Towards a better understanding of Prunus (Rosaceae): molecular and morphological notes on Prunus subgenus Cerasus in Turkey

Abstract

In this study, we explore the relationships among taxa of Prunus subgenus Cerasus in Turkey using a combination of macromorphology, micromorphology and molecular techniques. We recorded qualitative and quantitative characteristics of flowers, leaves, seeds and endocarp for 96 populations representing 14 taxa of Prunus subgenus Cerasus. ITS, matK, ycf1, trnL-trnF DNA sequence data were used to reconstruct phylogenetic trees for this group. Haplotype analyses results found 11 haplotypes among the samples used in this study. We also note that the infraspecific taxa of Cerasus angustifolia (=P. albicaulis), P. mahaleb, P. microcarpa and P. prostrata, as described by previous authors, do not exhibit significant differences even among the features of foliar morphology that are so often used to distinguish them from one another. As a general rule, our analyses indicate that the highly variable features of leaf micromorphology and endocarp and seed micromorphology are not taxonomically significant. In contrast, although floral morphology is often overlooked in systematic studies of Prunus subgenus Cerasus, we found several taxonomically useful flower characters, including sepal indumentum, hypanthium shape and indumentum and petal shape. This work clarifies the infraspecific relationships of members of Prunus subgenus Cerasus in Turkey and solves some taxonomic problems.

haplotype analysis, ITS, macromorphology, micromorphology, phylogenetic analysis, plastid DNA, Prunus alidagensis

INTRODUCTION

Rosaceae have a worldwide distribution, with members occupying many ecological niches and being forbs, shrubs or trees. Many species are economically important, especially apple, apricot, strawberry, peach and nectarine, and some have a long history of medicinal and cosmetic use. Phylogenetic relationships within the family and with other angiosperm lineages continue to be a topic of investigation (e.g. Hummer & Janick, 2009).

Prunus L. belongs to Rosaceae subfamily Amygdaloideae and comprises c. 220 species. One of the most widely accepted classifications defines five subgenera, sometimes treated as the genera Prunus s.s. (synonym: Prunophora Neck.), Amygdalus L., Cerasus Mill., Padus Mill. and Laurocerasus Duhamel, and further divides these subgenera into sections (Rehder, 1940). Ingram (1948) split Prunus into three subgenera [Prunus subgenera Cerasus (Mill.) A.Gray, Padus (Mill.) A.Gray and Lithocerasus Ingram (=Prunus)]. Other classifications, such as Browicz (1972), elevated subgenera such as Cerasus to the genus level. Plastid and nuclear regions have previously been used to investigate the phylogeny of Prunus (Bortiri et al., 2001; Ohta et al., 2007; Gilani et al., 2011). Prunus has been found to be monophyletic based on ITS and trnL-trnF sequence data, but it can also be separated into two groups, which suggests a need for further investigation targeting Prunus section Microcerasus (Spach) C.K.Schneid. (Bortiri et al., 2001). Another study tackling the phylogeny of Prunus using ITS and ndhF regions separated section Microcerasus from subgenus Cerasus (Wen et al., 2008). Shi et al. (2013) conducted the most comprehensive molecular phylogenetic study to date using 12 plastid and three nuclear regions to explore the infrageneric relationship in Prunus, which resulted in three subgenera and seven sections. A more detailed outline of the various treatments of Prunus is given in Table 1.

Table 1.

Systematic timetable of Prunus and its subgenera

De Tournefort, 1700	Linnaeus, 1753	Linnaeus, 1754	Bentham & Hooker, 1865	Rehder, 1940	Ingram, 1948	Hutchinson, 1964	Potter et al., 2007	Shi et al., 2013
6 genera	3 genera	4 genera	1 genus, 7 sections	1 genus, 4 subgenera	1 genus, 3 subgenera	3 genera	1 genus	1 genus, 3 subgenera
Amygdalus (L.)Focke Armeniaca (Scop.) Turcz. Cerasus (Mill.) Focke Laurocerasus (Duhamel) Rehder Persica (Mill.) Nakai Prunus s. s.	Amygdalus L. Padus Mill. Prunus L. s. s.	Armeniaca Scop. Cerasus Mill. Padus Mill. Prunus L. s. s.	genus Prunus L. sections Amygdalopsis Carr. Amygdalus (L.) Benth. & Hook.f. Armeniaca Juss. Cerasoides Sieb. & Zucc. Cerasus Juss. Laurocerasus Tourn. Prunus	genus Prunus L. subgenera Amygdalus (L.) Focke Padus (Moench) Koehne Cerasus Pers. Laurocerasus Koehne	genus Prunus L. subgenera Cerasus Pers. Padus Koehne Lithocerasus Ingram	Laurocerasus Duhamel Padus Mill. Prunus L. s. s.	Prunus L.	genus Prunus L. subgenera Prunus Padus Cerasus

Macromorphology of Rosaceae has been studied extensively (e.g. Robertson, Phipps & Rohrer, 1992; Evans & Dickinson, 1999; Khadivi-Khub, Zamani & Fatahi, 2012; Song, Roh & Hong, 2020). Scanning electron microscope techniques have also been used and have proved to be taxonomically valuable for Rosaceae (Song et al., 2020; Khadivi et al., 2022). Leaf features, such as trichomes and epicuticular waxes, may be useful in assigning genera into tribes in this family (Song & Hong, 2016), but standard methodologies may be insufficient for some systematic complexes. Prunus shows extensive morphological and/or anatomical similarities to other infrageneric groups (Bortiri, Heuvel & Potter, 2006). Similarly, molecular markers may be incongruent for certain taxa of Rosaceae because of hybridization events or rapid radiation, necessitating more supporting data (Campbell et al., 2007).

