Next Article in Journal
Enantioselective, Decarboxylative (3+2)-Cycloaddition of Azomethine Ylides and Chromone-3-Carboxylic Acids
Next Article in Special Issue
Volatile Organic Compound Profiles of Cystoseira corniculata (Turner) Zanardini 1841 and Ericaria amentacea (C.Agardh) Molinari and Guiry 2020 (ex. Cystoseira amentacea (C.Agardh) Bory de Saint-Vincent, 1832)
Previous Article in Journal
Wet End Chemical Properties of a New Kind of Fire-Resistant Paper Pulp Based on Ultralong Hydroxyapatite Nanowires
Previous Article in Special Issue
Integrating Enhanced Profiling and Chemometrics to Unveil the Potential Markers for Differentiating among the Leaves of Panax ginseng, P. quinquefolius, and P. notoginseng by Ultra-High Performance Liquid Chromatography/Ion Mobility-Quadrupole Time-of-Flight Mass Spectrometry
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Molecules
Volume 27
Issue 20
10.3390/molecules27206810
molecules-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Chemotaxonomic Markers for the Leaf Buds of Common Finnish Trees and Shrubs: A Rapid UHPLC MS Fingerprinting Tool for Species Identification

by
Marianna Manninen
,
Maarit Karonen
* and
Juha-Pekka Salminen
Natural Chemistry Research Group, Department of Chemistry, University of Turku, FI-20014 Turku, Finland
*
Author to whom correspondence should be addressed.
Molecules 2022, 27(20), 6810; https://doi.org/10.3390/molecules27206810
Submission received: 13 September 2022 / Revised: 6 October 2022 / Accepted: 7 October 2022 / Published: 11 October 2022
(This article belongs to the Special Issue Chromatographic Science of Natural Products III)
Browse Figures
Versions Notes

Abstract

:
In this study, a chemotaxonomic tool was created on the basis of ultra-high-performance liquid chromatography–mass spectrometry (UHPLC–MS) for the identification of 13 common Finnish deciduous trees and shrubs from their leaf bud metabolites. The bud extracts were screened with UHPLC–ESI–QqQ–MS and UHPLC–ESI–Q–Orbitrap–MS to discover suitable markers for each species. Two approaches were tested in the marker selection: (1) unique species-specific markers to obtain selective fingerprints per species and (2) major markers to maximise the sensitivity of the fingerprints. The markers were used to create two selected ion-recording-based fingerprinting tools with UHPLC–ESI–QqQ–MS. The methods were evaluated for their selectivity, repeatability, and robustness in plant species identification by analysing leaf buds from several replicates of each species. The created chemotaxonomic tools were shown to provide unique chromatographic profiles for the studied species in less than 6 min. A variety of plant metabolites, such as flavonoids, triterpenoids, and hydroxycinnamic acid derivatives, were found to serve as good chemotaxonomic markers for the studied species. In 10 out of 13 cases, species-specific markers were superior in creating selective and repeatable fingerprints.

1. Introduction

The chemical diversity of plants and the special chemical features of different plant families, genera, and species provide exciting possibilities for their discrimination and chemotaxonomic classification. Such discrimination is useful for not only educational [1] and species identification purposes [2] but also quality control [3,4,5] as different types of plant-based natural products increase in their popularity. Depending on the need, know-how, and available equipment, different types and levels of tools are needed to execute these discriminative or chemotaxonomic actions.
The development of modern liquid chromatographic (LC), mass spectrometric (MS), and LC–MS fingerprinting tools for plant extracts offer a vast amount of information that can be used in multiple ways to reveal species similarities and differences. For instance, HPLC–UV was used by Lahtinen et al. [2] to distinguish Betula pubescens- and Betula pendula-type birch species, and HPTLC (high-performance thin-layer chromatography) was used by Melnyk et al. [3] for the identification of lime flowers. Direct Analysis in Real Time (DART) mass spectrometry has been shown to be a fast and reproducible tool in distinguishing Datura species from one another via their seeds and in differentiating several Salvia species from each other [6,7]. Furthermore, the LC–MS data of 74 medicinal plant extracts and machine learning were employed by Kharyuk et al. [8] for the identification of plant species and plant organs, with over 90% accuracy in classification.
For common Finnish deciduous trees, there is a wide range of structurally different compounds described in the scientific literature that could serve as chemotaxonomic markers. Populus tremula is a good example, with phenolic acid glycerols characteristic for its leaf buds [5], whereas anthocyanins have been suggested as chemotaxonomic markers for the anthers of different Populus species [9]. In addition, Abreu et al. and Julkunen-Tiitto found that salicylate-like simple phenolic glucosides could serve as chemotaxonomic markers at the genus level for both Populus and Salix species [10,11]. Species-specific, genus-specific, and family-specific chemotaxonomic markers have also been identified for other common Finnish deciduous trees. Flavonoids have been shown to be a useful fool for distinguishing B. pubescens- and B. pendula-type birch species [2], as well as for characterising the flowers of different Tilia species [3]. Diarylheptanoids are characteristic of the genus Alnus, and diarylheptanoids of bark extracts of Alnus incana and Alnus glutinosa have been shown to be reliable indicators for identification and discrimination between the species [12]. Triterpenoids have potential to serve as chemotaxonomic markers in the genus of Sorbus, since some Sorbus species have provided structurally novel compounds and a substantial part of crude plant material (especially of fruits) is constituted by triterpenoids [13]. Almost all the genera and species of the family of Oleaceae contain iridoids, and the occurrence of iridoids from the different biosynthetic pathways correlate well with phylogenetic classification [14].
The identification of a tree species from a leaf bud is not a trivial task, but LC–MS/MS could be a useful tool for this identification task. However, only one study describing an LC–MS-based method for the identification of B. pubescens and B. pendula from leaf buds has been published thus far [1]. The primary aim of the current study was to compare two different approaches for the selection of the most suitable chemotaxonomic markers for the leaf buds of 13 common Finnish deciduous trees. Another aim was to utilise these markers in a simple, rapid, and repeatable LC-ESI-QqQ tool that produces such species-specific fingerprints that are easy to interpret.

2. Results and Discussion

2.1. Two Approaches in the Development of the LC–MS Fingerprinting Method and Selection of Marker Candidates

We used different steps in the method development, which are visualised as a flow chart in Figure 1. The leaf bud extracts were screened with two UHPLC–MS instruments to detect all potential marker candidates for each species. Two approaches were applied for the marker selection: one for the most species-specific markers and one for the most species-sensitive markers (described in more detail in Section 2.1.1. and Section 2.1.2.). The selected markers were used to create selected ion recording (SIR) methods for species fingerprinting with UHPLC–QqQ–MS. The individual SIR methods of a species were grouped together to enable the rapid acquisition of species-specific sum traces of all the selected SIR chromatograms. The repeatability and specificity of the fingerprints were estimated by analysing replicate leaf buds from 4–10 plant individuals per species and qualitatively comparing the acquired fingerprints.

2.1.1. Method I: Species-Specific Markers from the High-Resolution MS Data Obtained with MZmine 2

High-resolution MS data were obtained for one replicate of each species and transferred for further processing into MZmine 2, which is an open-source software for mass spectrometric data processing. It was used to create extracted ion chromatograms (EICs) for all ions with an intensity above 1×105 and to align all features (i.e., detected variables with a retention time and an m/z ratio) into a feature list. The feature list enabled the comparison of the presence of detected ions in different leaf bud extracts. The data also revealed differences in the chemical diversity of the leaf bud extracts. The results showed that the chemical diversity between species from the same plant genus was similar. Both Alnus species produced a high number of detected features, making them the most chemically diverse species among the studied species (Figure 2). The smallest number of detected features was obtained for the Tilia species. In addition to the Alnus and Tilia species, the numbers of detected features from both Sorbus species were similar. Additionally, these species were similar in terms of the number of features in different categories according to the peak areas.
The marker candidates were chosen from the aligned feature list by excluding features that could be detected in multiple species. For each species, between three and seven marker candidates with the highest peak area were chosen for further testing. Since the goal was to create a simple fingerprinting method for QqQ–MS, the presence of the candidates in the QqQ–MS data was confirmed from the full scan MS data. To maximise the ion intensity of the chosen markers, a preliminary SIR test was performed with different cone voltages using four replicate leaf buds of the same plant individual. The cone voltage producing the highest ion intensity was used in the method for the testing of repeatability and specificity.
With this approach, the markers specific to S. aucuparia and present in the QqQ–MS data could not be found, but common markers to both S. hybrida and S. aucuparia were found. However, markers specific to S. hybrida were discovered, so two SIR methods were created for Sorbus species. The first method was able to distinguish Sorbus species from other species, while the second method was able to identify S. hybrida species, thereby enabling us to separate S. hybrida and S. aucuparia samples from each other.

