Introduction

Birch trees (Betula L.) are among the most common deciduous trees in the boreal and temperate zones of the Northern Hemisphere. According to different authors, in these zones there are more than 65 species of the genus Betula, however, European forests are dominated by two species: silver birch (Betula pendula Roth) and downy birch (Betula pubescens Ehrh.) (Hynynen et al. 2010). High morphological variability greatly complicates the distinction between these two species (Keinänen et al. 1999; Lahtinen et al. 2006; Migalina et al. 2010), and it is further complicated by their hybridization (Brown et al. 1982; Kennedy and Brown 1983).

Both species belong to the section Albae, subgenus Betula (Keinänen et al. 1999); however, they are distinguished by ploidy level: B. pendula is a diploid (2n = 28), whereas B. pubescens is a tetraploid (2n = 56). According to some authors, downy birch is an ancient allotetraploid, formed by a cross between a B. pendula-like species and another diploid species (Johnsson 1945; Howland et al. 1995). In Poland, silver birches bloom in April, approximately 10 days earlier than downy birch; in Finland and Karelia, birches bloom in early May, and the difference between species in flowering periods between the species is 7–10 days. In spite the difference in ploidy as well as flowering times, these two species are capable of forming hybrid forms.

The occurrence of plants that are morphologically intermediate between different birch species has been reported in many publications (Scoggan 1978; Atkinson 1992; Anamthawat-Jónsson and Thórsson 2003). For this reason, natural hybridization of different species of plants in the family Betulaceae has been and still remains the objective of research by specialists from different countries (Johnsson 1945; Natho 1959; Stern 1963; Brown et al. 1982; Anamthawat-Jónsson and Thórsson 2003; Banaev and Bažant 2007; Vetchinnikova and Titov 2017). The study of this phenomenon allows us to better understand the evolutionary role of natural hybridization, and it allows us to solve a number of theoretical and practical problems, such as the level of compatibility for natural crosses between the two birch species or the genetic improvement of birches.

A strong specificity of the secondary metabolite composition has been previously established for extracts from buds of B. pendula and B. pubescens (Isidorov et al. 2014, 2016), which makes it possible to use this distinction in chemotaxonomy. The extent to which the chemical compositions of different species of parent plants are transmitted to the progenies from crosses would be interesting to discover. To clarify this issue, we conducted a comparative study of the chemical composition of volatile organic compounds (VOCs) emitted into the gas phase by artificial hybrids and parental birch species using modern high-performance analytical techniques, which consisted of a combination of solid-phase microextraction with chromatography-mass spectrometry (HS-SPME/GC–MS) (Agelopoulos and Pickett 1998). The study of the VOC composition of birch buds and their volatilomes is of particular interest because these volatiles show anti-microbial and anti-herbivore activity and can serve to protect valuable reproductive parts of plants, such as buds (Holopainen 2004; Dudareva et al. 2004, 2006; Peñuelas and Munne-Bosch 2005; Rennenberg et al. 2006; Schwab et al. 2008; Karl et al. 2008).

Materials and methods

Plant material

The buds were collected from artificial hybrids of B. pendula Roth and B. pubescens Ehrh. growing at the field experimental station (FES) of the Forest Research Institute of the Karelian Research Centre of the Russian Academy of Sciences (KarRC RAS) near Petrozavodsk (61°45′N, 34°20′E). Hybrids (F1) were produced in 1969 in a series of crosses involving B. pendula seed parents and B. pubescens pollen parents and vice versa (Ermakov 1975, 1986). Crossing was carried out on trees in natural populations located in the southern part of the Republic of Karelia (North-West Russia) (61°19′N, 35°29′E). Besides, buds of downy birch and silver birch were gathered from 3‒5 trees belonging to the same populations in April‒May and August‒September 2018 at the FES. In addition, buds were gathered from December 2017 to January 2018, from trees growing in the Leningrad region (59°45′N, 30°04′E), Latvia (57°17′N, 26°34′E), and north-eastern Poland (53°32′N, 22°43′E) (Fig. 1). A previously described method based on the nuclear DNA isolation and sequencing was used to identify the birch species (Isidorov et al. 2014). Voucher specimens were deposited with the herbarium of the Forest Research Institute KarRC RAS (PTZ) and the Department of Pharmacognosy, of the Medical University of Bialystok, Poland (nos. BP-17034 and BO-17035).

