![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Drug Metabolism and Pharmacokinetics Research Laboratories (M.K., T.T.) and Advanced Technology Research Laboratories (H.H., Y.I.), Daiichi-Sankyo Co., Ltd., Tokyo, Japan; and Research Laboratories, Kyoto Pharmaceutical Industries, Ltd., Kyoto, Japan (K.T.)
(Received March 5, 2008; Accepted June 3, 2008)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
-1 position of the N-octyl indoline/indole group, were relatively distant from the aromatic moiety. A pactimibe analog with an N-decyl chain was similarly labile against CYP2D6; however, analogs with N-hexyl or N-dodecyl chains were stable to CYP2D6. An induced fit docking of the ligands with an X-ray crystal structure of substrate-free CYP2D6 (Protein Data Bank code 2F9Q) indicated the involvement of an electrostatic interaction between the carboxyl group and Arg221 and hydrophobic interaction between the aromatic moiety and Phe483. The docking model correctly positioned the site of metabolism above the heme. The effect of the N-alkyl chain length of pactimibe analogs on their CYP2D6 metabolic stability was plausibly explained by the docking model. In conclusion, we report herein a novel CYP2D6 binding mode for the acidic substrates pactimibe and R-125528. Further investigation, such as a site-directed mutation, will be necessary to directly demonstrate the involvement of Arg221 in CYP2D6 binding.
). CYP2D6 is one of the best known P450 isoforms and leads to large interindividual variations in drug concentration, drug response, therapeutic outcome, and/or toxicity because of its polymorphic nature (Brynne et al., 1998; Molden et al., 2002; Fux et al., 2005). Although CYP2D6 is a minor component (
2%) of the total P450 content in the human liver, many drugs currently in clinical use are metabolized by CYP2D6, such as antiarrhythmics, antidepressants, antipsychotics, β-blockers, and analgesics (Jones et al., 1997). The variability in CYP2D6 activity is mainly determined genetically, and 5 to 10% of Caucasians are reported to express a poor CYP2D6-metabolizing phenotype, resulting from the inheritance of two mutant null alleles (van der Weide and Steijns, 1999). Reducing clinical failures derived from polymorphic interindividual variations in drug response is very important. Therefore, several efforts have been directed to predicting the binding affinity against CYP2D6 (Strobl et al., 1993; de Groot et al., 1999a,b; Ekins et al., 2003).
CYP2D6 substrates generally exhibit basic nitrogens that have been predicted to interact with one of two acidic amino acid side chains, Asp301 or Glu216, of the enzyme. In early models of CYP2D6, a characteristic common to the vast majority of CYP2D6 substrates is the presence of at least one basic nitrogen atom. Pharmacophore models suggest that the site of oxidation is located approximately 5 to 7 Å away from the basic nitrogen (Koymans et al., 1992). Ellis et al. (1995) demonstrated that Asp301 plays an important role in determining the substrate specificity and activity of CYP2D6 and provided experimental evidence for electrostatic interaction between the basic nitrogen in CYP2D6 substrates and the carboxylate group of Asp301. Guengerich et al. (2003) and Paine et al. (2003) suggested the importance of Glu216 in addition to Asp301 as a key determinant factor for the substrate specificity and product regioselectivity in CYP2D6.
Another chemical structural property of typical CYP2D6 substrate is a flat aromatic moiety near the site of oxidation. Two aromatic phenylalanines, Phe120 and Phe483, and several hydrophobic residues such as Leu213 and Val308 (Keizers et al., 2004; Flanagan et al., 2004; de Graaf et al., 2007) have been proposed as active site residues. The hydrophobic pocket defined by these hydrophobic residues is thought to interact with the hydrophobic regions in the substrate.
