1. Introduction
Polyscias filicifolia (C. Moore ex E. Fourn) L.H. Bailey is a representative of the Araliaceae family.
P. filicifolia is a species that shares the genus
Polyscias with between 116 [
1] and 150 others [
2]. It is an evergreen, six-meter-tall shrub that naturally occurs in Southeast Asia, but is also available as a decorative potted plant [
3].
Polyscias spp. extracts are known to possess a number of pharmacological properties, including antimicrobial, immunostimulant and cytotoxic activities [reviewed in 1]. Approximately 97 compounds of different chemical classes have been identified in selected
Polyscias species; however, the main groups of compounds responsible for pharmacological effects are believed to be oleanolic triterpene saponins, present in the leaves and roots, and polyacetylenes, which only occur in the roots [
4,
5].
Polyscias spp. extracts may also contain polyphenols, flavonoids, glycosides, saponins, cyanogenic glycosides, sterols and alkaloids [
1,
4,
6,
7]. In Southeast Asia, plants of
Polyscias genus are broadly used in traditional medicine; for example, it is recorded in the official Vietnamese Pharmacopoeia as an anti-fatigue and cardiac drug [
8].
As
Polyscias plants demonstrate relatively slow growth in natural conditions, and their natural range is limited to tropical and subtropical areas, there is a need for biotechnological approaches to provide a more sustainable source of plant material. Śliwińska et al. [
3] developed an efficient method of micropropagation of
P. filicifolia via somatic embryogenesis from the apical meristems of the shoots. In addition, polyphenolic compound production was found to be enhanced in the shoot cultures treated with elicitors such as salicylic acid (SA) and/or methyl jasmonate (MeJA) [
9]. In addition, shoot extracts derived from in vitro cultures were found to reduce cancer cell viability and increase their mortality, and to exert a protective effect on healthy cells. In subsequent investigations, shoot extracts of
P. filicifolia derived from in vitro cultures also exhibited anti-genotoxic and anti-photogenotoxic activities [
10].
In vitro culture could serve as a reliable source of plant material for diverse applications in science and industry, thus protecting natural resources and ensuring reproducibility of desired specific genetic and physiological features. For many medicinal and endangered plants, various methods of regeneration have been described [
11,
12,
13]. The most frequently used methods of plant regeneration are somatic embryogenesis and organogenesis. The former allows a vast number of regenerants to be obtained; however, it may result in genetic changes such as chromosome breakage and rearrangement, increased ploidy or the exchange of bases in the DNA sequence [
11,
13,
14,
15,
16]. It is also characterized by a greater chance of epigenetic modifications, e.g., at the DNA methylation level [
17,
18]. Moreover, epigenetic reprogramming of embryogenic cells may result in hereditable, but potentially reversible changes, leading to chemical modification of DNA and histone proteins, including chromatin remodeling. It is believed that although epigenetic changes are not encoded by DNA, they can still be passed on to the next generation [
13,
19].
Various markers such as Simple Sequence Repeat (SSR), Randomly Amplified Polymorphic DNA (RAPD), Restriction Fragment Length Polymorphism (RFLP), Inter Simple Sequence Repeats (ISSR) and Amplified Fragment Length Polymorphism (AFLP) are used to assess the homogeneity of plant material obtained by somatic embryogenesis and/or organogenesis [
20]. Of these markers, AFLP is recommended because it allows a large number of polymorphisms distributed throughout the genome to be tested; it is also a relatively easy method, and requires only small amounts of DNA. This method has been used for the detection of somaclonal variants, among others [
21,
22,
23,
24,
25].
Other methods for detecting changes at the genetic and epigenetic level are Methylation-Sensitive Amplification Polymorphism (MSAP) [
20] and Methylation-Sensitive Amplified Fragment Length Polymorphism (metAFLP) [
26]; the latter allows simultaneous analysis of changes in genetic, i.e., in DNA base sequence, and epigenetic, i.e., base methylation, properties. This method is based on the action of two restriction enzymes which differ in their sensitivity to DNA methylation at the recognized site:
Acc65I, which is sensitive to methylation, and
KpnI, which is not. Based on the resulting metAFLP profiles produced by the two enzymes (
KpnI/
MseI and
Acc65I/
MseI), it is possible to identify methylation; in addition, the
KpnI/
MseI enzymes can determine variation at the sequential level.
