Observation of a hybrid random ping-pong mechanism of catalysis for NodST: A mass spectrometry approach

Authors


Abstract

An efficient enzyme kinetics assay using electrospray ionization mass spectrometry (ESI-MS) was initially applied to the catalytic mechanism investigation of a carbohydrate sulfotransferase, NodST. Herein, the recombinant NodST was overexpressed with a His6-tag and purified via Ni-NTA metal-affinity chromatography. In this bisubstrate enzymatic system, an internal standard similar in structure and ionization efficiency to the product was chosen in the ESI-MS assay, and a single point normalization factor was determined and used to quantify the product concentration. The catalytic mechanism of NodST was rapidly determined by fitting the MS kinetic data into a nonlinear regression analysis program. The initial rate kinetics analysis and product inhibition study described support a hybrid double-displacement, two-site ping-pong mechanism of NodST with formation of a sulfated NodST intermediate. This covalent intermediate was further isolated and detected via trypsin digestion and Fourier transform ion cyclotron resonance mass spectrometry. To our knowledge, these are the first mechanistic data reported for the bacterial sulfotransferase, NodST, which demonstrated the power of mass spectrometry in elucidating the reaction pathway and catalytic mechanism of promising enzymatic systems.

Sulfotransferases are enzymes that install the sulfuryl group from 3′-phosphoadenosine 5′-phosphosulfate (PAPS), a ubiquitous sulfate group donor, to numerous acceptor substrates including proteins (Farzan et al. 1999; Kehoe and Bertozzi 2000), carbohydrates (Bowman and Bertozzi 1999; Habuchi 2000), and small molecules. This particular reaction, referred to as sulfuryl group transfer, is widely observed from bacteria to mammals and plays a key role in a variety of cellular communication events (Bowman and Bertozzi 1999; Armstrong and Bertozzi 2000). For example, the tyrosine sulfation in chemokine receptor CCR5 is important in the HIV infection process, and the sulfation of estradiol, catalyzed by estrogen sulfotransferase (EST), can activate the estrogen receptor (ER) and is believed to underlie the genesis of breast cancer. The interaction between L-selectin and its sulfated glycoprotein ligands contributes to the leukocyte immigration into inflamed tissues and results in acute and chronic inflammation (Bistrup et al. 1999; Hiraoka et al. 1999). Thus, sulfotransferases play a significant role in modulating normal and pathogenic biological processes and are considered to be a novel class of therapeutic targets (Bowman and Bertozzi 1999; Armstrong and Bertozzi 2000).

The catalytic mechanism of sulfuryl group transfer has recently become the focus of intense interest due to its importance in understanding the role of sulfotransferases in vivo. Crystallographic and mutational analyses have provided critical information about the active-site residues involved in substrate binding and sulfotransferase catalysis (Kakuta et al. 1997, 1998a,b, 1999; Bidwell et al. 1999; Ong et al. 1999; Perdersen et al. 2000). For example, structural studies of the N-sulfotransferase domain of heparan sulfate N-deacetylase/N-sulfotransferase (HNDST-1) complexed with 3′-phosphoadenosine 5′-phosphate (PAP) have helped to unravel the central catalytic roles of Lys and Thr in the 5′PSB-loop, and Ser and Tyr in the 3′PB motif (Kakuta et al. 1999). In another study, the EST-PAP-estradiol structure along with site-directed mutagenesis work further verified Lys, Thr, Ser, and His as highly conserved residues in the active sites of sulfotransferases (Kakuta et al. 1997). Furthermore, the relative orientation of the bound vanadate and PAP molecules in the crystal structure of the ternary complex, EST-PAP-vanadate (Kakuta et al. 1998b), indicated a characteristic transition state for an in-line sulfuryl group transfer mechanism (Bartolotti et al. 1999).

Direct kinetic analyses have been performed to investigate the mechanism of bacterial arylsulfate sulfotransferases (ASSTs; Kwon et al. 2001) and cytosolic sulfotransferases. Varin and coworkers determined that phenol sulfotransfer-ases (PSTs) and flavonol sulfotransferases (FSTs) followed an ordered Bi-Bi mechanism, using product inhibition and initial rate experiments (Varin and Ibrahim 1992). A random Bi-Bi mechanism was elucidated for EST (Zhang et al. 1998) and insect sulfotransferase, retinol dehydratase (Vakiani et al. 1998) via similar kinetic analyses. Although a number of kinetic studies point towards a sequential mechanism for cytosolic sulfotransferases, a ping-pong mechanism with formation of a sulfated enzyme intermediate has been reported for a bacterial ASST using kinetic measurements (Kwon et al. 2001) and stereochemistry studies (Chai and Lowe 1992). Unfortunately, the structural and mechanistic information for mammalian carbohydrate and protein sulfotransferases is sparse, and no detailed kinetic analyses for the mechanisms of those Golgi-resident sulfotransferases have been reported to date. Considering the central role of Golgi-resident sulfotransferases in numerous disease states, it would be very useful to explore their catalytic mechanism via kinetic analyses. Insights gained from research in this area will reinforce our knowledge of sulfotransferases and greatly facilitate sulfotransferase-targeted drug design (Radzicka and Wolfenden 1995).