The primary objective of this study is to investigate and explore relationships among the taxa of Prunus subgenus Cerasus in Turkey, where these taxa are weakly delimited into subspecies and varieties. Furthermore, no comprehensive molecular or systematic study has reassigned all of these taxa to Prunus, so we refer to them as Cerasus Duhamel sensuBrowicz (1972). We apply a combined genetic, micromorphological and macromorphological approach to tease out information on relationships for a historically controversial, but economically significant group of plants.

MATERIAL AND METHODS

Sampling

Specimens of 14 taxa from genus Prunus subgenus Cerasus were collected from 96 populations in Turkey in 2017–2019. The material studied, Cerasus angustifolia (Spach) Browicz (=Prunus albicaulis Koehne ex Bornm.) var. angustifolia, C. angustifolia var. sintenisii (C.K.Schneid.) Browicz (=Prunus albicaulis), Prunus avium (L.) L., P. brachypetala Walp., P. cerasus L., P. hippophaeoides (Bornm.) Bornm., C. incana (Pall.) Spach [=Prunus incana (Pall.) Batsch], C. incana var. velutina Browicz (=Prunus incana), P. mahaleb L. var. mahaleb, P. mahaleb var. alpina Browicz, P. microcarpa C.A.Mey. subsp. microcarpa, P. microcarpa subsp. tortuosa (Boiss. & Hausskn.) Browicz, P. prostrata Labill. var. prostrata and P. prostrata Labill. var. glabrifolia (Moris) Browicz), includes the infraspecific taxa described by Browicz (1972). The specimens were identified using the treatment in Flora of Turkey and the East Aegean Islands (Browicz, 1972). Material is deposited in ISTF (Istanbul University Faculty of Science Herbarium) and collection data are available in Supporting Information, Appendix S1.

Morphological Studies

Morphological studies examined ten sun leaves and dissected eight to ten flowers from three to five individuals per population, except in rare cases when fewer individuals were found. We took numeric measurements with electronic callipers for larger parts, including leaves, and used an Olympus SZX7 stereomicroscope using Argenit Kameram v.3.1.0.0 software for observation of the flowers, fruit and seeds.

Quantitative morphological data collected were leaf dimensions, leaf width/height ratio, petiole length, pedicel length, hypanthium length, sepal length, petal dimensions, petal width/height ratio, style length, filament length, drupe dimensions and drupe width/height ratio. Qualitative characters are provided in Table 2. Characters were selected and modified from Rehder (1940), Ingram (1948) and Browicz (1972), and described according to Stearn (1996) and Harris & Harris (2009).

Table 2.

Qualitative morphological and micromorphological characters investigated, with all character states observed in the study. Character states that showed variation in the material studied were subject to one-way ANOVA and those found significant were used as categorical data in further analyses. AB: leaf abaxial surface, AD: leaf adaxial surface, AW: anticlinal wall, PW: periclinal wall

Character	Character states observed
Leaf shape	elliptic–falcate–lanceolate–oblanceolate–oblique–oblong–obovate–orbiculate–ovate–rhomboid–spatulate–thin obovate–thin ovate
Leaf apex	acuminate–acute–apiculate–obtuse–retuse–rotund–subacute–tridental
Leaf base	acuminate–acute–angustate–asymmetric–attenuate–cuneate–oblique–obtuse–orbiculate–retuse–short angustate–thin acute- truncate
Leaf margin	biserrate–crenulate–serrate–serrulate–triserrate
Extrafloral nectary location	lamina–leaf base–marginal teeth–petiole
Inflorescence	aggregate–racemose–solitary/binary
Pedicel trichome state	glabrous–lanate–pubescent–scabrous–villous
Hypanthium shape	campanulate–cylindrical–urceolate
Hypanthium inner trichome	glabrous–lanate–pubescent–scabrous–tomentose–villous
Hypanthium outer trichome	glabrescent–glabrous–pubescent–scabrous–villous
Sepal margin	dentate–entire–serrate
Sepal inner trichome	fimbriate–glabrous–lanate–pubescent–tomentose–villous
Sepal outer trichome	fimbriate–glabrous–lanate–pubescent–villous
Petal shape	broadly obovate–elliptic to ovate–narrowly obovate–obovate–ovate
Petal colour	pale pink–pink–white
Style trichome	glabrous–lanate–pubescent–sericeous–strigose–villous
Ovary trichome	glabrous–lanate–sericeous–villous
Drupe trichome	glabrous–hirsute
Drupe colour	black–claret red–orange–red–yellow
AB surface sculpture	irregularly wrinkled–rugose
AB AW shape	undulate–wavy
AB AW thickness	thick–thin
AB AW level	grooved–raised
AB PW level	convex
AB PW texture	rugose–rugose striate–scurfy–striate–striate scurfy
AB cell shape	isodiametric–isodiametric rectangular
AB trichome	absent–lanate–pilose
AB trichome ornamentation	absent–smooth–striate
AB trichome type	absent–simple
AB stomata	anomocytic
AD surface sculpture	semireticulate–reticulate–irregularly wrinkled–rugose
AD AW shape	straight–wavy–wavy undulate
AD AW thickness	thick–thin
AD AW level	grooved–raised
AD PW level	concave–convex
AD PW texture	irregularly striate–lineate–rugose–striate to rugose–smooth to rugose–verruculate
AD cell shape	isodiametric–isodiametric rectangular
AD trichome	absent–lanate–pilose–pubescent–villous–villous to lanate
AD trichome ornamentation	absent–lineate–verruculate
AD trichome type	absent–basic
AD stomata	absent–cyclocytic
Seed surface sculpture	colliculate
Seed AW shape	wavy
Seed AW thickness	thick–thin
Seed AW level	raised
Seed PW level	concave–convex
Seed PW Texture	irregularly striate–irregularly wrinkled–rugose–rugose to verruculate–striate to verruculate–verruculate
Seed cell shape	isodiametric
Endocarp surface sculpture	colliculate–lacunose–ruminate to lacunose
Endocarp AW shape	straight–wavy
Endocarp AW thickness	thick–thin
Endocarp AW level	grooved–raised
Endocarp PW level	concave–convex–slightly concave
Endocarp PW texture	reticulate–rugose-ruminate–ruminate- scrobiculate–scrobiculate to rugose–slightly scrobiculate–strongly scrobiculate
Endocarp cell shape	elongated–isodiametric–isodiametric hexagonal–isodiametric pentagonal–isodiametric rounded

