racemic mixture of all four stereoisomers. The purity of the soman analog was greater than 90% based on gas chromatography analysis . Two mL reaction assays contained buffer (50mM MOPS, 200mM NaCl, pH 7.0), 0.2mM CoCl2, and 0.3mM p-nitrophenyl soman. The reactions were conducted at three different temperatures (35°C, 50°C and 70°C). The enzyme and metal were incubated at the specified reaction temperature 5min prior to the start of the reaction. Absorbance of the product p-nitrophenolate was measured at 405nm over a 5min range.To calculate activity, the extinction coefficient for p-nitrophenolate of 10,101 M−1cm−1 was used.
Previous studies characterizing P. horikoshii prolidase homolog 1 (Ph1prol, PH0974) demonstrated that it has higher catalytic activity over a broader pH range, higher affinity for metal, and is more thermostable than either P. furiosus prolidase (Pfprol, PF1343) or P. horikoshii prolidase (Phprol, PH1149) when assayed with the dipeptide substrate Met-Pro . Based on these favorable attributes for Ph1prol when reacting with its natural substrates, Xaa-Pro dipeptides, there was interest in determining the relative activity of recombinant Ph1prol compared to Pfprol and Phprol against G-type nerve agent simulants DFP and soman analog, p-nitrophenylsoman. As indicated in Figure 1, DFP exhibited the greatest hydrolysis with Ph1prol. Ph1prol had a relative activity that was 843% higher than Pfprol and 817% higher than Phprol at 35°C and 1870% higher than both Pfprol and Phprol at 50°C (Figure 1). In contrast, the trends with the soman analog were very different as shown in Figure 2. The relative activity of Ph1prol was only 70%, 63%, and 68% of thePfprol activity at 35°C, 50°C, and 70°C, respectively (Figure 2). These results indicate that Ph1prol has a preference for DFP and does not exhibit high activity with the soman analog. Differences in the protein structures likely play a role in the substrate preference since Pfprol and Ph1prol share only 55% amino acid residue similarity . By altering the Ph1prol structure further using a random mutagenesis approach, it would be possible to isolate Ph1prol variants that show even greater hydrolysis of DFP and/or improved activity against the soman analog.
Since Ph1prol showed the most favorable properties including higher activity with DFP, it was selected for further mutagenesis using an error-prone PCR strategy, which employs a mutated polymerase. Transformed E. coli JD1 (λDE3) cells were used to select for Ph1prol variants on minimal media plates that were supplemented with 20μM Leu-Pro and grown at 20°C. Colonies that were visible in 3–7 days were plated on minimal and rich (LB) media. Ph1prol variants were screened using small-scale expression cultures (10mL) induced with IPTG. Four Ph1prol variants out of over 200 screened were selected for sequencing after showing two-fold higher activity with Leu-Pro at 30°C and somewhat reduced activity at 100°C as compared to wild type. The increased activity at the lower temperature of 30°C and variation at the higher temperature of 100°C is indicative of a mutation in a thermophilic protein, which can compromise activity or stability at higher temperatures but could create more flexibility and increased catalysis at the lower temperatures.
Prior to sequencing, the four Ph1prol mutants were numbered (10, 19, 35, and 72) based solely on the order in which they had been isolated. Sequencing of the variants revealed the locations of the amino acid substitutions for each mutant (Figure 3). Mutant no.10 has two mutations: one at position 195 in which there is a change from alanine to threonine (A195T) and the second at amino acid residue 306 in which glycine is changed to serine (G306S). Both of these mutations reside in the C-terminal region of Ph1prol. Mutant no.19 has two mutations: one at position 301 in which there is a change from tyrosine to cysteine (Y301C) and the second at amino acid residue 342 which has a substitution of lysine with an asparagine (K342N). Both of these mutations are in the C-terminal region of the enzyme. Mutant no.35 contains two mutations: one at position 127 in the α-helical linker region in which there is a change from glutamate to glycine (E127G) and the second in the C-terminal region at position 252 with a substitution of glutamate for aspartate (E252D). Mutant no.72 is the only mutation in the N-terminal region at position 36 with a change from glutamate to valine (E36V) in Ph1prol. The mutations are remote from the active site pocket, which is shown in Figure 3 as being located between two 310 helixes (red helices, residues 191–195, and 281–284). Therefore, the mutations are not likely directly changing the active site chemistry. Rather the mutations such as E36V, E127G, and Y301C may be affecting prolidase dimerization as those residues are located along the dimerization interface. Furthermore, the mutations may be affecting the conformational dynamics of the enzymes since some of the mutations are located in the loop and linker regions. The change in conformation dynamics is giving the variants better activity over a broader range of temperatures as indicated in Section 3.4.