However, Guengerich et al. (2002) identified a CYP2D6 substrate, spirosulfonamide, devoid of basic nitrogen, raising a question about the current CYP2D6 model based on a critical electrostatic interaction. Because of the absence of a basic nitrogen atom in its structure, no hydrogen bonds are formed to the negatively charged residues, Asp301 and/or Glu216. Thereafter, Kemp et al. (2004) docked spirosulfonamide positioned between the two phenylalanine residues, Phe120 and Phe483, in a CYP2D6 homology model.
|
). In human clinical drug-drug interaction study, the AUC0–72 h of pactimibe and R-125528 were increased 1.7- and 5.0-fold, respectively, by cotreatment with quinidine (Kotsuma et al., 2008a). Interestingly, their sites of metabolism are relatively far from the aromatic moiety (the -1 position of the N-octyl chain in Fig. 1B), where oxidation often occurs in typical CYP2D6 substrates. Pactimibe sulfate (formerly named CS-505) (Fig. 1A), which was discovered by Kyoto Pharmaceutical Industries, Ltd. (Kyoto, Japan) is a novel water-soluble antioxidative acyl coenzyme A:cholesterol acyltransferase inhibitor for the treatment of hypercholesterolemia and atherosclerotic diseases (Kamiya et al., 1997, 2000; Kitayama et al., 2006a,b,c; Nissen et al., 2006).
We describe herein the physicochemical properties and CYP2D6 affinity of the novel acidic substrates pactimibe and R-125528. In addition, the CYP2D6 metabolic stability of pactimibe analogs was examined. Very recently, the X-ray crystal structure of substrate-free CYP2D6 was resolved at a resolution of 3.0 Å (pdb code 2F9Q) (Rowland et al., 2006). To obtain the possible CYP2D6 binding mode, pactimibe, R-125528, and their structural analogs were docked to substrate-free CYP2D6.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-1 oxidized form of R-125528) was synthesized at Chemtec Labo., Inc. (Tokyo, Japan). The chemical structure of pactimibe sulfate is shown in Fig. 1A. KV-2908, KV-2909, and KV-2915 were synthesized at Kyoto Pharmaceutical Industries, Ltd. Propafenone hydrochloride, desipramine hydrochloride, propranolol hydrochloride, amitriptyline hydrochloride, imipramine hydrochloride, dextromethorphan hydrobromide monohydrate, NADP, glucose 6-phosphate, and glucose-6-phosphate dehydrogenase were purchased from Sigma-Aldrich (St. Louis, MO). MgCl2 · 6H2O was purchased from Wako Pure Chemicals (Osaka, Japan). Acetonitrile of high-performance liquid chromatography grade and 1 N HCl of volumetric analysis grade were purchased from Wako Pure Chemical. Other reagents and solvents were of analytical grade and were used without further purification. Microsomes Expressing Human P450 Isoforms. CYP2D6*1 + P450 reductase Supersomes and Insect Cell Control Supersomes prepared from baculovirus-infected insect cells were purchased from BD Gentest (Woburn, MA). The control microsome was used as a control for the native activities. Microsomes were stored at –80°C until use.
Determination of Km Values for CYP2D6. All microsomal incubations were carried out at 37°C using 100 mM potassium phosphate buffer (pH 7.4) and an NADPH-generating system containing final concentrations of 1.25 mM NADP, 1 mM MgCl2 · 6H2O, 12.5 mM glucose 6-phosphate, and 0.5 U/ml glucose-6-phosphate dehydrogenase.
Because of the absence of authentic samples for the metabolites of imipramine, desipramine, amitriptyline, propranolol, dextromethorphan, and propafenone, the substrate depletion method (Obach and Reed-Hagen, 2002) was used for determination of the Michaelis constant (Km) values for typical CYP2D6 substrates. The final CYP2D6 protein concentration and substrate concentrations were set at 10 pmol/ml and 0.1, 0.3, 1, 3, 10, and 30 µM, respectively. Thirty microliters of incubation mixture was added to 90 µl of ice-cold methanol containing 0.5 µM of d6-pactimibe as an internal standard at 0, 3, 6, and 9 min to stop the reaction. The mixture was centrifuged at 2,000 rpm for 10 min at 4°C (himac CF7D2, rotor no. RT3S3-A808; Hitachi Koki Co., Ltd., Tokyo, Japan). Ten microliters of the supernatant fraction was injected into the liquid chromatography/mass spectrometry (LC/MS) system for analysis.
