Abstract
Plants, in particular trees with specific habitat demands are excellent indicators of climate state. Vegetation successions in subrecent and deep geologic time is recorded in fossil macro-remains or pollen accumulating in geological archives like limnic and marine sediments, peat bogs and mires. Birch trees in Europe form a major part in plant successions and constitute the dwarf species Betula nana and Betula humilis representing cold-adapted habitats or climates and two tree birches, Betula pubescens and Betula pendula characteristic for temperate habitats or climates. These birch species exhibit highly similar pollen shape and size, preventing their unambiguous application as paleoclimate/paleovegetation proxies. We here present a chemotaxonomic differentiation of the four European birch species based on their epicuticular wax lipids. The dominating lipid classes in epicuticular birch waxes were found to be n-alkanes (in the range of n-C23 to n-C33), straight-chain primary alcohols and fatty acids (in the range of n-C20 to n-C32), and long-chain wax ester (in the range of n-C38 to n-C46) in variable amounts and distributions. When preserved in geological archives these lipids may serve in paleovegetation/paleoclimate reconstruction. Long-chain wax esters are susceptible to hydrolysis and upon diagenesis the release of ester-bound alcohols and fatty acids may modify the distribution pattern of the corresponding primary free lipids. Quantitative analysis of the hydrolyzable wax ester proportion revealed primary distribution patterns of birch lipids not to change substantially upon release of bound analogues. The specific composition and abundance of epicuticular wax lipids facilitates unambiguous chemotaxonomic separation of the four European birch species. Wax lipid-based discrimination in field application, however, is complicated by mixing of alkyl lipids derived from different birch species and contribution of wax lipids from other plants. In cases, where palynology indicates a high contribution of Betula species to European vegetation associations, wax lipids may serve for differentiation of the species contributing.
Introduction
Environmental demands, in particular climate, govern the present-day habitat and distribution of trees, which in turn facilitates determination of climate regimes in the contemporaneous as well as in the fossil domain. Taxonomy of present-day trees is based on anatomical, morphological, genetic and biochemical studies, whereby such features in the fossil record are best preserved in pollen distributions, due to a higher recalcitrance of pollen versus other plant organs, e.g. leaves and rare findings of other macro-remains, e.g. fruits. Under special conditions of preservation though fossil remains can be found in sediments dating back to the Eocene [1]. Paleovegetation and related paleoclimate reconstruction thus heavily relies on palynology, accompanied by analysis of marco-remains when present. Genetic [2–4] and biochemical approaches [5–8] are rare, except for molecular and isotopic investigation of leaf/needle wax, the latter rarely conducted on the species level.
Birches (Betula L., Betulaceae) are common broadleaf trees and shrubs occurring in diverse habitats of the boreal and the cold-temperate zones of the Northern Hemisphere [9,10]. Ranging from temperate zones in Northern America over Eurasia to East Asia and the circumpolar regions, birches populate different habitats including forests, swamps, tundra and mountainous terrains [11]. The number of species belonging to the genus Betula and their relationships is still under debate ranging from 30 to 60 different taxa within 4 to 6 subgenera [3,12–15]. Of these, four Betula species are endemic to Europe. The two tree birches, Betula pubescens (downy birch) and Betula pendula (silver birch) occur throughout most of Europe, whereby Betula pubescens has a more northerly and easterly distribution, while Betula pendula reaches more southern regions such as the Iberian Peninsula, Italy and Northern Greece [10,15]. Two dwarf/shrubby birch species, namely Betula nana (arctic dwarf birch) and Betula humilis (dwarf birch) thrive in Europe as well but are confined to a much smaller growth range. Betula nana is preferentially located at higher elevations like the Alps and Carpathian Mountains or in perennial colder regions like Northern Europe from Iceland over northern Scotland to Scandinavia [10,15]. Betula humilis has a wide distribution from western Germany to eastern Siberia and Korea, but its occurrence is very scattered with only a few habitats left in Central Europe [15]. As an important component of plant succession found in highly contrasting climate and growth regimes, these Betulaceae and their respective habitat demands are sensitive (paleo)climate and (paleo)environmental indicators, provided that they can be taxonomically differentiated.