Both the wild type Ph1prol and the four variants were most active at 100°C (Figure 4). WT-Ph1prol and E36V-Ph1prol had very high specific activities (3,955U/mg and 4094U/mg, resp.) with 4mM Met-Pro, both of which are twice as high as that of Pfprol at 100°C (2,154U/mg) . A195T/G306S- and E127G/E252D-Ph1prol were slightly less active (2,307U/mg and 2,831U/mg, resp.) than WT Ph1prol, and Y301C/K342N-Ph1prol showed the lowest level of activity in comparison to the other mutants at 1,842U/mg. Activity was reduced by more than half at 70°C for all of the variants; however, WT-Ph1prol, A195T/G306S-, and E36V-Ph1prol all had specific activities close to 1,000U/mg (973U/mg, 1000U/mg, and 913U/mg, resp.), whereas the specific activity of Pfprol at 70°C was 806U/mg . At 50°C, Y301C/K342N-Ph1prol had a higher specific activity than any of the other variants and the wild type (450U/mg) and was roughly three times more active than Pfprol at 50°C .
At the lower temperatures, 35°C, 20°C and 10°C, Y301C/K342N-Ph1prol out-performs the other Ph1prol variants, the WT-Ph1prol, the WT-Pfprol and R19G/G39E/K71E/S229T-Pfprol (the highest performing Pfprol mutant at lower temperatures) . At 35°C, Y301C/K342N-Ph1prol (298U/mg) has relative activity that is 121% that of WT-Ph1prol, 489% that of WT-Pfprol, and 244% that of R19G/G39E/K71E/S229T-Pfprol (246U/mg, 61U/mg, and 122U/mg, resp.). At 20°C, Y301C/K342N-Ph1prol (241U/mg) has a relative activity 184% that of WT-Ph1prol, 964% higher than WT-Pfprol, and 482% higher than R19G/G39E/K71E/S229T-Pfprol (specific activities of 131U/mg for WT-Ph1prol, 25U/mg for WT-Pfprol and 50U/mg for R19G/G39E/K71E/S229T-Pfprol) . The greatest improvement in performance for the Ph1prol mutants is seen when assayed at 10°C. Y301C/K342N-Ph1prol (109U/mg) has a relative activity that is 166% higher than that of WT-Ph1prol, 1,982% higher than that of WT-Pfprol, and 396% higher than R19G/G39E/K71E/S229T-Pfprol (specific activities of 66U/mg for WT-Ph1prol, 5.5U/mg for WT-Pfprol, and 27.5U/mg for R19G/G39E/K71E/S229T-Pfprol) .
Substrate specificity of WT-Ph1prol is shown in Table 1 along with the specific activities of A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol, which are reported as a percentage relative to the activity of the wild type. WT-Ph1prol was most active with the dipeptide Val-Pro (4,602U/mg), while mutants A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol had much lower activity at 33%, 36%, 46%, and 56% of the WT Ph1prol, respectively. A195T/G306S-Ph1prol showed the highest activity with Met-Pro at 143% that of the wild type (2,809U/mg), while Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol preferred Ala-Pro with specific activities of 183%, 324%, and 556% that of WT-Ph1prol (1,452U/mg). WT-Ph1prol seems to prefer the most hydrophobic amino acids, while the four variants have the highest activity with a less hydrophobic amino acid in the N-terminal position of the dipeptide substrate. While alanine is considered to be a hydrophobic amino acid, it is less hydrophobic than valine, methionine,phenylalanine, and leucine and is most similar in structure to glycine. WT-Ph1prol has much lower activity with Gly-Pro than with Ala-Pro (369 compared to 1,452U/mg). A195T/G306S-, Y301C/K342N-, and E127G/E252D-Ph1prol have less or similar activity with Gly-Pro as compared to WT-Ph1prol (369U/mg) while E36V-Ph1prol has 444% higher activity.
