ABSTRACT
Carbamate kinases catalyze the conversion of carbamate to carbamoyl phosphate, which is readily transformed into other compounds. Carbamate forms spontaneously from ammonia and carbon dioxide in aqueous solutions, so the kinases have potential for sequestrative utilization of the latter compounds. Here, we compare seven carbamate kinases from mesophilic, thermophilic, and hyperthermophilic sources. In addition to the known enzymes from Enterococcus faecalis and Pyrococcus furiosus, the previously unreported enzymes from the hyperthermophiles Thermococcus sibiricus and Thermococcus barophilus, the thermophiles Fervidobacterium nodosum and Thermosipho melanesiensis, and the mesophile Clostridium tetani were all expressed recombinantly, each in high yield. Only the clostridial enzyme did not show catalysis. In direct assays of carbamate kinase activity, the three hyperthermophilic enzymes display higher specific activities at elevated temperatures, greater stability, and remarkable substrate turnover at alkaline pH (9.9 to 11.4). Thermococcus barophilus and Thermococcus sibiricus carbamate kinases were found to be the most active when the enzymes were tested at 80°C, and maintained activity over broad temperature and pH ranges. These robust thermococcal enzymes therefore represent ideal candidates for biotechnological applications involving aqueous ammonia solutions, since nonbuffered 0.0001 to 1.0 M solutions have pH values of approximately 9.8 to 11.8. As proof of concept, here we also show that carbamoyl phosphate produced by the Thermococcus barophilus kinase is efficiently converted in situ to carbamoyl aspartate by aspartate transcarbamoylase from the same source organism. Using acetyl phosphate to simultaneously recycle the kinase cofactor ATP, at pH 9.9 carbamoyl aspartate is produced in high yield and directly from solutions of ammonia, carbon dioxide, and aspartate.
IMPORTANCE Much of the nitrogen in animal wastes and used in fertilizers is commonly lost as ammonia in water runoff, from which it must be removed to prevent downstream pollution and evolution of nitrogenous greenhouse gases. Since carbamate kinases transform ammonia and carbon dioxide to carbamoyl phosphate via carbamate, and carbamoyl phosphate may be converted into other valuable compounds, the kinases provide a route for useful sequestration of ammonia, as well as of carbon dioxide, another greenhouse gas. At the same time, recycling the ammonia in chemical synthesis reduces the need for its energy-intensive production. However, robust catalysts are required for such biotransformations. Here we show that carbamate kinases from hyperthermophilic archaea display remarkable stability and high catalytic activity across broad ranges of pH and temperature, making them promising candidates for biotechnological applications. We also show that carbamoyl phosphate produced by the kinases may be efficiently used to produce carbamoyl aspartate.
INTRODUCTION
Carbamoyl phosphate (CP) is an important metabolic intermediate. It is produced in anabolic pathways through the action of carbamoyl phosphate synthetases, consuming two equivalents of ATP, and is normally degraded in catabolic pathways through the action of carbamate kinases (CKs) for the generation of one equivalent of ATP (1). The kinase activity is reversible, however, and CKs are able to catalyze CP formation when supplied with ATP and carbamate. Carbamate forms spontaneously in aqueous solutions of ammonia and carbon dioxide (Fig. 1) (2).
Carbamate-kinase-catalyzed interconversion of carbamoyl phosphate and carbamate, formed in aqueous solution in equilibria between CO2/HCO3− and NH4OH/NH3.
CK was first isolated from Enterococcus faecalis, where it catalyzes ATP production as part of the arginine dihydrolase fermentation pathway (3). The kinase gene has since been identified in various other species that derive energy from arginine degradation (4–6). CKs have also been purified from pathogens of medical interest such as Mycoplasma penetrans (7), Giardia lamblia (8), and Trichomonas vaginalis (9). In the hyperthermophilic archaeon Pyrococcus furiosus, an enzyme originally described as a CK-like carbamoyl phosphate synthetase was subsequently characterized to be a CK (1, 10, 11). It is proposed to naturally catalyze anabolic CP formation (12, 13). This proposal is associated with the absence of the arginine dihydrolase pathway in Pyrococcus furiosus and the relatively high concentrations of ammonia and carbon dioxide, and therefore carbamate, in hydrothermal vents where this organism thrives (11, 12).
