Specialized metabolites of the genus Zephyranthes Herb.: a critical review on taxonomy and phytochemistry

Abstract

This article is a critical review of the taxonomy and phytochemistry of the genus Zephyranthes Herb., a group of plants known for their beautiful flowers and traditional medicinal uses. The present review summarizes the occurrence, isolation, and identification of specialized metabolites, which have recently been frequently studied because of their important biological activities. Among the accepted 203 species, only 27 have been phytochemically investigated. This paper provides an overview of the different types of specialized metabolites identified in these plants and considers problematic taxonomic evolution within this species. The differences between two internationally recognized databases, which classify only 41% of the species in the same way, are briefly summarized. In addition, there are many reports on their metabolites, especially alkaloids, but some of the data in the literature are occasionally inaccurate and sometimes even erroneous. This critical review aims to discuss, summarize, and evaluate up-to-date (up to July of 2023) information about metabolites of the genus Zephyranthes, focusing on phytochemistry and taxonomy.

Graphical abstract

Introduction

The Amaryllidaceae family, a group of monocotyledonous species consisting of approximately 85 genera and 1100 species, is one of the 20 most important alkaloid-containing plant families. Species are distributed widely in tropical and warm regions of the world and have been cultivated as ornamental plants for their colorful flowers and fragrant oils and their extensive usage as folk medicines against various diseases in many countries and areas. The medical properties of these plants were already known in the fourth century BC when Hippocrates of Kos used oil from the daffodil, Narcissus poeticus L., for the treatment of uterine tumors. Up to now, more than 600 structurally diverse Amaryllidaceae alkaloids (AAs) have been isolated from plants of this family with a wide range of interesting biological properties, including antitumor, antibacterial, antifungal, antimalarial, antiviral, analgesic, and acetylcholinesterase inhibitory activities (He et al. 2015). The present review summarizes phytochemical studies carried out on the genus Zephyranthes Herb., focusing on the occurrence, isolation, and identification of its specialized metabolites.

The genus Zephyranthes Herb., family Amaryllidaceae, is an American-Antillean genus with about 200 species. These plants are native to tropical and subtropical America. Several species are cultivated due to their gorgeous flowers and are known by plant breeders as „rain-lilies “, owing to their tendency to flower after the rainy period (Fernández Alonso and Groenendijk 2004). Plants of the genus Zephyranthes have become naturalized in other places such as Hawaii, Indonesia, Thailand, and South Africa. From spring through fall, rain-lilies can produce flushes of star-shaped, crocus-like actinomorphic white, pink, or yellow flowers, depending on the species. Rain-lilies flower in spring, summer, or fall, depending on the species. Each six-petalled, funnel-shaped flower is perched at the top of a stem that ranges in height from 5 cm to more than 30 cm and has a single, upward-facing, or slightly nodding flower on each stem (Z. drummondii rarely has two flowers on one stem). Each flower lasts only a day or two, depending on sunlight and temperature, but typically, new flowers continually develop for several days, creating flushes of flowering (Knox 2009).

This article is organized into two main parts. The phytochemical isolations of secondary metabolites are described in the following text, where the plants are arranged alphabetically. As the nomenclature has undergone changes, we have decided to use accepted names in the Plants of the World Online database (POWO 2023), and, therefore, some cited articles used a different name for the same plant (given in synonyms). The last section presents the species of Zephyranthes that have only been subjected to gas chromatography-mass spectrometry (GC–MS) profiling. Therefore, the manuscript lists compounds of this genus in Tables and Figures only if they have been isolated and their structure properly analyzed. A summary of all compounds, even those detected by MS, is given in the Supplementary Information (SI, Table S1), which is then shown in the compound numbering as an S descriptor. Additionally, all synonyms and accepted names listed by POWO and World Flora Online (WFO 2023) can be found in the SI (Table S2).

Genus Zephyranthes: Classification and taxonomic confusion

• Domain—Eukaryota.

• Kingdom—Plantae.

• Phylum—Tracheophyta.

• Class—Liliopsida (Monocotyledons).

• Order—Asparagales Link.

• Family—Amaryllidaceae J.St.-Hil.

• Subfamily—Amaryllioideae.

• Tribe—Hippeastreae.

• Genus—Zephyranthes Herb.

In 1821, William Herbert described Z. atamasca (L.) Herb. as the first species of the genus Zephyranthes (Herbert 1821). This genus of the tribe Hippeastreae originally belonged to the Liliaceae family, and through the years, the taxonomic classification changed many times. However, recent advances in classification methods and more detailed studies confirmed that this genus belongs to the family Amaryllidaceae, tribe Hippeastreae (García et al. 2019). The taxonomy within this tribe has changed over the years. It is still very dynamic depending on the relatively complicated phylogeny of the Hippeastreae tribe, which is more likely represented as a kind of network rather than a tree as is the usual presentation for morphological evolution (Cornwell and Nakagawa 2017). External morphological characters intertwine within the tribe and cannot be taken definitively. From a taxonomic point of view, the safest way to accurately classify a given plant species is to use molecular methods based on nuclear and chloroplast DNA analysis. However, even in this area, we encounter problems where the vast majority of described plants have not been analyzed using these methods. As part of the review, we came across problematic cases in which a single article describes a plant classified as one or two different species depending on the source database. Another challenging field in the taxonomy of this tribe is that the knowledge and methods of classification do not set clear rules in classification that are based on the frame of net-relations. Even nuclear topology did not bring a clear distinction between individual species of this tribe because of the presence of allopolyploidy in the genus Zephyranthes (García et al. 2014).

We provide a thorough overview of possible Zephyranthes synonyms for each species in the SI (Table S2) to disclose problematic nomenclature. A small part of this genus has been subjected to phytochemical analysis (Fig. 1), which is discussed below.

Fig. 1

General overview of phytochemical research on the genus Zephyranthes (accepted species in POWO). 87% of the genus has not yet been studied

As this article is built on the POWO taxonomy, comparison with the WFO nomenclature is straightforward, which showed an interesting result (see Fig. 2). In conclusion, from these databases, 203 different species are listed as Zephyranthes in either the accepted name or its synonym. Of these, only 88 species have the same accepted name in POWO and WFO. Another 95 species are classified as Zephyranthes sp. in POWO, but as another genus in WFO. Furthermore, 10 species are classified in a different genus, but with Zephyranthes listed as a synonym (POWO and WFO). In addition, 10 species are registered only in POWO but not yet recognized by WFO. Last, but not least, 12 species are waiting for classification in both databases.

Fig. 2

Comparison of POWO and WFO taxonomies that classify Zephyranthes as accepted name or synonyms. Only 88 species are equally classified as Zephyranthes in both databases

Phytochemistry and chemical constituents of genus Zephyranthes

Of the 203 accepted species of Zephyranthes (POWO 2023), only 15 have undergone phytochemical isolation studies (summarized in Table 1); another 12 were subjected to GC–MS profiling only (Fig. 1). Major attention within the reported phytochemical studies has been paid to alkaloids since they are the most studied constituents of Amaryllidaceae species. However, other components have also been studied (Fig. 3), such as flavonoids, flavans, gibberellins, phospholipids, sterols, lectins and ceramides (Katoch and Singh 2015).