Recent birch species can be distinguished by their leaf and catkin/fruit shape [16,17] even when fossilized in sediments, whereby their evolutionary relation and phylogeny is difficult to assess due to intensive hybridization in nature. Extensive hybridization and introgression has been investigated not only within a subgenus but also across Betula species of different subgenera [18,19]. Precise classification is complicated by an spatial overlap of natural habitats for European birch species enabling natural hybridization [20]. The identification of birch remains in geological archives for paleovegetation reconstruction is limited by a common lack of well-preserved leaves or catkins occurring in statistically relevant quantities. However, the identification of birch vegetation over time is of great interest, especially during the Late Glacial and Early Holocene (approx. 15,000 – 9,000 a BP), to understand early colonization and forest/shrub expansion in deglaciated landscapes as well as the adaptation of vegetation to climate variability and perturbation. Macrofossil findings revealed a dominance of tree birches, like Betula pubescens, during temperate phases, whereas upon glacial stages Betula nana was the most abundant birch species, serving as a tundra indicator [21].
Most commonly, paleovegetation reconstruction in sedimentary archives like lakes, peats and bogs is based on palynological approaches, since macro remains are mostly absent [22–30]. However, differentiation of birch species by pollen is challenging, due to similar morphological traits, e.g. shape, diameter and depth of pores in pollen within the genus Betula, which leads to overlap in pollen-size distribution [31–35]. In addition, pollen morphology can also be effected by the chemicals used upon preparation, pollen maturity, type of mounting medium, but also a latitudinal and altitudinal effect cannot be excluded (reviewed in 32). In this study we present the epicuticular leaf wax composition as a complementary proxy for recognition and differentiation of the four European birches, with the potential to employ the wax distribution patterns in paleovegetation reconstruction.
Terrestrial plant leaves are covered by a hydrophobic barrier to protect against the loss of water due to evaporation, mechanical damage, ultraviolet radiation and bacterial or fungal pathogens [36–39]. This barrier consists of an epicuticular wax layer, composed of long-chain alkyl compounds including amongst others n-alkanes, n-alcohols, n-alkanoic acids, n-alkyl esters, n-aldehydes, n-ketones and others [37,40]. Their composition is highly variable in quality and quantity across plant species and therefore has high chemotaxonomic potential [41,42,51,43–50]. n-Alkanes with carbon chain-length between C25 and C35 carbon atoms are associated with higher plants with a strong odd-over-even predominance (expressed as the carbon preference index-CPI), while their shorter chain homologues (<C20), especially nC17, are mainly found in aquatic microorganisms [52,53]. Intermediate chain-length n-alkanes with C23 and C25 predominance are primarily found in aquatic macrophytes and in mosses of the genus Sphagnum [49,54,55].
Both, n-alcohols and n-alkanoic acids in contrast to the n-alkanes hold an even-over-odd predominance in carbon chain-lengths, which typically ranges from C20 to C32 [37]. Alkyl esters consist of even-numbered n-alkanoic acids, which are esterified to even-numbered n-alcohols, generating long-chain aliphatic compounds with on average 38 to 52 carbon atoms [56]. Among these lipid classes, the n-alkanes abundances are most frequently reported in vegetation reconstruction, since they are very robust against alteration processes due to the lack of functional groups. Betula-derived n-alkanes can be found in a variety of geological archives including peats, soils, limnic and marine sediments of different ages ranging from modern times up to several million years [45,57–59] and are easily extracted from sediments and leaves by geochemical methods [37]. Therefore, most studies conducted to investigate paleovegetation history employing leaf wax lipids are based on n-alkanes [60]. However, the use of several lipid classes instead of a single will increase the discriminative power and representativeness of the wax lipid composition. Studies involving n-alkanoic acids, n-alcohols or n-alkyl ester for paleovegetation reconstruction are highly underrepresented [61,62], as these lipids are usually not reported from modern plant homologues and their degree of preservation in natural archives may vary [63]. Following incorporation into soil or sediment, wax esters can be hydrolysed, releasing free n-alcohols and n-alkanoic acids. These in turn can be converted into n-alkanes by decarboxylation and reduction, respectively [60]. Diagenetic fate of functionalized lipids may vary depending on various factors such as oxygen availability and pH affecting microbial reworking, but the potential of functionalized lipid classes in paleobotany has been proven for sediments of up to Miocene age [63,64].