Micromorphological Studies

Mature seed, endocarp and leaf samples were fixed to sample stubs using carbon tape and labelled with specimen numbers. Leaf samples were fixed face up and face down to examine both adaxial and abaxial surfaces.

All samples were loaded into an EmiTech SC7620 Sputter Coater to be coated with gold/palladium at 800 Pa pressure and a plasma current of 10 mA for 60 seconds for seed and endocarp and 45 seconds for leaves. Seeds and endocarp were examined and photographed at 100×, 300×, 500× and 1000× magnification and leaves at 100×, 300×, 600× and 1000× magnification.

Leaves were bleached to allow for a clearer view of surface features and placed under a light microscope to observe the anticlinal walls (Lersten & Horner, 2000). Characters and character states observed are provided in Table 2. Surface features were described using the terminology of Barthlott (1981), Stearn (1996) and Harris & Harris (2009).

DNA Extraction and Purification

Leaves were dried using silica gel and stored at −80 °C before DNA extraction. We used a GeneSpin DNA extraction kit (Eurofins) according to the manufacturer’s instructions to extract DNA from 10 mg of leaf material homogenized with liquid nitrogen in a mortar and pestle. Extraction was performed twice, producing two elutions of 25 μL (total 50 μL). DNA content and purity were measured using a Thermo Scientific NanoDrop 2000/2000 °C at 260 and 280 nm wavelengths. Specimens deemed inadequate were subsequently processed using a modified cetyltrimethylammonium bromide extraction method based on that of Porebski, Bailey & Baum (1997) or a HiPurA 96 SuperPlant DNA Purification Kit (HiMedia) designed for plants with high polyphenol content, according to the manufacturer’s instructions. Samples were stored at −20 °C before polymerase chain reaction (PCR) processing.

PCR Amplification and Sequencing

We targeted the ITS, matK, ycf1 and trnL-trnF regions for sequencing. Relevant literature was consulted for the selection of primers for each: ITS (Potter et al., 2007), matK (Yu, Xue & Zhou, 2011), ycf1 (Dong et al., 2015) and trnL-trnF (Taberlet et al., 1991). The amplification procedures were carried out in a final volume of 25 μL containing 1 µL of DNA, 12.5 μL of PCR mix, 11 μL of distilled water and 0.5 μL of each primer. The PCR conditions for the amplification of ITS, matK, ycf1 and trnL-trnF were as follows: denaturation at 94 °C for 3 min, 40 cycles of denaturation at 94 °C for 30 s, annealing at 55–60 °C for 30 s, extension at 72 °C for 1 min and a final extension at 72 °C for 10 min.

We used a QIAquick Purification Kit (Qiagen) according to the manufacturer’s instructions to prepare samples for PCR and checked the quality of the final products on a 1% agarose gel. Sanger sequencing was carried out using an ABI 3730XL Sanger Sequencer (Applied Biosystems, Foster City, CA, USA) and BigDye Terminator v.3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s instructions.

Data Analysis

Statistical analysis of morphological data

All statistical analyses were performed using averages ± standard deviation. Quantitative morphological data were analysed using one-way analysis of variance (ANOVA) to determine significant characters and Tukey’s test to determine significant similarities between closely related taxa on GraphPad Prism v.8.0 software. Principal coordinate analysis (PCoA) was performed using all significant morphological and micromorphological data using PAST v.4.08 software (Hammer, Harper & Ryan, 2001).

Sequence editing and alignment

Chromatogram control was performed manually on forward and reverse sequences using Chromas v.2.6.5 and Geneious Prime programs. The DNA obtained sequences were manually aligned and edited using PhyDE (Phylogenetic Data Editor).

Haplotype analysis

The haplotype map based on the aligned sequence data was constructed with a TCS network approach (Clement et al., 2000) using PopART v.1.7 (Leigh & Bryant, 2015). The haplotype map was interpreted graphically using InkScape Portable v.0.92.

Phylogenetic analysis

Prunus tomentosa Buch.-Ham. ex D.Don (NCBI: MF624726.1) was selected as the outgroup for plastid DNA analysis. Bayesian Markov chain Monte Carlo (MCMC) phylogenetic analysis (Yang & Rannala, 1997) was performed using the plastid sequence data with MrBayes v.3.2.6. The MCMC algorithm was applied over 1 000 000 generations and a tree was constructed at every 1000 samples. The final phylogenetic tree was visualized and edited using FigTree v.1.4.3. and MEGA v.7. Analysis of the ITS region was performed with the same procedure using P. tomentosa Maxim. (NCBI: AF318746) as the outgroup.

