Abstract
• Premise of the study: The current study aimed at developing nuclear microsatellite markers for the relict tree species Zelkova carpinifolia, which is threatened in its natural range in the South Caucasus.
• Methods and Results: Pyrosequencing of an enriched microsatellite library on the Roche FLX platform using the 454 Titanium kit produced 86,058 sequence reads, most of which contained short tandem repeats. Eighty microsatellite loci identified using the software package QDD version 1 were selected and tested for proper PCR amplification. Of these, 13 allowed proper amplification and were shown to be polymorphic among a sample of 25 Z. carpinifolia specimens from various geographic origins.
• Conclusions: The set of microsatellite markers will be suitable for the assessment of genetic diversity in Z. carpinifolia. They will allow for an examination of phylogeographic patterns as well as of population structure and gene flow within this species.
Keywords: conservation genetics, microsatellites, phylogeography, pyrosequencing, Ulmaceae, Zelkova carpinifolia
Zelkova carpinifolia (Pall.) K. Koch. (Ulmaceae) is a representative species of the Arcto-Tertiary relict flora (Wang et al., 2001; Milne and Abbot, 2002; Kozlowski et al., 2012). It occurs naturally in Hyrcan and Colchic forests in Iran, Azerbaijan, Georgia, and Turkey (Browicz and Zielinski, 1982; Denk and Grimm, 2005; Kvavadze and Connor, 2005; Akhani et al., 2010). Its valuable timber has led to uncontrolled logging, which has resulted in a general population decline. Therefore, Z. carpinifolia was listed under the International Union for Conservation of Nature (IUCN) category “Near Threatened” (Güner and Zielinski, 1998). To plan conservation strategies for Z. carpinifolia, investigation of its genetic diversity is of great importance (Kozlowski et al., 2012). The phylogeography of the species needs to be understood as there may be specific refugia within the range of the species that harbor relatively high genetic diversities and regionally adapted genotypes may occur.
The first studies of genetic diversity in Zelkova Spach employed inter-simple sequence repeat (ISSR) and amplified fragment length polymorphism (AFLP) markers (Fineschi et al., 2004). Primers targeting microsatellite loci in Z. serrata (Thunb.) Makino (Fukatsu et al., 2005) did not amplify well in Z. carpinifolia and revealed only limited polymorphism (unpublished data). The current study therefore aimed to develop polymorphic microsatellite markers for Z. carpinifolia using 454 pyrosequencing.
METHODS AND RESULTS
Plant material and DNA extraction
Zelkova carpinifolia samples were collected from six populations (10–20 individuals per population) in Azerbaijan and Iran (Appendix 1). Genomic DNA was extracted from silica gel–dried leaves using either the NucleoSpin Plant II kit (Macherey Nagel, Düren, Germany) following the manufacturer’s protocol or a three-fraction cetyltrimethylammonium bromide (CTAB) method (Borsch et al., 2003).
Microsatellite marker development
Construction of the enriched microsatellite libraries was carried out according to the protocol of Glenn and Schable (2005) with modifications. Genomic DNA was digested in 25-μL reactions containing 1× NEBuffer 3.1, 0.1 mg/mL bovine serum albumin (BSA), 0.05 M NaCl, 20 U EcoRI, 10 U MseI restriction enzymes (New England Biolabs, Frankfurt, Germany), and 2.4 μg genomic DNA during a 2-h incubation at 37°C. Digested DNA was run on a 2% agarose gel, and 200–400-bp fragments were extracted using the Gel/PCR DNA Fragment Extraction Kit (Avegene, Zollikofen, Switzerland) following the manufacturer’s protocol. To prepare double-stranded adapters, forward (EcoRI: 5′-CTCGTAGACTGCGTACC, MseI: 5′-GACGATGAGTCCTGAG) and reverse (EcoRI: 5′-AATTGGTACGCAGTCTACGAG, MseI: 5′-TACTCAGGACTCATCGTC) adapters (Metabion, Martinsried, Germany) were mixed in equimolar amounts (10 μM), incubated at 95°C for 5 min, and allowed to slowly cool down to room temperature. Adapters were ligated to the restriction fragments in 40-μL reactions containing 0.8 μM of each of the double-stranded EcoRI and MseI adapters, 1× T4 DNA ligase buffer, and 600 U T4 DNA ligase (New England Biolabs). Reactions were performed at 16°C for 12 h and successful ligation was tested by PCR amplification in 50-μL