Specific activities were consistently low with both Pro-Ala and Val-Leu-Pro (2U/mg, WT-Ph1prol) for the wild type Ph1prol and the four mutants. Due to the nature of the prolidase enzyme and its unique ability to cleave the bond between Xaa-Pro dipeptides, it was expected that the enzyme would not show any notable activity with a Pro-Xaa dipeptide or a tripeptide. While two of the four variants (mutants A195T/G306S- and E36V-Ph1prol) show increased activity with Pro-Ala when compared to wild type, it must be noted that specific activity with Pro-Ala for WT-Ph1prol was extremely low, at only 14U/mg. Y301C/K342N-Ph1prol had 88% WT activity with Val-Leu-Pro; however, WT-Ph1prol-specific activity was only 2U/mg.
The catalytic activities of WT-Ph1prol and its mutants were tested at 70°C with Leu-Pro (Table 2). All Ph1prol mutants had higher kcat values than the WT-Ph1prol suggesting that the amino acid changes in the mutant enzymes did have an effect on structure and ultimately the catalytic activity of the prolidase with the substrate Leu-Pro. Although the kcat values were higher, the overall enzyme turnover rates were not for some of the mutants compared to WT-Ph1prol. All the Ph1prol mutants showed an increased turnover rate, kcat/Km, with Leu-Pro except for Y301C/K342N-Ph1prol. This could be due to the increase in the Km of this mutant, which is almost three times higher than WT-Ph1prol.
To determine whether the amino acid substitutions in the four Ph1prol variants had any effect on thermostability, the mutants were incubated in sealed vials at 90°C, anaerobically, for 72h. Samples were taken every 24h to measure catalytic activity with Met-Pro (4mM) as the substrate. In Theriot et al. 2010, it was shown that WT-Pfprol was more thermostable than its mutants and had lost 50% activity with Met-Pro by 21 hours at 90°C . After 40h at 90°C, WT-Ph1prol had lost 50% activity. Mutants Y301C/K342N-Ph1prol and E127G/E252D-Ph1prol both demonstrated increased thermostability at 90°C compared to wild type and had lost 50% activity after 58 hours of incubation. Mutants A195T/G306S-Ph1prol and E36V-Ph1prol displayed a 50% loss of activity after 32 and 35 hours at 90°C, respectively. Both WT-Ph1prol and the four Ph1prol mutants were stable at 90°C for a significantly longer time period than Pfprol and its mutants .
Pot-life activity was monitored over the course of 21 days with samples taken every seven days until day 14 and then again on days 16 and 21 while the samples were incubated anaerobically at 70°C (Figure 5). Pot-life experiments with WT-Pfprol and its mutants as reported in Theriot et al., 2010 were conducted over a 48h period at 70°C . In preliminary pot-life experiments with WT-Ph1prol and its mutants, the specific activities with Met-Pro were still remarkably high after 48h (no obvious decrease in specific activity was detected; data not shown), so the experiments were continued until the enzymes were considered to be no longer active (21 days). Initial specific activities for WT-Ph1prol, A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol were 3,150U/mg, 3,400U/mg, 1,250U/mg, 2,200U/mg, and 3,600U/mg, respectively, with 4mM Met-Pro as the substrate. WT-Ph1prol had lost 50% activity by day 12 of incubation at 70°C. Mutants Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol had 50% activity remaining by days 12, 13, and 14, respectively. A195T/G306S-Ph1prol was at 50% activity after 10 days at 70°C. By 21 days at 70°C, all five prolidases were at or below 25% of the initial activity.
As reported in Theriot et al., 2010, the specific activities of WT-Pfprol and its three mutants (G39E-, R19G/K71E/S229T-, and R19G/G39E/K71E/S229T-Pfprol) were 1,083U/mg, 599U/mg, 722U/mg, and 4,496U/mg, respectively, with 4mM Met-Pro after 48h of incubation at 70°C . In contrast, after 7 days (or 168h) at 70°C, WT-Ph1prol, A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol had specific activities of 2,950U/mg, 3,050U/mg, 1,230U/mg, 2,650U/mg, and 2,400U/mg, respectively, with 4mM Met-Pro. After 48h at 70°C, WT-Pfprol had a higher specific activity than any of its mutants (1,083U/mg). After 7 days at 70°C, Y301C/K342N-Ph1prol showed the lowest specific activity of the Ph1 prolidases; however, it still had activity 114% that of WT-Pfprol at 48h .
