Juniperus excelsa

Hojjati, Fatemeh, Kazempour-Osaloo, Shahrokh, Adams, Robert Philip & Assadi, Mostafa, 2018, Molecular phylogeny of Juniperus in Iran with special reference to the J. excelsa complex, focusing on J. seravschanica, Phytotaxa 375 (2), pp. 135-157 : 155

The Juniperus excelsa complex

The present molecular analyses suggested that 24 individual specimens from SE Iran, which were considered members of the broadly defined J. excelsa by Assadi (1998), formed their own lineage distinct from the typical J. excelsa . Those samples are morphologically more similar to J. seravschanica , which is characterized mainly by the thickest ultimate branchlets. Other diagnostic characters are: seed cones 8–10 mm, scale leaves with clear, ellipsoid glands, often ruptured, with a clear exudate (Adams 2014). Moreover, some representatives (R2, R4, U5, U9, K2, K7) of this complex that were thought to be new taxa ( Adams et al. 2014) were analyzed in this study utilizing additional DNA sequences. Our phylogenetic analyses of 8428 nucleotide sites of nrDNA ITS, four plastid intergenic spacers (pet Npsb M, trn D- trn T, trn L- trn F and trn S- trn G) and five single copy nuclear genes (LHCA 4, maldehy, myb, CnAIP 3 and 4CL) show that these individuals from Iran belong to the J. seravschanica clade and the differences are just at the inter- and/ or infra-population level ( Fig. 6 ).

The Juniperus plants from NE Iran are recognized as J. polycarpos var. turcomanica with thin ultimate branchlets. The Juniperus plants from NW Iran are recognized as J. polycarpos var. polycarpos with ultimate branchlets being thicker than found in var. turcomanica . These results are consistent with the findings of previous studies on the Iranian J. excelsa complex based on isoenzyme, essential oils and DNA sequencing data ( Hojjati et al. 2009; Adams & Shanjani 2011; Adams & Hojjati 2012, 2013; Adams et al. 2014). Accordingly, there are no Juniperus samples from Iran that should be recognized as J. excelsa . Juniperus excelsa is distinguished from J. polycarpos in having very thin ultimate branchlets (vs. thicker branchlets in J. polycarpos ). Other diagnostic characters to distinguish these taxa, such as scaleleaf size and seed cone size, are discussed in Adams (2014).

The major unresolved problem in the Iranian Juniperus is the nature of samples from SW Iran (group X). These junipers are morphologically very diverse. In fact, Adams & Hojjati (2012) recognized a population from Fasa, SW Iran, as J. polycarpos var. turcomanica . Analyses of leaf essential oils data indicated that the low cedrol trees (three trees from Kuhbanan and 2 trees from Fasa but none of the samples from Khabr or Rabor) were loosely grouped with the samples of J. polycarpos var. turcomanica ( Adams & Hojjati 2013) . Furthermore, Adams et al. (2014) recognized the plants from Fasa and two samples from Kuhbanan as J. polycarpos var. turcomanica . However, the minimum spanning network analysis of nrDNA ITS and plastid sequences in that study indicated that an individual from Fasa (F5) might be a hybrid. Also based on the nrDNA ITS sequences, Adams et al. (2014) indicated that most of the trees from Kuhbanan appear to be hybrids. Clearly the dynamics of Juniperus taxa in S Iran are complex and implied either hybridization or incomplete lineage sorting. We utilized LHCA4, a single copy nuclear gene, to address the evolutionary history of the group X. In the LHCA 4 tree ( Fig. 5 ) group X is placed in the J. seravschanica clade, but in plastid pet Npsb M tree group X is outside that clade. Although these clades are weakly supported, the geographical distributions of the species infer the hybrid nature for group X. The Bayesian tree based on concatenated nrDNA ITS, pet N -psb M and LHCA 4 sequences does not group clearly group X with any species ( Fig. 3 ). However, they are morphologically more similar to J. seravschanica and J. polycarpos . Our examination of Figs. 4 and 5 suggests a hybrid nature for group X because it is grouped differently in plastid and LHCA 4 trees.

The examination of habitats of the taxa reveals that J. seravschanica is distributed in Central Asia, Pakistan to SE Iran with an arid climate, J. polycarpos var. turcomanica grows in mesic areas of the Elburz and Kopet mountains of Iran and Turkmenistan, J. polycarpos var. polycarpos occurs in the Caucasus and NW Iran with a temperate climate, and J. excelsa is distributed in Europe with a Mediterranean climate.

Hence, three species of this complex are clearly distributed in three very different climates. Thus, the existence of these species in SW Iran seems unlikely because the climate in SW Iran is different from climates in SE and N Iran. In fact, the climate is somewhat intermediate between climates in SE and N Iran. The intermediate climate in SW Iran would provide an intermediate habitat for group X of junipers that are hybrids (or derived from hybrids) between J. polycarpos from N Iran and J. seravschanica from SE Iran as supposed. Fig. 7 shows a distribution pattern of the J. excelsa complex members in Iran.

Figures

FIGURE 3. Fifty percent majority rule consensus Bayesian tree based on nrDNA ITS, plastid petN-psbM intergenic spacer and single copy nuclear gene LHCA4 combined data. Numbers above branches are Bayesian posterior possibilities (before slashes) and maximum likelihood bootstrap support values (after slashes), while numbers under branches are maximum parsimony bootstrap support values. Values <50 % are not shown. CI= 0.654 RI= 0.933.

FIGURE 4. Fifty percent majority rule consensus Bayesian tree based on plastid petN-psbM intergenic spacer data. Numbers above branches are Bayesian posterior possibilities (before slashes) and maximum likelihood bootstrap support values (after slashes), while numbers under branches are maximum parsimony bootstrap support values. Values <50 % are not shown. CI= 1.000 RI= 1.000.

FIGURE 5. Fifty percent majority rule consensus Bayesian tree based on single copy nuclear gene LHCA4 data. Numbers above branches are Bayesian posterior possibilities (before slashes) and maximum likelihood bootstrap support values (after slashes), while numbers under branches are maximum parsimony bootstrap support values. Values <50 % are not shown. CI= 0.909 RI= 0.989.

FIGURE 6. Fifty percent majority rule consensus Bayesian tree based on nrDNA ITS, plastid petN-psbM, trnD-trnT, trnL-trnF,trnS-trnG intergenic spacers and single copy nuclear genes (LHCA4, maldehy, myb, CnAIP3 and 4CL) combined data. Numbers above branches are Bayesian posterior possibilities (before slashes) and maximum likelihood bootstrap support values (after slashes), while numbers under branches are maximum parsimony bootstrap support values. Values <50 % are not shown. CI= 0.813 RI= 0.938.

FIGURE 7. Distribution of Juniperus polycarpos var. polycarpos, J. polycarpos var. turcomanica, J. seravschanica and hybrids in Iran.