We were interested in the potential in vitro use of CKs in the anabolic direction for the conversion of ammonia and carbon dioxide into CP. CP is readily transformed, biologically and chemically, into other useful compounds (14–18, 38). Therefore, this application of CKs is of interest as a method for the biocatalytic capture and reuse of waste ammonia. The application would involve relatively high concentrations of ammonia, and therefore alkaline conditions. Consequently, we were particularly focused on pH stability and activity, as well as overall robustness, of CKs from different source organisms.
Here, along with the previously reported enzymes from Pyrococcus furiosus (CK-Pf) and the mesophile Enterococcus faecalis (CK-Ef), previously unreported CKs from Thermococcus sibiricus (CK-Ts), Thermococcus barophilus (CK-Tb), Fervidobacterium nodosum (CK-Fn), Thermosipho melanesiensis (CK-Tm), and Clostridium tetani (CK-Ct) were each produced. All enzymes were recombinantly expressed using an optimized protocol, which led to 1,000-fold higher protein yields compared to that for CK-Pf isolated from Pyrococcus furiosus (10). The activity of CK-Pf was previously measured using a coupled enzyme assay, which indicated a pH optimum of around 8.0 (10). We developed a direct assay of substrate turnover by CKs, applicable for use with any ATP-dependent enzyme. This assay instead showed that CK-Pf displays remarkable stability and activity at alkaline pH (>11), with maximum catalysis at above pH 9.4. The difference between our results and those reported previously is attributable to complications associated with the earlier protocol. Both CK-Ts and CK-Tb showed stability and activity at alkaline pH similar to those displayed by CK-Pf. This activity under alkaline conditions was not displayed by the mesophilic and thermophilic CKs tested here and is atypical of intracellular enzymes (19).
The utility of the production of CP by these CKs depends on having the ability to use the CP to make other compounds. Although CP decomposes in alkaline aqueous solutions (20), for example, with a half-life of 10 min at pH 9.9 and 37°C, as a metabolic intermediate it is processed in vivo by transcarbamoylases (21), which limit CP breakdown through metabolic channeling (22) and conformational restriction (23). Here, we establish the viability of using a similar approach in vitro, through in situ conversion of CP produced by CK-Tb into carbamoyl aspartate by Thermococcus barophilus aspartate transcarbamoylase (ATC-Tb). Using Escherichia coli lysate and acetyl phosphate (AP) to recycle the kinase cofactor ATP (24) during this process, a solution of ammonia, carbon dioxide, and aspartate, at pH 9.9, efficiently produces a high yield of carbamoyl aspartate.
RESULTS AND DISCUSSION
Selection and recombinant expression of carbamate kinases.To investigate the biocatalyst space, in addition to the already reported hyperthermophilic CK-Pf and mesophilic CK-Fn, CKs from five alternative source organisms were selected. Organisms of the genus Thermococcus are hyperthermophilic archaea that form clusters separate from those of the genus Pyrococcus in the order Thermococcales (25). For Thermococcus sibiricus (26) and Thermococcus barophilus (25), genomic analysis shows a lack of arginine dihydrolase and the presence of pyrimidine and arginine biosynthetic enzymes. This suggests that CK-Ts and CK-Tb may play a similar anabolic physiological role to CK-Pf. Accordingly, CK-Ts and CK-Tb were selected to compare their activity with that of CK-Pf, with which they each share 77% amino acid sequence similarity. CK-Fn and CK-Tm from the thermophilic bacteria Fervidobacterium nodosum and Thermosipho melanesiensis, each with 48% sequence similarity to CK-Pf, were also targeted to more generally examine the behavior of CKs from thermophiles. Finally, CK-Ct from Clostridium tetani was chosen to include another CK from a nonthermophilic source. CK-Pf shares 42% and 50% sequence similarity with CK-Ef and CK-Ct, respectively. An alignment of the amino acid sequences of the seven carbamate kinases investigated is presented in Fig. 2.