Fig. 1
figure 1

Map giving the position of birch bud sampling

Full size image

HS-SPME sampling and GC–MS determination of VOCs

The previously described analytical procedure, HS-SPME/GC–MS, was used for VOC investigation (Isidorov et al. 2012, 2014). In these works, the comparison of PDMS 100, carboxen/PDMS (CAR/PDMS) and divinylbenzene/carboxen/PDMS (DVB/CAR/PDMS) sorption fibres was performed, and the best effectiveness of the extraction‒desorption cycle was obtained by the later. The chromatograms registered after the exposition of PDMS 100 fibres demonstrated hardly any peaks of compounds with low boiling temperatures (light carbonyls and alcohols). In turn, fibres CAR/PDMS do not completely return high-boiling-point components, such as sesquiterpenes, during the desorption stage, (Isidorov et al. 2012).

Harvested buds (0.5 g) were placed into a of 16 mL headspace vial and immersed into a thermostat at 50 °C. The membrane of the screw-cap was pierced by the needle with DVB/CAR/PDMS fibre and exposed to a headspace gas phase. After 50 min of exposition, the fibre was introduced for 10 min into the injection port of the GC–MS apparatus. The latter was operated at 250 °C in the splitless mode. The helium flow rate through the column was 1 mL min−1 in constant flow mode. The initial column temperature was 40 °C and rose to 220 °C at a rate of 3 °C min−1. The MSD detector acquisition parameters were as follows: the transfer line temperature was 280 °C, the MS source temperature was 230 °C and the MS quad temperature was 150 °C. The electron impact mass spectra were obtained at 70 eV of ionization energy. Detection was performed in the full scan mode. After integration, the fraction of separated components in the total ion current (TIC) was calculated.

To identify the components, both mass spectral data and the calculated retention indices were used. Mass spectrometric identification was carried out with an automatic system of GC–MS data processing supplied by NIST mass spectra library, as well as by computer search libraries containing the mass spectra and retention indices from Adams’ (2007) and Tkachev’s (2008) collections.

To determine the retention times of reference compounds, a SPME fiber was inserted for 2‒3 s into the headspace vial with a mixture of C5–C18n-alkanes, which were separated under the conditions described previously. The linear temperature-programmed retention indices (RI) were calculated from the equation:

$$ {\text{RI}}\, = \, 100\left[ {n\, + \,\left( {t_{x} - t_{n} } \right)/\left( {t_{n + 1} - \, t_{n} } \right)} \right], $$

where tx is the retention time of the analyte, tn is the retention time of the n-alkane eluting directly before the analyte, and tn+1 is the retention time of the n-alkane eluting directly after the analyte. Calculated retention indices of the registered components were compared with the above-mentioned collections, as well as with the NIST (2013) collection. The identification was considered reliable if the results of the computer-based search of the mass spectra library were confirmed by the experimental RI values, i.e., if their deviation from the averaged published values did not exceed ± 10 u.i. Mass spectrometric identification not confirmed by the retention index was considered as putative.

Results and discussion

The chemical composition of bud VOCs

The chemical compositions of VOCs emitted into the gas phase by buds of artificial hybrids of silver and downy birch species (12 samples) growing in Karelia were determined with the aid of HS-SPME/GC–MS. VOC compositions were also studied with the same technique for eleven samples of silver birch and eleven samples of downy birch buds. The geography of these samples encompasses a latitudinal interval of northeastern Europe from 61°N to 53°N (meridian distance ca. 890 km).

On the obtained chromatograms were registered 224 peaks which belonged to C1‒C18 organic compounds of different classes. The number of compounds detected in the volatiles of both birch species was nearly the same: 157 in silver birch emissions and 156 in downy birch emissions. Hence, the composition of their volatiles overlapped only partially; some components were found in the emission of only one species. It is remarkably that VOCs of both crossed variants were substantially less diverse: only 88 compounds in the B. pendula ♀ × B. pubescens ♂ emissions and 87 compounds in the B. pubescens ♀ × B. pendula ♂ emissions (107 compounds in total). The difference in the VOC compositions of downy and silver birch buds and their hybrids is demonstrated in Fig. 2.