Unfortunately, the Golgi-resident sulfotransferases are membrane-bound and difficult to express at high levels. We therefore initially focused our attention on a functionally related sulfotransferase, the GlcNAc-6-O-carbohydrate sulfotransferase, NodH, from the nitrogen fixing bacterium, Rhizobium meliloti (Ehrhardt et al. 1995). NodH, also called NodST, acts as a host-specific nodulation switch by catalyzing the transfer of a sulfuryl group from PAPS to the 6-hydroxyl group of the reducing terminal GlcNAc residue of a lipochitooligosaccharide (Roche et al. 1991). The resulting sulfated lipochitooligosaccharide, or “Nod factor,” is critical for root nodulation and bacterial infection (Freiberg et al. 1997). NodST can also utilize the simple disaccharide chitobiose (1) as substrate, producing chitobiose-6-OSO3(2) as the product (Scheme 1; Lin et al. 1995; Schultze et al. 1995). This enzyme can be generated in large quantities via bacterial overexpression (Burkart et al. 2000) and shows GlcNAc-6-O-sulfotransferase activity similar to that of some mammalian enzymes of therapeutic interest, making it an ideal model sulfotransferase for our preliminary mechanistic studies.

Soft ionization methods such as electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS) have been demonstrated to be complementary techniques to conventional spectrophotometric methods for enzyme kinetic studies (Newton et al. 1997, 1999; Zechel et al. 1998; Bothner et al. 2000; Ge et al. 2001; Pi et al. 2002; Gao and Leary 2003). The assay described herein is a facile and broadly applicable kinetics assay for sulfotransferases using ion trap (IT)-ESI-MS (Ge et al. 2001; Pi et al. 2002). The assay does not require chromophogenic substrates or products, and the total analysis time is comparable to those of traditional methods. Enzyme kinetic parameters such as the Michaelis-Menten constant KM, inhibition constant Ki, and the reaction turn over number kcat for some enzymes determined by the ESI-MS assay were in excellent agreement with that of the traditional spectrometric assay, thus showing proof of principle (Ge et al. 2001; Pi et al. 2002; Gao and Leary 2003). In particular, the KM value for PAPS and the Ki value for PAP were determined using this ESI-MS method (Pi et al. 2002).

Herein we demonstrate the initial application of this ESI-MS assay for determining the catalytic mechanism of the NodST catalyzed sulfuryl group transfer reaction, in which the sulfate group is transferred from PAPS to chitobiose, yielding PAP and chitobiose-6-OSO3 as products (Scheme 1). NodST was generated via bacterial overexpression and purified as a histine-tagged protein using Ni-NTA chromatography. Utilizing a chondroitin disaccharide, α-ΔUA-[1→3]-Gal-NAc-6S (ΔDi-6S), as an internal standard (Fig. 2A), a single-point normalization factor between the product and the internal standard was obtained and used for product quantification. The catalytic mechanism of NodST was subsequently determined by initial rate kinetic analysis and product inhibition study using the ESI-MS assay, and confirmed by MS identification of the covalent enzyme intermediate. These are the first mechanistic data reported for GlcNAc-6-O-carbohydrate sulfotransferase NodST, and the data are highly suggestive of a hybrid double-displacement ping-pong mechanism. To our knowledge, this is also the first time that mass spectrometry has been shown to unambiguously allow for enzyme mechanistic studies on a class of sulfotransferases.