Clustal Omega alignment of investigated carbamate kinase amino acid sequences.
The reported isolation of CK-Pf gave a low protein yield of only 15 μg per g of cell pellet. A higher yield has been obtained through recombinant expression in E. coli, but only by following an elaborate protocol with multiple protein purification steps (11). This also required the simultaneous overexpression of archaeal tRNA in the bacterial host due to direct use of the archaeal DNA sequence. For the present work, the CK genes from Pyrococcus furiosus and the other selected source organisms were instead codon-optimized for expression in E. coli. Each was encoded with an N-terminal His6 (N-His6) tag to allow easy purification through Ni-affinity chromatography. That resulted in production of 16 mg of CK-Pf per g of cell pellet. This protein yield represents a 1,000-fold improvement compared to isolation of the enzyme from the source organism (10) and a 3-fold improvement with a much-simplified protocol compared to the earlier recombinant coexpression (11). CK-Ts, CK-Tb, CK-Fn, CK-Tm, CK-Ef, and CK-Ct were each expressed in similar yield to CK-Pf. SDS-PAGE analysis following Ni-affinity chromatography confirmed the presence of purified, overexpressed protein in all cases (Fig. 3). Differences between the mobilities of the CKs on the gel may be attributed to a range of factors, such as variations between the protein aliphatic indices and hydropathy (GRAVY) values, which ranged from 88.69 to 97.80 and −0.113 to −0.350, respectively, when calculated using ProtParam (ExPASy server).
SDS-PAGE analysis of recombinant proteins. Molecular mass markers are an NEB unstained protein ladder (10 to 200 kDa).
Determination of carbamate kinase activity.The previously reported assay of CK-Pf activity used a coupled enzyme method, with indirect measurement of the ADP formed to assess CP production (10). It relied first on the reaction of ADP with phosphoenolpyruvate to produce ATP and pyruvate, catalyzed by pyruvate kinase, and second on the lactate-dehydrogenase-catalyzed, NADH-dependent reduction of the pyruvate to lactate. The decrease in NADH concentration was then monitored through measuring the ultraviolet (UV) absorbance at 340 nm. In our study, a high-performance liquid chromatography (HPLC)-UV method was instead developed to directly analyze ADP formation and CK activity. It is based on the separation and detection of ATP, ADP, and AMP, which all absorb at 259 nm. Therefore, in principle, it may be used to monitor any ATP-, ADP-, or AMP-dependent enzyme and is suitable for automation. The assay was used to study the behavior of the recombinantly expressed CKs under a range of conditions, with variations between our observations and those reported earlier for CK-Pf being a direct consequence of our being able to avoid the confounding effects of coupled enzymes.
Initially, CK catalysis was studied in aqueous solutions prepared using 0.2 M NaHCO3, 0.2 M NH3, 10 mM ATP, and 10 mM MgCl2, adjusted to between pH 7.9 and 11.4 with 6 M NaOH or HCl. Carbamate then formed spontaneously according to the equations shown in Fig. 1. Studies were not carried out at lower or higher pH, as this is beyond the reliable buffering capacity of ammonia and bicarbonate. Within this range, the pH of solutions varied by less than 0.1 during the course of the assays. Representative results from assays under these conditions at 40°C are illustrated in Fig. 4. Of the seven overexpressed CKs, only CK-Ct did not show activity. Thus, like the genes for CK-Pf and CK-Ef, those for CK-Ts, CK-Tb, CK-Tm, and CK-Fn were established to encode active kinases. Surprisingly, CK-Pf was found to be active over the entire pH range studied, with optimum activity at pH 9.9 and above (Fig. 4). CK-Ts and CK-Tb showed higher specific activities but the same increase in activity with increase in pH (Fig. 4). This activity at high pH is not shared by the nonhyperthermophilic CKs. CK-Fn exhibits a pH optimum of 9.9 but a sharp decrease in activity at higher pH, with little activity at pH 10.9 or above (Fig. 4). Initial screening showed that the pH activity profiles of CK-Tm and CK-Ef (data not shown) are similar to those of CK-Fn, so those kinases were not studied in detail.
pH activity profiles of recombinant carbamate kinases from different organisms, at 40°C in solutions prepared using 0.2 M ammonia and bicarbonate. Values are the averages of three replicates and error bars are standard deviations.
