Phosphodiesterase type 5 (PDE-5) inhibitors, which are clinically indicated for the treatment of erectile dysfunction (ED), are widely available on the illegal market and as undeclared adulterants.1-5 Some products have been found to be adulterated with approved PDE-5 inhibitors as well as their unapproved synthetic analogues, which contain minor structural modifications compared to the approved compounds. The presence of PDE-5 inhibitors or their analogues in herbal supplements could pose a significant risk to the public health.6 PDE-5 inhibitors can cause several adverse effects such as headaches, facial flushing, nasal congestion, dyspepsia, visual disorders, and back pain.7-9 Furthermore, patients taking nitrate medications should not use PDE-5 inhibitors, as this combination may provoke potentially life-threatening hypotension.10 However, the widespread use and popularity of PDE-5 inhibitors have led to an increase in the prevalence of illicit sexual performance enhancement products in many countries.11-15 The presence of various analogues of PDE-5 inhibitors in dietary supplements has increased tenfold over the past decade, potentially posing a serious health risk to humans.
A number of analytical techniques have been developed for the detection and determination of PDE-5 inhibitors, such as immunoassays,16 atomic emission spectrometry,17 ion mobility spectrometry,18,19 micro-Raman spectroscopy,20 high-performance liquid chromatography (HPLC) with ultraviolet or fluorescence detection,21-23 gas chromatography–mass spectrometry (GC-MS),24,25 and liquid chromatography-mass spectrometry (LC-MS).26-31 Liquid chromatography-quadrupole-time of flight mass spectrometry (LC-QTOF-MS) carried out in full scan mode can be used to elucidate the accurate mass of target and nontarget compounds, as well as that of unknown compounds, while the multiple reaction monitoring (MRM) mode can be used only to detect the mass of target compounds. Fragmentation patterns can be used to determine the structure of unknown components by interpreting the product ions. Accurate mass information is important for confirming the identities of product ions and it plays a key role in the elucidation of fragments.
In the present study, a convenient and economic method was established to detect the analogues of PDE-5 inhibitors. Fast and accurate analysis of these analogues was performed by evaluating the MS spectra and fragmentation pathway of 58 PDE-5 inhibitor analogues using LC-QTOF-MS. Explaining the fragmentation pathway would be very useful for monitoring and identifying the parents and their metabolites. Furthermore, to the best of our knowledge, no detailed fragmentation pathway research on cyclopentyltadalafil, homotadalafil, acetaminotadalafil, propoxyphenylthioaildenafil, propoxyphenylthiohomosildenafil, propoxyphenylthiohydroxyhomosildenafil and propoxyphenylthiosildenafil has been reported to date. The structural similarities and differences among the analogues are also presented.
2.1. Reagents and standards
Dapoxetine and icariin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Acetaminotadalafil, aminotadalafil, chloropretadalafil, demethyltadalafil, n-butyltadalafil, n-octylnortadalafil, tadalafil, tadalafil E, acetyl acid, benzylsildenafil, carbodenafil, chlorodenafil, cinnamyldenafil, cyclopentynafil, dimethylsildenafil, dimethylacetildenafil, gendenafil, homosildenafil, hongdenafil, hydroxychlorodenafil, hydroxyhomosildenafil, hydroxythiohomosildenafil, mirodenafil, nitrodenafil, norneosildenafil, oxohongdenafil, thiohomosildenafil, thiosildenafil, udenafil, acetylvardenafil, hydroxyvardenafil, imidazosagatriazinone, norneovardenafil, pseudovardenafil, vardenafil, avanafil, desmethylcarbodenafil, and lodenafilcarbonate were purchased from TLC PharmaChem (Vaughan, On, Canada). Dimethylthiosildenafil, sildenafil, thioquinapiperifil, desulfovardenafil, yohimbin, xanthoanthrafil, dichlorodenafil, hydroxyhongdenafil, piperidinohongdenafil, dimethylhongdenafil, epi-aminotadalafil, methylhydroxyhomosildenafil (homothiomethylsildenafil), homotadalafil, propoxyphenylthiohomosildenafil, propoxyphenylthiohydroxyhomosildenafil, propoxyphenylthiosildenafil, propoxyphenylthioaildenafil, and cyclopentyltadalafil were obtained from the Korea Ministry of Food and Drug Safety. To prepare the stock solutions, all standards were dissolved in methanol from Sigma-Aldrich (MeOH, HPLC grade) at a concentration of about 1 mg/mL. The stock solutions were stored in a refrigerator (4 °C) and diluted before analysis. Acetonitrile (ACN, HPLC grade), methanol (HPLC grade), and formic acid (analytical reagent grade) were supplied by Merck (Darmstadt, Germany), and Sigma-Aldrich. The water was purified with a Milli-Q system (18.2 mΩ) by Millipore (Billerica, MA, USA).