Fig. 2
figure 2figure 2

Typical chromatograms of VOCs emitted from buds of B. pendula (a), B. pubescens (b), B. pendula ♀ × B. pubescens ♂ (c), and B. pubescens ♀ × B. pendula ♂ (d) hybrids

Full size image

Table 1 presents the averaged semi-quantitative composition (TIC fraction) of bud volatiles detected in emissions of all 34 samples that were investigated. It also contains some analytical parameters that were used to confirm the identification results: experimental (RIExp) and literature (RILit) values of retention indices, m/z values of the most intensive ions in the mass spectra (in order of decreasing intensity), and mass numbers of molecular ions, M+, if registered in the mass spectra. The content of the compounds in Table 1 vary widely from traces (< 0.01% of TIC) to tens of percent. These variations can be explained by the circumstances under which the samples were collected: different seasons and populations growing in different climatic zones.

Table 1 Qualitative and semi-quantitative (% of TIC) composition of VOCs emitted by buds of two birch species and their hybrids
Full size table

The VOCs in Table 1 are divided into 11 groups according to their chemical structures; each group lists compounds in order of their retention indices. Although genes and enzymes specifically involved in the synthesis of these birch bud volatiles have not yet been investigated, some considerations according to their synthesis pathways can be presented based on data available in the literature.

The most comprehensive group of the birch bud VOCs is formed by terpenoids: 22 mono- and 93 sesquiterpenoids. Moreover, terpenoids are characterised by the higher content (80‒90% of TIC). It is known that all terpenoids derive from isopentenyl diphosphate and dimethylallyl diphosphate (Croteau et al. 2000). Terpene synthases are responsible for catalyzing the formation of terpenoids from these substrates. Many other enzymes participate in the transformation of initial products by their oxidation, hydroxylation, dehydrogenation and other processes (Croteau et al. 2000; Dudareva et al. 2004; Qualley and Dudareva 2009).

The majority of monoterpene compounds belong to minor components; only limonene, β-ocimene isomers and linalool oxide isomers were at quantities higher than 1% of TIC. Camphene, β-pinene, and myrcene were only registered in the VOCs of downy birch. In turn, 1,8-cineole was characteristic of the emissions of silver birch.

The assignment of some compounds to C15H24O, C15H24O2 and C15H26O2 sesquiterpenoids was based only on the MS data and should be considered putative. The majority of these substances belong to minor components, which share in the total ion current of the chromatogram not exceed 1% of TIC, as a rule. Qualitative composition of the positively identified sesquiterpenoids in birch bud VOCs is highly contrasted. Figure 3 presents the chemical structures of the main components of volatile emissions from buds of both birch species. It was determined that the sesquiterpenoids of B. pendula are structurally diverse; the main components of VOCs are monocyclic germacrene D, bicyclic compounds with cadaline (cadinenes, amorphenes, calocorenes and muurolenes) and guaiazulane (guaia-6,9-diene, α-guaiene, salviadienol) skeletons, as well as tricyclic (α-, β-copaenes, aromadendrenes, β-bourbonene and α-ylangene) compounds. In contrast, principal VOCs in B. pubescens emissions belong to bicyclic compounds with nor-caryophyllane (birkenal, birkenol, and des-4-methylcaryophyl-8(14)-en-5-one) and caryophyllane skeletons. Monocyclic germacrene D, as well as tricyclic α- and β-copaenes, β-bourbonene and α-ylangene were also present in VOCs of B. pubescens, but their share of the TIC is much smaller. In turn, β-caryophyllene and its 6- and 14-hydroxy derivatives belong to the minor components of B. pendula and to the main components of B. pubescens emissions.

Fig. 3
figure 3

Sesquiterpenoids and nor-sesquiterpenoids typical for VOCs of B. pendula (a) and B. pubescens (b) buds

Full size image

Among the volatiles of B. pubescens, we registered for the first time a C13H22 hydrocarbon tentatively identified (based on the MS data) as 4,8,8-trimethyl-2-methylene bicyclo[5.2.0]nonane (Fig. 3). If we accept the idea that the above-mentioned nor-caryophyllanes, birkenal, birkenol and des-4-methylcaryophyl-8(14)-en-5-one in the buds of B. pubescens are the ring-contracted products of β-caryophyllene (Klika et al. 2004), we can speculate that the C13H22 bicyclic hydrocarbon is the product of further ring-contracting.