Table 1 Overview of alkaloids isolated from the genus Zephyranthes

Fig. 3

Composition of specialized metabolites isolated from the Zephyranthes genus (type of secondary metabolite; number of compounds; overall percentage)

AAs are derived from the aromatic acids, phenylalanine, and tyrosine, which are used to produce the key intermediate 4′-O-methylnorbelladine (Bastida et al. 2011; Kilgore and Kutchan 2016), after which the biosynthetic pathway is named. AAs biosynthesis has recently been described in detail in several review articles (Berkov et al. 2014; Cahlíková et al. 2019; Desgagné-Penix 2014, 2021; Nair and van Staden 2013), and thus this review does not cover it.

The first phytochemical study of the genus Zephyranthes was carried out in 1941 revealing the presence of lycorine (1) in bulbs of Z. texana Hook. (Greathouse and Rigler 1941). However, due to the dynamic taxonomy of this species, it is nowadays recognized as Habranthus tubispathus (L'Hér.) Traub by WFO (WFO 2023) and Zephyranthes tubispatha Herb. by POWO (POWO 2023). Until now, over 150 alkaloids have been isolated from this genus (Fig. 4).

Fig. 4

Distribution of structural types in AAs isolated from Zephyranthes (structural type; number of alkaloids)

Since this article is a critical overview, Fig. 5 shows that some structural types tend to be predicted by GC–MS analysis while others are not. In particular, alkaloids belonging to plicamine-, secoplicamine-, and pancratistatin-types were isolated rather than only detected on GC–MS. But for more common structural types, such as lycorine-, galanthamine-, and haemanthamine-/crinine-type, there is a higher possibility of detection, but their existence is not subsequently proven by isolation from the extract.

Fig. 5

Comparison of the distribution of detected and isolated structural types. Each structural type is numbered by isolated alkaloids (white number) and detected alkaloids (black number)

The next chapter is the numbering system of AAs. Many authors use different approaches (by chemical nomenclature, biosynthetic, or other order; example in Fig. S1 in the SI). Therefore, when naming an undescribed compound, including a locant in the name can confuse. Because there are many ways of numbering, we have tried to choose one that makes sense for the names of most compounds of structural type. But of course, there will be discrepancies. Giving you the tip of the iceberg—galasine type, a rare structural type with only five known alkaloids described in two experimental works, is included. The first discovered alkaloid of this structural type was named galasine in 1995 (Latvala et al. 1995). The structure was determined not only by nuclear magnetic resonance spectroscopy (NMR), circular dichroism (CD), and MS but also by X-ray analysis. Later, Wang and co-workers identified other molecules of this structural type following the numbering established 12 years ago (Wang et al. 2007). Although only this numbering was known for this structural type, Berkov and co-workers used their own new and different numbering and renamed these alkaloids in their review (Berkov et al. 2020). More details are given in Fig. S2 in the SI, and we list the structural frameworks numbering of the isolated alkaloids belonging to lycorine type (Fig. 6), homolycorine type (Fig. 7), haemanthamine/crinine type (Fig. 8), galanthamine type (Fig. 9), pancratistatin type (Fig. 10), pretazzetine type (Fig. 11), plicamine and secoplicamine type (Fig. 12), and other structural types (Fig. 13).

Fig. 6

Lycorine type alkaloids isolated from Zephyranthes sp.

Fig. 7

Homolycorine alkaloids isolated from Zephyranthes sp.

Fig. 8

Haemanthamine and crinine type alkaloids isolated from Zephyranthes sp.

Fig. 9

Galanthamine type alkaloids isolated from Zephyranthes sp.

Fig. 10

Pancratistatin type alkaloids isolated from Zephyranthes sp.

Fig. 11

Pretazzetine type alkaloids isolated from Zephyranthes sp.

Fig. 12

Plicamine and secoplicamine type alkaloids isolated from Zephyranthes sp.

Fig. 13

Other structural types of alkaloids isolated from Zephyranthes sp.

Phytochemical studies of Zephyranthes ananuca (Phil.) Nic. García

Synonyms: Hippeastrum ananuca (Baker) Sealy

The very first plant in this review is classified as Zephyranthes ananuca in POWO, but isolation was performed on it under the name Hippeastrum ananuca. This nicely represents the complexity of the species classification problem.

All the work reported on this species comes from one research group, and was published between 1978 and 1986. In summary, three articles presented an isolation of six alkaloids, one of which was new, 17-epihomolycorine (46), an epimer of homolycorine (Pacheco et al. 1978; Perez et al. 1986). The correctness of the identified structure 46 was later confirmed by X-ray analysis (Gopalakrishna et al. 1978). Futhermore, one of their papers presents three proanthocyanidins (but the structure was disclosed for only two—172, 178), and daucosterol (201) was isolated from the non-alkaloidal part of the extract of this Chilean plant (Pacheo et al. 1981).

Phytochemical studies of Zephyranthes bifida (Herb.) Nic. García & Meerow

Synonyms: Rhodophiala bifida (Herb.) Traub

In 1968, Wildman and Brown (Wildman and Brown 1968b) described an isolation of 11.9 g of montanine (153), 3.1 g of haemanthamine (47), 1.2 g pancracine (154), and 0.5 g of vittatine (48) from 11.5 kg of Rhodophiala bifida bulbs. The identifications were based on a thorough data analysis of carbon, hydrogen, and nitrogen analysis (CHN), ultraviolet spectroscopy (UV), MS, ¹H NMR, melting points, and optical rotations. Several years later, Wildman and Feinstein (Feinstein and Wildman 1976) presented a proposal for montanine biosynthesis based on an isotope-labeled [³H] vittatine, which was introduced to the bulbs before growing. After 3 weeks, a harvest provided 437 g of bulbs, from which montanine (153) and haemanthamine (47) were isolated.

Later, Brazilian research groups studied this plant from different angles (Castilhos et al. 2007; de Andrade et al. 2016; Farinon et al. 2017; Reis et al. 2019; Schwedersky et al. 2020). Their works are summarized in several papers, but only two presented the actual isolation of a single alkaloid (Castilhos et al. 2007; Farinon et al. 2017). More or less, all articles including GC–MS profiling have the same result as the work of Wildman and co-workers, i.e. that it is a plant suitable for montanine (153) isolation. It is not clear to us if there was a new extract for every published paper or if there is some connection between these works. Not all works contain transparent information about how the plant material was obtained, recognized, and subsequently processed. As an example, the article by Farinon et al. describes, among other things, the isolation from dried powdered bulbs, but without any information about any weight or the final purification process. Additionally, they refer to the Supplementary data for additional confirmation by NMR. However, upon closer examination, only a poor description of the ¹H NMR spectrum is found, together with a picture of the spectrum showing nothing else but one signal at 4 ppm.

As an exception, Schwedersky and co-workers identified 153 and, surprisingly, nangustine (S180), which none of the other articles mentioned (but their paper provides incorrect structure drawings). Moreover, rarely occurring S180 seems to be problematic for GC–MS analysis, and there are differences in the spectra of the synthetic and isolated compound (Kokas et al. 2008). Since, S180 is only an isomer of pancracine (S179), which has actually been isolated from this plant (Wildman and Brown 1968b), a question mark hangs over this result.