Results

Initial velocity kinetic analysis

Prior to the initial rate kinetic study, optimum enzymatic reaction conditions were determined. Using a plot of the formation of product versus time at a constant NodST concentration, it was determined that a reaction quench time of 4 min was suitable and was within the linear region of the progress curve (data not shown). Initial rates were obtained using data within the low reaction conversion in order to avoid complicated reaction conditions such as substrate depletion, product inhibition, and reverse reaction, all of which commonly arise from the steady-state assumption and Michaelis-Menten equation. The initial velocity (V0) was then calculated by dividing the product concentration by the reaction time (Tq) as shown in equation 1.

equation image((1))

In order to obtain accurate and comprehensive mechanistic information for NodST using our kinetic method, the appropriate substrate concentration ranges were chosen to be 0.2–5.0 KM based on previously published Michaelis-Menten constants (Pi et al. 2002). The concentration of the donor substrate PAPS was varied between 1.25 μM and 25 μM according to its known apparent KM value of 6.7 μM at a chitobiose concentration of 1 mM, whereas the concentration of the sulfate acceptor chitobiose extended from 0.05 mM to 1 mM based on its apparent KM value of 0.28 mM at a PAPS concentration of 25 μM. Preliminary experiments were performed to make sure that inhibition of both substrates was negligible within the chosen concentration ranges. For each substrate, five different concentrations were used in the ESI-MS assay. The initial reaction rate was determined as a function of PAPS concentration at different fixed chitobiose concentrations and as a function of chitobiose at different PAPS concentrations (see Materials and Methods). The resulting kinetic data was fit to two mechanistic models (sequential mechanism and ping-pong mechanism) of bisubstrate reaction in the SAS program, and the best fit was obtained in the case of a ping-pong mechanism model. The two double reciprocal plots shown in Figure 3A,B were each an average of four replicate experiments, and both of them resulted in an array of five parallel lines.

Product inhibition study using PAP

To obtain more information about the catalytic mechanism of NodST, we evaluated the product inhibition patterns of PAP with respect to both substrates. In our earlier studies, PAP was shown to inhibit PAPS binding competitively with a Ki value of 1.80 μM (Pi et al. 2002). Herein, the inhibition study of PAP to the sulfate acceptor substrate, chitobiose, was performed. As depicted in Figure 4, the best data fit was obtained with a noncompetitive inhibition model. The five lines representing five different PAP concentrations in the double reciprocal plot shared a common X-intercept, unambiguously indicating a noncompetitive inhibition mode of PAP in regards to chitobiose. The apparent inhibition constant, Ki, was determined to be 5.9 (± 1.1) μM. These results were generated from three replicate experiments.

Identification of sulfated NodST intermediate

Incubation of an enzyme that follows a ping-pong mechanism with the donor substrate results in the formation of a covalent bound enzyme intermediate if the acceptor substrate is absent. Satishchandran et al. (1992) reported the characterization of a phosphorylated adenosine 5′-phosphosulfate (APS) kinase intermediate by incubating the enzyme with [γ-32P] adenosine 5′-triphosphate (ATP) followed by gel filtration removal of the nucleotides. They showed that phosphoryl linkage in the phosphorylated enzyme intermediate was quite stable and that the covalently bound intermediate can survive the gel filtration process. The sulfuryl group transfer reaction catalyzed by the nucleotide diphosphate (NDP) kinase proceeds via a ping-pong mechanism, and the sulfated NDP intermediate was also isolated based on the premise that the intermediate is stable without the addition of the second substrate (Peliska and O'Leary 1991).

To identify the intermediate formed via the ping-pong mechanism of NodST, the enzyme was incubated with PAPS, denatured at room temperature, and digested with trypsin to yield the soluble peptides. Previous studies have shown that sulfated peptides may fragment in positive mode (Seibert et al. 2002). Therefore, negative ion mode detection was used in the ESI-FT-ICR mass spectrometric analysis in this case. The digestion product was infused into the mass spectrometer in an NH4OAc buffer (pH 7.5) to avoid hydrolysis of the sulfated peptide observed in low pH conditions (Seibert et al. 2002). As seen in Figure 5A, when NodST was incubated with PAPS and then digested, an ion at m/z 1364.6402 with a molecular weight of 4096.9425 was identified to be the −3 charge state ion corresponding to [T2–3–3H]3− of the tryptic peptide T2–3 with a mass error of 0.51 ppm. The T2–3 peptide corresponds to the amino acid sequence 19–53 (TGTHYLEELVNEHPNVLSNGELLNTYDTNWPDKER) with a missed cleavage site at Lys51. This Lys residue is included in the amino acid sequence DKE, which has been shown to exhibit missed cleavages when treated with trypsin (Thiede et al. 2000). Because PAPS was prepared as the lithium salt, a small amount of lithium adduct was also observed as [T2–3+Li-4H]3−. In the spectrum of Figure 5A, another −3 charge state ion that is m/z 26.66 or 79.98 mass units higher than [T2–3–3H]3− was clearly observed in the spectrum, but missing from the spectrum resulting from digestion of the enzyme alone (Fig. 5B). In addition, the pair of peptides T2–3 and [T2–3+79.98] were also detected as −4 charge state ions. The mass increment of 79.98 is expected for a covalently bound sulfate, and is attributed to the sulfated T2–3 peptide resulting from the tryptic digest of the sulfated NodST intermediate. As a control experiment, the enzyme was also incubated with (NH4)2SO4. No sulfated intermediate was detected in this experiment, thus ruling out the possibility of a gas phase artifact.