Material and Methods
Leaf samples and collection
Fresh leaf samples were collected in the Botanical Garden at Kiel University in September 2017. Three leaves from each birch species were taken from branches at different sites of the tree from a height between 1 and 3 m. To avoid contaminations during sampling gloves were worn and leaves were stored in glass container or aluminium foil until extraction. Subsequently to sampling leaves were dried at 35°C in an oven for 48h. The mean annual air temperature for Kiel-Holtenau is about 9.34°C and the total annual precipitation is about 744 mm (1987-2018) (DWD, 2019)
Lipid extraction
Lipids were extracted by immersing a single leaf sample (0.02 – 0.27 g) for 60 s in a 30-50 ml hexane/dichloromethane solution (1:1 v/v). The resulting solution was filtered through NaSO4, evaporated under vacuum at 50°C in a Büchi solvent evaporator and transferred into pre-weighted vials. Per species three leaves were extracted to calculate standard deviation of lipid concentration and composition. Prior to analyses, an aliquot of the total lipid extract (TLE) was treated with 35 μl N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and 5 μl pyridine at 70°C for 1 h to convert the n-alkanoic acids and n-alcohols to their corresponding trimethylsilyl (TMS) derivates. 10 μg of perdeuterated tetracosane (C24), octadecanol (C18), and eicosanoic acid (C20) were added as internal standard for quantification. All samples were analysed by gas chromatography-mass spectrometry (GC/MS).
Gas chromatography-mass spectrometry (GC-MS)
The wax lipids were analysed using an Agilent 7890A (GC) equipped with an Agilent DB-5 column (30m × 0.25mm × 0.25μm) coupled to an Agilent 5975 B (MS). The oven program started at 60°C for 4 min, followed by a ramp to 140°C at 10°C/min and subsequently to 325°C at 3°C, followed by an isothermal period of 45 min. The MS operated with a scanning mass range of m/z 50-850 at an ionization energy of 70 eV. All compounds were identified by using authentic standards, NIST 14 library or their specific fragmentation pattern.
Leaf wax characteristic calculations
The n-alkane content of plant species was calculated as μg/g dry weight (d.w.) of leaf based on mean values of triplicate analysis with standard deviation (Fig. 1).
Average chain length (ACL) for i-alkanes with 23 to 33 carbon atoms was calculated as: with Cn as relative abundance of n-alkanes with the chain length n [65]. This proxy is used as weighted mean of n-alkane carbon chain length, which supposed to vary with climate. The carbon preference index (CPI) outlines the relative abundance of odd-over-even carbon chain lengths, whereby values >1 indicate a predominance of odd carbon chain lengths homologues. CPI values for n-alkanes with 24 to 34 carbon atoms were calculated according to:
Wax ester quantification
For total wax ester quantification, the respective wax ester peaks in the GC-MS chromatograms were integrated and quantified against an internal standard (deuterated tetracosane). For a wax ester of the type RCOOR’ the diagnostic fatty acid ion is RCOOH2+, indicative for the alkanoic acid chain length. R’ – 1+ derives from the corresponding alcohol moiety, however, its intensity is low and difficult to detect. Therefore, the diagnostic acid fragments (RCOOH2+) of peaks containing co-eluting alkyl esters of identical total mass but variable combinations of alcohol and alkanoic acid moieties were integrated to determine the percentage of the respective isomer contribution. Multiplication of isomer percentages by analogue abundances led to the proportion of the esterified acids. The proportional amount of esterified alcohol was obtained by subtracting the esterified acid from the corresponding total wax ester homologue.
Results and discussion
Alkyl lipid distribution of plants from Botanical Garden of Kiel University
The focus of this study lies on the epicuticular wax composition of the four birches endemic to Europe. The leaf wax compound classes n-alkanes, n-alcohols, n-alkanoic acid and n-alkyl ester were present in all species at variable composition distributions (S1). Epicuticular alkyl lipid abundances are reported as μg/g dry weight (d.w.) of leaf.
The lowest total amount of epicuticular waxes was observed in the artic dwarf birch B. nana with 538.3 μg/g d.w., followed by B. pendula with 1293.8 μg/g d.w., 3131.7 μg/g d.w. in B. pubescens and B. humilis with 4187.9 μg/g d.w..