Phytochemical studies of Zephyranthes brachyandra (Baker) Backer et Beknopte

Synonyms: Habranthus brachyandrus (Baker) Sealy

In 1968, the first alkaloidal structure from this plant was reported as habranthine (79) (Wildman and Brown 1968a). Another phytochemical study comes from 2009, when Jitsuno and co-workers isolated 18 secondary metabolites from Z. brachyandra bulbs (Jitsuno et al. 2009). The extraction was initiated with hot methanol (MeOH), followed by passage through a Diaion HP-20 column with eluents composed of MeOH, ethanol (EtOH), and ethyl acetate (EtOAc). The obtained fractions were further separated using silica gel column chromatography (CC) and different solvents to obtain pure compounds. Of those compounds, eight were flavan derivatives with newly described (2R,3R)-3,7-dihydroxyflavan (169), (2R,3R)-3,4′,7-trihydroxyflavan (170), (2R,3R)-3,4′-dihydroxy-7-methoxyflavan (171), and (2R,3R)-3,7-dihydroxy-3′,4′-methylenedioxyflavan (177); one new hydroxybutyric acid glucoside (206); three phenolic compounds (203–205); and six alkaloids (6, 8, 47, 50, 53, 92), with the highest yield of (+)-bulbispermine (53) (150 mg from 2.5 kg of fresh bulbs) (Jitsuno et al. 2009).

The latest study from 2022 describes the isolation and GC–MS analysis of alkaloids from 200 g of bulbs, which were initially processed by maceration in 2% H₂SO₄ in an ultrasonic bath for 4 h. The extract was centrifuged, and the supernatant was defatted with diethyl ether (Et₂O). After the alkalization of the water layer with 10% NaOH to pH 11–12, the alkaloids were extracted with dichloromethane (DCM) and the solution was dried with anhydrous sodium sulfate (the yield of the alkaloidal extract was ca. 0.1%). For isolation, a portion of the alkaloidal extract was roughly separated by flash chromatography on silica gel to obtain three fractions, which, after separation on Sephadex LH-20 and thin-layer chromatography (TLC), yielded ismine (149), tazettine (110), hippeastidine (65), and 3-epimacronine (113). Most recently, an article based on analysis of GC–MS spectra detected the content of eight additional alkaloids in minor amounts that had not been successfully isolated from this species (36, 38, 62, 83, 100, 110, 111, 145, 149) (Martinez-Peinado et al. 2022).

Phytochemical studies of Zephyranthes candida (Lindl.) Herb.

Zephyranthes candida is one of the most phytochemically studied plants of the genus Zephyranthes, and is currently also one of the most widespread species of this genus. The species originates from South America, including the states of Paraguay, Uruguay, Argentina, and Brazil (Tapia-Campos et al. 2012). Furthermore, Z. candida has also been domesticated in China, South Africa, and Korea (Luo et al. 2012).

Since alkaloids are the most active constituents of Z. candida, most of the phytochemical studies have led to either the isolation or detection of about 79 AAs from bulbs, flowers, and whole plants. A first pilot study was reported in 1955 by Boit and co-workers, who isolated and identified lycorine (1, Fig. 6), nerinine (38), tazettine (110), and haemanthamine (47) (Boit and Ehmke 1955). Later trans-dihydronarciclasine (93) was purified from 18 kg of fresh bulbs using non-acid–base extraction (Pettit et al. 1990).

Furthermore, in recent years, Yao's research group has published six papers focusing on the full phytochemical work of this plant leading to the isolation of several new skeletons of AAs and, in addition, new flavans. Of course, they also reported derivatives of well-known skeletons such as galanthamine-, plicamine-, lycorine-, pretazettine-, haemanthamine-, secoplicamine-, and miscellaneous structural types (Table 1). In general, the phytochemical procedures were quite standard, using the whole plant collected in China, which was dried and macerated in 95% EtOH. The studies differ in maceration conditions (2% HCl, 2% acetic acid, or without acid), and liquid–liquid extraction (LLE) with chloroform (TCM) and water phase with pH adjusted to 7(!) or 10. Considering the pKa of AAs, the neutral pH does not seem favorable for the extraction, but according to these articles, it led to the isolation of alkaloids, even with pH = 7. Right from the first article, published in 2012, isolation artifacts were described as alkaloids. Starting with ethyl esters 107 and 108, extraction in acidic aqueous EtOH at room temperature followed by evaporation in a rotavapor would lead to the esterification of the carboxyl of the parent acid. However, this does not change the fact that these are interesting structures. Similarly, the acid–base environment produced another alkaloid, N-methylated haemanthidine, as a mixture of 6-epimers (63—spelled in the article as N-methylhemaenthidine). Although the quaternary salt may seem a little suspicious because the hydroxyl is more nucleophilic than amine, the authors supported it with X-ray crystallography, and 2 years later, published a study focused only on its cytotoxicity (Guo et al. 2014). Moreover, after another 2 years, the same compound, which is a mixture of two diastereomers (problematic in drug development), was presented as “a novel agent beneficial to patients with acute myeloid leukemia” (Ye et al. 2016). However, this compound was not new at the time, as Wildman and Bailey reported this structure in 1969 (Wildman and Bailey 1969). Their comprehensive study presented acid–base interconversions of 6-hydroxy-5,10b-ethanophenanthridines to [2]benzopyrano[3,4-c]indoles and vice versa. For example, haemanthidine converts to tazettine, which is a known isolation artifact, in an alkaline environment, while acid-facilitated rearrangement of pretazettine produces haemanthidine methyl halide. We assume that 63 was formed during isolation and is, therefore, not a product of the plant metabolism. In addition, the paper also reports the isolation of nigragillin (156), a piperazine alkaloid, but we believe that it was introduced into the plant material by a fungal infection and is thus a co-isolated contaminant and not a product of the plant metabolism (Luo et al. 2012).

In 2016, Zhan and co-workers reported other alkaloids and flavans from Z. candida. They first published a study dealing with the processing of 50 kg of the dried whole plant, which was macerated in 95% EtOH (r.t., no acid added), the extract evaporated and partitioned between TCM and water adjusted to pH 10 (Zhan et al. 2016b). Subsequent multiple separations led to the isolation of 24 derivatives belonging to the plicamine, secoplicamine, and galanthamine groups of AAs. Indolo[3,3a-c]isoquinoline represents the plicamine's framework, while secoplicamine has only a spiroatom connecting cyclohexene with a tetrahydroisoquinoline scaffold in the 1,4′ position. Such alkaloids occur in plants only in trace concentrations (e.g., from 50 kg of the whole plant of Z. candida only 11.4 mg of N-isopentyl-5,6-dihydroplicane (124) and 2.8 mg of N-methyl-11,12-seco-5,6-dihydroplicane (144) were isolated) (Zhan et al. 2016b). Four alkaloids were described as amine oxides of known alkaloids (90, 91, 131, 132), which were determined by chemical shifts in NMR spectra of nitrogen surrounding carbons. The electron deshielding effect induced by a positive charge on nitrogen results in a shift of approx. 10 ppm in the ¹³C and approx. 0.8 ppm in the ¹H NMR spectra when compared to the corresponding tertiary base of such alkaloids. The second published work of this research group in 2016 focused on three new flavonoids, flavan-3-ol and flavan glycosides, isolated from non-alkaloidal fractions of the very same extract as the previous paper (Zhan et al. 2016c).

For a change, the third work published in 2016 presented a phytochemical study carried out on only 10 kg of dried whole plant of Z. candida (Zhan et al. 2016a). The plant material was left macerating in 95% EtOH with 2% HCl at room temperature. Surprisingly, the authors reported obtaining only 3.5 mg of a new alkaloid zephycandidine A (157), which is an AA with a rare imidazo[1,2-f]phenanthridine scaffold and nothing else. As stated in the article, the structure was already known at the time. Its first report presents this easily crystallizing tetracycle from acetone (Ace) in a one-pot synthesis with a focus on the discovery of new cytotoxic drugs with DNA binding properties (Parenty et al. 2006). Therefore, since the NMR data were known at the time, the reported structure elucidation seems quite excessive.