Discussion

An enzyme-catalyzed sulfuryl group transfer reaction can proceed via two basic mechanisms. One is a direct sulfate group transfer between the two bound substrates at the enzyme active site in a ternary complex. This is called a sequential Bi-Bi mechanism. The other one, a ping-pong Bi-Bi mechanism, involves a modified enzyme intermediate in a double-displacement pathway. Our results in the initial rate kinetic study are a family of parallel lines (Fig. 3), which supports a ping-pong (substituted enzyme) catalytic mechanism for NodST reaction. Thus, the kinetic pathway observed for NodST is different from that reported for cytosolic sulfotransferases, which were suggested to follow a sequential mechanism reflected in the initial velocity pattern as a family of intersecting lines. Although contradictory to the mechanisms observed for most sulfotransferases, the data generated herein are very compelling and quite supportive of a ping-pong mechanism. The ping-pong Bi-Bi mechanism indicated by our initial velocity patterns is further substantiated by the identification of a sulfated enzyme intermediate when NodST was incubated with PAPS in the absence of chitobiose.

Although all of the evidence discussed above is consistent with a classical ping-pong Bi-Bi mechanism, the product inhibition patterns obtained are not. The classical ping-pong mechanism of NodST (Scheme 2A) requires the initial sulfation of NodST by the donor substrate PAPS. A sulfated NodST intermediate is generated after PAP leaves the enzyme active site. Chitobiose, the acceptor substrate, binds to the sulfated-NodST intermediate, and the enzyme-bound sulfate group is transferred to chitobiose to produce chito-biose-6-OSO3. Thus, PAPS and chitobiose-6-OSO3 bind to a common site on the free enzyme form, whereas PAP and chitobiose also bind to the same common site but with the modified enzyme form, the sulfated NodST. Consequently, the product inhibition pattern of PAP with respect to PAPS will be noncompetitive, and that of PAP in regards to chitobiose will be competitive. However, the opposite product inhibition pattern was observed in our product inhibition studies, in which PAP was determined to be a competitive inhibitor with respect to PAPS and a noncompeti-tive inhibitor in regards to chitobiose. This result is consistent with a hybrid ping-pong rapid equilibrium random two-site mechanism (Northrop 1969; Wong and Wong 1983). In this hybrid ping-pong mechanism (Scheme 2B), PAPS and chitobiose bind independently and randomly at two different sites on NodST. PAPS and PAP compete for one site in rapid equilibrium, while chitobiose and chitobiose-6-OSO3 compete for the other site. A flexible domain which presumably contains the 5′PSB-loop of NodST will be sulfated by PAPS bound in the first site, yielding the substituted enzyme intermediate. This is followed by the transfer of the sulfuryl group from the mobile domain to the chitobiose bound in the second binding site, and PAP will be released in a random fashion before or after the addition of the second substrate, chitobiose. In this fashion, the reaction is not restricted to an exclusive formation of a binary complex or to a compulsory formation of a central ternary complex. The product inhibition pattern of this hybrid ping-pong mechanism will be completely reversed with regards to that of the classical ping-pong mechanism. Furthermore, in a classical ping-pong mechanism, the enzyme will accommodate both substrates using the same binding site, which requires the two substrates to be very similar in structure. However, the two substrates for NodST, PAPS and chitobiose, are distinctly different in geometry, size, and charge state.

The trypsin digestion experiment further supports our hypothesis of a ping-pong mechanism. The MS data show that the T2–3 tryptic peptide, which corresponds to amino acid sequence 19–53 of NodST, is sulfated when NodST is incubated with PAPS. Because the 5′-phosphosulfate binding (5′PSB) motif of NodST was identified as the amino acid sequence 17–24 in the putative PAPS binding pocket (Fig. 6), the flexible domain used to transfer the sulfate group in NodST's hybrid two-site ping-pong mechanism might reside in the PAPS/PAP binding site, but not the chitobiose/chitobiose-6-OSO3 binding site. The possible sulfation site of NodST might be located in the overlapping part of NodST's 5′PSB-loop sequence and the T2–3 tryptic peptide sequence, which starts at Trp19 residue and ends at Leu24 residue. Thus, the possible sulfate-modified nucleophilic amino acids of NodST include Thr19, Thr21, His22, and Tyr23. Ongoing investigations are focusing on the characterization of the sulfated NodST intermediate and identification of the sulfated residue at the active site of the enzyme.