For the determination of the Km value of R-125528, the formation of M-2 was determined. The final CYP2D6 protein concentration and substrate concentrations were set at 10 pmol/ml and 0.3, 1, 3, 10, 30 and 100 µM, respectively. Under the conditions used, the formation of M-2 was linear with respect to the incubation time and protein concentrations. To stop the reaction, 60 µl of incubation mixture was added to 120 µl of ice-cold solution A (acetonitrile-1 N hydrochloride-2 mM dithiothreitol, 98:1:1, v/v/v) containing 0.5 µM d6-pactimibe as an internal standard and 60 µl of solution A at 7.5 and 15 min. The mixture was centrifuged at 2000 rpm for 10 min at 4°C (himac CF7D2, rotor no. RT3S3-A808). Ten microliters of the supernatant fraction was injected into the LC/MS system for analysis.
Metabolic Stability of Pactimibe Analogs in CYP2D6-Expressing Microsomes. Microsomal incubations were carried out at 37°C using 100 mM potassium phosphate buffer (pH 7.4) and an NADPH-generating system containing final concentrations of 2.5 mM NADP, 10 mM MgCl2 · 6H2O, 25 mM glucose 6-phosphate, and 1 U/ml glucose-6-phosphate dehydrogenase. The final CYP2D6 protein concentration was set at 40 pmol/ml. The final concentrations of pactimibe, R-125528, KV-2908, KV-2909, and KV-2915 were set at 1 µM. To stop the reaction, 100 µl of incubation mixture was added to 200 µl of ice-cold solution A containing 0.5 µM d6-pactimibe as an internal standard at 0, 10, 20 and 30 min. The mixture was centrifuged at 2000 rpm for 10 min at 4°C (himac CF7D2, rotor no. RT3S3-A808). Ten microliters of the supernatant fraction was injected into the LC/MS system for analysis.
Determination of pKa. The pKa values for pactimibe and R-125528 were determined using a Sirius GLpKa meter.
LC/MS Analysis. An LC/MS method was applied for determination of the target analyte. A Waters Alliance 2795 high-performance liquid chromatography system (Waters, Milford, MA) coupled to a Quattro liquid chromatograph (Micromass UK Ltd., Manchester, UK) was used for the detection. The following ions (m/z) were monitored: 267 for desipramine, 278 for amitriptyline, 342 for propafenone, 260 for propranolol, 272 for dextromethorphan, 281 for imipramine, 417.1 for pactimibe, 415.1 for R-125528, 389 for KV-2908, 445 for KV-2909, 473 for KV-2915, and 431.4 for M-2.
Docking Methodology. The molecular docking of bufuralol, pactimibe, R-125528, KV-2908, KV-2909, and KV-2915 to CYP2D6 was performed using the crystal structure of substrate free CYP2D6 (pdb code 2F9Q) (Rowland et al., 2006). The substrates were prepared using the LigPrep module of Maestro (Schrödinger, LLC, New York, NY). Starting with two-dimensional structures, LigPrep produces a three-dimensional structure with ionization states at pH 7. In the case of pactimibe analogs, the carboxyl group carried a formal change of –1. Induced fit docking (IFD) was carried out using the Prime and Glide module (Sherman et al., 2006) with the default parameter set. In the IFD protocol, the side-chain degrees of freedom in the protein are sampled while allowing small backbone conformational changes through minimization. Docking is carried out using rotamer libraries to sample energetically reasonable side chains and to eliminate conformations with steric crashes. The docked protein-substrate complexes are ranked according to the IFD score. The IFD score is a composite score made up of the protein/ligand interaction energy (GlideScore) and the protein molecular mechanics energy (Prime energy) in an implicit solvation model (Sherman et al., 2006).
Data Analysis. In the substrate depletion experiments (Obach and Reed-Hagen, 2002), the analyte/internal standard peak area ratios were determined and normalized to the value obtained at t = 0. The percentage remaining versus the time at each substrate concentration was fitted to first-order decay functions to determine the initial substrate depletion rate constants (kdep). If the substrate decline demonstrated nonlinearity on the log percentage remaining versus time curves, only those initial time points at which log linearity was observed were used to determine the depletion rate constants. The Km values from the substrate consumption experiments were determined by plotting the kdep versus the substrate concentration on a linear-log plot using the following equation:
 |
The Km value of R-125528 was calculated by fitting the data to the following equation:
 |
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
). The Km values of typical CYP2D6 substrates were determined by the substrate depletion method or were derived from articles. The molecular weight and Km values of the atypical CYP2D6 substrates spirosulfonamide, pactimibe, and R-125528 are compared with those of typical CYP2D6 substrates.