The AFLP profiles produced by metAFLP can be used to identify somaclonal changes induced by in vitro cultures and calculate quantitative characteristics. This technique has been used to identify somaclonal variation of
Hordeum vulgare [
18,
26],
Gentiana pannonica [
27],
Gentiana cruciata [
28], Triticale (x
Triticosecale spp.) regenerants [
29,
30] and
Arabidopsis thaliana regenerants [
31].
This is the first study to assess the morphological, genetic and epigenetic variation of Polyscias filicifolia plants derived by three regeneration methods: indirect somatic embryogenesis (PSE), direct somatic embryogenesis from primary somatic embryos (DSE) and direct somatic embryogenesis from secondary somatic embryogenesis (TSE). Genetic and epigenetic variations were investigated using AFLP and metAFLP. Further, biochemical variation of resulting regenerants cultivated in vitro was studied by the analysis of the triterpenoids oleanolic acid, ursolic acid and hederagenin and was compared to nonembryogenic and embryogenic callus tissues, TSE regenerants cultivated ex vitro as well as donor plant using gas chromatography.
3. Discussion
Somatic embryogenesis is an alternative, vegetative method of clonal multiplication used in many seed-grown plants that are difficult to propagate, as well as endangered species and those with medicinal properties or are otherwise useful [
12,
32,
33]. As plant regeneration by somatic embryogenesis is an asexual process that only involves mitotic cell division, the expected result is to obtain genetically uniform plants. However, some plants obtained by in vitro culture procedures nevertheless demonstrate some degree of somaclonal variability (SV) due to changes arising in cell culture. Hence, somatic embryogenesis, especially procedures based on indirect somatic embryogenesis, carries the risk of occurrence of phenotypic, genetic and epigenetic changes in the received regenerants.
Therefore, to confirm the efficacy of embryogenesis, it is necessary to conduct a morphological assessment of the regenerants, together with and paying attention to the structural features of the leaves, flowers and fruit. In addition, the plant material should be analyzed at the genetic level to identify changes in DNA sequence. However, as changes in the regenerant genome may also occur following changes in the level of DNA methylation, these should also be confirmed using methylation-sensitive markers [
15,
21,
34,
35].
This work describes the first such evaluation of
Polyscias filicifolia, with the aim of determining the extent to which plant material from a donor plant can change after being introduced into in vitro culture. Our PCA analysis found that the samples from PSE, DSE and TSE groups could be separated into three groups based on methylation levels. The PSE regenerants demonstrated more variation than those from the DSE and TSE groups. This phenomenon could be explained by the fact that PSE regenerants originated from callus tissue which is known to be less genetically stable than organized tissues of somatic embryos, the source of explants for DSE and TSE. This can be attributed to the selection of competent cells involved in embryo formation using a specific genetic program [
36,
37] and the reduction in variation could be due to the so-called bottleneck effect [
14].
In the current study, the molecular data and morphological data are consistent, and suggest that the PSE, DSE and TSE regenerants generally form distinct groups; even so, some degree of variability and intermixing of regenerants was still observed. As molecular markers are not necessarily related to morphological traits or may not reflect genetic (or epigenetic) variation, it is understandable why such markers resulted in less pronounced differentiation between groups. However, the PSE, DSE and TSE were also subjected to molecular assessment using the AFLP method, which has been successfully used in studies of the homogeneity of plants regenerated by somatic embryogenesis and organogenesis in vitro [
21,
22,
23,
25,
38,
39,
40].
In the present study, the AFLP analysis was applied using eight pairs of selective primers, which allowed the identification of 428 bands, with polymorphic bands representing only 1.2%. A similarly high level of homogeneity (98.8%) was previously obtained in plant culture obtained by somatic embryogenesis of
Bamboo nutans [
25], where AFLP based on six selective primers led to the identification of 407 bands, with polymorphic bands accounting for 1.17%. AFLP testing has also revealed very high levels of genetic homogeneity in
Fressia hybrida regenerants [
41], as well as in
Coffea arabica plants for two hybrids, using embryogenic suspension (ESP) and secondary somatic embryogenesis (SCE) as propagation systems [
39]. The ESP-derived plants showed no AFLP polymorphism compared to mother plants and ranges of 0.003%.