Recently solved X-ray crystal structures of cytosolic sulfotransferase complexes have shed light on the catalytic mechanisms. Studies on the mEST-PAP-estradiol ternary complex (Kakuta et al. 1997), mEST-PAP-vanadate complex (Kakuta et al. 1998b), and SULT2A3-PAP binary complex (Perdersen et al. 2000) support a transition state structure for the reaction of in-line sulfate group transfer, which suggests a sequential catalytic mechanism (Armstrong and Bertozzi 2000; Negishi et al. 2001). In addition, kinetic analyses of human EST (Zhang et al. 1998), human PST (Varin and Ibrahim 1992), plant FST (Varin and Ibrahim 1992), and an insect sulfotransferase further verified that cytosolic sulfotransferases proceed via a sequential mechanism which requires the formation of a ternary complex. However, earlier crystal structures of mouse estrogen sulfotransferase showed similarities between kinase and sulfotransferase, and suggested that at least partial sulfation may proceed via a ping-pong mechanism with the formation of a sulphohistidine intermediate (Kakuta et al. 1997). These findings suggest that our discovery of a hybrid two-site ping-pong mechanism may be operative in the NodST system. Although this mechanism involves two independent substrate binding sites instead of a single binding site as in the classical ping-pong mechanism, a flexible domain on NodST is used to accommodate the sulfate and form the substituted enzyme intermediate which differentiates the hybrid ping-pong mechanism from the sequential mechanism adopted by cytosolic sulfotransferases. When the amino acid sequences of cytosolic sulfotransferases are aligned with NodST (Fig. 6), there are two possible explanations for the different observed mechanisms for NodST and typical cytosolic sulfotransferases. Both structural information and site-directed mutagenesis studies have demonstrated that Lys, Thr, and Ser in the PAPS binding domain and His in the sulfate acceptor binding domain of several cytosolic sulfotransferases are highly conserved residues and are determining factors for sulfotransferase catalysis (Kakuta et al. 1997, 1998b, 1999; Bidwell et al. 1999; Ong et al. 1999; Perdersen et al. 2000). For example, the positively charged side-chain of the conserved Lys48 in the 5′PSB-loop of mEST is coordinated to both the bridging oxygen of the 5′-phosphate group of PAPS and an equatorial oxygen of the sulfate group in the proposed transition state (Kakuta et al. 1998b). This means that the conserved Lys48 residue plays a critical role in stabilizing the noncovalent complex between EST and PAPS/PAP in the sequential mechanism. In addition, Lys48 facilitates the sulfate group in-line transfer by acting as a general acid catalyst (Kakuta et al. 1997, 1998b). However, NodST has an Arg19 at the position corresponding to Lys48 in mEST (Ong et al. 1999). This Arg residue has two amine groups and can provide a pair of parallel dipoles in a fixed orientation for strong electrostatic interactions with two oxygens of the sulfate group. Hence, the Arg in NodST's 5′PSB-loop is more suitable for sulfate group stabilization than Lys48 in mEST, which increases the likelihood of the formation of a sulfated NodST intermediate and could be a possible reason for the related hybrid two-site ping-pong mechanism. Actually, it has been reported that the K48R mutant of mEST exhibited a five-fold lower kcat value compared to the wild-type enzyme (Kakuta et al. 1998b), making this mutant a much less efficient enzyme for sulfate group transfer via a sequential mechanism. In other words, Arg19 could be a key amino acid residue for NodST's hybrid ping-pong mechanism. In addition, the Ser138 residue within the 3′-phosphate binding (3′PB) motif of mEST (Fig. 6) is critical for the catalysis and is conserved in all cytosolic sulfotransferases (Perdersen et al. 2002). The amino acid sequence in NodST's 3′PB motif was compared with that of the cytosolic sulfotransferases (Fig. 6), but no significant similarities were found. The 3′PB motif of NodST lacks the Ser residue critical for sequential mechanism, which suggests that NodST may use another catalytic mechanism for sulfate group transfer.

Conclusions

An ESI-MS assay was developed and used to delineate the catalytic mechanism of NodST. This assay, which is particularly useful in characterizing enzymes for which no spectrophotometric assay is feasible, is demonstrated to be a valid and efficient method for enzyme mechanism investigation. In the present study, a hybrid ping-pong rapid equilibrium random two-site mechanism is suggested for the GlcNAc-6-O-carbodydrate sulfotransferase, NodST, using initial velocity kinetic analysis and product inhibition study. In addition to very compelling kinetic data, a sulfated NodST intermediate was also identified by FT-ICR mass spectrometry. The amino acid sequence in the catalytic active sites of NodST and cytosolic sulfotransferases were closely compared, and the lack of Lys and Ser residues in NodST's PAPS binding domain is suggested as a possible reason for the observed hybrid ping-pong catalytic mechanism. We are currently sequencing the T2–3 tryptic peptide and investigating its properties by isotope labeling study to identify the exact modified residue in the sulfated NodST intermediate.