reactions containing 4 μL of the adapter-ligated DNA, 1× Taq DNA polymerase buffer, 1× PCR enhancer solution, 150 μM of each dNTP, 2 U Taq DNA polymerase (PeqLab, Erlangen, Germany), 35 ng/μL BSA, 0.5 μM EcoRI primer (5′-CTCGTAGACTGCGTACCAATTC), and 0.5 μM MseI primer (5′-GACGATGAGTCCTGAGTAA) (Metabion) using the following temperature profile: 95°C for 2 min; 20 cycles of 95°C for 20 s, 60°C for 20 s, 72°C for 1.5 min; and 72°C for 10 min. Two mixtures of 3′-biotinylated oligonucleotides [1: (AG)12, (TG)12, (AAG)8, (AAT)12, (ACT)12 and 2: (ACAG)6, (ACCT)6, (ACTC)6, (ACTG)6] were used to generate two separate libraries of adapter-ligated genomic DNA enriched for repetitive motifs. Hybridization of biotinylated oligonucleotides and adapter-ligated genomic DNA, followed by the capturing of hybridized DNA on Dynabeads (Life Technologies, Darmstadt, Germany) and subsequent washing and resuspension, followed the protocol of Glenn and Schable (2005). Recovery PCR was performed with 2 μL of enriched DNA in the same reaction mixture and the temperature profile as described above, but with 25 cycles. The enrichment procedure was repeated and the second recovery PCR was carried out with an EcoRI oligonucleotide primer extended with a Roche 454 pyrosequencing adapter (5′-GCCTCCCTCGCGCCATCAG-3′) and a specimen-specific barcode sequence and an MseI oligonucleotide primer extended with a Roche 454 B sequencing adapter (5′-GCCTTGCCAGCCCGCTCAG-3′; Appendix 2). Two enrichment libraries were prepared for each of six Z. carpinifolia specimens (population ZE053_IR, Appendix 1). Enrichment libraries were purified using the Amplicon Library Preparation Protocol (Roche, Branford, Connecticut, USA) according to the manufacturer’s instructions. Lib-L Kits and the GS FLX Titanium Sequencing Kit XLR70 (Roche) were used for the library sequencing on the Roche GS FLX System. The resulting DNA sequences were analyzed using the R package SeqinR (Charif and Lobry, 2007). The 454 pyrosequencing yielded 86,058 total reads ranging from 40 to 577 bp with an average length of 138 bp.
Microsatellite screening
Screening for microsatellite loci and primer design was performed using QDD software version 1 (Meglécz et al., 2010), which identified 136 microsatellite loci containing di-, tri-, and tetranucleotide repeat motifs. A total of 80 microsatellite loci with 5–45 repeats, including three loci with tetranucleotide repeats, 17 loci with trinucleotide repeats, and 60 loci with dinucleotide repeats, were selected randomly, and oligonucleotide primer combinations with a GC content of 35–60% and melting temperature (Tm) ranging between 57°C and 60°C were tested for proper PCR amplification on genomic DNA extracted from three specimens (Table 1). A gradient PCR was carried out for each primer pair in a 25-μL reaction volume containing 20–40 ng DNA, 0.4 μM each of forward and reverse primer (Eurofins MWG Operon, Ebersberg, Germany), 1× TaqBuffer S, 1.5 mM MgCl2, 250 μM of each dNTP, and 0.75 U Hot Taq polymerase (PeqLab). The temperature profile was as follows: 96°C for 2 min, 54 ± 8°C for 30 s, 72°C for 30 s; followed by 34 cycles of 95°C for 30 s, 54 ± 8°C for 30 s, 72° for 30 s; and 72°C for 15 min. PCR performance was assessed by gel electrophoresis. Of the initial 80 loci tested, 40 primer pairs failed to amplify a product. The reactions with the other 40 primer pairs were repeated in 50-μL reaction volumes. PCR products were run on 1.5% agarose gel and extracted using the NucleoSpin Gel and PCR Clean-up Kit (Macherey-Nagel) following the manufacturer’s protocol. PCR products were sequenced at Macrogen (Seoul, Korea) using standard Sanger sequencing to confirm locus-specificity of the oligonucleotide primers. Only the 20 primer pairs that specifically amplified the targeted microsatellite loci were retained for further analyses. Polymorphism assessment used genomic DNA extracted from 25 Z. carpinifolia specimens (five specimens per population; Appendix 1) and fluorescently labeled forward or reverse primers (Applied Biosystems, Warrington, United Kingdom), and the same reaction conditions as in the previous step. Two primer pairs did not amplify a product in all individuals and were excluded from further analysis. The rest of the PCR products were sent to Macrogen for fragment analysis.