n-alkane of Betula epicuticular wax
The carbon atom chain-lengths of n-alkanes varied from nC23 to nC33 and maximized either at nC27 or nC31. A predominance of nC27 was noted for B. humilis with 413.2±91.1 μg/g d.w.. In relative proportion, nC27 was the dominant n-alkane homologue in B. nana, though it occurred in minor absolute concentrations (7.8±2.5 μg/g d.w.) only. The n-alkanes of two tree birches, B. pubescens and B. pendula, maximised at nC31 with 799.0±133.3 and 183.0±103.8 μg/g d.w., respectively, whereby the latter species additionally contained large proportions of nC25 and nC27 alkanes. The average chain-lengths for odd-carbon-numbered n-alkanes in the range from nC23 to nC33 (ACL23-33) varied from 26.7 in B. humilis to 29.7in B. pubescens, while B. pendula und B. nana had intermediate ACL values of 28.0 and 28.5, respectively. As expected, in the recent leaves a typical odd-over-even predominance was detected, which is expressed in high CPI values. Highest CPI values were found in B. pendula averaging 41.3, followed by B. pubescens with 36.4 and B. nana with 11.1. The CPI for B. humilis could not be calculated due to the lack of even-numbered n-alkanes. The n-alkane CPI values of the three species were high, indicating that there was no significant contamination by diagenetic or petroleum-derived n-alkanes.
The birch trees examined in this study grew under identical environmental conditions in the Botanical Gardens of Kiel University, thus a possible influence of climate or soil condition on the wax lipid distribution should be identical for each species. The n-alkane compositions in all four Betula species exhibited a dominance of longer-chain lengths homologues (nC27, nC29, nC31) with a mean ACL of ca. 28, as typical of deciduous trees [49]. However, some species had only one dominant homologue (B. pubescens, B. humilis), whilst others had a more bimodal distribution with two dominant homologues (B. nana, B. pendula) (Fig. 1). The ACL has been used in geological archives as a climate and plant species indicator. For example, it has been recognized that higher ACL values correlated with higher temperatures in sediments of Lake Malawi (east Africa), potentially as protection against heat to increase the melting point of epicuticular waxes [66]. However, Diefendorf et al. 2017 in their compilation study could not confirm a significant temperature control on ACL neither in C3 woody plants nor in C3/C4 grasses. Higher ACL values, due to a higher proportion of nC31, nC33 and partially nC35, were observed in C4 grasses (poaceae) of arid zones in Africa, which distinguished them from C3 species from Peru and Australia [67]. A correlation between the preferred habitat of the four birches from our study (arctic/alpine vs. temperate zone) and ACL is not noticeable, since especially the two cold tolerant dwarf birches showed markedly different n-alkane distributions. However, the measured ACL values were in accordance with the expected values, since no extended long-chain n-alkanes with 33 or even 35 carbon atoms, typically for arid grasses, were found.
n-alcohols of Betula epicuticular wax
As with the n-alkane distribution, a typical terrestrial higher plant pattern was observed, yielding a strong even-over-odd dominance in carbon chain-lengths [68]. The n-alcohol chain-lengths ranged from nC16 to nC32 with a predominance in long-chain alcohols (>nC20) in all birches. As short chain-lengths n-alcohol homologues (<nC20) are primarily synthesised by microbes or algae [69], we based epicuticular wax analysis on the long-chain homologues. The primary n-alcohols showed the closest range in overall concentrations of all lipid classes (Fig. 1). The cumulative n-alcohol concentration extended from 168.5±22.0 μg/g in B. pendula to 436.4±84.7 μg/g in B. humilis. The three birches B. humilis, B. pubescens and B. pendula revealed a gradually decreasing concentration with increasing chain-lengths, whereby B. humilis maximized at nC20 with 209.0±40.2 μg/g d.w. and the latter two at nC22 with 153.5±12.9 and 79.2±10.0 μg/g d.w., respectively. B. nana was characterized by a narrower distribution, peaking at nC28 with 99.7±23.3 μg/g d.w.. Our results demonstrate that the concentration of n-alcohols either increased or decreased gradually with increasing carbon chain-length with only one homologue being dominant (Fig. 1). Both tree birches, B. pubescens and B. pendula, revealed similar alcohol patterns and differed in absolute concentrations only.
In contrast to the tree birches, however, the two shrub birches showed a markedly different n-alcohol distribution and were well distinguishable.