|
The pKa values of pactimibe and R-125528 were quite low among those of typical CYP2D6 substrates. As shown in Table 1, the pKa values of typical CYP2D6 substrates are generally higher than 8. On the other hand, pKa1 (for nitrogen in the indoline ring) and pKa2 (for the carboxyl group) of pactimibe were 3.39 and 4.34, respectively. The nitrogen in the indole ring of R-125528 was not protonated over the entire pH range and for the carboxyl group only pKa was determined as 4.45, suggesting that they are not positively charged in a physiological condition. Spirosulfonamide has also been reported as a nonpositively charged CYP2D6 substrate at neutral pH (Guengerich et al., 2002).
The molecular weights of pactimibe and R-125528 were relatively high compared with those of typical CYP2D6 substrates. Those of almost typical CYP2D6 substrates ranged from 200 to 300. On the other hand, those of pactimibe and R-125528 were 416.6 and 414.6, respectively.
Interestingly, however, pactimibe and R-125528 showed relatively high affinity for CYP2D6. The Km values of pactimibe and R-125528 for CYP2D6 were determined to be 25.1 and 1.74 µM, respectively. These values were comparable with those of typical CYP2D6 substrates. The Km values of bufuralol, debrisoquine, imipramine, desipramine, nortriptyline, amitriptyline, propranolol, dextromethorphan, metoprolol, and propafenone, were 3.4 (Emoto et al., 2006), 73.7 (Shen et al., 2007), 2.68, 1.26, 2.08 (Venkatakrishnan et al., 1999), 1.61, 1.48, 0.58, 21.6 (Uttamsingh et al., 2005), and 0.01 µM, respectively.
CYP2D6 Metabolic Stability of Pactimibe Analogs. We examined the effect of the N-alkyl chain length of pactimibe analogs on their metabolic stability against CYP2D6. The chemical structures of pactimibe analogs and their metabolic stability in CYP2D6-expressing microsomes are shown in Table 2. After a 30-min incubation in CYP2D6-expressing microsomes, the metabolic stabilities (percentage of 0 min) of pactimibe (C8), R-125528 (C8), KV-2908 (C6), KV-2909 (C10), and KV-2915 (C12) were determined to be 27.5, 0.8, 97.0, 38.1, and 98.4, respectively. Although substitution of the N-octyl chain to the N-decyl chain did not change the stability against CYP2D6, the N-hexyl or N-dodecyl chain improved the CYP2D6 metabolic stability.
|
CYP2D6 Docking of Pactimibe Analogs. At first, we docked a typical CYP2D6 substrate, bufuralol, into the crystal structure of substrate-free CYP2D6 (pdb code 2F9Q) (Rowland et al., 2006) using the induced fit docking mode of the Prime and Glide module for a feasibility assessment. Electrostatic interaction between basic nitrogen and Glu216 was observed (Fig. 2A). In addition, the aromatic moiety of bufurarol was located in the hydrophobic region defined by hydrophobic residues such as Phe120, Phe483, Leu213, and Val308 etc. The docking model correctly identified the reported site of the metabolism of bufurarol (de Groot et al., 1999a; de Graaf et al., 2007).
|
The docking model of R-125528 is shown in Fig. 2B. The docking positioned the indole ring of R-125528 close to Phe483 by hydrophobic interaction. In addition, the docking indicated electrostatic interaction between the carboxyl group of R-125528 and the basic residues of Arg221. Interestingly, in contrast to bufurarol docking, the aromatic moiety was not located close to the heme. Instead, an N-octyl chain was inserted into the active site, exhibiting interactions between hydrophobic residues, such as Phe120, Leu213, and Leu484. Hence, the docking correctly positioned the site of the metabolism of R-125528 at the -1 position of the N-octyl chain (Kotsuma et al., 2008b), above the heme. The distance between the -1 position of the N-octyl chain and the heme was 4.4 Å.