Zhan's article from 2017 presented another three new alkaloids belonging to the galanthindole type: zephycandidine I (146), mesembrine type: zephycandidine II (152), and ismine type: zephycandidine III (150). Paradoxically, no other alkaloids were reported when working with 50 kg of dried whole Z. candida (Zhan et al. 2017a). Compared to the first work of 2016, the experimental part differs only in pH adjustments. While this article mentioned only pH = 7 for LLE, the first one reported pH = 10. Surprisingly, the TCM extract weighed the same (40 g) in both articles. What is the relationship between these works is not clear. In addition, tentative biogenetic pathways for 158, 146, and 150 starting from tyramine and isovanillin have been described (Zhan et al. 2016a, 2017a).

The most recent work of this research group was built on the extraction from 50 kg of the air-dried whole plant (Zhan et al. 2023). The plant material was macerated in 95% EtOH acidified by 2% acetic acid (pH not defined), producing 3.78 kg of crude extract, which was later processed by acid–base LLE (1. step pH to 2 with HCl → H₂O/TCM, 2. step pH to 10 with NH₄OH → H₂O/TCM, and then H₂O/n-butanol (n-BuOH)) yielding 40 g of TCM alkaloidal extract and 215 g of n-BuOH alkaloidal extract. Although one would expect that the twelve first reported alkaloids for the first time should be found in trace amounts and that the substances previously reported should be isolated in higher yield, this correlation cannot be found. Moreover, this report does not mention the alkaloids they reported in previous years. Regarding the isolation artifact, five compounds with a 6-hydroxycrinane skeleton were reported as an inseparable mixture of 6-epimers (50, 55–58) with a diastereomeric ratio of 75:25; another compound was found as an ethoxy derivative of a known alkaloid (40). In terms of the yields, the epimers taken all together out-weighed the others by a ratio of 56:44. On the other hand, despite the harsh extraction conditions already producing artifacts, the authors were still able to isolate a 2-O-glycoside of lycorine (3). In addition, two known amides, N-trans-feruloyltyramine and its cis isomer (159, 160), were reported. All new alkaloids were elucidated by MS, NMR, UV, infrared spectroscopy (IR), CD analysis, and polarimetry. The authors also proposed an absolute configuration using computational CD spectra, but, in principle, this does not compare very well with the experimental data of the mixture of the two epimers in the sample.

Considering that 6-epimers of the 6-hydroxycrinane scaffold are usually isolated and synthesized together, as epimerization is favorable, Shitara and co-workers reported a successful separation (Shitara et al. 2014). During the preparation of the crude alkaloidal extract, conditions were different from those in the rest of the work on Z. candida. Only 411 g of the aerial parts (not dried) was macerated with MeOH. The extract was partitioned between EtOAc and 3% aqueous tartaric acid. Then the pH of the aqueous layer was converted to 10 using Na₂CO₃. LLE with TCM gave 310 mg of crude extract from which 100 mg of lycorine (1) was later isolated. The first separation utilized an amino Si₂O CC (hexane/EtOAc → TCM/MeOH), and the next column was the regular Si₂O CC (TCM/MeOH). The final step was reversed-phase high-performance liquid chromatography (RP HPLC) (C-18 column, 5 µm, 10 × 250 mm; isocratic elution with 23% acetonitrile and 0.1% trifluoracetic acid in water) affording 2-hydroxyalbomaculine (41) (homolycorine type, Fig. 7) in 9.6 min, 6α-hydroxyhippeastidine (67, Fig. 8) in 19.2 min, its 6-epimer (68) in 23.2 min, and 10-deoxy-6β-hydroxyhippeastidine (69) in 20.8 min. Selectivity or epimerization in an acidic environment was apparently not an issue. Although the amounts were low (1.7–5.0 mg), the yields were quite good for the amount of starting material.

Besides the article published by Zhan and co-workers focused on flavans (Zhan et al. 2016c), other research groups have also reported the isolation of flavonoids from this plant (Table 2). The first work presenting flavonoids in Z. candida reported a flavonoid glycoside (179) isolated from 700 g of fresh petals digested with hot MeOH (Nakayama et al. 1978). The only other study on rain-lily flowers presented the identification of volatile components using only GC–MS analysis, but did not provide much information (Chen and Bi 2022). The last work reporting other compounds in addition to alkaloids records 7-hydroxy-3′,4′-methylenedioxyflavan (173), together with lycorine and trisphaeridine (100) from 800 g of the whole plant using MeOH (Oluyemisi et al. 2015). However, because no chiroptic technique was used to analyze chiral compounds and the description of the experimental procedure is so vague, this paper appears to be quite questionable. On top of that, everything is crowned by completely confused structure drawings.

On the other hand, a detailed isolation from 2.1 kg of air-dried powdered bulbs (maceration for 3 days in EtOH) led to the isolation of compounds with different physicochemical properties. Although the yields were not high, two new ceramides were identified, candidamide A (194) (33 mg) and candidamide B (195) (28 mg), in addition to 12 known compounds, such as alkaloids, flavonoids, sterols, and a volatile macrocyclic lactone (202). In comparison, the most abundant compounds were β-sitosterol (200) (207 mg) and lycorine (1) (79 mg) (Wu et al. 2009b). Another four new ceramides (196–199) were reported by the same group while scaling up the primary quantity to 3.5 kg (Wu et al. 2009a) and 10.2 kg (Wu et al. 2010).

Regarding other work reported for this plant, in silico studies of known alkaloids based only on GC–MS profiling of Z. candida, Z. rosea, and Z. robusta (syn. H. robustus) were published in 2021 (Shawky et al. 2021). To complete the overview of the extensive works dedicated to Z. candida, we also must mention studies with a different isolation approach. Purification methods used in protein separation yielded new lectins, which were isolated from the plant bulbs (Kaur et al. 2007; Wu et al. 2006).

Phytochemical studies of Zephyranthes carinata Herb.

A rich source of common AAs, as well as unique ones, is Zephyranthes carinata, which is native to Mexico and naturalized in South China as an ornamental plant (Tapia-Campos et al. 2012). At first sight, considering several studies, Z. carinata constituents appear to be thoroughly investigated. Lycorine (1), galanthine (6), tazettine (110), and haemanthamine (47) were already reported in 1957 from this plant (Boit et al. 1957). The next work presented an isolation from 6.4 kg of fresh bulbs using acid–base LLE and chromatography for separation (CC and preparative TLC), followed by analytical methods such as ¹H NMR, MS, melting point, CHN, UV, optical rotatory dispersion, and polarimetry (Kobayashi et al. 1977). Thus, pretazettine (109) and carinatine (8) were added to the list of Z. carinata alkaloids. In 1998, pancratistatin (92) with its ester (94), and glycoside (95) were isolated, without using a standard acid–base extraction (Kojima et al. 1998). An amount of 0.8 kg air-dried bulbs was extracted under a reflux condenser with solvents of different polarities (TCM, Ace, MeOH). Then, the crude extract was subjected to either subsequent CC or preparative HPLC. Three years later, the same research group reported several more alkaloids obtained in exactly the same way as previously (Mutsuga et al. 2001). Surprisingly, the authors did not report any 1 in these isolations. The most abundant alkaloid was tortuosine (32) (270.9 mg), a quaternary amine salt. The structure of one new zwitterionic alkaloid (33) was reported, but the proposed structure has been proved to be different based on current information (discussed below).