The thermal stability of wild type Ph1prol and variants was determined by differential scanning calorimetry (DSC) experiments. Table 3 shows the denaturation temperature of the wild type and variant enzymes. The mutations that improved catalytic activity of the Ph1prol at lower temperatures did not adversely affect the temperature stability of the enzymes.
Substrate specificity of WT-Ph1prol with DFP is shown in Figure 6 along with the specific activities of A195T/G306S-, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol, which are reported as a percentage relative to the activity of the wild type. WT-Ph1prol was most active with DFP at 35°C and 50°C with a specific activity of 4U/mg and 10U/mg, respectively. The mutants A195T/G306S, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol had significantly lower activity with DFP; even at 50°C the activity was only 59%, 25%, and 55% of WT-Ph1prol activity. However, it should be noted that the Ph1prol mutants have 808%, 183%, and 402% (A195T/G306S, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol, resp.) of the DFP activity compared to WT P. furiosus prolidase and also compared favorably to the highest DFP activity reported for the R19/G39E/K71E/S229T Pfprol mutant, which was shown to have 398% higher activity than WT Pfprol .
Figure 7 reveals a different trend, where the mutations in WT-Ph1prol increased the specific activity with the soman analog, p-nitrophenyl soman. WT-Ph1prol showed the highest activity with the soman analog at 70°C, with a specific activity of 0.56U/mg. The mutant A195T/G306S-Ph1prol had a similar specific activity to WT-Ph1prol when incubated at 35°C, 50°C, and 70°C. Mutants Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol showed increased activity with the soman analog over WT-Ph1prol at each assay temperature. The most significant specific activities with the soman analog were seen with Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol at 70°C, which correlated to 125%, 186%, and 157% increase over WT-Ph1prol. Furthermore, the activities for the Ph1prol mutants against p-nitrophenyl soman (0.7, 1.0, and 0.9U/mg for Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol, resp.) compare favorably to the improved soman analog activities reported for the P. furiosus prolidase mutants (0.86, 1.02, and 1.7U/mgfor G39E-, R19G/K71E/S229T-, and R19G/G39E/K71E/S229T-Pfprol, resp.) . When looking at the substrate specificity of the WT-Ph1prol and variants with proline dipeptides, it was noticed that there was a shift in preference from more hydrophobic to less hydrophobic amino acids among the mutants. This is also seen with the OP nerve agents, where there is a shift in substrate specificity from DFP to the soman analog. The WT-Ph1prol prefers DFP as a substrate over the soman analog, while the Ph1prol variants show decreased activity with DFP and increased activity with the soman analog.
Current biodecontamination formulations for degradation of OP nerve agents that incorporate Alteromonas prolidases (OPAA) and PTE have limitations when used in the field . Long-term stability of the enzyme is not attainable in a formulation mixture that includes other solvents and denaturing solutions, and the need to add excess metal to reach maximum activity poses further complications for an enzyme-based detoxification system. A highly active enzyme that is stable over the long term and requires very little metal addition would be best suited for this application. The wild type and mutant prolidases characterized from P. horikoshii show promising enzymatic properties that make them potential candidates for future optimization studies for OP nerve agent degradation. Compared to Pfprol, Ph1prol and the four Ph1prol mutants show higher activity, higher affinity for the substrate, and significantly lower metal requirement for catalysis. Two of the variants, Y301C/K342N- and E127G/E252D-Ph1prol are thermostable for nearly three times as long as Pfprol and double the time of Ph1prol. A195T/G306S-Ph1prol has 808% of the DFP activity compared to wild type P. furiosus prolidase and is superior to any of the improved P. furiosus prolidase mutants . Furthermore, Y301C/K342N-, E127G/E252D-, and E36V-Ph1prol all have improved activities against p-nitrophenyl soman relative to WT-Ph1prol and also compare favorably to the best performing P. furiosus prolidase mutants . The Ph1prol variants have the potential to significantly improve upon current biodecontamination strategies. Their increased thermostability and pot life and activities against OP nerve agent analogs warrant further study into large-scale production and purification of these prolidases.
The authors would like to thank Dr. Sherry Tove for her helpful comments on the paper and to acknowledge Dr. Nathaniel Hentz and Jessica Weaver at the NCSU Biomanufacturing Training and Education Center (BTEC) for contributing their time and expertise for the purification and characterization of the prolidases. They would like to thank Dr. James Carney and Patricia Buckley for performing the DSC experiment. Support for this study was provided by the Army Research Office (Contract no. 44258LSSR).