Sulfadiazine Masquerading as a Natural Product from Scilla madeirensis (Scilloideae).

The structure of 2,4-(4'-aminobenzenamine)pyrimidine (1), a pyrimidine alkaloid previously isolated from the bulbs of Scilla madeirensis (Asparagaceae, synonym Autonoë madeirensis), has been revised. These conclusions were met via comparison of reported NMR and EIMS data with those obtained from synthetic standards. The corrected structure is the antibiotic sulfadiazine (2), which has likely been isolated as a contaminant from the site of collection. The reported bioactivity of 1 as an α1-adrenoceptor antagonist should instead be ascribed to sulfadiazine. Our findings appear to show another example of an anthropogenic contaminant being identified as a natural product and emphasize the importance of considering the biosynthetic origins of isolated compounds within a phylogenetic context.

The reported isolation of the analgesic tramadol from the roots of the African medicinal plant Nauclea latifolia Sm. (Rubiaceae) in 2013[1] garnered substantial attention from the scientific community in what would later be called “The Tramadol Wars” by medicinal chemist and columnist Derek Lowe.[2] The authors were met with skepticism by their conclusion that tramadol was being produced naturally,[3] and, even after a convincing biosynthetic proposal,[4] it was ultimately determined that the tramadol was an anthropogenic contaminant arising from its off-label administration to livestock in the region.[5] The tramadol story has since served as a warning to natural products chemists to be wary of the potential isolation of synthetic contaminants from natural sources. This lesson is especially important in an age where environmental drug contamination, most notably from those used in the agricultural sector,[6] is very high.[7]

The lessons learned from the tramadol incident were considered by Hassan et al. in 2017 after their isolation of the antidiabetes drug metformin from a specimen of Seidlitzia rosmarinus (Amaranthaceae).[8] The authors quickly determined that the metformin was an anthropogenic contaminant by analyzing its 14C content, the same method used to prove the synthetic origin of tramadol from N. latifolia in 2016.[5] The examples of tramadol and metformin notwithstanding, to the best of our knowledge, no other cases involving the attribution of synthetic drug contaminants as natural products have been reported. Herein, we report what appears to be another example of the isolation of an anthropogenic environmental contaminant (sulfadiazine) and its mistaken identity as a plant natural product from Scilla madeirensis (Asparagaceae, synonym Autonoë madeirensis). These conclusions were met via comparison of the reported spectroscopic data with those obtained from synthetic standards.

The Scilloideae (Asparagaceae) is a subfamily of bulbous plants containing approximately 900 species across 70 genera. The chemistry of the Scilloideae has been recently reviewed, revealing it to be particularly rich in homoisoflavones, bufadienolides, cardenolides, and steroidal glycosides. In terms of alkaloids, pyrrolizidines, pyrrolidines, and piperidines are particularly common.[9] However, the structure of 2,4-(4′-aminobenzenamine)pyrimidine (1), an alkaloid purportedly isolated from the bulbs of S. madeirensis,[10] appeared to us as slightly unusual, with no structurally similar compounds previously reported from the Scilloideae.

Upon further investigation, we were unable to corroborate the proposed structure of 1 with the reported NMR spectroscopic data (Figure ; Table ). Of particular note is the deshielded chemical shift of C-2′/C-6′ (δC 130.2) compared to C-3′/C-5′ (δC 112.5), which is not consistent with a p-phenylenediamine moiety.[11] Moreover, C-4′ (δC 153.4) is too deshielded to be an amine-bound aromatic carbon para to an amino group.[12] One explanation for the highly uneven chemical environments of C-2′/C-6′ and C-3′/C-5′ is a sulfonamide group linking C-1′ and C-2. This also justifies the deshielded chemical shift of C-4′.[12] The presence of a sulfonamide group at this position would identify 1 as the antibiotic sulfadiazine (2).

Figure 1

Proposed (1) and revised structure of sulfadiazine (2).

Table 1

Comparison of the Previously Reported NMR Data of 2,4-(4′-Aminobenzenamine)pyrimidine (1)[10] with Those of Synthetic 1 and Commercially Acquired Sulfadiazine (2) (All Data Are Reported in DMSO-d6)