2.1.2. Method II: Main Ions from the QqQ Full Scan Spectra

As a comparison to the previously described approach, a more straightforward approach for choosing marker candidates was applied. The full scan MS screening was conducted for all replicates of the studied species. The results showed that the full scan spectra were repeatable within species (data not shown). For each species, from three to six main ions from the full scan spectrum were chosen as the marker candidates. We omitted the time range of 0.00–0.25 min from the full scan spectrum to rule out the most polar and early eluting compounds such as sugars and other primary metabolites that are likely to be present in many species. This also enhanced the uniqueness and species-specificity of the fingerprints since most of the species would have otherwise had SIR detection at the same retention time window.
The full scan spectra of both Sorbus species were similar in terms of the main ions (Supplementary Materials, Figures S9 and S10). However, the intensity ratios of the main ions were different between the species, indicating that the combined SIR traces would have different profiles. Thus, only one joint method was created for these species. Neither of the two approaches used were able to find markers to reliably differentiate Tilia cordata and Tilia × europaea. The same main ions were present in both species (Supplementary Materials, Figures S12 and S13), and the marker candidates from the MZmine data could not be reliably detected from the QqQ–MS data. Instead, marker candidates for the differentiation of Tilia species from other species were discovered with both approaches. Consequently, only two methods that were specific to both Tilia species were created, one using the marker candidates from MZmine and the other one using the main ions as markers.

2.1.3. Repeatability of the Fingerprints

Fingerprint repeatability was evaluated by comparing the 4–10 replicate fingerprint profiles within species. The fingerprints were considered repeatable if the main peaks were the same in all replicates and no additional peaks were detected. A slight difference in the intensities of the main peaks compared to each other was considered acceptable. The fingerprints of all replicates of all species with the species-specific methods I and II can be found in the Supplementary Information.
As was expected on the basis of the good repeatability of the full scan mass spectra, good repeatability was obtained for all species with method II. For example, all replicates of S. hybrida, A. glutinosa, and S. phylicifolia produced the same main peaks with the fingerprinting methods using the main ions (Figure 3). With method I, fingerprints of 10 species were repeatable. Some challenges were detected with the remaining three species (S. hybrida, A. glutinosa, and S. phylicifolia). However, A. glutinosa and S. phylicifolia could still be correctly identified due to other peaks in the fingerprint.
Method I for separating Sorbus species from other species provided similar results for all S. aucuparia and S. hybrida samples. However, the method for S. hybrida provided the expected result for 8 out of 10 S. hybrida samples (Figure 4A). The fingerprints of the other two S. hybrida samples were noisy and lacked the characteristic peaks. Thus, the reliable identification of S. hybrida was not possible. An additional three leaf buds of both of these exceptional plant individuals were analysed, and two out of three replicates produced the correct fingerprint with the S. hybrida method. The repeatability issue could therefore be related to the low concentration of the marker compounds in the leaf buds.
The fingerprints of all replicates of A. glutinosa obtained with method I exhibited the same two main peaks. However, the fingerprints of two replicates showed additional peaks at a later retention time (Figure 4B). Therefore, the method was not 100% reliable. Identification was still possible, as the other species did not produce a fingerprint that could have been misinterpreted as A. glutinosa. Similarly, the fingerprints of all S. phylicifolia samples had the same main peak, which was not present in the fingerprints of any other species. The fingerprints of two replicates had an additional peak at 1.57 min, which was not observed in all the replicates (Figure 4C).

2.1.4. Specificity of the Fingerprints

The specificity of the fingerprints was evaluated by comparing the fingerprint of a selected species to the fingerprints of other species with the same species-specific method. The specificity of the fingerprints was studied from two perspectives: (1) is the intensity of the peaks in the fingerprint of the studied species significantly higher compared to the other species and (2) does the profile of the fingerprint clearly differ from the fingerprint profiles of other samples with the same method? Good specificity regarding intensity would enhance the readability of the results, as scaling the y-axes of all samples according to the most intensive peak of the studied species would lead to the species in question standing out from other species. Good specificity regarding the fingerprint profile meant that the fingerprint profile of the species in question was unique, which made it easier to compare the fingerprints of different species.
For example, method I for S. vulgaris was more specific regarding intensity, as the traces of other species dropped to baseline when the y-axes were scaled according to the most intensive peak of S. vulgaris (Figure 5A). Method I for S. vulgaris was also more specific regarding the fingerprint profile. S. vulgaris was the only species that produced three clear peaks, while the fingerprint profiles of other species were noisier and exhibited none of the main peaks of the fingerprint of S. vulgaris (Figure 5B). With method II for S. vulgaris, the fingerprints of other species were more dominant (Figure 6A). Furthermore, the profile-based specificity was poorer compared with that of method I, as most of the fingerprints of other species had the same peak at 2.73 min, which was also present in the fingerprint of S. vulgaris (Figure 6B).

2.2. Comparison between Different Fingerprinting Methods

Information on the repeatability and specificity of different fingerprinting methods is summarised in Table 1. For most of the species, both methods were valid. For some species, even the same ions were chosen as markers with both methods. The better specificity of method I resulted in its selection as the final method for most species. However, method II produced clearer and more repeatable results for both Sorbus species, and for that reason, method II was chosen as the final method for Sorbus species. Similarly, method II was chosen as the final approach for A. glutinosa due to its better repeatability.
The final fingerprinting method resulted in unique chromatographic profiles for each plant species with the species-specific SIR method (Figure 7). The results reinforced the observation of the chemical similarity of the two Sorbus and Tilia species, which was also noted from the number of features in the MZmine data. In contrast, no difficulties were encountered in finding markers to Alnus species, indicating that, even though the chemical diversity of both species was significant, qualitative differences could easily be found.

2.3. Identification of the Markers Based on the UHPLC–DAD–QOrbitrap–MS/MS Data

Among the studied species, the Orbitrap data of the markers led to the identification of 42 individual compounds that were analysed in detail. The characterisation was based on the comparison of UV spectra, mass spectra, and MS/MS fragmentation with published data. The data revealed a diverse range of phenolic compounds and triterpenoids. Table 2 summarises the UHPLC–DAD–QOrbitrap-MS/MS data of the markers including UV maxima, retention times, exact masses, and main fragment ions. However, these data could not reveal the exact positions for the substituents and functional groups nor the stereochemistry, so other isomers are possible. The Supplementary Information includes the MS/MS spectra for all markers and molecular formulas for the main fragment ions, along with the proposed fragmentation patterns for the suggested compounds.

2.3.1. Flavonoids

Some of the markers were identified as flavonoids. The main peak in the LC–MS fingerprints of S. phylicifolia samples was marker 3 at m/z 319. It was identified as dihydromyricetin, which has been previously reported in S. phylicifolia [15,16]. The fragments in the MS/MS spectrum at m/z 125 and 193, resulting from the loss of the B ring, were consistent with those reported for dihydromyricetin [17].
A similar UV spectrum to that of dihydromyricetin was obtained for marker 4, which was the main peak in the LC–MS fingerprints of Tilia samples. The exact mass matched dihydroquercetin glucoside, the main fragments at m/z 285 could correspond to the loss of glucose and water, and the fragment at m/z 151 could correspond to the loss of the A ring, all of which have been previously reported for dihydroquercetin derivatives [18,19]. The flowers of different Tilia species have been shown to contain different flavonoids such as flavonols and flavanones [20], but dihydroflavonols have not been reported in Tilia species.
Another proposed flavonoid derivative is marker 21, which was a minor peak in the fingerprints of P. padus. The MS/MS spectrum indicated a flavanone aglycone naringenin with the [A–H]− ion at m/z 271. The other fragments at m/z 119 and 151 could be attributed to the loss of the B and C rings and the loss of the A ring, respectively [21]. Naringenin and its derivatives have been previously found in different plant parts of the Prunus species [22,23].

2.3.2. Hydroxycoumarins

Marker 5 was a minor peak in the fingerprints of F. excelsior samples. The MS/MS spectrum showed fragments at m/z 163, 191, 206, and 221, which were consistent with those reported for fraxidin derivatives [24], but no further conclusions could be drawn about the side group. Fraxidin derivatives were previously found in F. excelsior [25].