Moving to the current experimental works, they describe mostly only GC–MS extract profiling followed by bioassays. Within these studies, 52 AAs have been reported; alkaloidal profiles have been dominated by 1, trisphaeridine (100), lycoramine (83), 6 and vittatine (48) /crinine (71) (Cortes et al. 2015a, b, 2019; Rojas-Vera et al. 2021). However, it is important to note that the GC–MS technique can be misleading without further confirmation such as NMR spectroscopy. Besides, some of the alkaloids do not ionize, spatial isomers cannot be determined from the GC–MS spectra at all, and structural isomers may have similar fingerprints as well. It is necessary to take the phytochemicals analyzed in these works with caution. Thus, we present alkaloids detected only by GC–MS in Zephyranthes in the SI (Table S1). To make it even more complicated and non-transparent, some works with GC–MS based profiling also provide docking studies of molecules that are assumed to be in the alkaloidal extract (Cortes et al. 2015a, 2019; Rojas-Vera et al. 2021).

Therefore, the most recent paper of Cortes' research group (Sierra et al. 2022) finally presents the isolated constituents extracted with the same procedure used in the works in 2015 and 2017 (Cortes et al. 2015a, 2018). However, the acid–base extraction from 1.8 kg of bulbs yielded only eight alkaloids—1, 6, 47, 83, 100, 110, hamayne (52), and finally enantiomers 48/71, whose stereochemistry was not elucidated in this study.

In 2017, Zhan and co-workers published a phytochemical study with structural elucidation working with an extract from 10 kg of whole dried plant (maceration in 95% EtOH at 40 °C for 2 days followed by evaporation and acid–base LLE) (Zhan et al. 2017b). The procedure used in this work resulted in 11 new compounds, mostly with the rare plicamine scaffold, along with 15 known alkaloids. Although a standard acid–base LLE procedure was used, five of the new compounds were N-oxides of known compounds. In conclusion, the described alkaloids were very similar to those reported from Z. candida by this research group in 2016 (Zhan et al. 2016b). They differed mainly in the oxidation state of carbons 6 and 11. 1 was, of course, the most abundant (1.2 g), and only another two, galanthamine (76) and 83, were obtained in more than 30 mg (78.9 and 64.3 mg, respectively). The structural analysis was well-handled by employing NMR and X-ray analysis together with experimental and computational CD. In addition, calculated ¹³C NMR chemical shifts verified zephycarinatine I (33), for which the inner salt was ruled out as an ionized isomer. So, it is the same alkaloid as reported in 2001 (Mutsuga et al. 2001) as the NMR data match. This work also proposed a biosynthetic pathway for the synthesis of plicamine and secoplicamine alkaloids, explaining different N-5 substitutions in these derivatives (essentially the condensation of benzaldehyde with different amino acids).

To cover all the works devoted to Z. carinata, Kaur and co-workers also tested the effect of crude lectin extract from its bulbs for antipoxviral and anticancer activity, with positive results (Kaur et al. 2007).

Phytochemical studies of Zephyranthes citrina Baker

Synonyms: Zephyranthes eggersiana Urb., Zephyranthes sulphurea Noter

Zephyranthes citrina, commonly known as “brujita amarilla” (yellow little witch), was described in 1882 and is native to South-East Mexico and from Cuba to Haiti (Herrera et al. 2001; Tapia-Campos et al. 2012). The first phytochemical study of 0.5 kg of Z. citrina bulbs describes the isolation of haemanthamine (47), galanthine (6), lycorine (1), and lycorenine (34) (Boit et al. 1957). Further phytochemical analysis was not described until many years later in 2001, when the isolation of eight AAs belonging to haemanthamine-/crinine- and lycorine- structural types was reported. Oxomaritidine (61) was first isolated from a natural source, previously known as an intermediate in the total synthesis of maritidine (Tomioka et al. 1977). From the bulbs collected in Caibarién in Cuba 1, 47, cavinine (74), joubertiamine (158), and zephyramine (75) have been isolated (Spengler Salabarria et al. 2001). The same research group in Cuba, a few years ago conducted a preliminary study of this species under the synonym of Z. eggersiana Urb. and detected seven alkaloids. However, they managed to isolate and identify only three of them—1, 47, and haemanthidine (50) (Trimino Ayllon et al. 1989).

A detailed phytochemical study of a concentrated alkaloidal extract prepared from 35 kg of fresh bulbs of Z. citrina was reported recently (Kohelová et al. 2021). Extensive chromatographic purification provided the isolation of more than twenty AAs of different frameworks, including crinine- (67–70), homolycorine- (39, 40), lycorine- (1, 4, 5, 6, 8, 16, 32), pretazettine- (110), haemanthamine- (47, 48, 50, 54, 59, 62,), and galanthamine- (83) types (Table 1). As this was an acid–base extraction, three alkaloids with a labile hemiaminal group (50, 67–69) were obtained in a mixture of 6-epimers. In addition, a thorough structure elucidation of narcieliine (155), a type of recently discovered narcikachnine, was discussed in this study. These two-nitrogen alkaloids are characterized by fused galanthamine-like and galanthindole moieties (Cahlíková et al. 2021). Narcieliine has two types of chirality. While galanthamine is defined by three stereogenic carbons, galanthindole contains a chiral axial bond connecting two sp² carbons in the 2,2′-disubstituted 1,1′-biphenyl. Therefore, a mixture of diastereomers was characterized for 155. During the NMR study, the increasing energy supplied by temperature (up to 115 °C) led to the overcoming of the sterically hindered bond, which was shown by the coalescence of individual signals in the ¹H NMR spectrum. So far, only five alkaloids of this structural type have been isolated from Narcissus sp. Although atropisomerism as a type of dynamic axial chirality is well known and is quite ubiquitous in medicinal chemistry, only Řezanka and co-workers have described it in the structural study of 6-hydroxygalanthindole in the past (Řezanka et al. 2010). Regarding the structures of AAs, atropisomerism is also found in the carltonine skeleton, in which carltonine A and B have galanthindole condensed with tyramine, while carltonine C-E consist of two galanthindole units (Al Mamun et al. 2020; Křoustková et al. 2022). Therefore, the dimeric derivatives are isolated in a mixture of atropodiastereomers (doubling of signals in the NMR spectra). Since carltonine A and B are considered drug-like butyrylcholinesterase inhibitors, it will be necessary to establish the effect of each enantiomer on the inhibition together with the stability. The energy barrier of rotation will always be an aspect of the evaluation of atropisomers and thus can be a stumbling block (Glunz 2018; Smyth et al. 2015; Toenjes and Gustafson 2018; Zask et al. 2013).

Additionally, the alkaloidal content of the species was studied under the name Zephyranthes sulphurea Noter, yielding known AAs 1, tazettine (110), 47, and maritidine (59) (Rao 1969; Rao and Rao 1979) (Fig. 13).

Phytochemical studies of Zephyranthes concolor (Lindl.) Benth & Hook. F.