Table 1.
Characteristics of 13 nuclear microsatellite loci for Zelkova carpinifolia.
| Locus | Primer sequences (5′–3′) | Fluorescent label | Repeat motif | Ta (°C) | Allele size range (bp) | GenBank/ENA accession no. |
| ZMS_1 | F: AAGGTTCATCCCATGAGCAC* | 6-FAM | (CTT)7 | 60 | 238–259 | KF408341 |
| R: GTCCAGTCCGGCATCTTTT | ||||||
| ZMS_2 | F: CAAGCTGCCAATACTCCACA* | VIC | (AGA)7 | 60 | 181–213 | KF408342 |
| R: GGTTTGGTTTCTGCTTGGAA | ||||||
| ZMS_3 | F: CCCAACAGCACTTGAAGCTA* | 6-FAM | (CA)8 | 60 | 136–138 | KF408343 |
| R: TTTGTGTAGTCTTTGGCCCC | ||||||
| ZMS_4 | F: ACCGTTGGAAACAAAACTCG* | NED | (AG)9 | 52 | 105–131 | KF408344 |
| R: AGGGCAGTAGAGCTTGATCG | ||||||
| ZMS_5 | F: TTTTGTTCGTTGTTGCATGTC* | PET | (GAA)6 | 60 | 99–114 | KF408345 |
| R: AGGCTTAGGTTGGGTGGTTT | ||||||
| ZMS_6 | F: TTTGGTGCTTAGCATGTGGA* | 6-FAM | (CT)8 | 58 | 101–120 | KF408346 |
| R: TCTTTTGTTTGCGATAAGTCG | ||||||
| ZMS_7 | F: TCTCTGTTATGATCACGTCTCCA | NED | (AC)9 | 60 | 120–128 | KF408347 |
| R: GAGCTAAACCCCATTTCAAATATAA* | ||||||
| ZMS_8 | F: ACGAACAAGATCGACAGAGG* | PET | (TC)12 | 60 | 188–204 | KF408348 |
| R: ATTGGATGGCACACGTAATTC | ||||||
| ZMS_9 | F: CAGGTGATTGGAATTGGAAAA* | 6-FAM | (AC)7 | 62 | 171–196 | KF408350 |
| R: CCTCACAGAGTACGAGGGATG | ||||||
| ZMS_10 | F: CTGCGAGGGCAGTAGGTAAG* | VIC | (GA)8 | 62 | 175–189 | KF408351 |
| R: TCCTAGATCCTTTGATATCTTTCTCA | ||||||
| ZMS_11 | F: AGTAGGCTTGGAAGCAAATGA | VIC | (GT)9 | 58 | 96–115 | KF408352 |
| R: TTGGAACTGCAATCGCTTT* | ||||||
| ZMS_12 | F: CTTCCAATCATTCCGAGAGC* | PET | (CA)7 | 58 | 100–104 | KF408353 |
| R: TGACATCTCAAAGGTTGCCA | ||||||
| ZMS_13 | F: AACAAAGAATCTTCATCCCC* | NED | (TG)9 | 58 | 110–116 | KF408354 |
| R: GCATTTGGTTTTACCTAAAGTTACAG | ||||||
| ZMS_14 | F: GCGGTTGAGAGGAGAAAGAA | 6-FAM | (AGG)10 | 57 | 105† | HG737351 |
| R: ACCAAACCCATCAACCATTG* | ||||||
| ZMS_15 | F: GCTCCTCTTCAGCCAAACAC* | 6-FAM | (AGG)12 | 60 | 128† | HG737352 |
| R: CTTCATCCTCGTTTGCATCA | ||||||
| ZMS_16 | F: GCCGGTACTGATTCCTTCAA* | 6-FAM | (AAG)5 | 60 | 233† | HG737353 |
| R: GGACCATCTCACCAACATCC | ||||||
| ZMS_17 | F: AGCCCGTGCTTTTGAATATG | 6-FAM | (GA)8 | 55 | 167† | HG737354 |
| R: GGATAAGGCTATTTTGGCCC* | ||||||
| ZMS_18 | F: GCCACGTAAACAGAGTAAGGG* | 6-FAM | (TG)7 | 58 | 103† | HG737355 |
| R: GCATGTTGGTGCTTTGAGAA |
Note: ENA = European Nucleotide Archive; Ta = optimal annealing temperature.