2.3.3. Glycerol Esters of Cinnamic Acid Derivatives

All four markers of P. tremula were found to be phenolic acid glycerols, and (except for 2-acetyl-1,3-di-feruloyl glycerol (24)) they have previously been reported in P. tremula leaf buds [5]. Marker 22 was identified as 2-acetyl-1,3-di-p-coumaroyl glycerol, and its MS/MS spectrum included fragment ions characteristic of coumaric acid at m/z 117, 119, 145, and 163. Marker 23 was 2-acetyl-3-p-coumaroyl-1-feruloylglycerol, and marker 24 was 2-acetyl-1,3-di-feruloylglycerol. Marker 23 shared the same fragments with marker 22, both containing coumaroyl moieties. The additional fragments of marker 23 at m/z 134, 160, 175, and 193, were found in the MS/MS spectrum of marker 24, which also contained feruloyl moieties. Marker 19 was identified as 2-acetyl-1,3-di-caffeoylglycerol, and it showed fragments characteristic of caffeic acid at m/z 133, 135, 161, and 179. The observed fragments and the UV spectra of the markers were also consistent with those reported in previous studies [5,26]. In the fingerprints of P. tremula samples, 2-acetyl-1,3-di-caffeoylglycerol was shown as a separate, more intensive peak, while the other three phenolic acid glycerols formed a less intensive hump at a later retention time.

2.3.4. Other Cinnamic Acid Derivatives

Marker 13 was identified as monocaffeoylquinic acid, and it was a minor peak in the fingerprints of the Sorbus samples. Due to the coelution of other compounds, a reliable UV spectrum could not be obtained. However, the MS/MS spectrum showed typical fragments to caffeoylquinic acids: a [quinic acid–H]− ion at m/z 191 and a [caffeic acid–H]− ion at m/z 179, as well as its fragments at m/z 135 and 161. Monocaffeoylquinic acids have been reported in many species of the genus Sorbus [13].
Marker 18, the main peak in the fingerprints of P. padus samples, could also include a coumaric acid moiety due to fragments at m/z 117, 145, and 163. The UV spectra with a maximum at 315 nm indicates the same. The molecular formula matched penta-O-acetyl-p-coumaroylsucrose, whose isomers have been identified in Prunus mume [27]. However, the identification is speculative since the fragments only evidence a coumaroyl moiety.
The marker of Q. robur (39) was also tentatively identified as a caffeic acid derivative on the basis of the similarity of its fragments to those of other compounds identified with a caffeic acid moiety. The late retention time at 3.55 min suggests that the compound has a side chain that increases the retention.

2.3.5. Salicylate-like Phenolic Glycosides

Marker 20 was identified as tremulacin, a known compound in Salix species [11,28,29]. The fragmentation matched with that reported by Kammerer et al. [28]. However, it was not present in all S. phylicifolia samples, and dihydromyricetin (3) was therefore the main marker for the species. Tremulacin has also been reported in P. tremula [10], and it was also observed in the P. tremula samples of this study, although at a lower intensity.

2.3.6. Secoiridoids

All four markers of S. vulgaris were secoiridoids, commonly known markers in Oleaceae [14]. Two of the markers, demethyloleuropein (10) and demethylligstroside (14), are oleuropein-like secoiridoids, and they have been characterised in S. vulgaris flower and fruit extracts [30]. The other two markers were 2″-epi-frameroside (16) and hydroxyframoside (17). Additionally, both of these have been observed in different parts of S. vulgaris previously [31]. Demethyloleuropein, demethylligstroside, and 2″-epi-frameroside were the three main peaks observed in the fingerprints of the S. vulgaris samples.

2.3.7. Phenylethanoids

Markers 9 and 12 were identified as calceolariosides A and B, respectively, which were shown in the F. excelsior LC–MS fingerprints as the main peak with a shoulder. In both cases, the MS/MS spectrum showed caffeic acid at m/z 179 and its fragments at m/z 133 and 161. Additionally, an [M–H–162]− ion at m/z 315 was observed, indicating the loss of a caffeic acid moiety. Eyles et al. reported calceolariosides A and B in other Fraxinus species, and according to their data, calceolarioside A elutes before B in reverse-phase chromatography [24]. Therefore, markers 9 and 12 are proposed to be calceolarioside A and calceolarioside B, respectively.

2.3.8. Benzoic Acid Glycosides

All four markers of A. platanoides were gallic acid and/or syringic acid glycosides. Glucosyringic acid (1) was observed as an [M–H]− ion at m/z 359, and it has been found in Acer saccharum Marsh. buds previously [32]. The MS/MS spectrum showed an [M–H–162]− or [syringic acid–H]− ion at m/z 197, and an [M–H–44–30]− ion at m/z 123 due to carboxyl and methoxyl losses. Markers 6, 7, and 8 shared some fragment ions with the glucosyringic acid. Additionally, the MS/MS spectrum showed an [M–H–198]− ion at m/z 313 from the cleavage of syringic acid, an [M–H–152–14]− ion at m/z 345, and a [gallic acid–H]− ion at m/z 169. Therefore, the compounds were tentatively identified as monogalloyl syringyl glucoses. They eluted as three not fully separated peaks at 0.70, 0.79, and 0.88 min, suggesting three isomers. Marker 11 also showed evidence of galloyl and syringyl moieties. The MS/MS spectrum showed an [M–H–152]− ion at m/z 511 from the loss of gallic acid, an [M–H–198]− ion at m/z 465 from the loss of syringic acid, and an [M–H–152–198]− ion at m/z 313 (loss of galloyl group and syringic acid). The compound was suggested to be digalloyl syringyl glucose.
Marker 15 at m/z 621 was the [M–2H]2– ion of heptagalloyl glucose previously identified in A. platanoides [33]. The MS spectrum also showed an [M–H]− ion at m/z 1243. It eluted at 1.15 min and had UV maxima at 221 and 278 nm. The MS/MS spectrum showed an [M–H–152–152–170]− ion at m/z 769, an [M–2H–152–152]2– ion at m/z 469, an [M–H–152–152–152–170]− ion at m/z 617, an [M–2H–152–152–152]2– ion at m/z 393, and a [gallic acid–H]− ion at m/z 169.
All four markers contributed to the LC–MS fingerprints of A. platanoides samples. Glucosyringic acid was the most intensive peak at the earliest retention time. The second peak with a shoulder was monogalloyl syringyl glucose, the third peak was digalloyl syringyl glucose, and the fourth peak was heptagalloyl glucose.

2.3.9. Triterpenoids and Triterpenoid Derivatives

Several markers of Sorbus samples were triterpenoids or triterpenoid derivatives. The UV and MS/MS spectra of markers 27 and 29 indicated that both of them included an additional hydroxycinnamic acid moiety. Marker 27 had a UV maximum at 309 nm and fragment ions characteristic of coumaric acid. The additional fragment ions of [M–H–18]− at m/z 615 resulting from the cleavage of water, [M–H–44] − at m/z 589 resulting from the loss of a carboxyl group, and [M–H–18–44]− at m/z 571 resulting from the loss of both could correspond for the triterpenoid part, which was tentatively identified as rotundic acid. De Tommasi et al. found similar triterpenoid derivatives in Eriobotrya japonica, including 3-O-trans-p-coumaroylrotundic acid [34]. Marker 29 had the same molecular formula but different UV and MS/MS spectra. The UV maximum was at 322 nm, and the fragment ions were characteristic of caffeic acid. The molecular formula matched that of 2-O-caffeoylmaslinic acid, as reported by Yang et al. in Hippophae rhamnoides [35]. Marker 27 was shown in the fingerprints of both Sorbus samples as a small shoulder before the main peak at 2.59 min. Marker 29 did not contribute to the overall fingerprint, as marker 30 eluted at the same retention time and produced a more intensive peak.
Marker 30 exhibited the first intensive peak in the fingerprints of both Sorbus species. It was tentatively identified as cashmirol B. The loss of carboxylic acid could be observed from the MS/MS spectrum by an ion at m/z 423 with a low intensity. Cashmirol B has been previously reported in Sorbus cashmiriana [36].
Markers 35, 36, and 38 were identified as either oleanane- or ursane-type triterpenoids, both of which are common in Sorbus species [13]. It has been shown that these types of pentacyclic triterpenes produce few fragment ions when negative ionisation is used [37]. However, a minor fragment ion at m/z 451 was observed for marker 35, which could have been due to the cleavage of the acetyl group. On the basis of the molecular formula, the compound could be oxoursolic acid acetate or oxooleanolic acid acetate. No fragments could be obtained for marker 36. The exact mass matched with oleanonic acid or ursonic acid. The fragment ion at m/z 437 in the MS/MS spectrum of marker 38 also indicated the cleavage of an acetyl group, and the compound could therefore be tentatively assigned as acetyl ursolic acid or acetyl oleanolic acid. Marker 36 exhibited one of the main peaks in the fingerprints of both Sorbus species. Marker 38 was the third main peak at 3.06 min in the fingerprints of S. hybrida samples, but it was only detected in the S. aucuparia samples as a small shoulder. Marker 35 was observed in the fingerprints of Sorbus hybrida as a small shoulder at 2.82 min, but it did not contribute to the overall fingerprint of the S. aucuparia samples, even though it could be observed in the EIC.
In addition to the Sorbus species, the markers of A. glutinosa were identified as triterpenoid derivatives. In the fingerprints of the A. glutinosa samples, markers 31 and 33 exhibited one intensive peak without a proper separation between the two isomers. They were tentatively identified as the isomers of curculigosaponin B. The MS spectrum showed both an [M–H]− ion at m/z 605.4 and an [2M–H]− ion at m/z 1211.8. The MS/MS spectrum showed an [M–H–132]− ion at m/z 473 resulting from the cleavage of arabinose and a minor [M–H–132–18]− ion at m/z 455 resulting from the additional cleavage of water. Although curculigosaponin B has only been previously reported in Curculigo orchioides [38], other tetracyclic triterpenes have been reported in several Alnus species [39]. Markers 32 and 34 did not contribute to the overall fingerprints of the A. glutinosa samples because of the higher intensity of the other markers. They were also identified as saponins, isomers, of alnustic acid arabinoside. The MS/MS spectrum showed aglycone at m/z 487. The loss of the carbon side chain yielding a fragment ion at m/z 389 and the following cleavage of carboxyl group or water resulting in the ions at m/z 345 and 371, respectively, were also detected in the MS/MS spectrum. Alnustic acid arabinoside was initially characterised in Alnus serrulatoides [40], and later discovered in some other Alnus species as well [41].
Compared with the other markers of A. glutinosa, the MS/MS spectrum of markers 25 and 26 had similar characteristics. The lack of a UV spectrum and similar retention times suggested that markers 25 and 26 could be structurally similar to markers 31, 32, 33, and 34, meaning that they could also be triterpenoid saponins, although with a different sugar part because the neutral loss of glucose or arabinose could not be detected from the MS/MS spectrum. Markers 25 and 26 produced the first main peak in the fingerprints of the A. glutinosa samples.