The stability and activity of the three hyperthermophilic proteins at high pH, which have potential relevance to processing ammonia from alkaline waste streams, instead prompted us to further examine the pH activity dependence of these CKs. Accordingly, assays were also carried out with solutions prepared using 0.05 M NH3 and NaHCO3, as well as 0.4 M concentrations of each reagent, in addition to the 0.2 M quantities discussed above. The results of these analyses at 40°C are presented in Fig. 5. The enzyme kinetics are complex, with the binding and reaction of both carbamate and ATP each probably pH dependent. Compounding this, the concentration of carbamate available for catalysis is dependent on the quantities of ammonia and bicarbonate used to prepare the assay solutions and on the pH-dependent equilibrium between the ammonia and bicarbonate, as well as that between carbon dioxide and ammonium hydroxide (Fig. 1). Given this complexity, only general conclusions should be drawn from the results. Even so, for each of the hyperthermophilic CKs, the shapes of the pH profiles at the different reagent concentrations (Fig. 5) are broadly similar. At each concentration, the CKs all show high activity at pH 9.4 and above.
pH activity profiles of carbamate kinases at 40°C in solutions prepared using 0.05 M (green), 0.2 M (blue), or 0.4 M (red) concentrations of both ammonia and bicarbonate. Values are the averages of three replicates and error bars are standard deviations.
The most obvious general effect of increasing and decreasing the concentrations of ammonia and bicarbonate is a matching rise and fall in enzyme activity. This is attributable to corresponding changes in the concentration of carbamate, showing that the kinases are not saturated with this substrate. Under these circumstances, it is even more remarkable that the hyperthermophilic kinases retain their activity at the highest pH values, where there is less carbamate. We used 13C nuclear magnetic resonance (NMR) spectroscopy to measure carbamate concentration as a function of pH in aqueous solutions prepared using NH3 and NaH13CO3. Apart from the substitution of the 13C-labeled bicarbonate, the spectroscopic conditions were identical to those used to measure CK activity, and the results are illustrated in Fig. 6. This shows that with 0.4 M NH3 and NaH13CO3, the amount of carbamate drops from about 56 mM at pH 9.9, corresponding to 14% of the bicarbonate used, to around 13 mM at pH 11.4, or less than 3.5%. Despite this 4-fold decrease in carbamate concentration, under the same conditions the activity of each of the hyperthermophilic CKs at pH 11.4 is only 15 to 25% less than that at pH 9.9.
Concentration of carbamate formed in solutions prepared using 0.4 M ammonia and 0.4 M 13C-labeled bicarbonate, measured using 13C NMR spectroscopy.
At any given pH, the equilibrium constant (K) of the reactions involved in the formation of carbamate (Fig. 1) is represented overall by the following equation:
As a further probe of CK robustness, enzyme behavior was also investigated at various temperatures. The results of representative analyses at pH 9.9, again with CK-Pf, CK-Ts, CK-Tb, and CK-Fn, in solutions prepared using 0.2 M NH3 and NaHCO3, at 20, 40, 60, and 80°C, are presented in Fig. 7. Temperature dependence introduces further complexity into the enzyme kinetics. Even so, all the hyperthermophilic CKs showed increasing activity with increasing temperature and maximum catalysis at 80°C. The thermophilic CK-Fn is most active at 40°C but shows decreased function at higher temperatures. CK-Ts and CK-Tb were found to be the most active at 80°C and maintained catalysis over the broadest temperature range. CK-Tb generally showed the highest activity of all the CKs tested. It is noteworthy that the activity of this enzyme, which normally functions at a physiological temperature above 80°C, is even comparable to that of CK-Fn (and of CK-Tm and CK-Ef; data not shown) at 20 to 40°C. This behavior is rare, as few hyperthermophilic enzymes have been found to show similar catalytic efficiency to their nonhyperthermophilic counterparts at these lower temperatures (29, 30). This property is particularly relevant, as lower temperatures may be preferable, or even required, for industrial processes utilizing these enzymes. It was also of interest to investigate whether the activity of the hyperthermophilic CKs at high pH and temperature correlates with stability over time, an important property in applied biocatalysis. There was no measurable change in catalysis with any of the hyperthermophilic CKs after incubation in buffer at 40°C for 60 h. In contrast, CK-Fn was inactivated by this treatment. In addition, CK-Pf and CK-Tb were stored at 4°C for 6 months and still retained full catalytic activity.