2.2. LC-QTOF-MS analysis
The mass of the PDE-5 inhibitor were accurately measured using HPLC combined with Agilent 6530 Accurate-Mass Quadrupole (QTOF-MS) equipped with ESI Jet Stream Technology (Agilent Technologies, Waldbronn, Germany). An Agilent XDB C18 column (150×2.1 mm2 I.D., 3.5 μm) kept at 35°C in an oven was used, and the mobile phases consisted of DW (v/v, A) and ACN (v/v, B) containing 0.1% formic acid. The gradient elution profile was as follows: 0-3 min (A:B = 80:20%), 3-13 min (A; 80%-40%, B; 20%-60%), 13-16 min (A:B = 40:60%), 16-18 min (A; 40%-0%, B; 60%-100%), 18-21 min (A:B = 0:100%), 21-22 min (A; 60%-80%, B; 100%-20%), and 22-25 min (A:B = 80:20%). The flow rate of the mobile phase was 0.25 mL min–1, and the injection volume was 3 μL. A Jet Stream ESI source was operated in the positive ionization mode; other acquisition conditions of QTOF-MS were as follows: gas temperature, 350°C; drying gas flow, 8 L min–1; nebulizer, 35 psi; sheath gas temperature, 350°C; and sheath gas flow, 11 L min–1. The scan source parameters were as follows: capillary voltage, 3500 V; nozzle voltage, 1000 V; and fragment voltage, 120 V. The collision energy was set at 15 V, and m/z 121.0506 and 922.0098 were selected as the reference masses to tune the QTOF-MS. The mass resolution was 10000-20000 at m/z 100-1000. The MS/MS spectra were generated in product ion scan mode at CID energies of 5, 10, 15, 20, 25, 30, 35, and 40 eV. Data acquisition and processing were conducted using Mass Hunter Workstation Software (Ver. B.02.01, Agilent Technologies).
2.3. Analysis of possible fragmentation pathway
The possible fragmentation pathways were reviewed and selected in accordance with the following criteria: (1) the tendency for the radical site to initiate a reaction in competition with the charge site generally parallels the radical site’s tendency to donate electrons: N > S, O, π, R· > Cl, Br > H, where π signifies an unsaturated site and R· is an alkyl radical; (2) in n-alkanes, the most easily cleaved bonds (σ-bonds between secondary carbons) have nearly equivalent bond strength; and (3) a number of the most significant differences arise owing to the low ionization energies of sulfur compounds, which are approximately 1 eV below those of the corresponding oxygen compounds.32
3. Results and discussion
Fifty-eight PDE-5 inhibitors and their analogues were analyzed by QTOF–MS; the analogues were classified into five groups: tadalafil, sildenafil, hongdenafil, vardenafil, and other analogues. The results are summarized in Table 1. The fragmentation pathways of the five types of PDE-5 inhibitors were studied using a combination of QTOF-MS. The protonated [M+H]+ molecule ions were selected as the precursor ions for the fragments in the MS/MS product ion spectra. To aid in the interpretation of the product ion spectra, accurate mass measurements of the product ions were conducted using LC-QTOF-MS. And the accurate mass measurement error for between the experimental and the theoretical mass of [M+H]+ ion was lee than 10 μg/kg. Identification of the compounds was based on retention time, protonated molecular ions, and fragment ions at individually selected collision energies (CEs). The CEs were set to 5, 10, 15, 20, 25, 30, 35, and 40 eV, and the optimal CE was determined for each PDE-5 inhibitor.