	2,4-(4′-aminobenzenamine)-pyrimidine (1) (reported by Dias et al.)[10]		2,4-(4′-aminobenzenamine)-pyrimidine (1) (synthetic)		sulfadiazine (2) (commercially acquired)
position	δC,a type	δHb (J in Hz)	δC,c type	δHd (J in Hz)	δC,e type	δHf (J in Hz)
1	N		N		N
2	157.6, C		160.4, C		157.2, C
2-NH	NH	11.26, brs	NH	9.07, brs	NH	11.24, brs
3	N		N		N
4	158.6, CH	8.47, d (4.8)	157.8, CH	8.34, d (4.6)	158.2, CH	8.47, d (4.8)
5	115.8, CH	7.00, t (4.8)	111.1, CH	6.67, t (4.6)	115.5, CH	6.99, t (4.8)
6	158.6, CH	8.47, d (4.8)	157.8, CH	8.34, d (4.6)	158.2, CH	8.47, d (4.8)
1′	125.2, C		130.0, C		124.9, C
2′	130.2, CH	7.61, d (8.7)	121.3, CH	7.32, d (8.4)	129.8, CH	7.62, d (8.8)
3′	112.5, CH	6.56, d (8.7)	114.3, CH	6.54, d (8.4)	112.1, CH	6.57, d (8.8)
4′	153.4, C		142.7, C		153.0, C
4′-NH2	NH2	6.00, s	NH2	5.18, brs	NH2	5.98, s
5′	112.5, CH	6.56, d (8.7)	114.3, CH	6.54, d (8.4)	112.1, CH	6.57, d (8.8)
6′	130.2, CH	7.61, d (8.7)	121.3, CH	7.32, d (8.4)	129.8, CH	7.62, d (8.8)

100 MHz.

400 MHz.

200 MHz.

600 MHz.

125 MHz.

500 MHz.

NMR analyses of commercially acquired sulfadiazine demonstrate that it shares identical spectroscopic data to those originally reported for 1 (Table ).[10] As attempts to obtain samples from the original publication were unsuccessful, we synthetically prepared a standard of 1. The acquired NMR data underscore the clear spectroscopic differences between synthetic 1 and sulfadiazine (Table ). The most prominent differences in the 13C NMR data are the chemical shifts of C-2′/C-6′ and C-4′, which are shielded by 8.5 and 10.3 ppm, respectively, in comparison to sulfadiazine.

We then sought to determine whether the EIMS data support our structural revision. The original authors reported a base peak observed in EIMS (70 eV) at m/z 185, which was ascribed as the [M – H]+ ion of 1, while the molecular ion was proposed to be at m/z 186 (rel int 61%).[10] Previously published EIMS data of sulfadiazine report that its molecular ion peak occurs at m/z 250 (rel int 0.3%), while the base peak is observed at m/z 185 [M – HSO2]+. A prominent peak also occurs at m/z 186 [M – SO2]+ (Figure ).[13,14] The low intensity of the molecular ion peak of sulfadiazine appears to have—understandably—complicated the assignment of its molecular formula. Our acquired EIMS data of sulfadiazine (Supporting Information) match with both previously published data for sulfadiazine[13,14] and the originally reported EIMS data of 1.[10] Interestingly, our acquired EIMS data (70 eV) of synthetic 1 show the authors’ predicted base [M – H]+ peak at m/z 185, while the molecular ion is observed at m/z 186 (rel int 96%) (Supporting Information).

Figure 2

Proposed ions of sulfadiazine (2) formed after the loss of SO2 (m/z 186) and HSO2 (m/z 185)[13]

Although we were unable to obtain any of the original isolated compound or plant material, the site of the plant collection, the Botanical Garden of Funchal, Madeira, Portugal, supplied two freshly collected samples of S. madeirensis bulbs. LC-MS analyses of their MeOH and CH3CN/H2O (1:4) extracts did not show any molecular ions associated with either 1, 2, or the main mammalian metabolites of sulfadiazine (4-hydroxysulfadiazine, N-acetylsulfadiazine).[15] The original collection appears to have been performed approximately 20 years ago,[10,16] and thus there are a litany of potential reasons (e.g., decomposition over time, new specimens being planted, changes in horticultural practices) that sulfadiazine was not observed in the current study. However, given that sulfadiazine has never previously been reported as a natural product, we were led to the hypothesis that this is another example of an anthropogenic contaminant being mistakenly identified as a natural product.

Sulfadiazine is a known environmental contaminant due to its widespread use in veterinary medicine.[6] Its low bioavailability leads to excretion in animal feces and urine, which can then be spread throughout the environment as fertilizer.[17] Studies on willow (Salix fragilis L.), maize (Zea mays L.),[18] and wheat (Triticum aestivum L.)[19] growing in sulfadiazine-spiked soil have demonstrated that sulfadiazine may accumulate in plant roots and even be translocated to stems and leaves. Natural products containing a sulfonamide moiety are also quite rare. They are predominantly produced by bacteria[20,21] but have also been isolated from marine invertebrates.[22,23] To the best of our knowledge, no sulfonamide-containing natural products have been isolated from or observed in terrestrial plants. Given this and the known capacity for sulfadiazine to bioaccumulate in plants, it is most likely that its isolation from S. madeirensis is a consequence of anthropogenic contamination at the site of collection. This is supported by our failure to redetect it from fresh plant samples.