Temperature activity profiles of carbamate kinases at pH 9.9 in solutions prepared using 0.2 M ammonia and bicarbonate. Values are the averages of three replicates and error bars are standard deviations.
Overall, the three hyperthermophilic CKs that show high sequence similarity, CK-Pf, CK-Ts, and CK-Tb, display high activity at pH 9.4 to 11.4 and, as mentioned above, are more efficient at processing the available carbamate at the top of this pH range. It is intriguing to note that the CKs that share these properties are all likely to operate physiologically in the anabolic direction. However, based on the available evidence, it is not possible to predict whether there is any physiological significance to this observation. It is not likely that the CKs have evolved to enable them to function at high pH, as CK is an intracellular enzyme. Instead, it is probable that the activity of these CKs under alkaline conditions is a coincidental consequence of their extreme thermal stability. Hyperthermophilic protein variants rely on an increased number of stabilizing effects to confer structural stability at higher temperatures. Even though disruption of some of these, including ion pairs that are thought to contribute to the thermal stability of CK-Pf (12), would be likely at high pH, other temperature-stabilizing interactions such as hydrophobic effects (30) would be expected to persist. Other hyperthermophilic enzymes have been shown to have increased robustness and withstand denaturing effects of organic solvents, high salt concentrations, and extremes of pH, as side effects of their thermal stability (31). However, pH stability at high or low pH is not an inherent property of all hyperthermophilic enzymes. Instead, it is more commonly associated with extracellular hydrolytic enzymes (32) that are required to function in harsh environments.
Irrespective of their basis, the properties of the hyperthermophilic CKs described above, and particularly those of the most active and robust, CK-Ts and CK-Tb, make them suitable for application in bioremediation and biosynthetic processes. Among uses that are envisaged, ammonium hydroxide solutions have already been proposed for carbon dioxide capture (2, 33, 34). Absorbed carbon dioxide will spontaneously form carbamate, which CK-Ts and CK-Tb would then convert to CP, en route to other useful compounds. The stability and activity of these CKs at alkaline pH mean that they are able to function directly in aqueous ammonia solutions, with 0.0001 to 1.0 M solutions having pH values of approximately 9.8 to 11.8, that will be somewhat reduced through carbon dioxide absorption. Alkaline ammonia solutions are preferable to those adjusted to lower pH values because they more efficiently absorb carbon dioxide and have a higher loading capacity (2). More broadly, we are currently investigating potential routes for further chemical and biochemical transformation of the CP formed by the hyperthermophilic CKs into value-added compounds. In aqueous alkaline solutions, CP decomposes to cyanate (20), but in ammonium hydroxide solutions cyanate is converted quantitatively to the fertilizer urea (35). This is the basis of Wöhler urea synthesis (36), reported in 1828 and widely recognized as a cornerstone of organic chemistry.