Materials and methods

General materials and methods

All chemical reagents were obtained from commercial suppliers and used without further purification. Chitobiose was purchased from Calbiochem. Coomassie plus protein assay reagent (950 mL) and albumin standard (10 × 1 mL) were purchased from Pierce. The Ni-NTA agarose (100 mL) was purchased from QIAGEN. All of the other substrates and internal standard were purchased from Sigma, including 3′-phosphoadenosine 5′-phosphosulfate (PAPS), 3′-phosphoadenosine 5′-phosphate (PAP), and α-ΔUA-[1→3]-GalNAc-6S (ΔDi-6S). E. coli was purchased from NEB. Agarose gel electrophoresis was performed using standard procedures. Sequencing-grade trypsin was purchased from Promega. All of the mass spectrometric kinetics assays were performed at 22°C in 10 mM NH4OAc (pH 8.0) (Buffer A).

Instrumentation

Ion trap mass spectrometry

A Finnigan LCQ ion trap mass spectrometer equipped with an ESI source, and an HPLC pump (Thermo-Finnigan) was used. The capillary temperature and the spray voltage were kept at 200°C and 3.2 kV, respectively. Approximately 20 μL of each sample solution was injected into a 5-μL injection loop and delivered via an LC pump at a flow rate of 20 μL/min. The product ion (m/z 503) and the internal standard ion (m/z 458) were monitored in the negative ion mode using selected ion monitoring (SIM). The signals for the ions of interest were optimized by using the automatic tuning option on the instrument. The optimized conditions were then applied in subsequent experiments. When the signal intensity for one sample decreased from approximately 5×105 detector counts per scan to 5×103 detector counts per scan, indicating the consumption of the sample, the next sample was injected. The chromatogram of the Qual Browser program was used to monitor the processing and ionization of the sample versus time, with each peak representing a different sample that was analyzed. An average of 17×3 = 51 scans were taken to obtain a spectrum list for each sample, which provided the absolute intensities for the monitored ions along with the relative abundance. The sums of the intensities within 0.8 mass units around the center of the product ion and internal standard ion were used to determine their intensity ratio (IP/IIS).

FT-ICR mass spectrometry

A Bruker FT-ICR mass spectrometer equipped with an actively shielded 7 Tesla superconducting magnet was used for analyzing tryptic digests of NodST and the NodST-PAPS complex. Solutions were infused into an Analytica electrospray source at a rate of 1 μL/min. The N2 nebulizing and drying gas pressures were maintained at 50 psi and 30 psi, respectively. All protein samples were analyzed using 20 mM NH4OAc (pH 7.5). The bias on the glass capillary was kept at 4600 V, and 102°C drying gas was used to assist the desolvation process. A throttle valve was installed at the nozzle-skimmer region, and the pressure was adjusted to ∼1×10−5 mbar. Ions were externally accumulated in a radio frequency-only hexapole for 1–2 sec before transfer into the ICR cell for mass analysis through a series of electrostatic ion optics. All samples were collected using gated trapping. Ions with high kinetic energy were cooled down and trapped by colliding with Ar pulsed into the cell to a pressure of 10−7 mbar. Eight ion injection loops were used per scan. A pump down time of 0.02 sec was used between each ion injection loop, and a further pump down of 1 sec was applied when all ion packages were injected. The trapping voltage was then lowered to ∼0.5 V, and a final stage 2 sec pump down was applied, allowing for high-resolution detection. Each spectrum was an average of 16–80 transients composed of 1024 k data points acquired.