2.3.10. Other Compounds

Markers 2, 28, 37, and 40–42 could not be reliably assigned to any of the compound classes discussed above. The molecular ion of marker 2 had an even m/z value, which suggests that it contains and odd number of nitrogen atoms. Marker 2 was a minor peak in the fingerprints of F. excelsior. Marker 28 was detected in all of the studied species. However, the combination of 28 and 37 detected as two peaks was not obtained for any other species than the Tilia species. Markers 40–42, found in A. incana, shared similar MS/MS characteristics, which could indicate that they belong to the same molecular family. The smallest fragments could have been derived from a coumaroyl moiety.

3. Materials and Methods

3.1. Chemicals

LC–MS-grade acetonitrile was purchased from Merck (KGaA, Darmstadt, Germany), and LC–MS-grade formic acid was purchased from VWR International (Fontenay-Sous-Bois, Paris, France). Water was purified with the Millipore Synergy water purification system (Merck KGaA, Darmstadt, Germany). Ethanol (99.5%, Aa grade) was purchased from Altia (Rajamäki, Finland).

3.2. Plant Samples

Leaf bud samples were collected from the Turku area, Southwest Finland, in spring of 2019, 2020, and 2021. After collection, the samples were frozen, lyophilised, and stored in a freezer. The extraction protocol was modified from Lahtinen et al. [2]. The extracts were prepared by dropping one leaf bud into 2 mL of ethanol–water (95/5, v/v) solution and shaking for 10 min. The extract was filtered with a 0.20 µm PTFE filter.

3.3. UHPLC–QqQ Full Scan Screening and Fingerprinting Analysis

The screening for markers and final fingerprinting analysis was performed using an Acquity UPLC system (Waters Corporation, Milford, MA, USA) coupled with a Xevo TQ triple−quadrupole mass spectrometer (Waters Corp.). The UPLC system consisted of a sample manager, a binary solvent manager, a column (Acquity UPLC BEH Phenyl 30 mm × 2.1 mm, 1.7 μm, Waters Corporation, Ireland), and a diode array detector. The mobile phase consisted of acetonitrile (A) and water/formic acid (99.9:0.1, v/v) (B). The elution profile was as follows: 0–0.3 min, 10% A in B; 0.3–3.1 min, 10–75% A in B (linear gradient); 3.1–3.5 min, 75% A in B; 3.5–3.6 min, 75–95% A in B; and 3.6–5.0 min column wash and stabilisation. The flow rate was 0.65 mL/min, and the injection volume was 5 μL. Mass analyses were performed using an ESI source and negative ionisation. The ESI conditions were as follows: capillary voltage, 1.8 kV; source temperature, 150 °C; desolvation temperature, 650 °C; desolvation and cone gas (N2), 1000 and 100 l/h, respectively; and collision gas, argon. The mass range for the full scan screening was set to m/z 150–2000. The selected ion recording (SIR) parameters are presented in Table 3.

3.4. UHPLC–QOrbitrap–MS/MS Analysis

A similar Acquity UPLC system was configured with a hybrid quadrupole–Orbitrap mass spectrometer (QExactive, Thermo Fisher Scientific GmbH, Bremen, Germany). The column and gradient were similar to the ones used in the QqQ analysis. The injection volume was 5 μL, and the flow rate was 0.65 mL/min. The heated ESI source (H-ESI II, Thermo Fisher Scientific GmbH, Bremen, Germany) was operated in the negative ion mode. The parameters were set at as follows: spray voltage, −3.0 kV; sheath gas (N2) flow rate, 60 (arbitrary units); aux gas (N2) flow rate, 20 (arbitrary units); sweep gas flow rate, 0 (arbitrary units); and capillary temperature, +380 °C. The in-source collision-induced dissociation energy was 30 eV. A resolution of 35,000 and an automatic gain control of 3×106 were used for full scan MS data. The mass range was set to m/z 150−2250. MS/MS data, namely, dd-MS2 (Top N) data, were acquired using a resolution of 17,500; an automatic gain control of 1×105; a TopN of 7; and stepped normalised collision energies (NCEs) of 30, 50, and 80. The calibration was performed with a Pierce ESI Negative Ion Calibration Solution (Thermo Fisher Scientific Inc., Waltham, MA, USA) and was the most accurate at m/z > 250. The data were processed with Thermo Xcalibur Qual Browser software (Version 4.1.31.9, Thermo Fisher Scientific Inc., Waltham, MA, USA).

3.5. Data Analysis by MZmine 2

The MS/MS data files were converted from the Thermo Scientific .raw data format to .mzXML with the MS Convert Software included in the ProteoWizard package [42]. The .mzXML files were submitted to the MZmine 2.53 [43]. The mass detection was performed with noise levels of 1×105 (for MS scans) and 1×104 (for MS/MS scans). MS chromatograms were built using the ADAP chromatogram builder [44], with a minimum group scan size of 5, a group intensity threshold of 1×104, a minimum highest intensity of 3×105, and an m/z tolerance of 0.01. The chromatographic deconvolution was achieved by the local minimum search algorithm with the following settings: m/z range for MS/MS scan pairing of 0.01, RT range for MS/MS scan pairing of 0.1 min, chromatographic threshold of 75%, search minimum in RT range of 0.05 min, minimum relative height of 1%, minimum absolute height of 1×105, minimum ratio of peak top/edge of 1.5, and peak duration range of 0.2–1.0 min. Chromatograms were deisotoped using the isotopic peaks grouper algorithm with an m/z tolerance of 0.01 and an RT tolerance of 0.01 min. The chromatograms were then aligned together into a feature list with the join aligner module at an m/z tolerance of 0.01 (weight for m/z = 75; weight for RT = 25; absolute RT tolerance = 0.1 min). The aligned feature list was gap-filled using the multithreaded peak finder module (intensity tolerance of 10%, m/z tolerance of 0.001, retention time tolerance of 0.05 min) and filtered using the peak filter module by area (5 × 105–1 × 1010) and height (1×105–1×1010). Finally, the feature list was filtered using the feature list rows filter to remove peaks with fewer than two peaks in an isotope pattern.