The docking models of pactimibe (C8), KV-2908 (C6), KV-2909 (C10), and KV-2915 (C12) are shown in Fig. 3. When pactimibe and KV-2909 (C10) were docked to CYP2D6, hydrophobic interaction between Phe483 and the indoline ring and electrostatic interaction between the carboxyl group and the basic residues of Arg221 were indicated (Fig. 3, A and C). The N-alkyl chain was inserted into the hydrophobic region defined by hydrophobic residues such as Leu213, Met374, and Leu484. The -1 position of the N-alkyl chain was located above the heme, and the distances between the -1 position of N-octyl chain and N-decyl chain and the heme were 5.0 and 5.3 Å, respectively. On the other hand, when KV-2908 (C6) and KV-2915 (C12) were docked, a different binding mode was suggested. The distance between the -1 position of the N-hexyl chain and the heme was 4.3Å. However, no hydrophobic interaction between the indoline ring of KV-2908 and Phe483 was observed (Fig. 3B). Regarding the docking model of KV-2915, the direction of the N-dodecyl chain was not toward the heme but toward the entrance of the enzyme (Fig. 3D).
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Pactimibe and R-125528 are novel acidic ligands for CYP2D6 in contrast to the vast majority of typical CYP2D6 substrates, which contain basic nitrogen in their structures. It has been suggested that an electrostatic interaction between a protonated basic nitrogen and a carboxyl group of Glu216 and/or Asp301 plays an important role in determining substrate specificity. As shown in Table 1, the pKa values of typical CYP2D6 substrates are generally more than 8, suggesting that the basic nitrogens are protonated at physiologically equivalent pH. However, the pKa value for a nitrogen in the indoline ring of pactimibe is 3.39. Moreover, nitrogen in the indole ring of R-125528 is not protonated over the entire pH range. If we consider that pactimibe and R-125528 possess a carboxylic acid in their structures, they are negatively charged in a physiological condition.
As far as we know, pactimibe and R-125528 are the first identified acidic ligands for CYP2D6 with a carboxyl group in their structures, even though there were several reports of nonbasic ligands for CYP2D6. Guengerich et al. (2002) identified a high-affinity ligand, spirosulfonamide, devoid of basic nitrogen. It has also been reported that peptide-mimetic compounds having no basic nitrogen (Kumar et al., 1996) and some steroids (Niwa et al., 1998) are CYP2D6 substrates. These findings suggest that the presence of basic nitrogen is not a common requirement for high-affinity CYP2D6 substrates.
To obtain a possible docking mode of acidic ligands against CYP2D6, we conducted an induced fit docking of the ligands with an X-ray crystal structure of substrate-free CYP2D6 (pdb code 2F9Q) recently reported by Rowland et al. (2006). At first, we tested a typical CYP2D6 substrate, bufuralol (Fig. 2A), for a feasibility assessment. The docking mode of bufuralol showed a typical CYP2D6 binding mode, indicating electrostatic interaction between a basic nitrogen of bufuralol and Glu216. The aromatic ring and its accompanying short alkyl chain, where the metabolism reaction occurs (de Graaf et al., 2007), were correctly located above the heme surrounded by a hydrophobic region defined by Phe120, Phe483, Leu213, and Val308. This observation coincided with the general finding that aromatic moieties of CYP2D6 substrates are located between Phe120 and several hydrophobic residues such as Leu213 and Val308 (de Graaf et al., 2007) and are further metabolized.
In contrast, pactimibe and R-125528 showed a novel binding mode (Figs. 2B and 3A). The aromatic indoline/indole ring was sequestered from the heme by electrostatic and hydrophobic interactions, respectively, with Arg221 and Phe483, which are located in the access channel. The sites of metabolism (-1 position of the N-octyl chain), which are relatively distant from the aromatic ring, were located close to the heme. Curiously, the basic residue of Arg221 was proposed to form a salt bridge with the carboxyl group of pactimibe/R-125528. This result is interesting because it has been generally suggested that electrostatic interaction between the basic nitrogen of a substrate and Glu216/Asp301 is important for the binding of CYP2D6 substrates (Guengerich et al., 2003; Paine et al., 2003). Further studies, such as a site-directed mutation, will be necessary to investigate the involvement of Arg221 in CYP2D6 binding.