Zephyranthes concolor is an endemic species to Mexico (Centeno-Betanzos et al. 2021). Its alkaloids were first described in a general rapid TLC/GC–MS analysis reported in 2008 (Berkov et al. 2008). Maceration of dried, powdered plants (3–5 g) of several different genera followed by acid–base LLE led to the detection of lycorine (1), galanthamine (76, Fig. 9), and chlidanthine (77) in the extract of Z. concolor. Other AAs were present in trace amounts (< 1% of TIC). Later, a detailed phytochemical investigation working with 3 kg of fresh bulbs macerated in EtOH for 15 days led to the isolation of already known alkaloids when using acid–base LLE and either crystallization or CC with silica gel 60 for the purification. Using the same procedures, 77, 76, and galanthamine-N-oxide (88) have also been isolated from 0.9 kg of fresh aerial parts, together with another galanthamine-type alkaloid epinorgalanthamine (78). Neither lycorine nor galwesine was detected in this part (Reyes-Chilpa et al. 2011). It is noteworthy that, except for 76 (272 mg in bulbs, 31 mg in aerial parts), the yields of alkaloids did not differ much according to the quantity of starting material (77—50 mg in bulbs, 30 mg in aerial parts; 88—83 mg in bulbs, 25 mg in aerial parts).

Phytochemical studies of Zephyranthes fosteri Traub

Another endemic plant of Mexico is Zephyranthes fosteri, where it is known as “mayitos”, “flor de mayo” or “quiebra platos” (Centeno-Betanzos et al. 2021). The only phytochemical analysis of Z. fosteri started with the preparation of MeOH extracts of different plant organs. The drying of 3.72 kg of fresh bulbs for 3 weeks gave 0.67 kg of material, which was then crushed and macerated in MeOH for 15 days (three times). Evaporation afforded 71 g of extract, which was subjected to acid–base LLEs providing 2.39 g of alkaloid fractions (EtOAc, EtOAc/MeOH). As expected, lycorine (1) crystallized directly. The only other isolated alkaloid was 3′-demethoxy-6-epimesembranol (151) (open CC and preparative TLC—both on Si₂O). The rest of the alkaloids in the EtOAc extract were detected only by GC–MS analysis. Other constituents in MeOH extracts of roots, bulbs, aerial parts, flowers, and leaves were analyzed by GC–MS (then fatty acids, sterols, and another two alkaloids in the EtOAc extract). Interestingly, although 151 was isolated from the bulbs' EtOAc extract, it was not detected in the MeOH extract of the bulbs (Centeno-Betanzos et al. 2021).

Phytochemical studies of Zephyranthes itaobina (Ravenna) Nic. García

Synonyms: Habranthus itaobinus Ravenna

A collective from Brazil and Spain described the isolation of five alkaloids from the MeOH extract of H. itaobinus (43, 48, 100, 109, 110), which was obtained from 1.7 kg bulbs (Cole et al. 2019).

Phytochemical studies of Zephyranthes minuta (Kunth) D. Dietr.

Synonyms: Zephyranthes grandiflora Lindl.

Zephyranthes minuta is native to Mexico and Guatemala, but widely cultivated as an ornamental in Hawaii, the Andaman Islands, the islands of the south-western Caribbean (Dietrich 1839; RHS 2023), and China (Katoch et al. 2013). The first report presented a bioassay-guided investigation focusing on the cytotoxicity of an extract that resulted in the isolation of pancratistatin (92) from 68 kg of bulbs(!); however, its structure was confirmed only by TLC and IR (Pettit et al. 1984). Two MS-based studies of bulbs proposed a tentative constitution of 20 AAs belonging to lycorine-, α-,β-crinane-, narciclasine-, pretazettine-, and galanthamine-type (Table S1) (Cahlíková et al. 2011; Katoch et al. 2012). Apart from misleading information in the proposed structures, and errors in the drawings and naming, these predictions can still be compared with those of the four works that actually isolated specialized metabolites and provided a proper structure elucidation. Starting in 2012, 250 g of crude extract of dried bulbs (2 kg) obtained by percolation with EtOH for 24 h was partitioned with solvents of different polarity without any acid or base, and the n-BuOH fraction was selected for the first CC (Katoch et al. 2013). Lycorine (1) crystallized directly, and further CC over silica gel using 5–7% of MeOH in TCM gave galanthine (6), lycoramine (83), hamayne (52), and haemanthamine (47). Ionized molecules ungeremine (28), tortuosine (32), and zephgrabetaine (31) were obtained through further silica gel CC with 10% of MeOH in TCM. A study by the same research group published in 2020 (Katoch et al. 2020) probably continued with the remaining fractions of the same plant material (however, now calling the species by a synonym, Z. minuta) and described a crystallization of 100 mg of previously reported 92, which seems to be the most abundant alkaloid in this species, and, therefore, explains the isolation target of the first work in 1984 (Pettit et al. 1984). As discussed above, the apparently mild conditions of the phytochemical procedures without acid or base substitution allowed the isolation of more labile components, i.e., the O-glycoside, narciclasine-4-O-β-d-xylopyranoside (98), and the ester, 1-O-(3-hydroxybutyryl)pancratistatin (94) (Fig. 10) (Katoch et al. 2020).

On top of that, there are two recent phytochemical studies by different research groups from China who worked with 10 kg of the whole plant (Yang and co-workers did not mention whether fresh or dried, which is important information since it can have a substantial impact on the composition of the isolated metabolites) and reported new compounds with plicamine and secoplicamine scaffolds (Wang et al. 2018; Yang et al. 2022). They are derivatives of alkaloids isolated from Z. carinata and Z. candida in previous years. Also, these works adopted the non-acid–base extraction. In particular, Wang and co-workers obtained over 50 mg of each new derivative, which is quite a good yield (Wang et al. 2018).

Comparing all the results regarding constituents of Z. grandiflora, the MS-based papers did not report any plicamine or secoplicamine alkaloids.

Phytochemical studies of Zephyranthes robusta (Herb.) Baker

Synonyms: Habranthus robustus Herb.

Zephyranthes robusta is believed to have originated from Rio Grande do Sul in Brazil. It is native to Brazil, Argentina, and Uruguay, but now is naturalized in Florida, Colombia, South Africa, and Mauritius (POWO 2023). Z. robusta is widely grown as an ornamental. It is one of the most prolific summer flowering rain-lilies. Pilot phytochemical studies described the presence of three common AAs, lycorine (1), haemanthamine (47), and maritidine (59) (Rao 1969). However, the identification at the time relied on TLC, IR, melting point, and optical rotation. Later, in 2010, GC–MS profiling detected seven AAs, and only 47 from the previous reports was listed (Cahlíková et al. 2010) (Table 1, Fig. 8). Subsequent phytochemical study on 12 kg of fresh bulbs involving acid–base LLE resulted in the isolation of fourteen AAs. Alkaloids suggested by GC–MS analysis were isolated, except for nerbowdine (S92). It is noteworthy that although the constitution of this alkaloid is also known as buphantine, nothing can be found when searching for reliable analytical data. Published specific rotation, melting point, IR, and UV spectra, however, have no informative value in this case. So, this just illustrates one of the problems with previously reported natural compounds. The main alkaloid was galanthine (6) (780 mg), which was found in twice the amount as lycorine (1) (350 mg) (Kulhánková et al. 2013). Since the isolation yielded hippeastidine (75) with the same molecular weight and similar structure as S92, perhaps the interpretation of the suggestion from the GC–MS database may have created such confusion. Another interesting thing is associated with 8-O-demethylmaritidine (62). Additionally, the name for 9-O-demethylgalanthine (8) has been revised along with the NMR data. This compound, first isolated from Habranthus brachyandrus (Baker) Sealy, was named 10-O-demethylgalanthine (Jitsuno et al. 2009). Recently, alkaloidal extracts of different parts of Z. robusta (leaves, bulbs and roots collected during flowering, and bulbs and roots collected in the pre-flowering stage) were analyzed by GC–MS (Shawky et al. 2021). Most of the identified alkaloids were in agreement with previous studies (Table S1).