Diefendorf et al. (2011, 2015) reported that the average concentrations of n-alcohols in leaves of deciduous angiosperm trees from the U.S. commonly were twice as high as those of n-alkanes. In our study B. nana exclusively revealed about 10 times higher concentrations of n-alcohols compared to n-alkanes. The other three species produced two to four times lower amounts of n-alcohols than n-alkanes. Previous studies have shown that grasses are mainly characterized by nC26, nC28 and nC32 alcohols [67,70]. This corresponds closely to generally low amounts of these homologues in favour of the high proportions of the shorter alcohols nC20 or nC22 in B. humilis, B. pubescens and B. pendula. Only the leaves of B. nana produced a higher proportion of nC26 and nC28 alcohols.
n-alkanoic acids of Betula epicuticular wax
The n-alkanoic acid abundances were characterized by a strong even-over-odd dominance in carbon chain-lengths in the range of nC12 to nC30. Similar to the n-alcohols, short-chain homologues (<nC20) were not significant for higher plants since these compounds are also produced by a variety of organisms like bacteria and algae, or derive from cellular membranes rather than waxes. Here, in all four birches species alkanoic acids in the range of nC12 to nC30 were observed to peak at nC16 or nC28. In the range of the long-chain fatty acids (>nC20), the four Betula species maximized exclusively at nC28, whereby their concentrations differed by two orders of magnitude (Fig. 1). Thus, B. nana yielded only 1.75±2.2 μg/g d.w. of nC28, while B. humilis produced 201.2±13.4 μg/g d.w. of C28. B. pubescens and B. pendula showed intermediate concentrations of 34.3±2.4 and 34.7±4.0 μg/g d.w., respectively. B. nana produced more short chain alkanoic acids than long-chain homologues, with highest concentration at nC16 with 44.3±6.1 μg/g d.w. and nC18 with 20.8±1.3 μg/g d.w..
The relative distribution patterns of long-chain alkanoic acids in the four birches were too similar to allow for differentiation. These basic findings were consistent with a litter and topsoil transect experiment in which deciduous forest sites also showed a dominance of nC28 alkanoic acids and differed from conifer (nC24) and grasslands sites (nC32 and nC34) [71]. Therefore, C28 alkanoic acid preponderance may serve to distinguish wax lipid inputs of birches from those of grasses and conifers, like pines. This may be applicable for sediments in periods such as the Late Glacial and Early Holocene in Central Europe, where successions were characterised by a minor diversity in plant species. In contrast, Diefendorf et al. (2011) observed that the n-alkanoic acid distributions across plant groups from the east coast of the USA, both angiosperms and gymnosperms, were similar with no dominant homologue present.
n-alkyl esters of Betula epicuticular wax
Wax esters are dimeric wax compounds build by a n-alcohol and a n-alkanoic acid moiety, whereby each n-alkyl ester isomer can be composed of several different combinations of n-alcohol and n-alkanoic acid homologues (S2) [72]. Saturated wax ester homologues were identified according to their characteristic molecular ions (M+) [73]. Straight-chain wax esters in the range of nC38 to nC48 were detected in all four species, which is typical for higher plants [74] (Fig. 1). Additionally, B. humilis produced minor quantities of the nC36 homologue. The alkyl ester composition of B. nana maximized at nC44 with 104.2±33.4 μg/g d.w. with an almost normal distribution. B. humilis peaked at nC36 with 1047.1±254.0 μg/g d.w., followed by a linear decrease in concentration of longer-chain wax esters up to nC48. Wax esters chain lengths of B. pubescens and B. pendula leaves ranged from nC38 to nC46, whereby the concentrations decreased with an increase in chain-lengths. Concentrations of nC38 wax ester of B. pubescens and B. pendula yielded 535.7±99.4 and 122.7±35.5 μg/g d.w., respectively.
Isomer distribution of wax esters and input to free n-alkanoic acids and n-alcohol amount
The mass spectral analysis of wax esters by GC/MS allowed to investigate their corresponding bound n-alcohol and n-alkanoic acids (Fig. 6, S3–S6 Tables).
In all four species, only even-chain alkanoic acids in the range of nC14 to nC28 and alcohol moieties ranging from nC18 to nC32 were observed resulting in even-chain alkyl esters. In B. nana, nC20 was the overall dominant ester-bound alkanoic acid moiety with 62.2 μg/g d.w. and nC24 the most prominent ester-bound alcohol with 58.9 μg/g d.w.. In the shorter nC38 and nC40 esters shorter ester-bound alkanoic acids with 14 and 16 carbon atoms as well as shorter ester-bound alcohols like nC22 occurred. The ester homologues of B. humilis, B. pubescens and B. pendula were dominated by short-chain nC16 bound alkanoic acid moieties (841.5, 368.8, 75.7 μg/g d.w.), while nC20 was the most dominant alkanoic acid homologue in the long-chain alkanoic acid fraction (190.6, 38.0, 33.4 μg/g d.w.). However, compared to the short-chain alkanoic acids, the long-chain homologues were less abundant in concentrations up to factor of 10. The major esterified alcohol within the three species varied significantly. B. humilis was dominated by nC20 (744.2 μg/g d.w.), B. pubescens by nC22 (346.6 μg/g d.w.) and B. pendula by nC24 (82.7 μg/g d.w.) and nC22 (81.2 μg/g d.w.) bound alcohols.