Labeled primer.
Expected size.
Microsatellite marker data analysis
Electropherograms from the fragment analysis were analyzed using GeneMarker software version 1.95 (SoftGenetics, State College, Pennsylvania, USA). Analyzed amplification products with unscorable peaks were not considered to be useful in the studies of variability due to the possible misinterpretation of data. Thirteen primer pairs produced polymorphic and heterozygous loci (accession numbers KF408341–KF408348, KF408350–KF408354). The Excel Microsatellite Toolkit (Park, 2001) was used for the calculation of the number of alleles, observed (Ho) and expected (He) heterozygosities, and polymorphism information content (PIC) values. Allele sizes ranged from two to eight alleles per locus across 25 individuals from five different populations, with Ho and He varying between 0.12 and 0.84, and 0.25 and 0.81, respectively (Table 2).
Table 2.
Results of initial primer screening of specimens from five populations of Zelkova carpinifolia.
| ZE001_AZ (N = 5) | ZE007_AZ (N = 5) | ZE013_AZ (N = 5) | ZE017_AZ (N = 5) | ZE021_AZ (N = 5) | Total (N = 25) | ||||||||||||||
| Locus | A | Ho | He | A | Ho | He | A | Ho | He | A | Ho | He | A | Ho | He | A | Ho | He | PIC |
| ZMS_1 | 2 | 0.60 | 0.4667 | 4 | 1.00 | 0.7778 | 2 | 0.80 | 0.5333 | 4 | 1.00 | 0.7111 | 4 | 0.60 | 0.6444 | 5 | 0.80 | 0.6294 | 0.5536 |
| ZMS_2 | 3 | 1.00 | 0.7111 | 4 | 1.00 | 0.7333 | 4 | 0.60 | 0.5333 | 4 | 0.80 | 0.7333 | 3 | 0.80 | 0.6444 | 8 | 0.84 | 0.7053 | 0.6439 |
| ZMS_3 | 2 | 1.00 | 0.5556 | 2 | 0.80 | 0.5333 | 2 | 0.80 | 0.5333 | 2 | 0.20 | 0.5556 | 2 | 0.20 | 0.4667 | 2 | 0.60 | 0.4971 | 0.3685 |
| ZMS_4 | 3 | 0.80 | 0.6000 | 4 | 0.80 | 0.8000 | 4 | 0.80 | 0.8000 | 5 | 1.00 | 0.8444 | 5 | 0.40 | 0.8000 | 8 | 0.76 | 0.8188 | 0.7811 |
| ZMS_5 | 3 | 0.40 | 0.3778 | 2 | 0.40 | 0.3556 | 1 | 0 | 0 | 1 | 0 | 0 | 3 | 0.60 | 0.5111 | 5 | 0.28 | 0.2596 | 0.2455 |
| ZMS_6 | 2 | 0 | 0.3556 | 2 | 0 | 0.3556 | 3 | 0 | 0.6222 | 5 | 0.40 | 0.8667 | 5 | 0.20 | 0.8222 | 8 | 0.12 | 0.8049 | 0.7599 |
| ZMS_7 | 3 | 1.00 | 0.6444 | 4 | 1.00 | 0.7778 | 4 | 0.60 | 0.6444 | 2 | 0.80 | 0.5333 | 3 | 0.80 | 0.6000 | 4 | 0.84 | 0.7094 | 0.6413 |
| ZMS_8 | 5 | 1.00 | 0.8222 | 4 | 1.00 | 0.7333 | 4 | 1.00 | 0.7333 | 3 | 0.80 | 0.6000 | 4 | 0.20 | 0.6444 | 7 | 0.80 | 0.7861 | 0.7351 |
| ZMS_9 | 3 | 0.40 | 0.6222 | 3 | 0 | 0.7111 | 4 | 0.60 | 0.7333 | 3 | 0.60 | 0.5111 | 3 | 0.20 | 0.5111 | 5 | 0.36 | 0.7127 | 0.6506 |
| ZMS_10 | 2 | 0.20 | 0.5556 | 4 | 0.60 | 0.7333 | 3 | 0.40 | 0.6000 | 2 | 0.20 | 0.4667 | 4 | 0.80 | 0.7778 | 5 | 0.44 | 0.6588 | 0.6065 |
| ZMS_11 | 1 | 0 | 0 | 3 | 0.40 | 0.7111 | 2 | 0 | 0.3556 | 5 | 0.40 | 0.8222 | 2 | 0 | 0.3556 | 6 | 0.16 | 0.5037 | 0.4526 |
| ZMS_12 | 2 | 0.60 | 0.4667 | 2 | 0.40 | 0.3556 | 2 | 0.20 | 0.2000 | 2 | 0.20 | 0.2000 | 1 | 0 | 0 | 3 | 0.28 | 0.2539 | 0.2316 |