However, some reports indicate the occurrence of higher levels of genetic variability during the development of somatic embryos. For comparison, an AFLP study of 23 rye regenerants obtained from somatic embryos allowed the identification of polymorphism frequency of 8.8%. High levels of polymorphism were obtained independently in plants from different cell lines, revealing potential mutational hot spots: independent mutational events were found to occur in the same genome regions of these plants [
22].
The AFLP method also proved to be a sensitive and reliable molecular marker for detecting possible somaclonal changes in the
Echinacea purpurea micropropagation system. Research on 40 regenerants obtained from leaf organogenesis and five donor plants, using eight primer pairs, detected 3805 fragments, of which 301 (9.40%) were polymorphic. The mean percentage of polymorphic fragments in the five donor groups ranged from 1.6% to 20.6% [
23]. Studies suggest that the variability observed in plants regenerated from somatic embryos may be due to changes in the level of DNA methylation; this process may lead to changes at the epigenetic level, which, unlike genetic changes, do not depend on the DNA sequence and may be responsible for changes in gene function or expression [
15,
16,
18,
19]. DNA methylation is one of the basic, hereditary epigenetic features that determine the silencing of specific DNA sequences.
The process of somatic embryogenesis is regulated by many factors, including DNA methylation [
16,
42,
43]. In embryogenic cells, the level of DNA methylation is modified by auxin activity and in vitro conditions; the methylation level in turn affects the expression patterns of the genes involved in SE. In the present study, the
P. filicifolia regenerants demonstrated a significant range of methylation, which may be associated with culture stress, particularly the addition of auxin, i.e., 2,4-D, and the number of passages, i.e., the duration of culture. A review of the literature shows that 2,4-D, and auxins in general, play an important role in the induction of somatic embryos in many plants, and their use can result in a range of epigenetic and genetic changes in cells, including methylation and mutations in DNA [
15,
44,
45,
46]. In addition, the presence of high concentrations of 2,4-D can contribute to numerous disorders in embryo development and disrupt normal genetic and physiological processes in cells [
46,
47,
48]. It has also been shown that 2,4-D and stressful environmental conditions posed by in vitro culture affect the processes of histone modification, chromatin structure rearrangement and DNA methylation, which in turn can contribute to changes in gene expression and thus interfere with the development of somatic embryos [
43,
45,
49]. In the present study, to obtain
P. filicifolia regenerants by indirect somatic embryogenesis, it was necessary to first induce a nonembryogenic callus using a relatively high concentration of 2,4-D and sucrose (6%).
The present study evaluated the variability of three groups of P. filicifolia regenerants, viz. PSE, DSE and TSE, and the donor plant based on the metAFLP method. It allows simultaneous detection of changes in DNA sequence and methylation level based on the properties of KpnI/MseI and Acc65I/MseI restriction enzyme pairs, which vary in cytosine methylation sensitivity. In the current study, metAFLP analysis allowed the identification of 849 bands for both pairs of restriction enzymes (Acc65I/MseI and KpnI/MseI), with the level of variability found to be 4.74% for the KpnI/MseI enzymes, and 3.51% for the Acc65I/MseI enzymes.
metAFLP analysis has also been used to detect epigenetic variability in various regeneration systems of plants grown in vitro, such as in
Hordeum vulgare [
50], as well as
Gentiana cruciata regenerated from somatic embryos that have previously undergone short-term and long-term cryopreservation [
28]. It was also used to examine tissue culture-induced variability of triticale regenerants derived from four different genotypes using androgenesis and somatic embryogenesis [
29]. It also identified changes induced by tissue culture conditions, including sequence variations and changes in methylation patterns, in
Gentiana pannonica plants derived from somatic embryogenesis; in this case, tissue culture-induced variation was found to be 16% [
27]. The metAFLP method was also effective at detecting variability at the DNA sequence and methylation levels in
Arabidopsis thaliana plants obtained from seeds, somatic embryos and modified transformed plants: the modified plants derived from somatic embryos were found to demonstrate a higher level of variation in methylation (7.5%) compared to controls (3.2%) [
31].