Our focus is biotransformations and, as a preliminary illustration of the utility of CK-derived CP, we have established the feasibility of in situ conversion by ATC-Tb of CP produced by CK-Tb into carbamoyl aspartate (Fig. 8). The catalytic subunit of the transcarbamoylase from Thermococcus barophilus was expressed (Fig. 3) using the methodology described above for the CKs and was found to produce carbamoyl aspartate when provided with CP and aspartate. When the CP was generated in situ using CK-Tb, with ATP in solutions of ammonia and bicarbonate, the yield of carbamoyl aspartate was >95%, based on ATP converted to ADP, indicating that the CP produced is efficiently converted in situ by ATC-Tb. Finally, to reduce the ATP requirement of the CKs, we applied the cost-effective recycling system we reported previously (24), using cheaply prepared AP and E. coli lysate as the catalyst. The multienzymatic reaction was carried out at 40°C. Using 0.4 M NH3 and 0.4 M NaHCO3 at pH 9.9, where CK-Tb displayed the highest specific activity, with aspartic acid (5 mM), AP (2 mM), ATP (0.1 mM), CK-Tb (25 μg), ATC-Tb (75 μg), and S30 cell-free lysate from E. coli BL21(DE3) (2% vol/vol; 10 μl), carbamoyl aspartate production was monitored directly using HPLC-mass spectrometry (HPLC-MS), and the results are shown in Fig. 9. The stability of the system is demonstrated by the steady production of carbamoyl aspartate during the first 30 min of this assay. After this time, most of the AP has been consumed and the reaction rate decreases, but after 60 min a final yield of ∼85% is reached, based on the amount of the limiting reagents used (AP and the catalytic amount of ATP). Furthermore, the ratio (18:1) of carbamoyl aspartate produced to ATP used establishes that the ATP recycling system is effective under these conditions, since the ATP must have been turned over 17 times. Most significantly, the conversion of CP to carbamoyl aspartate is only one step in the multienzyme cascade using either the stoichiometric quantity of ATP or the catalytic amount of ATP recycled with AP. Therefore, the respective yields of carbamoyl aspartate of >95% and ∼85% from these processes correspond in each case to the minimum proportion of the CP that undergoes in situ biotransformation, even in these nonoptimized systems. These results thus establish proof of concept that hyperthermophilic carbamate kinases may be used to transform ammonia and carbon dioxide to produce CP directly in aqueous ammonia solutions. Furthermore, the CP may be used not only through chemical transformation to urea but also for in situ biotransformations into stable products using only catalytic amounts of ATP.
Multienzymatic reaction scheme for the synthesis of carbamoyl aspartate from ammonia and bicarbonate through the coupled activity of carbamate kinase and aspartate transcarbamoylase, with ATP recycling catalyzed by E. coli S30 cell lysate.
Carbamoyl aspartate formation in aqueous ammonia (0.4 M) and bicarbonate (0.4 M), at pH 9.9 and 40°C, with aspartic acid (5 mM), acetyl phosphate (2 mM), and ATP (0.1 mM), catalyzed by Thermococcus barophilus carbamate kinase and aspartate transcarbamoylase, and S30 cell lysate from E. coli BL21(DE3). Values are the averages of three replicates and error bars are standard deviations.
MATERIALS AND METHODS
Strains and recombinant expression vectors.All DNA cloning cell work was carried out with E. coli DH5α and AN1459 (37). Protein expression was carried out with E. coli BL21(DE3). The amino acid sequences of CKs from the hyperthermophiles Pyrococcus furiosus (UniProt accession no. P95474), Thermococcus sibiricus (UniProt C6A5J0), and Thermococcus barophilus (UniProt F0LK57), the thermophiles Fervidobacterium nodosum (UniProt A7HNY8) and Thermosipho melanesiensis (UniProt A6LPA8), and the mesophiles Enterococcus faecalis (UniProt P0A2X7) and Clostridium tetani (UniProt Q890W1), and of the aspartate transcarbamoylase (ATC) catalytic subunit sequence from Thermococcus barophilus (UniProt F0LI09) were retrieved from the UniProt Knowledgebase. Amino acid sequence similarity of the CKs was determined on the UniProt platform using Clustal Omega software, and alignments were formatted using CLC Sequence Viewer v6.5.3 (CLC bio A/S) (Fig. 2). Genes encoding the proteins, prepared through reverse translation using codon optimization for expression in E. coli, were synthesized commercially (GeneArt, Germany) in the pMK transport vector, which is flanked by NdeI and EcoRI restriction sites. Using these restriction sites, each gene was subcloned into the pETMCSIII expression vector (30), which allows expression of the coding region under the control of a T7 promoter with an N-His6 tag and contains the ampicillin resistance gene (ampR).