Table 1.RT*: Retention time CE**: Collision energy
For the MS/MS studies, the mass spectrometric parameters were optimized in order to obtain the best mass fragmentations. Accurate mass values were observed for the product ions in MS/MS, and the element compositions were derived from the measured m/z values and the general concept that fragmentation should involve logical neutral losses. Additionally, the product ion data were used to construct the fragmentation pathways.
3.1. Fragmentation of tadalafil and tadalafil analogues
Tadalafil analogues contain different functional groups on the tadalafil backbone. For example, aminotadalafil contains an amino group in the R position, while demethyltadalafil lacks a methyl group in the R position, and cyclopentyltadalafil contains a cyclopentyl moiety in the R position. Tadalafil exhibits an m/z 390.14 as [M+H]+, and fragment ions at m/z 268.11, 197.07, 169.07, and 135.04 were observed in the QTOF–MS/MS mode. Fragments of m/z 268.10 [M+H-C7H6O2]+ and m/z 135 [M+H-C14H14N3O2]+ could be suggested from [M+H]+ in CE value of 20 eV. A peak at m/z 197 was observed owing to the elimination of C3H5NO from m/z 268.10 [M+H-C7H6O2]+, and a peak at m/z 169.08 was observed due to the elimination of a carbonyl group (Fig. 1). Among the fragments and MS/MS spectra of the tadalafil analogues, common ions at m/z 169.08 and 135.05 were observed. Although tadalafil analogues have different molecular weights, they exhibit common ions at m/z 169.08 and 135.05, and show similar patterns in their MS/MS spectra and fragmentation pathways.
Fig. 1.(a) Chemical structures of tadalafil and tadalafil analogues, (b) typical spectra of tadalafil obtained by LC–QTOF–MS in positive electrospray mode, and (c) fragmentation pathway of [M+H]+ from tadalafil in positive ion mode.
3.2. Fragmentation of sildenafil, sildenafil analogues, and hongdenafil analogues
Sildenafil analogues were sorted into 18 categories based on the functional group; several examples are shown in Fig. 2. The structures of sildenafil and hongdenafil were found to contain the sulfonyl group. Substituting an oxygen atom on R2 with a sulfur atom in sildenafil results in the formation of thiosildenafil. Common product ions of sildenafil and sildenafil analogues were observed at m/z 311.15 and 283.12, respectively; major product ions of thiosildenafil analogues were found at m/z 327.12 and 299.09 in the QTOF-MS/MS mode. The peak at m/z 377.12 was attributed to a methyl piperazinedetached molecule from [M+H]+, whereas the peak at m/z 311.15 was attributed to a methyl sulfonyl piperazine-detached molecule from [M+H]+. Because of the fragmentation pathway, the peak at m/z 283.12 could be ascribed to the backbone structure of sildenafil and that at m/z 311.15 originated from the separation of the methylpiperazine-detached molecule from [M+H]+. In contrast, hongdenafil analogues were synthesized from sildenafil by substituting the sulfonyl group with a carbonyl group. Hongdenafil analogues exhibited similar ions as sildenafil analogues at m/z 311.15 and m/z 297.13. The detailed pathways of the formation and structure of the compounds is shown in the Fig. 3. As with tadalafil analogues, sildenafil analogues also showed similar patterns in their MS/MS spectra and fragmentation pathways, even though their molecular weights varied due to the presence of different functional groups.
Fig. 2.(a) Chemical structures of sildenafil and sildenafil analogues, (b) typical spectra of sildenafil obtained by LC-QTOF–MS in positive electrospray mode, and c) fragmentation pathway of [M+H]+ from sildenafil in positive ion mode.