RESULTS

Micromorphology

Micromorphological characters of seed, endocarp and leaf surfaces were examined for Prunus subgenus Cerasus. Seed surface examinations revealed the same surface sculpture (colliculate), anticlinal wall shape (wavy), level (raised), thickness (thick) and cell shape (isodiametric) features among all taxa studied. Endocarp cells have raised, wavy anticlinal walls in all taxa with the exception of P. microcarpa, which has grooved, wavy cell walls, and P. incana, which has raised, straight walls.

The adaxial surface micromorphology differs between populations of the same taxon; however, they have extremely stable abaxial surface morphology (Fig. 1). Trichomes on leaf adaxial surfaces are simple in all taxa. Adaxial stomata were found only in P. microcarpa. The abaxial leaf anticlinal walls are grooved in all taxa except P. cerasus, which has raised walls. Periclinal walls are convex in all taxa. Surface features could not be characterized in taxa with dense trichome cover.

Figure 1.

Abaxial (A) and adaxial (B–D) leaf surface characteristics observed in the studied taxa. Rugose in A, Prunus mahaleb (CP75); B, reticulate in P. avium (CP53); C, semireticulate in P. microcarpa subsp. tortuosa (CP143) and D, irregularly wrinkled in P. prostrata (CP148).

Other characteristics of leaf, seed and endocarp surfaces were found to be more or less identical between infraspecific taxa, with the exception of C. incana var. velutina and C. incana var. incana. Additionally, some P. microcarpa and C. angustifolia specimens (CP140 and CP136, respectively) with the same haplotype showed similar surface characteristics (Fig. 2). The characters are given for each taxon in Supporting Information, Appendix S2.

Figure 2.

Although the taxa are roughly consistent concerning finer structure of surface characters on an infraspecific level, some specimens of different taxa with the same haplotype show similar leaf adaxial surface characteristics as can be seen in the figure. A, Prunus microcarpa (CP140) and B, Cerasus angustifolia (CP136).

Macromorphology

Comprehensive qualitative characters are provided for each taxon in Supporting Information, Appendix S2. The quantitative data are summarized in Figure 3 for ease of comparison. The ANOVA results for quantitative characters are given in Table 3. Leaf structure may differ significantly even within the same individual according to placement and time of emergence on the plant. Quantitative leaf macromorphology data show no significant differences between most of the infraspecific taxa, with the exception of C. incana var. velutina.

Table 3.

Results of one-way ANOVA test for the quantitative macromorphological characters from 96 populations of taxa of Prunus subgenus Cerasus in Turkey

	F	P
Leaf width	297.7	< 0.0001
Leaf length	299.2	< 0.0001
Leaf length/width ratio	316.0	< 0.0001
Petiole length	191.1	< 0.0001
Pedicel length	59.56	< 0.0001
Hypanthium length	27.6	< 0.0001
Sepal length	13.69	< 0.0001
Petal length	37.38	< 0.0001
Petal width	62.99	< 0.0001
Petal length/width ratio	26.17	< 0.0001
Style length	16.88	< 0.0001
Filament length	27.44	< 0.0001
Drupe length	84.24	< 0.0001
Drupe width	129	< 0.0001
Drupe length/width ratio	16.93	< 0.0001

Figure 3.

Comparative quantitative data from leaves, flowers and fruits of Prunus subgenus Cerasus taxa in Turkey. Whiskers show the maximum and minimum, lines inside boxes show average values.

The importance of floral morphology is generally underestimated in descriptions of these taxa (Browicz, 1972). Reasons for this include the fragile nature of the flowers and a lack of flowers in herbarium material, a fate shared by most of the previously collected specimens in our herbarium (ISTF). For this reason, we arranged separate trips to ensure collection of both flowering and fruiting specimens in the present study. On examination of the flowers (Supporting Information, Appendix S2), we found several taxonomically useful characters, including sepal indumentum, hypanthium shape and indumentum and petal shape. Figure 4 shows a morphological comparison of the flowers in the examined taxa.

Figure 4.

Comparative floral morphology of members of Prunus subgenus Cerasus in Turkey. A, P. cerasus; B, P. avium; C, P. microcarpa; D, P. microcarpa subsp. tortuosa; E, P. prostrata; F, P. prostrata var. glabrifolia; G, C. angustifolia var. angustifolia; H, C. angustifolia var. sintenisii; I, C. incana var. velutina; J, C. incana; K, P. hippophaeoides; L, P. brachypetala; M, P. mahaleb var. alpina and N, P. mahaleb. Scale bar 1 cm. Drawings by R. Mollman.

The PCoA resulted in four distinct groups based on morphology (Fig. 5). Prunus avium, P. cerasus, P. mahaleb and P. microcarpa are distinct, whereas the other taxa form a close group. One notable outcome is the grouping of subspecies of P. angustifolia and P. hippophaeoides with the other taxa, although they are readily distinguished with the naked eye because of their narrower leaves.

Figure 5.

Principle coordinate analysis results obtained from morphological and micromorphological data showing four major groups. Cerasus incana var. velutina falls far from all these groups, including that which contains C. incana.

Genetics

Haplotype analysis

A haplotype analysis performed on 65 individuals found 11 different haplotypes (Fig. 6), including up to three different haplotypes per taxon (P. mahaleb and P. cerasus) and up to five taxa sharing the same haplotype (P. hippophaeoides, C. angustifolia var. sintenisii, P. prostrata var. glabrifolia, P. prostrata var. prostrata and P. brachypetala). The haplotypes with their corresponding specimens and taxa are given in Table 4.