4. Conclusions

In this study, two alternative approaches were applied in the selection of marker candidates for the leaf buds of 13 common Finnish tree species. The approach using MZmine was found to be suitable for discovering species-specific markers, and it demonstrated differences in the chemical diversity of the studied species. The main ions chosen by the second approach were less specific to the species, but they produced repeatable fingerprints. The final UHPLC–MS fingerprinting tool utilised a combination of markers obtained with the different approaches. The fingerprinting tool enabled species identification from a single leaf bud in less than 6 min. The high-resolution mass data revealed that the markers belonged to many different plant metabolite subclasses such as flavonoids, hydroxycinnamic acid derivatives, and triterpenoids. The structural diversity of the markers qualitatively demonstrated the diversity in the leaf bud chemistry of different species.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/molecules27206810/s1, Figures S1–S13: Orbitrap and QqQ full scan spectra for all species and the repeatability of fingerprinting. Figures S14–S60: Fingerprints of all studied species with species-specific methods I and II. Figures S61–S95 and Tables S1–S36: MS/MS spectra, fragmentation patterns, m/z values, and other mass spectrometric information of the fragments of the markers (1–42) used in the final fingerprinting methods.

Author Contributions

Conceptualisation, J.-P.S.; methodology, M.K. and J.-P.S.; software, M.M.; validation, M.M., M.K. and J.-P.S.; formal analysis, M.M.; investigation, M.M.; resources, J.-P.S.; data curation, M.M.; writing—original draft preparation, M.M.; writing—review and editing, M.M., M.K. and J.-P.S.; visualisation, M.M.; supervision, M.K. and J.-P.S.; project administration, J.-P.S.; funding acquisition, J.-P.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Alfred Kordelin Foundation, Finnish Cultural Foundation, Palomaa-Erikoski, and the Turku University Foundation (research grants for M.M.). The Strategic Research Grant (Ecological Interactions) enabled the purchase of the UPLC-DAD-MS/MS instrument.

Data Availability Statement

The data presented in this study are available on request from the corresponding author.

Acknowledgments

Asmo Aro-Heinilä, Niko Luntamo, and Suvi Vanhakylä are acknowledged for their assistance in the sample collection. The members of the Natural Chemistry Research Group are thanked for scientific discussions.

Conflicts of Interest

The authors declare no conflict of interest.

Sample Availability

Samples of the compounds are not available from the authors.