Another previously undescribed characteristic feature of B. pubescens is the emission of C13H20 hydrocarbons, which were identified as megastigmatriene isomers based on mass spectral and chromatographic characteristics. In the volatile emissions of the B. pubescens buds, these hydrocarbons are represented by five isomers, the total content of which in VOC was 0.9–3.2%. These hydrocarbons were found for the first time among the volatile components of purple passion fruit (Passiflora edulis) juice. The biosynthetic pathway of their formation was postulated to occur successively through β-ionone and β-ionol, which quickly undergoes dehydration (Camisir et al. 1981; Simkin et al. 2004). However, β-ionol itself was not detected among the volatile components of the downy birch buds (as well as among the VOCs of passion fruits). Therefore, the assumption of their origin, as well as the role of C13H20 hydrocarbons, requires further study.

The next comprehensive group of VOCs was formed by carbonyl compounds, which included 27 aliphatic saturated and unsaturated C2‒C10 aldehydes and ketones, as well as two aromatic aldehydes: benzaldehyde and benzene acetaldehyde. The short-chain aldehydes play important roles in the plant defence strategies. Volatile unsaturated aldehydes, as well as some unsaturated alcohols and their esters [e.g. (Z)-3-hexenol and (Z)-3-hexenyl acetate] originated from unsaturated fatty acids, which have been found in birch buds in significant amounts (Vetchinnikova 2004; 2005; Vedernikov and Roshchin 2009; Isidorov et al. 2018). In their biosynthesis through the oxylipin pathway (Matsui 2006; Stumpe and Feussner 2006), several different enzymes are involved: lipooxygenases (LOX), hydroperoxyde lyase (HPL), isomerase, and alcohol dehydrogenase.

LOX catalyse the regio- and enantio-selective dioxygenation of linoleic and α-linolenic acids. We assumed that only 13-LOX enzymes participate in hydroperoxidation in birch buds and lead to the 13-hydroperoxy derivatives, (Z)-3-hexenal and products of its former transformations as listed in Table 1: (E)-2-hexenal, hexanal, and (E)-2-hexen-1-ol, (Z)-3-hexen-1-ol and (Z)-3-hexen-1-ol acetate (Matsui 2006). However, we did not find typical products of 9-LOX hydroperoxidation such as isomeric nonadienals and nonadienols among birch bud VOCs.

It can be assumed that products created by HPL in birch buds are subjected to Z,E-isomerisation and reduction to unsaturated alcohols by alcohol dehydrogenases (ADH). Unsaturated and saturated C6 alcohols are capable of producing a wide range of esters by alcohol acyl transferases (AAT). We observed (Z)-3-hexenyl acetate, (Z)- and (E)-hexenyl butanoates, and (Z)-3-hexenyl pentanoate among bud volatiles (Table 1).

Interestingly, the lists of carbonyl compounds and esters identified in the emissions of hybrids were much shorter than those of the parent birch species. For example, out of 27 aldehydes and ketones detected from parent species, only 14 were presented in detectable amounts among the VOCs of hybrids, and out of 17 esters, only two (methyl and ethyl acetate) were registered in emissions of hybrids.

Saturated short chain linear carbonyls, alcohols acids and their esters were detected among VOCs of birch buds. They could be derived from saturated fatty acids through repeated β-oxidative cycles (Schwab et al. 2008). Branch-chain volatile carbonyls and alcohols can also be formed in the degradation process of branch-chain aliphatic amino acids. In addition, aromatic compounds such as 2-phenylethanol and 2-phenylacetaldehyde can be derived from phenylalanine (Kaminaga et al. 2006).

In this investigation, we established that the chemical composition of VOCs emitted from the buds of silver and downy birches is species-specific regardless of both the geographical origin of the trees (at least within the boundaries of the boreal and mid-latitude zone), as well as the time of bud collection. This difference is evident in the dendrogram in Fig. 4. The left part of the picture (dendrogram A) is clearly divided into two groups; one of which includes VOCs from buds of silver birch (BB-11–BB-36), whereas the other shows the VOCs of downy birch buds (BO-11–BO-35). To construct the dendrograms, we used the data of Table 2, which reflected a generalized composition of the main bud VOCs, i.e., sesquiterpenoids belonging to different groups: compounds with a bicyclic caryophyllane- and norcaryophyllane-type skeleton, with cadaline and muurolane structures (cadinenes and muurolenes) and tricyclic copaenes, as well as C13H20 hydrocarbons.