| ZMS_13 | 3 | 0.60 | 0.6444 | 3 | 1.00 | 0.6444 | 2 | 0.80 | 0.5333 | 2 | 0.20 | 0.2000 | 2 | 0.20 | 0.2000 | 4 | 0.56 | 0.5102 | 0.4431 |
Note: A = number of alleles; He = expected heterozygosity (Nei’s unbiased gene diversity); Ho = observed heterozygosity; N = sample size; PIC = polymorphism information content.
CONCLUSIONS
In the current study, we used 454 pyrosequencing to obtain sequence reads for microsatellite primer development in Z. carpinifolia. After thorough evaluation, we present 13 primer pairs for amplifying polymorphic microsatellite loci in Z. carpinifolia. These 13 loci are proposed to be used for the microsatellite screening of naturally distributed Z. carpinifolia populations, but they may also cross-amplify in closely related species. Analysis of the obtained allele frequencies and knowledge of the individual genotypes will help us to understand the population structure and relationships between and within different populations of Zelkova species.
Appendix
Appendix 1.
Locations and voucher information for populations of Zelkova carpinifolia used in this study. Vouchers are deposited at the herbarium of the Botanischer Garten und Botanisches Museum Berlin-Dahlem (B), Berlin, Germany.
| Population | Locality | Geographic coordinates | Collection no. |
| ZE001_AZ | Zuvand Mtn., Lerik, Azerbaijan | 38°43′33″N, 48°25′53″E | EM01 |
| ZE007_AZ | Azfilial, Lankaran, Azerbaijan | 38°40.938′N, 48°47.222′E | EM13 |
| ZE013_AZ | Tezekend, Lankaran, Azerbaijan | 38°39.091′N, 48°48.284′E | EM20 |
| ZE017_AZ | Zungulesh, Astara, Azerbaijan | 38°27.175′N, 48°47.575′E | EM28 |
| ZE021_AZ | Khanbulan, Lankaran, Azerbaijan | 38°39.809′N, 48°46.649′E | EM33 |
| ZE053_IR* | Gorgan, Golestan, Iran | 36°43′35″N, 54°34′49″E |
Used for construction of the genomic library.
Appendix 2.
EcoRI-A and MseI-B primers used for PCR recovery of the enriched DNA library.
| Primer name | Primer sequence (5′–3′)a |
| EcoRI_A001 | CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG ACG AGT GCG TCT CGT AGA CTG CGT ACC AAT TC |
| EcoRI_A002 | CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG ACG CTC GAC ACT CGT AGA CTG CGT ACC AAT TC |
| EcoRI_A003 | CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG AGA CGC ACT CCT CGT AGA CTG CGT ACC AAT TC |
| EcoRI_A004 | CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG AGC ACT GTA GCT CGT AGA CTG CGT ACC AAT TC |
| EcoRI_A005 | CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG ATC AGA CAC GCT CGT AGA CTG CGT ACC AAT TC |
| EcoRI_A006 | CCA TCT CAT CCC TGC GTG TCT CCG ACT CAG ATA TCG CGA GCT CGT AGA CTG CGT ACC AAT TC |
| MseI_B | CCT ATC CCC TGT GTG CCT TGG CAG TCT CAG GAC GAT GAG TCC TGA GTA A |
Specimen-specific barcodes (tags) are shown in bold.