References

  1. Manninen, M.; Vesterinen, V.M.; Vainio, A.K.; Korhonen, H.; Karonen, M.; Salminen, J.P. Identification of Tree Species by Their Defense Compounds: A Study with Leaf Buds of White and Silver Birches. J. Chem. Educ. 2021, 98, 973–981. [Google Scholar] [CrossRef]
  2. Lahtinen, M.; Lempa, K.; Salminen, J.P.; Pihlaja, K. HPLC Analysis of Leaf Surface Flavonoids for the Preliminary Classification of Birch Species. Phytochem. Anal. 2006, 17, 197–203. [Google Scholar] [CrossRef] [PubMed]
  3. Melnyk, N.; Pawłowska, K.A.; Ziaja, M.; Wojnowski, W.; Koshovyi, O.; Granica, S.; Bazylko, A. Characterization of Herbal Teas Containing Lime Flowers – Tiliae flos by HPTLC Method with Chemometric Analysis. Food Chem. 2021, 346, 128929. [Google Scholar] [CrossRef]
  4. Ding, S.; Dudley, E.; Plummer, S.; Tang, J.; Newton, R.P.; Brenton, A.G. Fingerprint Profile of Ginkgo biloba Nutritional Supplements by LC/ESI-MS/MS. Phytochemistry 2008, 69, 1555–1564. [Google Scholar] [CrossRef] [PubMed]
  5. Bertrams, J.; Müller, M.B.; Kunz, N.; Kammerer, D.R.; Stintzing, F.C. Phenolic Compounds as Marker Compounds for Botanical Origin Determination of German Propolis Samples Based on TLC and TLC-MS. J. Appl. Bot. Food Qual. 2013, 86, 143–153. [Google Scholar] [CrossRef]
  6. Giffen, J.E.; Lesiak, A.D.; Dane, A.J.; Cody, R.B.; Musah, R.A. Rapid Species-Level Identification of Salvias by Chemometric Processing of Ambient Ionisation Mass Spectrometry-Derived Chemical Profiles. Phytochem. Anal. 2017, 28, 16–26. [Google Scholar] [CrossRef]
  7. Lesiak, A.D.; Cody, R.B.; Dane, A.J.; Musah, R.A. Plant Seed Species Identification from Chemical Fingerprints: A High-Throughput Application of Direct Analysis in Real Time Mass Spectrometry. Anal. Chem. 2015, 87, 8748–8757. [Google Scholar] [CrossRef]
  8. Kharyuk, P.; Nazarenko, D.; Oseledets, I.; Rodin, I.; Shpigun, O.; Tsitsilin, A.; Lavrentyev, M. Employing Fingerprinting of Medicinal Plants by Means of LC-MS and Machine Learning for Species Identification Task. Sci. Rep. 2018, 8. [Google Scholar] [CrossRef] [Green Version]
  9. Alcalde-Eon, C.; García-Estévez, I.; Rivas-Gonzalo, J.C.; Rodríguez De La Cruz, D.; Escribano-Bailón, M.T. Anthocyanins of the Anthers as Chemotaxonomic Markers in the Genus Populus L. Differentiation between Populus nigra, Populus alba and Populus tremula. Phytochemistry 2016, 128, 35–49. [Google Scholar] [CrossRef] [Green Version]
  10. Abreu, I.N.; Ahnlund, M.; Moritz, T.; Albrectsen, B.R. UHPLC-ESI/TOFMS Determination of Salicylate-like Phenolic Gycosides in Populus tremula Leaves. J. Chem. Ecol. 2011, 37, 857–870. [Google Scholar] [CrossRef]
  11. Julkunen-Tiitto, R. Phenolic Constituents of Salix: A Chemotaxonomic Survey of Further Finnish Species. Phytochemistry 1989, 28, 2115–2125. [Google Scholar] [CrossRef]
  12. Vidakovic, V.; Novakovic, M.; Popovic, Z.; Jankovic, M.; Matic, R.; Teševic, V.; Bojovic, S. Significance of Diarylheptanoids for Chemotaxonomical Distinguishing between Alnus glutinosa and Alnus incana. Holzforschung 2017, 72, 9–16. [Google Scholar] [CrossRef] [Green Version]
  13. Sołtys, A.; Galanty, A.; Podolak, I. Ethnopharmacologically Important but Underestimated Genus Sorbus: A Comprehensive Review. Phytochem. Rev. 2020, 19, 491–526. [Google Scholar] [CrossRef]
  14. Jensen, S.R.; Franzyk, H.; Wallander, E. Chemotaxonomy of the Oleaceae: Iridoids as Taxonomic Markers. Phytochemistry 2002, 60, 213–231. [Google Scholar] [CrossRef]
  15. Vihakas, M.; Pälijärvi, M.; Karonen, M.; Roininen, H.; Salminen, J.P. Rapid Estimation of the Oxidative Activities of Individual Phenolics in Crude Plant Extracts. Phytochemistry 2014, 103, 76–84. [Google Scholar] [CrossRef] [PubMed]
  16. Rank, N.E.; Köpf, A.; Julkunen-Tiitto, R.; Tahvanainen, J. Host Preference and Larval Performance of the Salicylate-Using Leaf Beetle Phratora vitellinae. Ecology 1998, 79, 618–631. [Google Scholar] [CrossRef] [Green Version]
  17. Fan, L.; Tong, Q.; Dong, W.; Yang, G.; Hou, X.; Xiong, W.; Shi, C.; Fang, J.; Wang, W. Tissue Distribution, Excretion, and Metabolic Profile of Dihydromyricetin, a Flavonoid from Vine Tea (Ampelopsis grossedentata) after Oral Administration in Rats. J. Agric. Food Chem. 2017, 65, 4597–4604. [Google Scholar] [CrossRef]
  18. Abad-García, B.; Garmón-Lobato, S.; Berrueta, L.A.; Gallo, B.; Vicente, F. A Fragmentation Study of Dihydroquercetin Using Triple Quadrupole Mass Spectrometry and Its Application for Identification of Dihydroflavonols in Citrus Juices. Rapid Commun. Mass Spectrom. 2009, 23, 2785–2792. [Google Scholar] [CrossRef]
  19. Fanzone, M.; González-Manzano, S.; Pérez-Alonso, J.; Escribano-Bailón, M.T.; Jofré, V.; Assof, M.; Santos-Buelga, C. Evaluation of Dihydroquercetin-3-O-Glucoside from Malbec Grapes as Copigment of Malvidin-3-O-Glucoside. Food Chem. 2015, 175, 166–173. [Google Scholar] [CrossRef] [Green Version]
  20. Ziaja, M.; Pawłowska, K.A.; Józefczyk, K.; Pruś, A.; Stefańska, J.; Granica, S. UHPLC-DAD-MS/MS Analysis of Extracts from Linden Flowers (Tiliae flos): Differences in the Chemical Composition between Five Tilia Species Growing in Europe. Ind. Crops Prod. 2020, 154, 112691. [Google Scholar] [CrossRef]
  21. Zhu, H.; Poojary, M.M.; Andersen, M.L.; Lund, M.N. Effect of PH on the Reaction between Naringenin and Methylglyoxal: A Kinetic Study. Food Chem. 2019, 298, 125086. [Google Scholar] [CrossRef] [PubMed]
  22. Agarwal, C.; Hofmann, T.; Vršanská, M.; Schlosserová, N.; Visi-Rajczi, E.; Voběrková, S.; Pásztory, Z. In Vitro Antioxidant and Antibacterial Activities with Polyphenolic Profiling of Wild Cherry, the European Larch and Sweet Chestnut Tree Bark. Eur. Food Res. Technol. 2021, 247, 2355–2370. [Google Scholar] [CrossRef]
  23. Mikulic-Petkovsek, M.; Stampar, F.; Veberic, R.; Sircelj, H. Wild Prunus Fruit Species as a Rich Source of Bioactive Compounds. J. Food Sci. 2016, 81, C1928–C1937. [Google Scholar] [CrossRef]
  24. Eyles, A.; Jones, W.; Riedl, K.; Cipollini, D.; Schwartz, S.; Chan, K.; Herms, D.A.; Bonello, P. Comparative Phloem Chemistry of Manchurian (Fraxinus mandshurica) and Two North American Ash Species (Fraxinus americana and Fraxinus pennsylvanica). J. Chem. Ecol. 2007, 33, 1430–1448. [Google Scholar] [CrossRef] [PubMed]
  25. Jensen, S.R.; Nielsen, B.J. A New Coumarin Fraxidin 8-O-β-d-Glucoside and 10-Hydroxyligstroside from Bark of Fraxinus exelsior. Phytochemistry 1976, 15, 221–223. [Google Scholar] [CrossRef]
  26. Okińczyc, P.; Widelski, J.; Szperlik, J.; Żuk, M.; Mroczek, T.; Skalicka-Woźniak, K.; Sakipova, Z.; Widelska, G.; Kuś, P.M. Impact of Plant Origin on Eurasian Propolis on Phenolic Profile and Classical Antioxidant Activity. Biomolecules 2021, 11, 68. [Google Scholar] [CrossRef] [PubMed]
  27. Zhang, X.; Lin, Z.; Fang, J.; Liu, M.; Niu, Y.; Chen, S.; Wang, H. An On-Line High-Performance Liquid Chromatography-Diode-Array Detector-Electrospray Ionization-Ion-Trap-Time-of-Flight-Mass Spectrometry-Total Antioxidant Capacity Detection System Applying Two Antioxidant Methods for Activity Evaluation of the Edible Flowers from Prunus mume. J. Chromatogr. A 2015, 1414, 88–102. [Google Scholar] [CrossRef]
  28. Kammerer, B.; Kahlich, R.; Biegert, C.; Gleiter, C.H.; Heide, L. HPLC-MS/MS Analysis of Willow Bark Extracts Contained in Pharmaceutical Preparations. Phytochem. Anal. 2005, 16, 470–478. [Google Scholar] [CrossRef]
  29. Meier, B.; Julkunen-Tiitto, R.; Tahvanainen, J.; Sticher, O. Comparative High-Performance Liquid and Gas-Liquid Chromatographic Determination of Phenolic Glucosides in Salicaceae Species. J. Chromatogr. A 1988, 442, 175–186. [Google Scholar] [CrossRef]
  30. Tóth, G.; Barabás, C.; Tóth, A.; Kéry, Á.; Béni, S.; Boldizsár, I.; Varga, E.; Noszál, B. Characterization of Antioxidant Phenolics in Syringa vulgaris L. Flowers and Fruits by HPLC-DAD-ESI-MS. Biomed. Chromatogr. 2016, 30, 923–932. [Google Scholar] [CrossRef]
  31. Wozniak, M.; Michalak, B.; Wyszomierska, J.; Dudek, M.K.; Kiss, A.K. Effects of Phytochemically Characterized Extracts from Syringa vulgaris and Isolated Secoiridoids on Mediators of Inflammation in a Human Neutrophil Model. Front. Pharmacol. 2018, 9, 349. [Google Scholar] [CrossRef]
  32. Geoffroy, T.R.; Stevanovic, T.; Fortin, Y.; Poubelle, P.E.; Meda, N.R. Metabolite Profiling of Two Maple-Derived Products Using Dereplication Based on High-Performance Liquid Chromatography-Diode Array Detector-Electrospray Ionization-Time-of-Flight-Mass Spectrometry: Sugar Maple Bark and Bud Hot-Water Extracts. J. Agric. Food Chem. 2019, 67, 8819–8838. [Google Scholar] [CrossRef] [PubMed]