Phe483 seems not only to interact with the indoline/indole ring by hydrophobic interaction but also to sterically hinder the access of the main body to the active site. This might be one of the reasons that pactimibe and R-125528 are good ligands for CYP2D6 despite their relatively high molecular weight (>400) compared with the other CYP2D6 substrates (Table 1). In our docking model, although the N-octyl chain was inserted into the active site, the bulky t-butyl and acetic acid moieties were positioned at the entrance of the enzyme. In fact, Smith et al. (1998) pointed out that the bulky phenyl ring of Phe483 is positioned across the channel mouth, thus limiting the size of the substrate that can access the active site.
This binding model provided a plausible explanation for the difference in the CYP2D6 metabolic stability of pactimibe analogs with N-alkyl chains of different lengths. As shown in Table 2, the metabolic stability of pactimibe (C8) against CYP2D6 was low and was followed by KV-2909 (C10). On the other hand, KV-2908 (C6) was stable against CYP2D6. One possible reason is that the site of metabolism is far from the heme. Actually, by induced fit docking (Fig. 3B), the -1 position of the N-hexyl chain was close to the heme (4.3 Å). However, in this docking model, the orientation of the indoline ring was different than that of pactimibe. Then, we docked KV-2908 (C6) using a constraint so that the indoline ring and carboxyl moiety of KV-2908 would be positioned where those of pactimibe are located. As a result, a binding mode similar to that of pactimibe was obtained (data not shown). The distance between the -1 position of the N-hexyl chain and the heme was 5.6 Å, which is larger than that between the -1 position of the N-octyl chain and the heme (5.0 Å).
KV-2915 (C12) was also metabolically stable against CYP2D6, probably because the site of metabolism could not be located in the active site. This docking model suggested that the direction of the N-dodecyl chain was not toward the heme but toward the entrance of the enzyme (Fig. 3D). The N-dodecyl chain seems to be too bulky to fit itself into the CYP2D6 active site. Even when we docked KV-2915 to the crystal structure, in which the residue at position 374 is valine (the actual residue in wild-type CYP2D6) instead of methionine (which has a larger residue than valine) as used in this study, the long N-dodecyl chain was not accommodated in the active site (data not shown). This observation is in accordance with the result of an in vitro metabolic study using wild-type CYP2D6 expression microsomes.
The CYP2D6 affinity of R-125528 (1.74 µM) was higher than that of pactimibe (25.1 µM). This might be due to the difference in lipophilicity between pactimibe and R-125528. The oxidative conversion from the indoline ring to the indole ring changes pactimibe (logP 4.68) into a more lipophilic metabolite, R-125528 (logP 5.83). The hydrophobic interaction between Phe483 and the indole ring is considered to be more significant than that between Phe483 and the indoline ring, even though they showed similar binding modes.
To conclude, we report herein a novel CYP2D6 binding mode for atypical acidic substrates, pactimibe and its metabolite, R-125528. Although they are negatively charged in a physiological condition, their affinities for CYP2D6 were comparable with those of typical basic CYP2D6 substrates. The induced fit docking to CYP2D6 (pdb code 2F9Q) indicated the involvement of an electrostatic interaction between the carboxyl group and Arg221 and a hydrophobic interaction between the aromatic moiety and Phe483. This docking model correctly positioned the site of metabolism (the -1 position of the N-octyl chain) above the heme. The effect of the N-alkyl chain length of pactimibe analogs on their CYP2D6 metabolic stability was plausibly explained by this molecular docking model. Further studies, such as a site-directed mutation, will be necessary to investigate the involvement of Arg221 in CYP2D6 binding.
|
Footnotes |
|---|
ABBREVIATIONS: P450, cytochrome P450; pdb, Protein Data Bank; M-2, -1 oxidized form of R-125528; LC/MS, liquid chromatography/mass spectrometry; IFD, induced fit docking.
Address correspondence to: Dr. Masakatsu Kotsuma, Daiichi Sankyo Co., LTD, 1-2-58, Hiromachi, Shinagawa-ku, Tokyo 140-8710, Japan. E-mail: kotsuma.masakatsu.gu{at}daiichisankyo.co.jp
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
) together with the protonated nitrogen atom (+).