Protein expression.For each CK, competent E. coli BL21(DE3) (New England BioLabs) cells were transformed through electroporation with the expression vector (pETMCSIII-CK) encoding the desired CK and grown overnight on agar plates with lysogeny broth medium supplemented with 0.5 mg · ml−1 ampicillin (LB-Amp). Single colonies were used to inoculate 100 ml of LB-Amp and cultures were grown for 18 h (37°C at 200 rpm). IPTG (isopropyl-β-d-thiogalactopyranoside) induction was not necessary, as basal levels of T7 RNA polymerase in the BL21(DE3) cells sufficed for the expression of high recombinant protein yields. Cells were then harvested through centrifugation (4,000 × g for 15 min) and lysed in a French pressure cell, and the lysate was clarified by high-speed centrifugation (20,000 × g for 1 h). Overexpressed N-His6-CK was then purified through Ni-affinity chromatography (His Gravi-Trap gravity flow column; GE Healthcare). Overexpression of purified recombinant proteins was confirmed through SDS-PAGE analysis and carried out using 20% acrylamide Laemmli gels and Laemmli buffers (Fig. 3). Purified protein concentrations were determined from absorbance at 280 nm using theoretically determined molar extinction coefficients (ProtParam; ExPASy server). These measurements were verified using Bradford reagent analysis and found to agree within 15%. The same approach was used to express N-His6-ATC.
Carbamate kinase activity assays.CK activity was measured through monitoring the catalytic conversion of ATP to ADP accompanying the conversion of carbamate to carbamoyl phosphate. Assays were conducted in 0.5-ml aqueous solutions of NH3 and NaHCO3, in the presence of 10 mM ATP and 10 mM MgCl2, and initiated through addition of CK (0.025 μM for CK-Tb and CK-Fn, 0.05 μM for CK-Ts, and 0.1 μM for CK-Pf). For determination of the CK activity at various pH values (Fig. 4) and NH3 and NaHCO3 concentrations (Fig. 5), aqueous solutions were prepared using 0.05, 0.2, or 0.4 M concentrations of both NH3 and NaHCO3. The pH of these equimolar solutions was initially around 9.5 to 10.0 and was adjusted to the desired value (7.9 to 11.4) through addition of either 6 M HCl or NaOH. Assay mixtures were incubated at various temperatures over 20 min, during which time 30-μl aliquots were withdrawn, quenched with an equal volume of 1.0 M HCl, rested for 20 min, and then mixed with an equal volume of 2.0 M NaOH.
Conversion of ATP to ADP in the assay samples was analyzed using reverse-phase HPLC with detection of absorbance at 259 nm. The tetrabutylammonium cation was used as an ion-pairing agent, allowing separation of the nucleoside phosphates based on the number of phosphate groups. In this way, ATP, with three ionized phosphates (buffered at pH 5), was retained longer by the tetrabutylammonium-coated stationary phase than was ADP, with two phosphates. More specifically, ion-pair HPLC separation of ADP and ATP was carried out on an Agilent 1100 HPLC system fitted with a peptone-dextrose agar (PDA) detector, using an Alltech Alltima HP C18 column (5 μm, 250 × 4.6 mm, with a 7.5 × 4.6 mm guard column). Samples were eluted (1 ml · min−1) with a gradient of 60 mM ammonium dihydrogen phosphate and 5 mM tetrabutylammonium phosphate in H2O (solvent A) and 5 mM tetrabutylammonium phosphate in methanol (solvent B), according to the following program: 0 to 18 min, 87% A; 19 to 23.2 min, 70% A; and 24 to 27 min, 87% A, with linear transitions between mixtures. Observed retention times were as follows: AMP, 6.9 min; ADP, 12.4 min; and ATP, 23.9 min. The dwell volume of the analytical system is approximately 1.2 ml, so ATP elutes with 70% solvent A before the initial solvent mixture is restored for the next analysis. Higher analytical throughput was obtained through monitoring 4 consecutive sample injections (1-min delay) in a single 27-min analysis window, without coelution of peaks from different samples.