LITERATURE CITED
- Akhani H., Djamali M., Ghorbanalizadeh A., Ramezani E. 2010. Plant biodiversity of Hyrcanian relict forests, N Iran: An overview of the flora, vegetation, palaeoecology and conservation. Pakistan Journal of Botany 42: 231–258 [Google Scholar]
- Borsch T., Hilu K. W., Quandt D., Wilde V., Neinhuis C., Barthlott W. 2003. Noncoding plastid trnT-trnF sequences reveal a well resolved phylogeny of basal angiosperms. Journal of Evolutionary Biology 16: 558–576 [DOI] [PubMed] [Google Scholar]
- Browicz K., Zielinski J. 1982. Chronology of trees and shrubs in south-west Asia and adjacent regions. Polish Academy of Sciences, Institute of Dendrology, Polish Scientific Publishers, Warsaw, Poland [Google Scholar]
- Charif D., Lobry J. 2007. SeqinR 1.0-2: A contributed package to the R project for statistical computing devoted to biological sequences retrieval and analysis. In U. Bastolla, M. Porto, H. E. Roman, and M. Vendruscolo [eds.], Structural approaches to sequence evolution, 207–232. Springer, Berlin, Germany. [Google Scholar]
- Denk T., Grimm G. W. 2005. Phylogeny and biogeography of Zelkova (Ulmaceae sensu stricto) as inferred from leaf morphology, ITS sequence data and the fossil record. Botanical Journal of the Linnean Society 147: 129–157 [Google Scholar]
- Fineschi S., Cozzolino S., Migliaccio M., Vendramin G. G. 2004. Genetic variation of relic tree species: The case of Mediterranean Zelkova abelicea (Lam.) Boisser and Z. sicula Di Pasquale, Garfì and Quézel (Ulmaceae). Forest Ecology and Management 197: 273–278 [Google Scholar]
- Fukatsu E., Isoda K., Hirao T., Takahashi M., Watanabe A. 2005. Development and characterization of simple sequence repeat DNA markers for Zelkova serrata. Molecular Ecology Notes 5: 378–380 [Google Scholar]
- Glenn T. C., Schable N. A. 2005. Isolating microsatellite DNA loci. In E. A. Zimmer and E. H. Roalson [eds.], Methods in enzymology, 202–222. Academic Press, San Diego, California, USA. [DOI] [PubMed] [Google Scholar]
- Güner A., Zielinski J. 1998. Zelkova carpinifolia In IUCN 2013–IUCN Red List of Threatened Species. Version 2013.1. Website http://www.iucnredlist.org [accessed 22 August 2013].
- Kozlowski G., Gibbs D., Huan F., Frey D., Gratzfeld J. 2012. Conservation of threatened relict trees through living ex situ collections: Lessons from the global survey of the genus Zelkova (Ulmaceae). Biodiversity and Conservation 21: 671–685 [Google Scholar]
- Kvavadze E. V., Connor S. E. 2005. Zelkova carpinifolia (Pallas) K. Koch in Holocene sediments of Georgia: An indicator of climatic optima. Review of Palaeobotany and Palynology 133: 69–89 [Google Scholar]
- Meglécz E., Costedoat C., Dubut V., Gilles A., Malausa T., Pech N., Martin J. F. 2010. QDD: A user-friendly program to select microsatellite markers and design primers from large sequencing projects. Bioinformatics (Oxford, England) 26: 403–404 [DOI] [PubMed] [Google Scholar]
- Milne R. I., Abbot R. J. 2002. The origin and evolution of tertiary relict floras. Advances in Botanical Research 38: 281–314 [Google Scholar]
- Park S. D. E. 2001. Trypanotolerance in west African cattle and the population genetic effects of selection. Ph.D. thesis, University of Dublin, Dublin, Ireland.
- Wang Y.-F., Ferguson D. K., Zetter R., Denk T., Garfi G. 2001. Leaf architecture and epidermal characters in Zelkova, Ulmaceae. Botanical Journal of the Linnean Society 136: 255–265 [Google Scholar]