  33. Haddock, E.A.; Gupta, R.K.; Al-Shafi, S.M.K.; Haslam, E.; Magnolato, D. The Metabolism of Gallic Acid and Hexahydroxydiphenic Acid in Plants. Part 1. Introduction. Naturally Occurring Galloyl Esters. J. Chem. Soc. Perkin Trans. 1 1982, 2515–2524. [Google Scholar] [CrossRef]
  34. De Tommasi, N.; De Simone, F.; Pizza, C.; Mahmood, N.; Moore, P.S.; Conti, C.; Orsi, N.; Stein, M.L. Constituents of Eriobotrya japonica. A Study of Their Antiviral Properties. J. Nat. Prod. 1992, 55, 1067–1073. [Google Scholar] [CrossRef] [PubMed]
  35. Yang, Z.G.; Li, H.R.; Wang, L.Y.; Li, Y.H.; Lu, S.G.; Wen, X.F.; Wang, J.; Daikonya, A.; Kitanaka, S. Triterpenoids from Hippophae rhamnoides L. and Their Nitric Oxide Production-Inhibitory and DPPH Radical-Scavenging Activities. Chem. Pharm. Bull. 2007, 55, 15–18. [Google Scholar] [CrossRef] [Green Version]
  36. Kazmi, M.H.; Ahmed, E.; Hameed, S.; Malik, A.; Fatima, I.; Ashraf, M. Cashmirols A and B, New Lipoxygenase Inhibiting Triterpenes from Sorbus cashmiriana. Chem. Biodivers. 2009, 6, 1471–1476. [Google Scholar] [CrossRef]
  37. Sun, Y.; Feng, F.; Nie, B.; Cao, J.; Zhang, F. High Throughput Identification of Pentacyclic Triterpenes in Hippophae rhamnoides Using Multiple Neutral Loss Markers Scanning Combined with Substructure Recognition (MNLSR). Talanta 2019, 205, 120011. [Google Scholar] [CrossRef]
  38. Xu, J.P.; Xu, R.S.; Li, X.Y. Glycosides of a Cycloartane Sapogenin from Curculigo orchioides. Phytochemistry 1992, 31, 233–236. [Google Scholar] [CrossRef]
  39. Ren, X.; He, T.; Chang, Y.; Zhao, Y.; Chen, X.; Bai, S.; Wang, L.; Shen, M.; She, G. The Genus Alnus, a Comprehensive Outline of Its Chemical Constituents and Biological Activities. Molecules 2017, 22, 1383. [Google Scholar] [CrossRef] [Green Version]
  40. Aoki, T.; Ohta, S.; Aratani, S.; Hirata, T.; Suga, T. The Structures of Four Novel C31-Secodammarane-Type Triterpenoid Saponins from the Female Flowers of Alnus serrulatoides. J. Chem. Soc. Perkin Trans. 1 1982, 1399–1403. [Google Scholar] [CrossRef]
  41. Aoki, T.; Ohta, S.; Suga, T. Triterpenoids, Diarylheptanoids and Their Glycosides in the Flowers of Alnus Species. Phytochemistry 1990, 29, 3611–3614. [Google Scholar] [CrossRef]
  42. Chambers, M.C.; MacLean, B.; Burke, R.; Amodei, D.; Ruderman, D.L.; Neumann, S.; Gatto, L.; Fischer, B.; Pratt, B.; Egertson, J.; et al. A Cross-Platform Toolkit for Mass Spectrometry and Proteomics. Nat. Biotechnol. 2012, 30, 918–920. [Google Scholar] [CrossRef] [PubMed]
  43. Pluskal, T.; Castillo, S.; Villar-Briones, A.; Orešič, M. MZmine 2: Modular Framework for Processing, Visualizing, and Analyzing Mass Spectrometry-Based Molecular Profile Data. BMC Bioinformatics 2010, 11, 395. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  44. Myers, O.D.; Sumner, S.J.; Li, S.; Barnes, S.; Du, X. One Step Forward for Reducing False Positive and False Negative Compound Identifications from Mass Spectrometry Metabolomics Data: New Algorithms for Constructing Extracted Ion Chromatograms and Detecting Chromatographic Peaks. Anal. Chem. 2017, 89, 8696–8703. [Google Scholar] [CrossRef] [PubMed]
Molecules 27 06810 g001 550
Figure 1. Flow chart of the method development of two different methods.
Figure 1. Flow chart of the method development of two different methods.
Molecules 27 06810 g001
Molecules 27 06810 g002 550
Figure 2. The number of features detected with MZmine 2 in the UHPLC-Orbitrap-MS data of the studied species. On the basis of the peak areas, the features were categorised into six categories that are denoted with different colours.
Figure 2. The number of features detected with MZmine 2 in the UHPLC-Orbitrap-MS data of the studied species. On the basis of the peak areas, the features were categorised into six categories that are denoted with different colours.
Molecules 27 06810 g002
Molecules 27 06810 g003 550
Figure 3. Within-species repeatability of the fingerprints in the different plant individuals of (A) Sorbus hybrida, (B) Alnus glutinosa, and (C) Salix phylicifolia recorded with the species-specific method II.
Figure 3. Within-species repeatability of the fingerprints in the different plant individuals of (A) Sorbus hybrida, (B) Alnus glutinosa, and (C) Salix phylicifolia recorded with the species-specific method II.
Molecules 27 06810 g003
Molecules 27 06810 g004 550
Figure 4. Within-species repeatability of the fingerprints in the different plant individuals of (A) Sorbus hybrida, (B) Alnus glutinosa, and (C) Salix phylicifolia recorded with the species-specific method I. The results were obtained for the same samples as in Figure 3.
Figure 4. Within-species repeatability of the fingerprints in the different plant individuals of (A) Sorbus hybrida, (B) Alnus glutinosa, and (C) Salix phylicifolia recorded with the species-specific method I. The results were obtained for the same samples as in Figure 3.
Molecules 27 06810 g004
Molecules 27 06810 g005 550
Figure 5. Fingerprints of all studied species with the species-specific method I for Syringa vulgaris. (A) The y-axes of all samples were scaled according to the most intensive peak of all samples, which was produced by S. vulgaris. (B) The y-axes were scaled to the most intensive peak of each sample. Plant species are as follows: (a) Acer platanoides, (b) Alnus glutinosa, (c) Alnus incana, (d) Fraxinus excelsior, (e) Populus tremula, (f) Prunus padus, (g) Quercus robur, (h) Salix phylicifolia, (i) Sorbus aucuparia, (j) Sorbus hybrida, (k) Syringa vulgaris, and (l) Tilia cordata. Tilia × europaea produced a similar fingerprint to that of Tilia cordata.
Figure 5. Fingerprints of all studied species with the species-specific method I for Syringa vulgaris. (A) The y-axes of all samples were scaled according to the most intensive peak of all samples, which was produced by S. vulgaris. (B) The y-axes were scaled to the most intensive peak of each sample. Plant species are as follows: (a) Acer platanoides, (b) Alnus glutinosa, (c) Alnus incana, (d) Fraxinus excelsior, (e) Populus tremula, (f) Prunus padus, (g) Quercus robur, (h) Salix phylicifolia, (i) Sorbus aucuparia, (j) Sorbus hybrida, (k) Syringa vulgaris, and (l) Tilia cordata. Tilia × europaea produced a similar fingerprint to that of Tilia cordata.
Molecules 27 06810 g005
Molecules 27 06810 g006 550
Figure 6. Fingerprints of all studied species with the species-specific method II for Syringa vulgaris. (A) The y-axes all samples were scaled according to the most intensive peak of all samples, which was produced by Populus tremula (e). (B) The y-axes were scaled to the most intensive peak of each sample. Plant species are as follows: (a) Acer platanoides, (b) Alnus glutinosa, (c) Alnus incana, (d) Fraxinus excelsior, (e) Populus tremula, (f) Prunus padus, (g) Quercus robur, (h) Salix phylicifolia, (i) Sorbus aucuparia, (j) Sorbus hybrida, (k) Syringa vulgaris, and (l) Tilia cordata. Tilia × europaea produced a similar fingerprint to that of Tilia cordata.
Figure 6. Fingerprints of all studied species with the species-specific method II for Syringa vulgaris. (A) The y-axes all samples were scaled according to the most intensive peak of all samples, which was produced by Populus tremula (e). (B) The y-axes were scaled to the most intensive peak of each sample. Plant species are as follows: (a) Acer platanoides, (b) Alnus glutinosa, (c) Alnus incana, (d) Fraxinus excelsior, (e) Populus tremula, (f) Prunus padus, (g) Quercus robur, (h) Salix phylicifolia, (i) Sorbus aucuparia, (j) Sorbus hybrida, (k) Syringa vulgaris, and (l) Tilia cordata. Tilia × europaea produced a similar fingerprint to that of Tilia cordata.
Molecules 27 06810 g006
Molecules 27 06810 g007 550
Figure 7. The LC–MS fingerprints of the studied species with the species-specific methods. (A) Acer platanoides, (B) Alnus glutinosa, (C) Alnus incana, (D) Fraxinus excelsior, (E) Populus tremula, (F) Prunus padus, (G) Quercus robur, (H) Salix phylicifolia, (I) Sorbus aucuparia, (J) Sorbus hybrida, (K) Syringa vulgaris, and (L) Tilia cordata. Tilia × europaea produced a similar fingerprint to that of Tilia cordata. The LC–MS fingerprints consisted of selected ion recording traces of 1–6 markers (Table 2).
Figure 7. The LC–MS fingerprints of the studied species with the species-specific methods. (A) Acer platanoides, (B) Alnus glutinosa, (C) Alnus incana, (D) Fraxinus excelsior, (E) Populus tremula, (F) Prunus padus, (G) Quercus robur, (H) Salix phylicifolia, (I) Sorbus aucuparia, (J) Sorbus hybrida, (K) Syringa vulgaris, and (L) Tilia cordata. Tilia × europaea produced a similar fingerprint to that of Tilia cordata. The LC–MS fingerprints consisted of selected ion recording traces of 1–6 markers (Table 2).
Molecules 27 06810 g007
Table