The bound n-alkanoic acid and n-alcohol moieties of wax esters might be released during hydrolysis upon incorporation of alkyl esters into soil or during early burial stages in sediments. As consequence, the amount of hydrolysis-released, previously ester-bound n-alkanoic acids and n-alcohols in a sediment will impact on the quantity and distribution pattern of free n-alkanoic acids and n-alcohols derived from leaf waxes [70] and needs to be considered in paleovegetation reconstruction. However, intact wax esters can survive in sediments and can be used for paleovegetation reconstruction [63,75,76] as well.
To test for the potential release of alcohols and acids from esters two different scenarios were calculated. In the first scenario, 50% of the esterified n-alkanoic acids and n-alcohols were released and added to their corresponding free homologues (Fig. 2). In the second scenario, the maximum release of 100% of the bound lipids and addition to the free analogues was used. Since the wax esters of all four birches consisted mainly of short-chain fatty acids, they do not significantly increase the pool of long-chain fatty acids (>nC20) typical for terrestrial higher plants. Solely in B. nana, the alkanoic acid distribution changed from a previously rather balanced distribution of nC20 to nC28 with 1 to 2 μg/g d.w. to a dominance of nC20 with >60 μg/g d.w., due to the release of the esterified homologues. A dominance of nC28 alkanoic acid was still observed for the other three birches when 50% of the esterified alkanoic acids had been released. Only upon 100% release of the bound long-chain fatty acids, a bimodal distribution maximizing at nC20 and nC28 could be noted for B. pubescens, which would complicate a source identification in sediments.
The bound n-alcohols of the alkyl esters in the four European birches ranged from nC20 to nC32. Sometimes, the dominant esterified n-alcohol was found to be the same as the dominant free homologue in the same species (Fig. 3). For example, in B. humilis nC20 and in B. pubescens nC22 were the dominant free and esterified n-alcohol, respectively. The release of 100 % bound n-alcohols of both species increased the total amount by about 25%. However, the relative distributions remained comparable. The previously identified dominance of the free nC22 alcohol in B. pendula was reduced by the release of a high proportion of nC24. In contrast, the distribution in B. nana changed from a dominance of nC28 to a bimodal distribution maximizing at nC24 and nC28 due to the addition of ester-bound n-alcohols. The dominance of free nC22 alcohol in B. pendula was reduced by the release of the high proportion of nC24.
Our model indicates that the decay of the n-alkyl ester can significantly affect the original free lipid composition of birch leaves, which in sediments may complicate an unambiguous assignment based on n-alcohol and n-alkanoic acids.
Wax lipids from Betula grown in Kiel compared with literature data
To consolidate the application of the lipid composition of the four European birches for paleovegetation reconstruction and chemotaxonomic differentiation, we compared our wax lipid data of birch trees grown in the Botanical Garden in Kiel with previously published data (Table 1). However, solely the n-alkane composition of B. nana, B. pubescens and B. pendula can be compared, as to the best of our knowledge, the other wax lipid classes were not examined in full in previous studies. To our best knowledge, there are no previous studies on the wax lipid composition of B. humilis leaves, thus we do not have any complementary data for comparison. For a better comparison of published data, we have recalculated the distributions given in absolute concentrations into relative abundances, to exclude variation in absolute concentration due to different extraction techniques or analytical protocols. Different extraction techniques and analytic procedures used in the cited studies are briefly described here:
- Extraction of epicuticular waxes by immersion of the hole leaf into solvent mixture with or without ultrasonic bath, e.g. DCM:hexane (1:1) or pure DCM (our study, [77]);
- Prior to extraction, grounding or milling of leaves into a fine powder, followed by ASE (extraction under elevated temperature and pressure) or Soxhlet extraction [78–80]; here, intra-cuticular waxes potentially have been extracted;
- Hydrolysis (10% KOH in ethanol) of ground leaves to extract bound lipids [81]; esterified alkanoic acids and alcohols were released and increased the pool of their free homologues.