Analysis of radiocarbon (14C) levels provides an experimental platform to distinguish between carbon atoms recently incorporated into a molecule of interest via photosynthesis and those derived from petroleum-based precursors, where 14C nuclides have long since decayed.[5,8] Unfortunately, the lack of access to the originally isolated sample or plant material precludes this analysis, and we cannot definitively rule out the possibility that sulfadiazine was biosynthesized by S. madeirensis or an associated microorganism.

The original authors also report that 1 is active as an α1-adrenoceptor antagonist.[10] This activity should instead be attributed to sulfadiazine. To the best of our knowledge, sulfadiazine has not been previously reported as an α-blocker. Several clinically used α-blockers (e.g., prazosin, terazosin, alfuzosin, doxazosin) contain a core 2,4-diaminoquinazoline functionality, and sulfadiazine shares some structural similarities with these drugs.[24]

These findings illustrate the importance of considering the biosynthetic origins of newly discovered natural products within a phylogenetic context. Our attention was drawn to 1 due to its structural dissimilarity to any other alkaloids isolated from the Scilloideae, which is typically known to produce alkaloids of the pyrrolizidine, pyrrolidine, and piperidine types.[9] These alkaloids are biosynthesized by a limited set of precursors (l-ornithine in the case of pyrrolizidine and pyrrolidine alkaloids; l-lysine for piperidines),[25] and it is challenging to propose a biosynthetic pathway for 1 within this framework. While the structures of new natural products should always be thoroughly validated, particular scrutiny must be paid to proposed structures that deviate considerably from those known to occur within a taxonomic group.

Experimental Section

General Experimental Procedures

NMR spectra were acquired at 298 K on a Bruker 500 MHz (TXO CRPHe TR-13C/15N/1H 5 mm-Z CryoProbe) and a Bruker 600 MHz (TCI CRPHe TR-1H and 19F/13C/15N 5 mm-EZ CryoProbe) spectrometer. Chemical shifts were referenced to the solvent peak for (CD3)2SO at δH 2.50 and δC 39.52. GC-MS analyses were performed on an Agilent Technologies 7890A GC system equipped with an Agilent Technologies 7693 autosampler and an Agilent Technologies 5975 Series MSD. A linear gradient from 70 to 300 °C was employed, and EI was set at 70 eV. The column used was a CP-SIL 8 CB Low Bleed (30 m × 0.25 mm) capillary column. High-resolution mass spectrometric measurements were acquired using a UPLC-QToF nanospray MS (Waters nanoAcquity, QToF Micro). The UPLC column used was a Waters ACQUITY UPLC M-Class Peptide BEH C18 column (1.7 μm, 130 Å, 75 μm × 150 mm). DMSO-d6 (99.8%) was from Teknolab Sorbent. Sulfadiazine (≥99.0%) was purchased from Sigma-Aldrich.

Plant Material

S. madeirensis material was collected and identified in September 2019 at the Botanical Garden of Funchal, Madeira, Portugal, by Francisco Manuel Fernandes. Voucher specimens (MADJ 14346 and MADJ 14347) are housed at the Botanical Garden of Funchal, Madeira, Portugal.

Extraction and LC-MS Analysis

Small portions (1 g) of oven-dried (40 °C) and ground S. madeirensis bulbs (100 g) were exhaustively extracted via sonication (20 min) in MeOH (20 mL) and then 1:4 CH3CN/H2O (20 mL).[15] The samples were evaporated, filtered (0.45 μm), and redissolved in 20% MeOH at a concentration of 5 mg/mL and subject to LC-MS analysis. The mobile phase used was a linear gradient from 1% to 90% CH3CN (0.1% FA) over 50 min at a flow rate of 0.3 μL/min. Data were analyzed using MZmine 2.[26]

2,4-(4′-Aminobenzenamine)pyrimidine (1)

Synthesis of 1 was performed using a method patented by Jautelat et al. (2008).[27] A suspension of 4-nitroaniline (28 mg, 0.20 mmol) and 2-chloropyrimidine (23 mg, 0.20 mmol) in acetonitrile (2 mL) was treated with HCl (4 M in dioxane, 50 μL) and water (50 μL). The reaction mixture was then refluxed for 16 h before cooling to room temperature. Triethylamine (34 μL, 0.24 mmol) was added, and the resulting solid washed with acetonitrile and water and then dried under vacuum overnight. The solid was then suspended in ethyl acetate and ethanol (1:1, 4 mL), and the vessel flushed with nitrogen. Pd/C 10% (5 mg) was added, and the reaction placed under a hydrogen atmosphere. After stirring for 20 h, the reaction mixture was filtered through Celite, and the filtrate concentrated to afford the title compound (26 mg, 69%).