Fig. 4
figure 4

Dendrograms of the chemical similarity of silver birch (BB) and downy birch (BO) volatiles (a), and hybrids B. pendula ♀ × B. pubescens ♂ (B × O) and B. pubescens ♀ × B. pendula ♂ (O × B) (b) (Hammer et al. 2001)

Full size image
Table 2 Relative group composition (% of TIC) of volatile compounds from buds of two birch species and their hybrids
Full size table

Taxonomic implications of sesquiterpenoid variation in the VOCs of hybrid birch buds

The differences we observed indicate the existence of some connection between the composition of VOCs and the genetic characteristics of the birch trees studied that had different ploidies. Notably, despite the difference in these birches’ number of chromosomes, many authors have demonstrated the possibility of interspecies crossings to produce fertile seeds (Natho 1959; Clausen 1963; Gardner 1984; Ermakov 1986; Vetchinnikova 2004, 2005; Vetchinnikova et al. 2013). For example, when the silver birch (♀) was crossed with the downy birch (♂) in Karelia (Russia), germination of seeds was about 27%. Conversely, when the downy birch (♀) was crossed with the silver birch (♂), up to 32% germination was achieved (Ermakov 1975, 1986). Allotetraploidy of downy birch, or the presence of the silver birch genome in the genotype, in all likelihood “facilitates” the hybridization between these species even with different ploidies. At the same time, morphological features of shoots and leaves characteristic of these species (hairy leaves and stems, resin glands on vigorous shoots, toothing of leaves) are often erased (Ermakov 1986; Vetchinnikova 2004, 2005). Consequently, these features cannot reliably be used to detect cases of natural hybridization between different birch species.

A natural question arises: if the composition of the bud VOCs of the parent plants are so different that they can serve a chemotaxonomy purpose, is it possible to distinguish their hybrids on this basis? This question becomes more important in connection with the fact that due to the warming and instability of the climate observed in past decades, the flowering times of both birches increasingly overlap, thus eliminating the phenological isolation typical of these species and thereby fostering hybridization.

The results of the study of hybrid offspring (F1) presented in Table 2 and on the right part of Fig. 4 (dendrogram B), indicate that the buds of hybrid plants have a characteristic composition of volatile substances, which allows a definitive conclusion about the hybrid origin in some cases. As can be seen on dendrogram B, all hybrids investigated are clearly divided into two groups depending upon the type of crossing. From the data of Table 2, it can also be seen that the ability of species to accumulate specific components in the buds to a greater or lesser extent was consistently preserved in the offspring. At the same time, hybrids clearly show dominant inheritance along the maternal line in both crossbreeding variants: the B. pendula ♀ × B. pubescens ♂ hybrid that resembled silver birch and the B. pubescens ♀ × B. pendula ♂ that resembled downy birch.

However, the pollen of tetraploid downy birches substantially affects the composition of the VOCs of the progenies; in volatile secretions from buds of plants, when the pollen donor was downy birch (hybridization variant of B × O), there were significant amounts of nor-caryophyllenes and C13H20 hydrocarbons, which were uncharacteristic for the maternal species, silver birch (Table 2). The mechanism of the phenomenon of gene transfer through pollen described for oaks is called “pollen swamping” (Petit et al. 2003). Another distinctive feature of this hybridization variant is a noticeable decrease in the relative content of components typical for silver birch, e.g., cadinanes, muurolanes, and copaenes. On the other hand, in the second crossing variant (O × B), the influence of the pollen donor (diploid silver birch) on the VOCs composition was nearly undetectable.

Conclusion

The study of the composition of volatile secretions of birch buds reliably reveals only those hybridization variants for which the pollen donor was the plant with the higher ploidy, i.e., downy birch. Identification of the second hybridization variant requires a more detailed study of the composition of secondary metabolites, including non-volatile compounds and those contained in other plant organs and tissues. For these purposes, the molecular cytological mapping of ribosomal genes and species-specific DNA may also be useful (Anamthawat-Jónsson and Thórsson 2003).

Author contribution statement

Conceptualization: VI, Sample collection and preparation for analysis: LV, MS, Acquisition of data: VI, MS, Analysis and interpretation of data: VI, LV, Writing—original draft: VI.