Most leaves from B. nana are characterized by a variable distribution of n-alkanes ranging from nC23 to nC33. Samples from very northern latitudes such as Greenland, Alaska, Siberia and Norway had a high proportion of nC27 to nC31 homologues [80–82]. Conversely, B. nana leaves from Siberia and Norway can be distinguished from the other B. nana samples as these showed significant abundances of mid-chain nC23 and nC25 alkanes depressing relative amounts of C29 and C31 homologues [83,84]. Leaf n-alkanes of B. pubescens from Scotland were bimodally distributed with maxima at nC27 and nC31 [85]. In contrast, a unimodal distribution with a maximum at nC27 was reported from Pagani et al. (2006), Ronkainen et al. (2015), and Balascio et al. (2018), with variable proportions of nC23 and nC25, respectively. A dominance of C27 alkane in leaves from B. pubescens has also been reported from Schwark et al. (2002) and Sachse et al. (2006). The latter author stated that only in birches (both B. pubescens and B. pendula) from northern Scandinavia nC27 is the dominant homologue, while species from southern Scandinavia and Germany maximized at nC31. This shift in n-alkane carbon chain-length was also expressed in an increasing ACL from North to South [87]. In contrast to this, Mayes et al. (1994) found a prevalence of C25 in leaves from Norway.
Similar to the leaves of B. pubescens, those of B. pendula contained high proportions of C25, if the distribution maximized at nC27 and then constituted only minor long-chain homologues with more than 29 carbon atoms [78,79,88,89]. Two B. pendula species from Estonia and the UK, when grown under artificial laboratory conditions had a distinct bimodal distribution with maxima at C27 and C31 [77,90].
Due to the complexity of the n-alkane patterns both within a species and between different species, we subdivided the data according to distributions of n-alkanes in each species into three groups to improve comparison. The published data of each species were subdivided into two groups (type I and type II) of similar composition and compared with the lipid distribution of the birches from Kiel (Fig. 4, Table 2). Type I of each species had a wax lipid composition similar to the Betula species from Kiel and was characterized by a dominance of long-chain n-alkanes (nC27, nC29, nC31). In contrast, type II of each species was defined by a high presence of mid-chain n-alkanes (nC23, nC25), but also nC27, and only minor quantities of long-chain n-alkanes with more than 29 carbon atoms.
All Betula tree species from Kiel and from globally distributed type I were distinct from grasses and shrubs by a prominent prevalence of the nC27 alkane (Fig. 1 and 4). Grasses are mainly dominated by very long-chain n-alkanes with nC31 and nC33 or even nC35, and therefore are characterized by a high ACL (>30) [44,49,67]. Typical pioneer shrubs of the late glacial period such as Artemisia sp. or Junipers sp. are also characterized by a dominance at nC31 to nC35, which is not prevalent in Betula species [45,91,92]. Other prominent species such as Pinus sp., which are probably the most widely spread conifer species in Europe, synthesize n-alkanes in the range from nC27 and nC31 [45,93]. However, the quantitative amounts are significantly lower, pointing to subordinate proportions of sedimentary n-alkanes originating from pines [47].
The type II birch species had n-alkane distributions similar to aquatic macrophytes and non-emergent (submerged and floating) aquatic plants, with a prevalence of mid-chain n-alkanes (nC23, nC25). These homologues can be a major source to geological archives, especially in lake sediments, often expressed in the Paq proxy [54,94]. It has been postulated that terrestrial plants correspond to Paq < 0.1, emergent macrophytes to Paq 0.1 – 0.4 and non-emergent macrophytes to Paq 0.4 – 1.0 to [54]. Each type II birch species, as well as B. humilis from our study, had a Paq value above 0.75 corresponding to a non-emergent macrophyte. Therefore, when applying the Paq proxy in the study of lake sediments receiving birch input, it must be considered that mid-chain n-alkanes may derive from either aquatic macrophytes or alternatively from birch trees.
n-Alkane abundance or chain-length based indicators like ACL have been used as a proxy for temperature, aridity, geographic location, or vapour pressure deficit [95–98]. The variation of the ACL compiled for B. nana did not show a trend with climatic drivers. The B. nana trees listed, mostly derived from cold environments such as Alaska, Greenland or Siberia and varied in their ACL between 26.4 and 28.9 [80,82,83]. Even two samples both originating from Norway varied by over 2 units in their ACL [81,84]. The B. nana investigated from Kiel, and therefore from the warmest region, did not reveal the highest ACL, but rather values between those of leaves from Greenland and Alaska. The ACL distributions of B. pubescens revealed higher values in samples from a moderate climate (Kiel, Scotland), whereas samples from colder regions (Siberia, Norway) had a lower ACL. Under certain assumptions, this can be attributed a geographical or temperature effect. The origin for the B. pubescens leaves from the study by Pagani et al. (2006) were not reported. Similar to B. nana, no latitudinal or temperature trend in the n-alkane distribution of B. pendula wax was observed.