Table 4.

The haplotypes and corresponding taxa with collection number

Haplotype 1	CP109 CP116 CP117 CP130	P. mahaleb var. mahaleb P. mahaleb var. mahaleb P. mahaleb var. mahaleb P. mahaleb var. mahaleb
Haplotype 2	CP58 CP59 CP62 CP66 CP194	P. avium P. avium P. avium P. avium P. avium
Haplotype 3	CP71 CP75	P. mahaleb var. mahaleb P. mahaleb var. mahaleb
Haplotype 4	CP57 CP63 CP67 CP68 CP112 CP168 CP187 CP188 CP192 CP193 CP195	P. avium P. avium P. avium P. cerasus P. avium P. avium P. cerasus P. cerasus P. avium P. avium P. avium
Haplotype 5	CP72 CP74 CP78 CP80 CP88 CP89 CP141	P. mahaleb var. mahaleb P. mahaleb var. alpina P. mahaleb var. alpina P. mahaleb var. mahaleb P. mahaleb var. mahaleb P. mahaleb var. alpina P. mahaleb var. mahaleb
Haplotype 6	CP189 CP196	P. cerasus P. cerasus
Haplotype 7	CP91 CP97 CP98 CP99 CP113 CP115 CP118 CP119 CP127 CP128 CP131 CP136 CP138 CP140 CP143 CP144	P. microcarpa subsp. microcarpa P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa P. microcarpa subsp. microcarpa P. microcarpa subsp. tortuosa P. microcarpa subsp. microcarpa P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa C. angustifolia var. angustifolia P. microcarpa subsp. tortuosa P. angustifolia var. angustifolia P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa
Haplotype 8	CP181 CP185	P. incana var. velutina P. incana var. incana
Haplotype 9	CP92 CP101 CP103 CP104 CP139 CP142	P. hippophaeoides C. angustifolia var. sintenisii C. angustifolia var. angustifolia C. incana var. incana C. angustifolia var. sintenisii C. angustifolia var. angustifolia
Haplotype 10	CP84 CP85 CP90 CP110 CP111 CP129 CP148 CP164	P. prostrata var. prostrata P. prostrata var. prostrata P. hippophaeoides C. angustifolia var. sintenisii P. prostrata var. prostrata P. hippophaeoides P. prostrata var. glabrifolia P. brachypetala
Haplotype 11	CP94 CP145	P. microcarpa subsp. tortuosa P. microcarpa subsp. tortuosa

Figure 6.

Haplotype analysis performed with 65 individuals results in 11 different haplotypes, including up to three haplotypes in a single taxon and up to five taxa sharing the same haplotype. The size of the circles indicates the number of individuals sharing a particular haplotype; colour codes of taxa are provided in the legend. Black dots represent missing haplotypes.

Phylogenetic analysis

The total length of the ITS region ranged from 543 to 551 base pairs (bp) for 53 accessions in this study. This dataset along with an outgroup species of P. tomentosa resulted in an alignment of 556 bp in length with 55 potentially parsimony-informative sites (10%). The resulting Bayesian inference (BI) phylogenetic tree using ITS data (Fig. 7) formed two clades, one containing the narrow-leaved taxa (both P. angustifolia taxa and P. hippophaeoides) and the other containing all others, with the exception of C. incana var. incana, which did not fall into either clade. Part of this latter clade has two subclades, one containing P. avium, P. cerasus and P. mahaleb and another containing P. microcarpa plus one C. angustifolia specimen (collected from in the distribution range of P. microcarpa); the remaining parts of the clade consist of the taxa of P. prostrata, P. brachypetala var. bornmuellerii and C. incana var. velutina.

Figure 7.

Bayesian inference tree obtained from plastid data of taxa of Prunus subgenus Cerasus in Turkey. The node labels indicate posterior probability.

The plastid data include three regions. The length of matK ranged from 646 to 651 bp, ycf1 was 820–849 bp and trnL-F was 848–850 bp. The resulting matrix length ranged from 2301 to 2308 bp and the final alignment length was 2608 bp. The BI tree based on plastid data resulted in a phylogenetic tree (Fig. 8) with four clades. The first clade contains the narrow-leaved taxa P. hippophaeoides, P. angustifolia taxa and P. incana taxa, the second contains taxa of P. prostrata and P. brachypetala, the third contains taxa of P. microcarpa and the fourth contains P. avium, P. cerasus and taxa of P. mahaleb.

Figure 8.

Bayesian inference tree obtained from ITS data with different leaf shapes of studied Prunus subgenus Cerasus taxa. The node labels indicate posterior probability. A, P. cerasus; B, P. avium; C, P. mahaleb; D, P. mahaleb var. alpina; E, P. microcarpa subsp. tortuosa; F, P. macrocarpa; G, C. incana var. velutina; H, P. prostrata; I, P. prostrata var. glabrifolia; J, C. incana; K, P. brachypetala; L, C. angustifolia var. sintenisii; M, C. angustifolia var. angustifolia and N, P. hippophaeoides. Scale bars 1 cm. Drawings by R. Mollman.

The two BI phylogenetic trees resulted in similar groupings for all taxa except C. incana var. incana, which produced polytomies on the backbone of the ITS tree (Fig. 7).