Table 1. Comparison of two fingerprinting methods. Repeatability is denoted as the number of replicates that produced similar fingerprints for all replicates. Intensity-based specificity was ranked as good if the species in question produced the most intensive signal, OK if one other species produced as intensive a peak as the species in question, and poor if the fingerprint of some other species was the most intensive. Profile-based specificity was ranked as good if the fingerprint profile of the species in question was unique and poor if the fingerprint profile of the species in question showed the same peaks as other species.
Table 1. Comparison of two fingerprinting methods. Repeatability is denoted as the number of replicates that produced similar fingerprints for all replicates. Intensity-based specificity was ranked as good if the species in question produced the most intensive signal, OK if one other species produced as intensive a peak as the species in question, and poor if the fingerprint of some other species was the most intensive. Profile-based specificity was ranked as good if the fingerprint profile of the species in question was unique and poor if the fingerprint profile of the species in question showed the same peaks as other species.
Method I (Species-Specific Ions)Method II (Main Ions)
SpeciesRepeatabilitySpecificityRepeatabilitySpecificityFinal Method
Intensity-BasedProfile-BasedIntensity-BasedProfile-Based
Acer platanoides8/8OKgood8/8poorpoorI
Alnus glutinosa6/8 *goodgood8/8goodgoodII
Alnus incana8/8goodgood8/8poorpoorI
Fraxinus excelsior9/9goodgood9/9poorpoorI
Populus tremula10/10goodgood10/10goodgoodI
Prunus padus5/5goodgood5/5poorpoorI
Quercus robur9/9OKgood9/9poorpoorI
Salix phylicifolia3/4 *goodgood4/4poorpoorI
Sorbus aucuparia †9/9poorgood9/9goodgoodII
Sorbus hybrida8/10goodgood10/10goodgoodII
Syringa vulgaris5/5goodgood5/5poorpoorI
Tilia cordata and Tilia × europaea15/15poorgood15/15goodpoorI
* = There was some deviation in the fingerprint profiles, but the identification of species was still possible; † = Method I produced a similar fingerprint for both S. aucuparia and S. hybrida.
Table
Table 2. Chromatographic, UV, and mass spectral data of the species-specific markers obtained from the UHPLC–DAD–ESI–QOrbitrap–MS.
Table 2. Chromatographic, UV, and mass spectral data of the species-specific markers obtained from the UHPLC–DAD–ESI–QOrbitrap–MS.
No.SpeciesUV Max (nm)Retention Time (min)Molecular Formula[M−H]−m/z Values of Main Fragment IonsExact Mass, CalculatedExact Mass, MeasuredError (ppm)
1Acer platanoides212, 257, 300 (sh)0.40C15H20O10359.09864123, 182, 166, 197360.10565360.105940.29
2Fraxinus excelsiornd.0.52*792.08522191, 248, 263, 298, 354, 422, 438, 453, 468, 630
3Salix phylicifolia205, 2900.66C15H12O8319.04573109, 125, 151, 193320.05322320.05303−0.19
4Tilia sp.230, 2910.67C21H22O12465.10378107, 151, 285, 303466.11113466.11108−0.05
5Fraxinus excelsior205, 225 (sh), 3290.67C18H22O12429.10425163, 191, 206, 221430.11113430.111550.42
6Acer platanoides2580.70C22H24O14511.1103125, 169, 182, 197, 313, 345512.11661512.117600.99
7Acer platanoides2580.79C22H24O14511.11011125, 169, 182, 197, 313, 345512.11661512.117410.8
8Acer platanoides2580.88C22H24O14511.11000125, 169, 182, 197, 313, 345512.11661512.117300.69
9Fraxinus excelsior198, 230, 290 (sh), 3280.94C23H26O11477.14055133, 161, 179, 315478.14752478.147850.335
10Syringa vulgaris195, 237, 282 (sh), 333 (sh)0.96C24H30O13525.16106121, 139, 165, 183, 209, 227, 249, 363, 389526.16865526.16836−0.285
11Acer platanoides225, 2780.98C29H28O18663.12130125, 169, 313, 465, 511664.12757664.128601.03
12Fraxinus excelsior203, 231, 295 (sh), 3301.00C23H26O11477.14055133, 161, 179, 315478.14752478.147850.335
13Sorbus sp.nd.1.04C16H18O9353.08783135, 161, 179, 191354.09509354.095130.045
14Syringa vulgaris193, 237, 278 (sh), 333 (sh)1.10C24H30O12509.16640121, 139, 165, 233, 277, 347510.17373510.17370−0.03
15Acer platanoides221, 2781.15C55H40O341243.13081 a125, 169, 393 b, 469 b, 617, 7691244.140111244.13990−0.21
16Syringa vulgarisnd.1.31C27H38O15601.21324153, 197602.22108602.22054−0.535
17Syringa vulgaris243, 278 (sh)1.44C32H38O14645.21930121, 139, 149, 165, 209, 275, 329, 413646.22616646.226600.44
18Prunus padus3151.56C31H38O18697.19783117, 145, 163698.20582698.20513−0.69
19Populus tremula193, 221, 243, 297 (sh), 3281.60C23H22O10457.11453133, 135, 161, 179, 235, 295, 397458.12130458.121830.53
20Salix phylicifolia240, 2701.60C27H28O11527.1560593, 109, 121, 137, 155, 405528.16317528.163350.185
21Prunus padusnd.1.75C28H26O11537.14196119, 151, 177, 271538.14752538.149261.745
22Populus tremula233, 290 (sh), 3141.89C23H22O8425.1250093, 117, 119, 145, 163, 219, 365426.13147426.132300.83
23Populus tremula236, 296 (sh), 3191.93C24H24O9455.13543117, 119, 134, 145, 160, 163, 175, 193, 395456.14204456.142730.695
24Populus tremula240, 293 (sh), 3251.97C25H26O10485.14578134, 149, 160, 175, 193, 207, 425486.15260486.153080.48
25Alnus glutinosand.2.35C35H58O10637.39620389, 477, 521, 605638.40300638.403500.5
26Alnus glutinosand.2.41C35H58O10637.39641389, 477, 521, 605638.40300638.403710.71
27Sorbus sp.3092.49C39H54O7633.38042117, 119, 145, 163, 571, 589, 615634.38696634.387720.765
28Tilia sp.nd.2.64C13H28O8311.16894119, 183 312.17842312.17624−2.18
29Sorbus sp.3222.66C39H54O7633.38031133, 135, 161, 179634.38696634.387610.655
30Sorbus sp.nd.2.67C30H46O4469.33235423470.33961470.339650.04
31Alnus glutinosand.2.72C35H58O8605.40623283, 345, 455, 473606.41317606.413530.36
32Alnus glutinosand.2.79C36H60O8619.42097343, 345, 371, 389, 487620.42882620.42827−0.55
33Alnus glutinosand.2.82C35H58O8605.40664283, 345, 455, 473606.41317606.413940.77
34Alnus glutinosand.2.88C36H60O8619.42214343, 345, 371, 389, 487620.42882620.429440.62
35Sorbus sp.nd.2.93C32H48O5511.34265451512.35018512.34995−0.225
36Sorbus sp.nd.3.00C30H46O3453.33786nd.454.34470454.345160.465
37Tilia sp.nd.3.04C18H32O4311.22300111, 155, 171312.23006312.230300.24
38Sorbus sp.nd.3.15C32H50O4497.36354437498.37091498.37084−0.07
39Quercus roburnd.3.55C27H44O4431.31652133, 161, 179432.32396432.32382−0.14
40Alnus incanand.3.81C53H80O11891.56326117, 145, 163, 271, 331, 399, 417, 491, 745892.57007892.570560.495
41Alnus incanand.3.85C57H83O8895.60993117, 119, 145, 163, 749896.61662896.617230.61
42Alnus incanand.3.89C46H76O6723.56173117, 145, 223, 267, 285, 297, 455, 705724.56419724.569034.84
a = Utilised in the fingerprinting method as a [2M–H]2– ion at m/z 621.06265. b = Doubly charged ion. * = The exact molecular formula could not be predicted, but the compound seems to contain an odd number of nitrogen atoms. sh = shoulder. nd. = not detected.
Table
Table 3. Selected ion recording parameters for the LC–MS fingerprinting methods.
Table 3. Selected ion recording parameters for the LC–MS fingerprinting methods.
Method IMethod II
Speciesm/z ValueCone Voltage (V)RT Range (min)m/z ValueCone Voltage (V)RT Range (min)
Acer platanoides359.0300.25–1.50359.0300.25–4.00
511.050469.030
621.030545.050
663.050712.650
Alnus glutinosa605.4502.00–3.40605.4502.00–3.40
637.450619.050
661.450637.450
Alnus incana723.5703.40–4.00439.4302.40–4.00
891.570453.430
895.670471.430
Fraxinus excelsior 0.30–1.30369.0300.20–3.50
429.010455.030
477.050471.030
792.050477.030
623.050
Populus tremula425.0301.00–2.25441.0301.20–2.25
455.030457.030
457.030471.030
485.030
Prunus padus 1.00–2.00455.0301.00–3.50
537.030471.030
697.050497.030
697.050
Quercus robur431.0500.25–3.80217.0300.20–4.50
289.030
431.050
439.030
Salix phylicifolia 0.40–1.90217.0300.30–4.50
319.030301.030
527.030503.430
712.530
Sorbus hybrida653.0500.40–3.20
1045.650
Sorbusaucuparia and
Sorbus hybrida		 0.60–2.80353.0300.80–3.50
453.430
425.030469.430
569.050497.430
511.530
633.530
Syringa vulgaris509.0300.60–1.70455.4300.80–3.20
525.030471.030
601.030539.030
645.030
Tilia sp. 0.50–3.25217.0300.25–4.50
311.030289.030
465.030475.430
503.630
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Manninen, M.; Karonen, M.; Salminen, J.-P. Chemotaxonomic Markers for the Leaf Buds of Common Finnish Trees and Shrubs: A Rapid UHPLC MS Fingerprinting Tool for Species Identification. Molecules 2022, 27, 6810. https://doi.org/10.3390/molecules27206810

AMA Style

Manninen M, Karonen M, Salminen J-P. Chemotaxonomic Markers for the Leaf Buds of Common Finnish Trees and Shrubs: A Rapid UHPLC MS Fingerprinting Tool for Species Identification. Molecules. 2022; 27(20):6810. https://doi.org/10.3390/molecules27206810

Chicago/Turabian Style

Manninen, Marianna, Maarit Karonen, and Juha-Pekka Salminen. 2022. "Chemotaxonomic Markers for the Leaf Buds of Common Finnish Trees and Shrubs: A Rapid UHPLC MS Fingerprinting Tool for Species Identification" Molecules 27, no. 20: 6810. https://doi.org/10.3390/molecules27206810

Article Metrics

Back to TopTop