Fig. 3.(a) Chemical structures of hongdenafil, (b) typical spectra of hongdenafil obtained by LC-QTOF-MS in positive electrospray mode, and (c) fragmentation pathway of [M+H]+ from hongdenafil in positive ion mode.
3.3. Fragmentation of vardenafil and vardenafil analogues
Vardenafil has six types of analogues; several examples are shown in Fig. 4. The piperazinyl ethanol-substituted analogue is called hydroxyvardenafil and the piperidine-substituted analogue is called pseudovardenafil. In the mass scan, vardenafil showed a peak at m/z 489.22 as [M+H]+, and fragment ions at m/z 376.12, 312.15, and 151.08 were observed. The peak at m/z 376.12 was attributed to the ethylpiperazine-detached structure from [M+H]+, while that at m/z 312.15 originated from the detachment of the ethylpiperazine and sulfonyl groups. The peaks at m/z 312.16 and 151.09 are common ion molecules, which were observed in all the six vardenafil analogues. The vardenafil analogues exhibited common ions at m/z 312.16 and 151.09 and showed similar patterns in their MS/MS spectra and fragmentation pathways.
Fig. 4.(a) Chemical structures of vardenafil and vardenafil analogues, (b) typical spectra of vardenafil obtained by LC-QTOF-MS in positive electrospray mode, and (c) fragmentation pathway of [M+H]+ from vardenafil in positive ion mode.
3.4. Fragmentation of other analogues
Fig. 5 shows the QTOF-MS spectra and fragmentation pathways of other representative analogues. Among the 18 types of other analogues, similarities or common ion peaks could not be found, except for tadalafil, sildenafil, and vardenafil. For example, in the mass scan, avanafil exhibited a peak at m/z 484.18 as [M+H]+, and fragment ions at m/z 375.12 and 221.10 were observed. Further, m/z 375.12 was observed due to the elimination of C5H9N3 from m/z 484.18 [M+H]+; m/z 221.10 was observed due to the elimination of C8H9ClO from m/z 375.12 [M+HC5H9N3]+ (Fig. 5).
Fig. 5.(a) Chemical structure of avanafil, b) typical spectra of avanafil obtained by LC-QTOF-MS in positive electrospray mode, and (c) fragmentation pathway of [M+H]+ from avanafil in positive ion mode.
3.5. Identification of unknown compounds: Potential application
A difference of 14 Da was observed between the peak at m/z 390.14 from [M+H]+ of tadalafil and the peak at m/z 404.00 from [M+H]+ of an unknown compound; therefore, the loss of a CH2 group could be expected. The unknown compound was predicted to be a tadalafil analogue, as it exhibited similar characteristic ions as tadalafil at m/z 135.05, 169.08, and 197.07 in the MS/MS spectrum.33 Furthermore, in regard to the fragmentation pathway, a difference of 14 Da between the peak at m/z 268.10 of tadalafil and that at m/z 282.12 of the unknown compound indicates that the attached group on the R position was an ethyl group, not a methyl group. The daughter ions generated from the parent ions help predict the fragmentation pattern of the molecule and are useful for confirming the target analytes. Therefore, the fragmentation pathway of homotadalafil could be represented as shown in Fig. 6.
Fig. 6.(a) Typical spectra of tadalafil and homotadalafil obtained by LC-QTOF-MS in positive electrospray mode and (b) fragmentation pathway of [M+H]+ from tadalafil and homotadalafil in positive ion mode.
Structural analysis of 58 PDE-5 inhibitor analogues was performed using LC-QTOF-MS. The 58 analogues were sorted into five groups: tadalafil, sildenafil, hongdenafil, vardenafil, and other analogues. By analyzing the structural characteristics and fragmentation pathways of the compounds, the common ion molecules were confirmed for each group. The developed analytical method could be utilized in the investigation and identification of new PDE-5 inhibitor analogues and can be very useful in the detection of adulteration. The developed method is expected to serve as a novel and facile method for the expeditious identification of unknown compounds.