DISCUSSION

Prunus is a complex and extensive genus that, due to its size and range, requires a ‘divide and conquer’ strategy for an ample treatment of its many members and their relationships. In the spirit of this endeavour, we set out to elucidate the morphological, micromorphological and molecular characteristics of Prunus subgenus Cerasus taxa represented in Turkey.

Plasticity complicates identification of these taxa from herbarium material and thus the taxonomic usefulness of leaf morphological features for this group. Indumentum is often used as a stand-alone character in the separation of infraspecific taxa in the Flora of Turkey treatment (Browicz, 1972). However, trichome presence and density show striking variation even on the same plant in this group, as we observed during field work. It has been established that trichome density varies according to climatic conditions, especially in narrow-leaved plants (Picotte et al., 2007). Many members of Rosaceae develop significant differences in leaf morphology based on seasonal changes in shoot or leaf type (Weber, 1964, Robertson et al., 1992, Wu, Potter & Cui, 2019). Changing environmental conditions over the growing season may explain different trichome densities across leaves of the same plant. We also observed that leaf apex shape, another character used by Browicz (1972), varies according to the developmental period in many species. As a result of this variation, studied taxa cannot be divided into varieties based on trichome or leaf apex characteristics alone. More detailed work is necessary to fully understand ecological and developmental conditions and their effects on such leaf features, which is beyond the scope of the current study. Suffice to say that relying on such traits confounds the creation of a natural classification in this subgenus.

Seed surface morphologies are overall stable, showing almost no variation among the taxa of Prunus subgenus Cerasus. Although the tertiary structure (epicuticular wax ornamentation) exhibits some variation, it is not adequate to make a clear distinction between taxa. On the other hand, endocarp surface features show variation even within taxa, with the notable exception of P. hippophaeoides, which is always ruminate. Thus, neither seed nor endocarp surface micromorphology provides taxonomically valuable characters.

Leaf micromorphology has proved to be helpful at the infrageneric level in some groups of Rosaceae (Faghir, Chaichi & Shahvon, 2014; Song & Hong, 2016). In this study, we find that adaxial surface micromorphology may differ between populations of the same taxa; however, they have extremely stable abaxial surface morphology.

At an infraspecific level, the taxa are also roughly consistent concerning finer structure of surface characters. As mentioned in the Results, some specimens of P. microcarpa and C. angustifolia with the same haplotype show similar surface characteristics, probably due to hybridization events. Hybrid plants tend to show surface characteristics similar to the maternal taxon (Wissemann, 2000). Since hybridization is known to occur in Prunus subgenus Cerasus, it is best to use surface characters with caution. Consequently, leaf surface micromorphological features alone cannot be used in separating the studied taxa. Further investigation is needed to clarify whether the determinant factor is based on ecological conditions or hybridization.

On the basis of the morphology PCoA results, the studied taxa fit neatly into four groups. The largest group is tightknit, with the exceptions of P. hippophaeoides and C. incana var. velutina. The consistently dense indumentum of P. hippophaeoides is its major distinguishing feature, showing that this character can be used in some limited applications where it is stable, although it is too inconsistent among the other taxa to be used in infraspecific taxon identifications. In the case of C. incana var. velutina, the higher leaf/width ratio and shorter petioles are the two major distinguishing characteristics. Of the leaf morphology features, leaf to width ratio is among the more consistent characters for the studied species.

The other surprising placement includes the gap between the two varieties of C. angustifolia. According to the PCoA loadings, this is due to differences between measurements in their floral morphologies. As can be seen in Figure 4, there is a clear difference in size, but not qualitative characteristics between these two varieties, with C. angustifolia var. sintenisii having significantly larger flowers. We believe this to be primarily due to sampling, as C. angustifolia var. sintenisii is represented in this study by many more specimens, which contributes to the larger range of measurements we observed in this taxon than in C. angustifolia var. angustifolia.

The PCoA of macro- and micromorphological characters shows correlation with the molecular results. Bortiri et al. (2006) placed P. avium and P. mahaleb in Prunus subgenus Cerasus section Cerasus and section Mahaleb, respectively, based on ITS and morphological data. Our ITS results also indicate they are sister taxa. However, in the aforementioned study Bortiri et al. (2006) also suggested P. microcarpa and P. prostrata should be placed in Prunus section Microcerasus; our ITS tree only weakly supports this classification. The ITS, plastid and morphological data indicate that Prunus section Microcerasus is paraphyletic. Although some groups are well supported on the phylogenetic tree, ITS alone is an inadequate marker for explaining the phylogenic relationships among taxa of Prunus section Microcerasus.

Two monographic studies have previously separated Prunus subgenus Cerasus into sections (Rehder, 1940; Ingram, 1948). Table 5 compares our classification to the aforementioned works. Our molecular and morphological findings corroborate Rehder (1940). Under his classification, Rehder (1940) characterized Prunus section Microcerasus as having flowers with short pedicels, and therefore included C. incana in this section. Both our molecular and morphological analyses place P. prostrata, C. angustifolia, P. brachypetala and P. hippophaeoides, with their short pedicellate flowers, into this section, with C. incana. Prunus section Pseudocerasus is characterized by non-clustering flowers, erect or spreading sepals and pointed teeth on the leaf (Rehder, 1940). According to morphological results of this study, P. microcarpa belongs in this group. With its racemose inflorescence and blunt leaf teeth, P. mahaleb, which is quite distinct both molecularly and morphologically, remained in Prunus section Mahaleb, whereas P. avium and P. vulgaris (Mill.) Schur (=P. cerasus) are found in Prunus section Cerasus with their reflexed sepals and obtuse leaf teeth, according to our results.