The concentrations of ADP in assay samples were quantified from the integrated chromatogram peak area (Chemstation; Agilent), using a standard calibration curve. For each experiment, ADP concentration was plotted against time, and initial velocity was measured as the slope of the line of best fit, as determined through linear regression. Enzyme specific activity was then calculated as μmol ADP formed per min per mg of CK added (Fig. 4, 5, and 7). The enzyme activity presented for each set of conditions is the average of three replicate assay experiments and the errors are standard deviations. Nonenzymatic controls were carried out to monitor background ATP and ADP hydrolysis, which was shown to be negligible. An example of production of ADP over time as measured with an enzymatic and a nonenzymatic assay is shown in Fig. 10.
Illustration of ADP formation in ammonia (0.4 M) and bicarbonate (0.4 M), at pH 11.4 and 40°C, in the presence and absence of Thermococcus barophilus carbamate kinase.
13C NMR spectroscopic analysis of carbamate concentration.Aqueous solutions prepared using equimolar amounts of 13C-labeled sodium bicarbonate and ammonia were adjusted to a pH ranging from 7.9 to 11.4 using either 6 M HCl or NaOH, then analyzed using 13C NMR spectroscopy (Avance 400MHz; Bruker), following the procedure of Mani and coworkers (2). At acidic pH, the only carbon resonance is observed at 159.5 ppm. It was assigned to the rapidly exchanging bicarbonate/carbonate pair. Increasing pH caused this peak to shift upfield, in accordance with the literature (2), and a second carbon resonance appeared at 165.6 ppm. This second peak did not shift with pH and was assigned to carbamate formed through the reaction of carbon dioxide (from bicarbonate) with ammonia (Fig. 1). Integration of these peaks was used to measure relative concentrations (Fig. 6).
Enzyme-catalyzed production of carbamoyl aspartate.Carbamoyl aspartate production was performed at pH 9.9 and 40°C in 0.5 ml water containing NH3 (0.4 M), NaHCO3 (0.4 M), l-aspartic acid (5 mM), ATP (0.1 mM), MgCl2 (25 mM), CK-Tb (1.5 μM), ATC-Tb (4.5 μM), and E. coli BL21(DE3) S30 lysate (24) (10 μl), with AP (24) (2 mM) added in 5 aliquots at 5-min intervals. Samples were withdrawn, quenched with an equal volume of HCl, neutralized with NaOH, and then diluted with water prior to analysis by HPLC-MS. Analysis was performed using a Waters Alliance 2695 separation module coupled to a Waters tandem quadruple detector (TQD). HPLC separation was performed using a Phenomenex Develosil RP-Aqueous-AR 5-μm C30 (250 × 4.6 mm) column heated at 35°C and eluting with 0.1% formic acid in H2O (0 to 1 min), followed by a linear gradient to 0.1% formic acid in ACN/H2O 30/70 (1 to 7 min) at a flow rate of 0.7 ml · min−1 (15:1 flow splitter). Carbamoyl aspartate, with a retention time of 5.5 min, was identified and quantified using integrated multiple reaction monitoring (MRM) and a calibration curve developed with authentic commercial material. The MRM parameters used were as follows: carbamoyl aspartate (parent [m/z], 176.9475), channel 1 (daughter [m/z], 45.9169; dwell [s], 0.025; cone [V], 20; collision [V], 24), channel 2 (daughter, 69.9541; dwell, 0.025; cone, 20; collision, 26), channel 3 (daughter, 73.917; dwell, 0.025; cone, 20; collision, 24), channel 4 (daughter, 87.9496; dwell, 0.025; cone, 20; collision, 18), and channel 5 (daughter, 45.9169; dwell, 0.025; cone, 20; collision, 24).
ACKNOWLEDGMENTS
We acknowledge financial support of this work by the Grains Research and Development Corporation (GRDC), the Australian Research Council (ARC), the Australian National University, and CSIRO.
FOOTNOTES
- Received 13 October 2017.
- Accepted 4 November 2017.
- Accepted manuscript posted online 17 November 2017.
- Copyright © 2018 American Society for Microbiology.