Our present metAFLP analysis using
EcoRI/
MseI,
KpnI/
MseI and
Acc65I/
MseI found a number of variations between the PSE, DSE and TSE regenerants and the donor plant. In total, 24 pairs of selective primers identified 1277 DNA fragments, of which 3.13% constituted polymorphic bands. This small number of obtained polymorphic bands could be explained by the way the cultures were conducted: the nonembryogenic and embryogenic callus cultures, somatic embryos and subsequent induction of secondary embryos were initiated by direct somatic embryogenesis from a single donor plant; this could, to a large extent, narrow the pool of variability among the in vitro regenerants. Gao et al. [
40] reported similar findings in a study of
Fressia hybrida cultures also initiated by somatic embryogenesis from a single donor plant to minimize existing heterozygosity or natural explant mutations; they conclude that the genetic and epigenetic variability observed in their regenerants was the result of tissue culture stress.
In the present study, the UPGMA and PCoA cluster analysis found that the donor plant and PSE, DSE and TSE regenerants formed distinct groups. Nevertheless, it should be noted that PSE regenerants obtained through the callus phase constituted a more diverse group, as did DSE regenerants obtained from PSE embryos by direct somatic embryogenesis. The most compact clusters were TSE regenerants, which were obtained from DSE embryos by direct somatic embryogenesis.
These findings correlate with the results of the AMOVA, which found that 21% of variance corresponded to the variability “between” groups of regenerants and 79% to variability “within” the groups. However, no such clear grouping of
Coffea arabica regenerants obtained by direct (DSE) or indirect somatic embryogenesis (ISE) was revealed in a previous UPGMA analysis by Sanchez-Teyer et al. [
50]. Based on an AFLP analysis of DNA stability in regenerants, their results indicate that the somatic embryogenesis induces rearrangements at the DNA level, and revealed discrepancies between the mechanisms involved in the process of indirect and direct somatic embryogenesis. In contrast, no such regenerant variation was observed in cultures of S
mallanthus sonchifolis [
51],
Olea europaea [
52] or
Coriandrum sativum [
53], despite the fact that the regeneration process also went through the callus phase.
In the present study, it is possible that, apart from growth regulators, the greater diversity in the PSE
P. filicifolia embryos could be due to the relatively long cultivation time: a total of 88 passages were performed between callus initiation and the creation of the somatic embryos. Indeed, plants regenerated by callus phase in long-term cultures have shown high variability in previous studies; for example, numerous embryos with one cotyledon or multiple cotyledons were observed in
Quercus suber culture [
54].
Somaclonal variation has also been observed in somatic seedlings derived from long-term cell cultures of
Coffea arabica, i.e., 11 and 27-month-old calli; although all plants derived from younger cell cultures, i.e., four months, exhibited normal phenotypes, the variation in the latter group affected almost all the regenerated plants. However, neither MSAP nor AFLP analyses of somatic seedlings indicated any changes on the genetic level. Similarly, no genetic or epigenetic change was demonstrated in
Miscanthus ×giganteus during long-term shoot cultivation [
55].
It should be noted that in long-term cultures, regenerants can demonstrate variation not only at the morphological or genetic level, but also at the epigenetic level. Studies of 12-month
Elaeis guineensi suspension cultures obtained from callus tissue originated from zygotic embryos found that cell proliferation in vitro induced DNA hypermethylation in a time-dependent manner [
56].
In addition, RAPD, REMAP (retrotransposon microsatellite amplified polymorphism) and MSAP analysis of
Humulus lupulus plants grown in vitro for 2 years did not reveal differences between control (field cultivated) and treated plants (cultivated in vitro) at the genetic level, even after 12 cycles of micropropagation; however, variation was detected at the epigenetic level between plants collected from field cultivation and plants growing in vitro. Among all the plants from the in vitro culture group, almost 30% of the detected fragments demonstrated the same pattern of change [
57]. The authors conclude that the variation detected at the epigenetic level is associated with in vitro culture conditions.
In turn, long-term secondary somatic embryogenesis cultures of
Theobroma cacao were found to demonstrate lowered embryogenic potential [
58]. Global DNA methylation levels appeared to increase in the older somatic embryos during long-term in vitro culture, and DNA methylation increased during SE expression. It was also found that aging embryos could regain lost embryogenic potential after treatment with 5-azaC at a certain concentration.
In the current study, analysis of the second group of DSE regenerants, i.e., those obtained by direct somatic embryogenesis from primary embryos (PSE), demonstrated relatively large diversity compared to the TSE group, i.e., those obtained by direct somatic embryogenesis from DSE.