Table 5.

Systematic classification of taxa of Prunus subgenus Cerasus with natural distributions in Turkey. The table compares the subgeneric and sectional placements of these taxa based on results of the current study to the two previous monographs, which included most of the taxa investigated in this study. * denotes taxa that were considered in previous treatments and were thus not assigned to a subgenus or section

	Subgenus			Section
Taxon	This study	Rehder (1940)	Ingram (1948)	This study	Rehder (1940)	Ingram (1948)
P. microcarpa	Cerasus	*	Lithocerasus (=Prunus)	Pseudocerasus	*	Amygdalocerasus
P. microcarpa subsp. tortuosa		*			*
C. incana var. velutina		Cerasus		Microcerasus	Microcerasus
C. incana
P. prostrata					*
P. prostrata var. glabrifolia					*
C. angustifolia		*	*		*	*
C. angustifolia var. sintenisii		*	*		*	*
P. hippophaeoides		*	*		*	*
P. brachypetala		*	*		*	*
P. mahaleb		Cerasus	Eucerasus(=Cerasus)	Mahaleb		Phyllomahaleb
P. mahaleb var. alpina
P. avium				Eucerasus (=Cerasus)		Eucerasus (=Cerasus)
P. cerasus

Remarks on systematic positions of infraspecific taxa

Prunus mahaleb L., Sp. Pl. 1: 474 (1753)

=Cerasus mahaleb var. alpina Browicz. In: Notes Roy. Bot. Gard. Edinburgh, 31 (2): 321. (1972). syn. nov.

Varieties of P. mahaleb are in need of validation for reasons outlined in Bostanci Ordu et al. (2021). Prunus mahaleb var. alpina was nested in P. mahaleb var. mahaleb in trees derived from both ITS and plastid genome data in our study. These two varieties differ significantly only in respect to leaf and petiole dimensions. Floral and drupe characters do not support the separation of these two taxa. Morphology is indeed an important part of systematic botany, but it is critical to determine the characters that are informative for taxonomy and recognize those that vary tremendously depending on ecological conditions. Because the aforementioned characters are inconsistent even within a single individual, we recommend that this species should not be further divided into varieties.

Prunus incana (Pall.) Steven in Mém. Soc. Imp. Naturalistes Moscou 3: 263 (1812)

Morphological and molecular data suggest that Cerasus incana var. velutina Browicz is distinct from Prunus incana (Pall.) Steven var. incana. The material used in this study was collected from the locus classicus on Talas Mountain. A detailed description is given for C. incana var. velutina (under the proposed name P. alidagensis nom. nov.). This taxon is endemic to Kayseri Province in Turkey, where it is under threat from overgrazing and its use as a rootstock for grafting commercially valuable taxa of Prunus subgenus Cerasus. This population, which has a limited distribution and low population density, is in need of protection. A description for this species is given below. This taxon differs from its close relative C. incana by its higher leaf to width ratio and shorter petioles.

Prunus alidagensis Erol, Ciftci & Yaprak nom. nov.

≡ Cerasus incana var. velutina Browicz in Notes Roy. Bot. Gard. Edinburgh 31: 321 (1972)

The epithet velutina is not available under Prunus because of Prunus velutina Batalin, Trudy Imp. S.-Peterburgsk. Bot. Sada xiv. (1895) 168.

Type: Turkey, Kayseri, Talas, Ali Dag, 1400 m, Balls 239. Holotype: E00011263!, isotype: K!.

Leaves 12.55–19.54 (mean 16.62 ± 2.58) × 23.77–34.99 (mean 28.83 ± 3.79) mm, leaf width/length ratio 1.58–1.89 (mean 1.74 ± 0.09), usually oblanceolate or obovate, rarely ovate; leaf apex acute or obtuse; base rounded cuneate or widely cuneate; margins serrate; petiole 0.38–2.88 mm (mean 1.62 ± 0.58). Flowers solitary, in pairs or aggregated; pedicel 0.24–2.34 mm (mean 0.88 ± 0.57), lanate; hypanthium 3.90–5.21 mm (mean 4.30 ± 0.31), cylindrical, villose on the inside, pubescent on the outside; sepals 1.94–2.59 mm (mean 2.23 ± 0.23), narrowly ovate, margins ± dentate, inner surface lanate, outer surface ± pubescent; petals 1.29–2.02 (mean 1.65 ± 0.22) × 3.08–5.74 mm (mean 4.52 ± 0.91), length/width ratio 2.03–3.95 (mean 2.75 ± 0.55), ovate, pink, sericeous along the midvein; filaments 1.65–2.70 mm (mean 2.18 ± 0.30); style 4.59–7.76 mm (mean 5.49 ± 0.82), villose, rarely lanate; ovary villose. Drupe 8.31–9.27 (mean 8.73 ± 0.34) × 6.94–8.51 mm (mean 7.66 ± 0.43), width/length ratio 0.89–1.04 (mean 0.94 ± 0.04); globose, sometimes cordate; rarely glabrous; red. Habitat: volcanic turf, scree slopes.

Kayseri: Talas: northern slopes of Alidağ, O. Erol, Y.C. Gercek CP181!

Prunus microcarpa C.A.Mey. in Verz. Pfl. Casp. Meer.: 166. 1831

≡Prunus microcarpa C.A.Mey. subsp. tortuosa (Boiss. & Hausskn.) Browicz in Arbor Kórnickie 13: 20. 1968 syn. nov.

=Cerasus microcarpa subsp. tortuosa (Boiss. & Hausskn.) Browicz in Arbor Kórnickie 13: 20. 1968 syn. nov.