Overall, a high variability of wax n-alkanes within the individual species was noted without a temperature or geographical trend. This may suggest that genetic differences between the populations control wax lipid composition, preferentially. It is conceivable that not only “pure-bred” birches of the respective species were examined in these studies, but also subspecies or varieties. For example, Betula pubescens has several varieties that occur naturally in a narrow space like var. pubescens, var. fragrans, var. litwinowii and var. pumila [15]. Wild hybridization may also affect leaf wax composition, whereby hybridization readily occurs between species with the same chromosome number like B. pendula × B. nana (diploid × diploid), but there are also reports of interploidy-level hybrids like pendula × B. pubescens (diploid × tetraploid) [99,100]. Therefore, future studies may address the association between ploidy level and wax lipid composition of Betula species to investigate species determination.
Analytical and extraction protocols used in previous Betula studies
Different analytic protocols have been used to extract the plant lipids as briefly descripted above. Previous studies have shown that the length of the extraction time, as well as the solvent used, had an influence on extraction yield or lipid extract composition [37,46]. Thus, n-alkanes and n-alcohols were extracted earlier than n-alkanoic acids and long-chain homologues earlier than the shorter ones [101]. Jetter et al. (2008) indicated extraction yields of n-alkanes depended on polarity of binary solvent mixtures. Moreover, saponification upon extraction [81], hydrolyzed wax esters leading to enhanced release of bound n-alcohols and n-alkanoic acids adding to the proportion of the free homologues. Since this comparative investigation used n-alkane distributions from different studies with different extraction methods, the results are not unequivocally comparable. For a better comparability of future work, the influence of the extraction method on the other lipid classes including n-alcohols, n-alkanoic acids and n-alkyl esters should be investigated and a standard extraction protocol established.
Conclusion
The leaves of four Betula species, B. nana, B. humilis, B. pubescens, B. pendula, which are endemic in Europe were studied, aiming to investigate their epicuticular wax lipid composition. The following conclusions can be drawn from this study. The n-alkane compositions in leaves of Betula species from Kiel were found to be specific, allowing unambiguous differentiation. Betula wax n-alcohol and n-alkyl ester composition allowed a distinction to be made between B. nana, B. humilis and the two birch trees, however the latter two cannot be easily distinguished from each other due to a similar fingerprint. The n-alkanoic acids seemed to be less suitable for species differentiation since all four species were dominated by the C28 alkanoic acid, however with variations in concentration of about two orders of magnitude. A flowchart (Fig. 5) provides a simple means for discrimination of epicuticular waxes from the four birches from Kiel University.
The n-alkyl esters consisted of different isomers with varying n-alcohol and n-alkanoic acid moieties. In the species B. humilis and B. pubescens, the dominant esterified alcohol also was the dominant free alcohol, therefore the n-alcohol patterns in sediments would not be disturbed by hydrolysis of the wax esters. In B. nana and B. pendula the n-alcohol distribution changed substantially upon ester hydrolysis, when bound homologues were released. Due to the preponderance of short-chain ester-bound alkanoic acids in wax esters, the distribution of free long-chain alkanoic acids was only slightly impaired. The ratio was influenced in B. nana only, as large amounts of bound C20 were released.
When comparing the n-alkane composition of the Betula waxes collected in Kiel with published data, no trend in geographical location or temperature could be identified. It appears that Betula wax composition is genetically controlled, and differences occur due to presence of plant hybrids or variants.
Acknowledgements
T. Martens and V. Grote are thanked for laboratory assistance during lipid extraction. S. Petersen is thanked for her advice during leaf sampling in the Botanical Garden of Kiel University.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.
- 7.
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.
- 14.
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.↵
- 31.↵
- 32.
- 33.
- 34.
- 35.↵
- 36.↵
- 37.↵
- 38.
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.
- 97.
- 98.↵
- 99.↵
- 100.↵
- 101.↵