cDNA cloning, expression, and purification of NodST

The enzyme was cloned and expressed essentially as described by Burkart et al. (2000) with the following exceptions. The nodH gene was amplified by PCR from R. meliloti genomic DNA using the primers: 5′-TCAGATCGTACATATGACCCATTCCACGCT GCC (forward) and 5′-TGGTATTCGATGGATCCGCGTCGTTA GCAAGCTCAAACAAC (reverse). The resulting PCR product was digested with NdeI and BamHI and ligated into similarly digested pET24B (Novagen). Expression of recombinant NodH was carried out using E. coli BL21 cells transformed with the pET24B-nodH vector. Cells were induced with 0.3 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at A600 = 0.5 and grown overnight at 18°C. The cell pellets were harvested by centrifugation and resuspended in lysis buffer (100 mM Tris/HCl, 0.3 M NaCl, 5 mM imidazole, 20 mM β-mercaptoethanol, and 10% glycerol [pH 8.0]). The resuspended cells were lysed by sonication before purification. Approximately 30 mL Ni(II)-agarose affinity column was equilibrated with lysis buffer. The cell lysate was applied to the top of the settled resin and washed with 3 columns of lysis buffer, 3 columns of wash buffer 1 (lysis buffer with 10 mM imidazole) and wash buffer 2 (lysis buffer with 25 mM imidazole). NodST was cleaved by 30 mL of elute buffer (lysis buffer with 250 mM imidazole). The fractions containing NodST determined by SDS-PAGE (Fig. 1) were dialyzed against 100 mM Tris (pH 8.0) containing 20 mM β-mercaptoethanol and 10% glycerol to remove imidazole and Na+ before use in the mass spectrometric study. The NodST concentration was determined using the Lowry assay. The expressed NodST was stored at −80°C after adding 50% glycerol. For the ESI-MS enzyme kinetics assay and mechanism study, NodST was further buffer-exchanged into 100 mM NH4OAc (pH 8.0) with an Ultrafree-4 Centrifugal Filter Unit from Millipore. Three 30-min washings were performed by centrifuging the filter at 5000g.

Enzyme kinetics

Methodology for product quantification

The single-point normalization factor R is determined from equation 2 below, through which the relative ion intensity ratio of the product and the internal standard (IP/IIS) is related to their concentration ratio (Ge et al. 2001; Pi et al. 2002; Gao and Leary 2003). R can be obtained by analyzing a mixture of the internal standard and the product of known concentration.

equation image((2))

In our kinetic study, a chondroitin disaccharide, α-ΔUA-[1→3]-GalNAc-6S (ΔDi-6S) (Fig. 2A) was chosen as an internal standard because of its similar molecular weight and chemical structure to the product, chitobiose-6-OSO3 (Fig. 2B). This ensures a linear response between the concentration and ion intensity ratios. The two ions monitored were [ΔDi-6S] and [chitobiose-6-OSO3], at m/z 458 and m/z 503, respectively. According to the previously described procedure (Pi et al. 2002), the average single-point normalization factor was determined to be 2.7 and only varied slightly (± 0.1) during the course of the study. This relatively minor standard deviation ensures the precision of product quantification using our ESI-MS assay.

For each sample analyzed, the product concentration can be calculated via equation 3 using the ESI-MS data (IP/IIS) and the normalization factor R determined above (Ge et al. 2001; Pi et al. 2002):

equation image((3))

Initial velocity kinetic analysis

All kinetic experiments were carried out at room temperature with gentle rocking. A stock solution of NodST (250 nM; NodST stock 1), five stock solutions of chitobiose (0.5, 0.75, 1, 2.5, 10 mM), and five stock solutions of PAPS (12.5, 25, 50, 100, 250 μM) were prepared in Buffer A. For each enzymatic reaction, a 10-μL pre-reaction solution was generated by mixing 5 μL PAPS and 5 μL chitobiose stock solutions at proper concentrations. Each reaction was initiated by addition of 40 μL NodST stock 1. During the initial reaction period, an aliquot of 10-μL reaction solution was quenched in 40 μL of MeOH with 6.25 μM ΔDi-6S, and the quenched sample was analyzed by ESI-MS. Using the internal standard, the amount of product in each sample was quantified through the single-point normalization factor. To determine the catalytic mechanism for NodST, the concentration of PAPS was varied from 1.25 to 25 μM, and the concentration of chitobiose ranged from 0.05 to 1 mM. Partial substrate inhibition by chitobiose is negligible at this concentration range. Velocities were determined under each of the 25 conditions defined by a 5 × 5 matrix of substrate concentrations. The kinetic mechanism of NodST was evaluated by plotting 1/V0 as a function of both 1/[PAPS] at different fixed chitobiose concentrations and 1/[chitobiose] at different fixed PAPS concentrations using the nonlinear regression analysis program SAS (version 1.2; SAS Institute).