Cerasus microcarpa subsp. microcarpa and C. microcarpa subsp. tortuosa, which are historically differentiated by leaf shape and indumentum, did not differ at the molecular level. However, this species shows significant quantitative variation in flower morphology (Fig. 4). The subspecies concept is problematic in systematic botany, and some authors view the subspecies as a ‘trash bin’ due to the tendency of many researchers to place enigmatic specimens at the subspecific rank (H. Kerndorff, pers. comm.). Nonetheless, in the most basic sense, a subspecies is a geographically separate population within a species. In light of the geographical, molecular, morphological and anatomical data collected in this study, we treat these two subspecies as synonyms of P. microcarpa to avoid an artificial classification.

Prunus albicaulis Koehne ex Bornm. in Feddes Rep. Beih. 89: 223 (1940)

=Cerasus angustifolia (Spach) Browicz in Fl. Turkey 4: 16 (1972) syn. nov.

=Prunus erzincanica (Yıld.) Yıld., Ot Sist. Bot. Dergisi 21(1): 37 (2015) (syn. nov.)

=Cerasus angustifolia var. sintenisii (C.K.Schneid.) Browicz in P.H.Davis (ed.), Fl. Turkey 4: 16 (1972) syn. nov.

Since the epithet angustifolia is not available under Prunus because of Prunus angustifolia Marshall, Arbust. Amer. 111 (1785), the earliest legitimate name for this taxon in the genus is Prunus albicaulis Koehne ex Bornm. A phenomenon of variation in leaf characteristics similar to that of P. mahaleb has been observed in C. angustifolia. One can identify a single C. angustifolia specimen as either variety depending on the orientation or emergence of the branch they have at hand. Therefore, we treat these two varieties (C. angustifolia var. angustifolia and C. angustifolia var. sintenisii) as synonyms of P. albicaulis.

Lectotype (designated here):

Turkey: ‘Armenia turcica, Sipikor, inter Szadagh et Arvschusch’, Sintenis, P., #Iter orientale 1890 3494, (JE), JE00000026!; isolectotype: (JE), JE00000031!; Syntype: Turkey: ‘Armenia turcica, Auschin [=Avsin] ad Euphratem, in montosis supra Biachbet’, Sintenis,P., #Iter orientale 1890 2200, (JE), JE00000032!

We observed Prunus erzincanica (Yild.) Yild. in its type locality and found it to be indistinguishable from P. albicaulis Koehne ex Bornm. Yildirimli (1993) used lengthwise lenticels as a diagnostic character, but we observed this character in young shoots of all taxa in Prunus section Microcerasus (P. incana, C. angustifolia, P. alidagensis, P. brachypetala and P. prostrata). Accordingly, we consider this species a synonym of P. albicaulis.

Prunus prostrata Labill. var. glabrifolia (Moris) Browicz, Dec. Pl. Syr. i. t. 6

Prunus prostrata var. glabrifolia was described from Sardinia, Italy, on the basis of its glabrous leaves. We collected this variety from Tahtali mountain, the only known locality in Turkey, but observed no differences between it and material from elsewhere. The Prunus specimens are missing from the Moris collection, including the type specimen of P. prostrata var. glabrifolia. Despite this, molecular and morphological data suggest that the Tahtali specimens do not differ from other specimens of P. prostrata collected from the rest of the country. We conclude that Cerasus prostrata var. glabrifolia (Moris) Browicz Fl. Turkey 4: 13 (1972) is not found in Turkey.

Prunus brachypetala (Boiss.) Walp. in Ann. Bot. Syst. 1: 272. 1848

Donmez & Yildirimli (1999) published a new record of Cerasus brachypetala var. brachypetala for Turkey, but did not present a clear distribution for the taxa. Neither our field work nor our laboratory observations confirm the existence of this variety in Turkey. Further problems arise because researchers of this taxon in Turkey (Browicz, 1972; Donmez & Yildirimli, 1999) distinguish them using leaf morphology, whereas notes on the type specimen left by Koehne in 1915 in JE herbarium focus on hypanthium morphology. Turkey constitutes the northern end of the distribution range of P. brachypetala. Examination of flower characteristics is nearly impossible with previously collected herbarium specimens because of their deciduous petals, which makes it difficult to determine the validity of this variety. Examination of populations across their distribution range is necessary to be sure. Our data suggest that P. brachypetala var. bornmuelleri is the only variety distributed in Turkey.

ACKNOWLEDGEMENTS

The authors would like to thank the curators of K, ISTE, KNYA, ISTO, JE, E, VANF, HUB, ANK, NGBB, DUOF, MARE, EGE, BULU and GOET. The molecular section is part of the PhD dissertation of Y.C.G. This study was funded by the Scientific and Technological Research Council of Turkey (TUBITAK) under grant number 116Z247 and the Scientific Research Projects Coordination Unit of Istanbul University under project number FDP-2018-31327 and FDK-2018-27427. Authors declare no conflict of interest.

DATA AVAILABILITY

The data underlying this article are available in the Supporting Information and in the GenBank Nucleotide Database at https://www.ncbi.nlm.nih.gov/, and can be accessed with ISTF accession numbers found in Appendix S1.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article at the publisher's web-site:

Appendix S1. Material studied genetically, morphologically and micromorphologically in this work.

Appendix S2. Leaf, flower, drupe and seed macro- and micromorphology qualitative features.

Author notes

Almila Ciftci and Yusuf C Gercek. These authors contributed equally to this work.