Regeneration of
Psidium guajava plants by direct somatic embryogenesis without callus phase was found to provide genetically stable material [
59]. In turn, Rani and Raina [
60] suggest that plants regenerated from meristems or by direct somatic embryogenesis maintain plant genetic integrity with the least risk of genetic variation. Moreover, Carra et al. [
61] report obtaining homogeneous plant material from
Anthurium andraeanum cv. Fantasia by direct somatic embryogenesis. Furthermore, Gao et al. [
41] report a much higher level of genetic variation in
Fresia hybrida plants obtained by direct (0.97%) rather than indirect somatic embryogenesis (0.27%), based on the AFLP method. In turn, MSAP analysis revealed changes in cytosine DNA methylation in both CG and CNG levels compared to the donor plant. The authors suggest that tissue culture induces stress that may cause some hereditary epigenetic variations in flowering plants. In addition,
Coffea arabica plants regenerated by somatic embryogenesis were found to demonstrate very high similarity with the mother plants, suggesting that the process of somatic embryogenesis is genetically stable and has a mechanism for selecting competent cells [
62].
In the present study, the second aim was to compare the biochemical variation of the triterpenoid compounds expressed by in vitro and ex vitro TSE with that of nonembryogenic, embryogenic callus tissue, as well as that of the donor plant. Previous phytochemical studies of
Polyscias spp. have mainly focused on the isolation and identification of the structure of isolated triterpene saponins and the identification of their aglycones. Further, production of phenolic acids as well as oleanolic acid in shoot cultures of
P. filicifolia was studied [
3,
10,
63]; however, UA and HE accumulation was not investigated yet.
As oleanolic acid is not present in its free state, the most efficient extraction conditions were determined. Previous studies indicate that both the type of solvent used for extraction and the method of extraction play a key role in the final assessment of the quality and quantity of the tested plant material [
64,
65,
66,
67]. The results indicating that methanol was definitely the most suitable extractant are consistent with Janicsák et al. [
68]. Methanol has frequently been used as an extractant of triterpene saponins and/or their aglycones [
65,
68,
69,
70,
71,
72]. In turn, Yin et al. [
73] reported that ethyl acetate was the most efficient extractant for triterpene acids in three types of
Cyclocarya paliurus plant tissue, while Marchev et al. [
74] indicated that acetone provided the most efficient extraction of triterpene acids in
Salvia scabiosifolia leaf and callus.
In the present study, extraction using a Soxhlet apparatus was found to yield higher recovery of investigated triterpenes than sonication. This superiority has been demonstrated in various tissues and plants by Janicsák et al. [
64,
68], and Wójciak-Kosior et al. [
67]. However, a study of various approaches by Wójciak-Kosior et al. [
67] found the most efficient method to be the use of a heat reflux condenser, while Soxhlet extraction provided slightly less performance.
Our present findings indicate that OA predominated among the investigated aglycones. Its yield was highest in TSE cultivated in vitro, and the lowest was in callus tissue; this difference may be due to the fact that not all the biosynthesis pathways of the compounds had been activated in the callus. It was previously reported that the higher productivity of plant in vitro cultures is connected with differentiation of cultivated cells reviewed in [
75]. This phenomenon was also described in respect to OA glycosides accumulation (expressed as the quantity of aglycone glycosides after hydrolysis) in in vitro cultures of
Calendula officinalis. In undifferentiated cell suspension cultures, the OA content amounted to 0.06 mg/g DW and increased up to 0.84 mg/g DW under treatment with 100 µM of methyl jasmonate [
69]. While in hairy root cultures even the initial level of OA was higher and ranged from 4.59 to 8.42 mg/g DW depending on root line [
70]. Further, OA content was significantly augmented in
C. officinalis hairy root cultures from 2.56 or 3.75 mg/g DW according to the root line to 52.52 and 41.18 mg/g DW, respectively, when elicitation with 100 µM of jasmonic acid was applied [
76]. In addition, higher content of investigated compounds in TSE cultivated in vitro then ex vitro could be explained by the fact that artificial conditions act as a stressor which could induce secondary metabolism. Nevertheless, in studies of
Cyclocarya paliurus, higher total tritepenic acid (TTA, i.e., OA, UA and betulic acid) content was found, in suspension culture (6.24%) rather than in callus (2.16%) or in the leaves of soil-grown plants (3.74%) [
73].
The results of the current study demonstrated that the somatic embryogenesis route could be efficiently applied to production of genetically stable Polyscias filicifolia plants capable of biosynthesizing triterpene compounds.