The research group of Mikšátková and co-workers elucidated the content of isoflavonoids in several Amaryllidaceae plants, including Z. robusta bulbs. Based on comparison with standards using HPLC–MS, several of the isoflavonoids were identified (Table 2, Fig. 14) (Mikšátková et al. 2014).

Table 2 Overview of other secondary metabolites isolated from Zephyranthes species

Fig. 14

Isoflavonoids isolated from Zephyranthes robusta

Phytochemical studies of Zephyranthes rosea Lindl.

Zephyranthes rosea is a species of rain-lily native to Peru and Colombia (Acevedo and Strong 2005). The plants are widely cultivated as ornamentals and have become naturalized in tropical regions worldwide. The alkaloids of this plant were studied in the 1950s, when Boit reported finding lycorine (1) and galanthamine (76) in 0.2 kg of bulbs (Boit et al. 1957). Later, in 1984, Ghosal and co-workers described the isolation of crinamine (51), haemanthamine (47), (+)-epimaritidine (60), and maritidine (59), when working with ca. 1 kg of fresh bulbs. On top of that, an epimerization of 59/60 was described as associated with prolonged heating in the presence of aqueous HCl, but no acid–base LLE was employed in this case (Ghosal and Ashutosh 1985). The most recent work confidently reports the detection of thirteen AAs by GC–MS analysis of leaves and bulbs of Z. rosea (Shawky et al. 2021). Not surprisingly, 76, and 47 were detected as major alkaloids of both extracts (Table 1). Mikšátková and co-workers in 2014 tried to elucidate the content of isoflavonoids in Z. rosea bulbs, but their content did not exceed the limits of quantification (Mikšátková et al. 2014).

Phytochemical studies of Zephyranthes tubispatha Herb.

Synonyms: Z. tubispatha–Z. texana Herb., Habranthus tubispathus (L'Hér.) Herb.

This Amaryllidaceae plant, found in southern Brazil (WFO 2023), was studied by Cavallaro and co-workers in 2014 (Cavallaro et al. 2014). A kilo of fresh bulbs identified as H. tubispathus was cut into pieces, macerated in EtOH for 2 weeks, and then boiled for 3 h. The dry extract, weighing 73 g, was further separated by acid–base LLE (2% HCl for acidifying, NaHCO₃ for taking to pH 9) and extracted with DCM to obtain an alkaloidal fraction (280 mg). Fresh bulbs (544 g) of H. jamesonii (Baker) Ravenna (accepted name in POWO: Zephyranthes jamesonii (Baker) Nic. García & S.C.Arroyo, discussed below in the GC–MS section) were treated in the same way to obtain a dry alkaloidal extract of 150 mg. A further 162 g of aerial parts of H. jamesonni was also processed in the same way to yield 24 mg of alkaloidal extract. From 226 mg of the alkaloidal extract of H. tubispathus, three alkaloids were isolated by CC over silica gel and preparative TLC. These were lycorine (1) (7.4 mg), hippeastidine (65) (5.1 mg), and 3-O-demethylhippeastidine (66) (3.6 mg). GC–MS analysis of the alkaloidal extracts indicated the presence of further alkaloids (trisphaeridine (100), galanthamine (76), sanguinine (81), vittatine (48), narwedine (S102), maritidine (59), montanine (153), haemanthamine (47), 3-O-demethylhippeastidine (66), anhydrolycorine (17), haemanthidine (50), 8-O-demethylhomolycorine (36), and pseudolycorine (7) and compared their percentage representation (bulbs of H. tubispathus × bulbs of H. jamesonii × aerial parts of H. jamesonii).

However, the older work dedicated to the isolation described the identification of 1, nerispine, powelline (72), and tubispacine (73) (Doepke 1963). For the latter, the structural elucidation was described 2 years later by Doepke (Doepke 1965). Since the articles did not disclose the structure of nerispine, we searched the literature but found nothing—no other experimental paper mentioned the structure of nerispine or any information about the compound. However, we can be sure that 1 was isolated as it was already known from a paper written by Greathouse and Rigler in 1941 presenting its isolation from bulbs (1–2 kg) of this plant under its synonym Z. texana (Greathouse and Rigler 1941). Additionally, narciprimine (103) and narciclasine (96) were isolated by Spengler and Trimino in 1989 (Spengler Salabarria and Trimino Ayllon 1989).

Phytochemical studies of Zephyranthes cv. Ajax

Hybridization is common in the Amaryllidaceae family. It occurs spontaneously or can be artificially induced driven by financial profit or purely out of curiosity. The cultivar Zephyranthes 'Ajax' is a cross of Z. candida × Z. citrina. Unfortunately, the phytochemicals cannot be simply compared with the parent species due to the lack of reports of its crossing. The only project dealing with a horticultural cultivar and not a botanical species of Zephyranthes was published in 2020, reporting specialized metabolites isolated from Z. cv. Ajax Hort. collected in Vietnam (Nguyen et al. 2020). Maceration of bulbs in MeOH (5.5 kg; it is not clear whether they were dried or fresh) gave 467 g of extract, which was further subjected to LLE. The increasing polarity of solvents used for the extraction yielded 120.3 g of DCM fraction, 126.8 g of EtOAc fraction, 43.5 g of n-BuOH fraction, and 140.3 g of water fraction. Separation by CC (silica gel, C-18 coated silica gel, Sephadex LH-20) and preparative reverse phase HPLC (C-18 column) resulted in the isolation of haemanthamine (47, 210 mg) as the only alkaloid reported. To our surprise, neither lycorine nor galanthamine was detected. Furthermore, this work additionally reports the isolation of eight flavonoids in very low yields: (2R,3R)-3-acetoxy-7-hydroxy-3′,4′-methylenedioxyflavan (176), 7-hydroxyflavan (181), 7,4′-dihydroxyflavan (182), 7,4′-dihydroxy-8-methylflavan (183), 7,3′-dihydroxy-4′-methoxyflavan (164), 5,4′-dihydroxy-7-methoxy-6-methylflavan (185), 7-hydroxy-3′,4′-methylenedioxyflavanone (186) (Table 2, Fig. 15). Interestingly, no flavonoid glycosides were obtained.

Fig. 15

Flavonoids isolated from Zephyranthes sp.

The composition of the metabolites obtained from this hybrid plant was different from those of either Z. candida or Z. citrina. However, the processing was different with no acid–base LLE.

Phytochemistry and chemical constituents of other species formerly classified in the genus Zephyranthes

Phytochemical studies of Pyrolirion flavum Herb.

Synonyms: Zephyranthes flava Baker

The only research group that has studied the constituents of this species published three papers between 1985 and 1987, in which the results using different parts of the plant were reported. Each work presents a different approach to extract processing. Various flavans and flavan-7-O-glucosides were reported when working with 0.5 kg of dried, milled bulbs (no acid–base LLE). Generally speaking, simple flavans having a 2S configuration are characterized by levorotatory negative optical rotation, which corresponds with the absolute configuration outlined in a report by de Rezende and co-workers (de Rezende et al. 2015). Surprisingly, among others, betaines (27–30) (Ghosal et al. 1986) and glycosides (2, 10) were obtained from 200 g of seeds after acid–base LLE and further purification. The last paper is dedicated to alkaloids from Z. flava flowers and describes interesting conjugates. Results from elemental, MS, and NMR analyses supported the structure elucidation of alkaloidal phospholipids obtained by CC, preparative TLC, and HPLC. All of them were essentially 2-O-lycorine esters (21–26). The most abundant alkaloidal phospholipid, of which 112 mg was isolated from 1 kg of fresh flowers, was the least substituted glycerophosphoryllycorine (21). Other derivatives had the glycerol hydroxyls substituted with palmitic, stearic, and/or oleic acid with corresponding molecular weights of around 1000 Da (Ghosal et al. 1987) (Figs. 16, 17).