Product inhibition study

A stock solution of NodST (50 nM; NodST stock 2) was prepared in Buffer A. Stock solutions of chitobiose (1–6 mM), PAPS (250 μM), and PAP (0–70 μM) were prepared in Buffer A. The reaction solution was prepared in 50 μL total volume. For each of the five PAP concentrations (0, 1, 3, 5, 7 μM), a series of five chitobiose concentrations ranging from 0.1 mM to 0.6 mM were used, while the PAPS concentration was kept constant at 25 μM. Hence, a total of 25 pre-reaction mixtures (25 μL each) were prepared by mixing 5 μL of 250 μM PAPS stock, 10 μL Buffer A, and 5 μL of chitobiose stock and PAP stock of appropriate concentrations. Each reaction was initiated by addition of 25 μL of NodST stock 2, and thus the NodST concentration in each reaction was 25 nM. A 20-μL aliquot of each reaction solution was quenched in 80 μL MeOH with 6.25 μM ΔDi-6S at an optimized reaction time. The 25 quenched samples were analyzed by ESI-MS, and the amount of product in each sample was quantified. The mode of inhibition was evaluated by analyzing the pattern of the double reciprocal plots using the GraFit program (version 4.0.12; Erithacus Software). At the same time, the Ki value of PAP in regards to the substrate chitobiose was obtained.

Analysis of the sulfated NodST intermediate

A solution of 33 μM NodST and 140 μM of PAPS (containing 25 μM PAP) was added to 100 μL of 20 mM NH4OAc buffer (pH 7.5), and the mixture was incubated on ice for 0.5 h. Then 0.4 μg of sequencing grade trypsin was added into the solution to perform limited digestion on NodST. The solution was then analyzed using FT-ICR MS and negative ion detection without further desalting or separation of the resultant tryptic peptides. The theoretical tryptic digestion pattern of NodST was generated using Masslynx 3.3 Build 004. In this experiment, the maximum missed cleavage sites allowed was four, and no cysteine was modified. The enzyme itself was also digested with trypsin, and the tryptic fragments were compared with those of the enzyme digested in the presence of PAPS.

Figure Figure 1..

SDS-PAGE analysis of Ni(II)-NTA column fractions. Lane (a) MW markers, (b) flow-through, (c,d) buffer wash without imidazole, (e) buffer wash with 10 mM imidazole, (f) buffer wash with 20 mM imidazole, (g) buffer elution with 200 mM imidazole (elution).

Figure Figure 2..

Structure of internal standard and monitored product. (A) Internal standard, α-ΔUA-[1→3]-GalNAc-6S (ΔDi-6S), with m/z = 458. (B) Monitored product, chitobiose-6-OSO3, with m/z = 503.

Figure Figure 3..

Double reciprocal plots for NodST-catalyzed sulfate group transfer from PAPS to chitobiose. (A) 1/V0 vs. 1/[PAPS] at different [chitobiose]. (Filled triangles) [chitobiose] = 0.05 mM, (open triangles) [chitobiose] = 0.075 mM, (filled circles) [chitobiose] = 0.1 mM, (empty circles) [chitobiose] = 0.25 mM, (squares) [chitobiose] = 1 mM. (B) 1/V0 vs. 1/[chitobiose] at different [PAPS]: (filled triangles) [PAPS] = 1.25 μM, (open triangles) [PAPS] = 2.5 μM, (filled circles) [PAPS] = 5 μM, (open circles) [PAPS] = 10 μM, (squares) [PAPS] = 25 μM.

Figure Figure 4..

Double reciprocal plots for inhibition of NodST by PAP at varied chitobiose concentrations. 1/V0 vs. 1/[chitobiose] at different [PAP]. [PAPS] was fixed at 25 μM. (Open circles) [PAP] = 0 μM, (filled circles) [PAP] = 1 μM, (open squares) [PAP] = 3 μM, (filled squares) [PAP] = 5 μM, (triangles) [PAP] = 7 μM.

Figure Figure 5..

Identification of sulfated NodST intermediate. (A) ESI-MS spectra of 33 μM NodST with 165 μM PAPS after trypsin digestion. (B) ESI-MS spectra of 33 μM NodST after trypsin digestion. (The mass range around m/z 1391 has been expanded to show isotopic resolution and magnified by 13-fold.)

Figure Figure 6..

Sequence alignment of NodST with cytosolic sulfotransferases.

Scheme Scheme 1..

NodST catalyzes the sulfation of chitobiose (1), generating PAP and chitobiose-6-OSO3(2) as products.

Scheme Scheme 2..

(A) Classical ping-pong mechanism for NodST. (B) Proposed hybrid ping-pong rapid equilibrium random two-site mechanism for NodST.

Acknowledgements

We thank the NIH (no. GM63521) for financial support, and Prof. Carolyn R. Bertozzi, Prof. Jack F. Kirsch, and Dr. David J. Vocadlo for many helpful discussions.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.

Ancillary

Advertisement