Fig. 16

Ceramides isolated from Zephyranthes candida

Fig. 17

Other non-alkaloidal compounds isolated from Zephyranthes sp.

GC–MS profiling of plants classified as Zephyranthes sp. or their synonyms

GC–MS profiling of Pyrolirion albicans Herb.

Huaylla and co-workers studied the alkaloidal profile of bulbs and leaves simply by using GC–MS (Huaylla et al. 2021). Dried and powdered plant material was macerated in MeOH, and the alkaloid content was separated from non-polar compounds by acidification with 2% H₂SO₄ and extraction with TCM. After adding NH₄OH solution, the alkaloids were partitioned in TCM. The highest content was found for montanine (153) in the bulbs, while other known alkaloids such as galanthamine (76) and lycorine (1) were also present in the leaves.

GC–MS profiling of Zephyranthes alba Flagg, G. Lom. Sm. & García-Mend.

Only one paper mentions the analysis of secondary metabolites of Zephyranthes alba (Centeno-Betanzos et al. 2022). It also describes the nooks and crannies of the complexity of the identification problems with the genus Zephyranthes. Morphological, genetic, and chemical markers, all together, are described to be a reliable source for species identification. The aim of the work was therefore to demonstrate the differences in the GC–MS profile of the alkaloidal extracts of Z. alba and Z. fosteri. The latter showed a complex AAs composition. On the other hand, in Z. alba, mainly galanthamine type AAs were detected. However, we have to take the report as a prediction rather than a rock-solid truth. Isolation and structure elucidation would be necessary to confirm the proposed alkaloidal profile of Z. alba.

GC–MS profiling of Zephyranthes filifolia Herb. ex Kraenzl., Zephyranthes gilliesiana (Herb.) Nic. García, Zephyranthes graciliflora (Herb.) Nic. García, Zephyranthes jamesonii (Baker) Nic. García & S.C.Arroyo

Synonyms: Z. gilliesiana–Rhodophiala mendocina (Phil.) Ravenna

• Z. graciliflora–Phycella herbertiana Lindl.

• Z. jamesonii–Habranthus jamesonii (Baker) Ravenna

Unfortunately, almost no phytochemical isolation has been reported for these species. There is only one GC–MS profiling report describing Z. filifolia and Z. jamesonii as a source for galanthamine (76) and tazettine (110). Extracts were prepared from 100 g of dried bulbs and processed by acid–base LLE using 2% H₂SO₄ followed by 25% NaOH, which are relatively harsh conditions (Ortiz et al. 2012). So far, no detailed analysis supported by the isolation of individual constituents has been reported.

When it comes to nomenclature, a problem appears for Z. gilliesiana. POWO and WFO take a different stance on this species. POWO recognizes it as two different plants (Zephyranthes gilliesiana (Herb.) Nic. García and Zephyranthes elwesii (C.H.Wright) Nic. García). WFO classify all the synonyms under one name—Rhodophiala gilliesiana (POWO 2023; WFO 2023).

GC–MS profiling of Phycella chilensis (L'Hér.) Grau ex Nic. García, Rhodolirium andicola (Poepp.) Ravenna, Zephyranthes araucana (Phil.) Nic. García, Zephyranthes montana (Phil.) Nic. García, Zephyranthes splendens (Renjifo) Nic. García, Zephyranthes bagnoldii (Herb.) Nic. García

Synonyms: Ph. chilensis–Rhodophiala pratensis (Poepp.) Traub

• R. andicola–Zephyranthes andicola Baker

• Z. araucana–Rhodophiala araucana (Phil.) Traub

• Z. montana–Rhodophiala montana (Phil.) Traub

• Z. splendens–Rhodophiala splendens (Renjifo) Traub

• Z. bagnoldii–Rhodophiala bagnoldii (Herb.) Traub

Zephyranthes andicola was, together with four other Chilean Amaryllidaceae plants, studied for its alkaloidal profile by GC–MS (Tallini et al. 2018). The aerial parts and bulbs were separated and freeze-dried before repeated MeOH extraction under sonification. The crude extracts were acidified by 2% H₂SO₄, while neutral compounds were removed with Et₂O. After basification with NH₄OH, the alkaloids were extracted with EtOAc ready for analysis. The results are quite surprising. In particular, Z. andicola stands out for its content of O-methyltazettine (S137), which exceeds even the lycorine- and haemanthamine-type compounds, whose content was dominant in the other species.

The alkaloidal content of Ph. chilensis was studied by GC–MS and is reported in another three articles (Lizama-Bizama et al. 2018; Trujillo-Chacón et al. 2019a, b). As far as alkaloids are concerned, nothing new was reported. However, our attention was caught by the fact that they were probably searching the same plant recognized in POWO as Phycella chilensis and in WFO as Rhodophiala chilensis. The articles used different names—(Trujillo-Chacón et al. 2019b): Rhodophiala pratensis and Rhodolirium speciosum; (Lizama-Bizama et al. 2018): Rhodophiala pratensis and Rhodophiala volckmanii. Analyzing the profiling results, only Trujillo-Chacón and co-workers report an alkaloidal content in which the samples differ slightly.

Conclusion

This review summarizes the complete phytochemical information about plants of the genus Zephyranthes and about genera that have been taxonomically included in this genus in the past and are currently synonymous with it. So far, 15 species of this genus have been fully studied phytochemically, and nearly 160 AAs of different structural types have been isolated and identified, together with non-alkaloidal metabolites. However, some articles need to be taken with a grain of salt. Especially those offering only GC–MS profiling. Some species have only been screened by this approach.

The presence and structures of some reported AAs should be re-evaluated, as they may be isolation artifacts and not naturally occurring compounds. However, we believe that researchers should not be afraid to publish compounds formed during isolation, as they can still bring new insights to drug discovery. But they should mention that it is an artifact and suggest the conditions of its formation. Since only conventional methods of extraction and isolation are mentioned in the Zephyranthes literature, it is difficult to determine which procedure leads to better results. Indeed, when using acid–base LLE or extraction under reflux, the probability of generating an artifact on the reactive side is higher. In any case, work using MeOH as a solvent in such a procedure is unlikely to identify potential artifacts, as methoxy groups are accepted substituents in natural products, unlike, for example, ethoxy or chloromethyl groups.

When it comes to structural analysis, we recommend using more supporting data than less. Building structure (known or unknown) only on GC–MS fragmentation brings more easily achievable data but has no added value because the existence of such a molecule is only hypothetical. We appeal to researchers to use NMR analysis supported by chiroptical methods (CD analysis, polarimetry) when characterizing a chiral compound, also for comparison reasons.

As several works have shown, the distribution of alkaloids in plant parts is similar. But bulbs are usually processed in larger quantities, so when working with large-scale material, the discovery of a new molecule is more likely.

In the light of the presented overview of scientific data, the genus Zephyranthes can be recognized as an interesting source of different structural types of AAs. Due to the complex nomenclature of this genus (and related genera), future research should, among other things, focus on the correct taxonomic classification, indicating which database it is based on. Nevertheless, more comprehensive studies of this genus are needed to better understand its content and further applications.