Pollen morphological traits analysis of eighteen 'Nepeta' species in Iran

Nepeta is one of the largest genera of Lamiaceae, which is widely distributed in Iran. Several species of the genus are traditionally used as culinary, industrial, and medicinal plants. In the present study, we examined fourteen qualitative and quantitative pollen morphological characters of twenty-four populations of eighteen Nepeta using light and scanning electron microscope. The pollen grains are radially symmetric, isopolar, and hexacolpate in all species. We have found that the quantitative features vary among the studied taxa. According to PCA analysis, the colpus length and width are the most variable traits. Polar view shapes in all the studied taxa are circular, while the equatorial view is prolate-spheroidal, subprolate, perprolate, and prolate. We determined microreticulate and bireticulate exine sculpturing, and the bireticulate type was dominant. We revealed that polar and equatorial axes length of pollen grains varied between populations of the same species, and they, therefore, should be ignored in taxonomical identification of species. The polar view was similar in all the studied pollen grains and could not be considered as a diagnostic trait. The equatorial shape was stable between populations of the same species, and this was taxonomically important, while we registered some infraspecific variations in exine sculpture. Our cluster of taxa based on the pollen morphology is not in concordance with the traditional sectional classification of the genus.


Introduction
The genus Nepeta L. is the largest in Lamiaceae in Iran. It is included in the Nepetoideae subfamily, tribe Mentheae (Cantino, 1992), being the infrageneric classification highly controversial. Rechinger (1982) has listed 63 species in Iran grouped into 13 sections. Recently, Jamzad (2012) identified more than 79 species in the country and classified them into six sections.
Nepeta species are widely used in traditional medicine because of their expectorant, antispasmodic, antitussive, diuretic, antiseptic, and antiasthmatic activities, while the inflorescences of some species have also been used as a sedative drug (Baser et al., 2000). Some species were used as culinary plants, for example, Nepeta cataria L. had been used as a tea beverage in Europe before the real tea was imported (Newall et al., 1996;Baser et al., 2000). Sammataro & Avitabile (1998) reported that bees use several species as a source of pollen and nectar.
Previous studies proved that frequent hybridization and introgression associated with the substantial age-related or phenotypic plasticity make Nepeta be considerably complex genus (Talebi et al., 2018). Until now, various patterns of infrageneric classifications were constructed in different Flora by Boissier (1879), Pojarkova (1954), Rechinger (1982), Hedge & Lamond (1982), Budantsev (1997) and Dirmenci (2003). These patterns are mainly based on the different morphological ARTICLES features of vegetative and reproductive organs, like plant habit, leaf morphology, inflorescence, calyx and corolla structure, and nutlet traits. Jamzad et al. (2003) conducted a phylogenetic study of this genus and reported that Nepeta is a monophyletic genus. Moreover, they proposed the revision of the previous infrageneric classification.
Palynology is one of the important tools in the taxonomy of Lamiaceae, and it provides valuable data towards the classification of genera and species within the family. According to Erdtman (1945), family Lamiaceae is divided into two subfamilies based on the pollen: 1) Lamioideae with tricolpate and bi-nucleate pollen and 2) Nepetoideae with hexacolpate and trinucleate pollen. The classification based on a number of nuclei and apertures in the pollen grains is widely used (Harley et al., 1992;Abu-Asab & Cantino, 1994).
The pollen morphological characteristics of some Nepeta species have been studied by several botanists (e.g. Ubera, 1982;Perveen & Qaiser, 2003). Moon et al. (2008) revealed that ornamentation of sexine is variable and could be significant at the generic level. Just a few authors studied pollen morphological features of Nepeta species in Iran, but these investigations restricted to members of one to two sections/groups, namely section Psilonepeta (Azizian et al., 2001), section Cataria (Ranjbari et al., 2004), and groups 2 and 4 (Razavi et al., 2017). Jamzad et al. (2000) studied pollen morphology of eleven annual species of two sections of the genus. Razavi et al. (2017) found variations in pollen characteristics among some taxa, and their species classification based on palynological traits were similar to Rechinger (1982) subgeneric classification of Nepeta.
The phylogenetic finding in the genus Nepeta has triggered the need to obtain phylogenetic patterns based on other features than inflorescence morphology and its sequence data. The highly diverse Nepeta flora of Iran shows a large coverage of the phenotypical variations and sections of the genus. While some species of the genus Nepeta growing in Iran have been poorly known palynologically, the present study confirmed the importance of pollen morphological traits and their infraspecific variations in the infrageneric classification.
Therefore, the main goal of this research was to: a) determine the infrageneric and infraspecific pollen morphological variations of eighteen Iranian Nepeta taxa including in five groups and b) to assess the implementation of the palynological data for classification at infraspecific and sectional levels within the genus.

Material and Methods
We examined the pollen grains of twenty-four populations representing eighteen recognized taxa of the genus Nepeta from Iran. Based on the geographical distribution, one to two population(s) of each taxon was collected during spring 2015-2016. The identification of collected samples was done according to descriptions provided in Flora Iranica (Rechinger, 1982) and Flora of Iran (Jamzad, 2012 Table 1.

Light microscopy
We prepared pollen grains for light microscopy (LM) according to the Acetolysis method (Erdtman, 1960). We measured twenty acetolyzed pollen grains per herbarium sheet with Olympus bx51 light microscope and digital camera dp71. Measurements include polar and equatorial axes length, colpus length and width, apocolpium and mesocolpium diameter. We used image tools (ver. 2) for the measurements.

Scanning Electron Microscopy
For scanning electron microscopy (SEM) examination, pollen grains were transferred directly to a stub covered with double-sided transparent tape. The studs were coated with gold in a JFC-1600 Auto Fine Coater and photographed in SU 3500 scanning electron microscope at 15 kV. We measured polar and equatorial shapes, pori diameter, lumen shape and dimensions by the electronic micrographs. Descriptive terms were, according to Moore et al. (1991) and Erdtman (1960). The pollen shape has been determined using the ratio between the polar axis and the equatorial diameter (P/E ratio). If P/E = ca. 1, the grains are termed spheroidal. If P/E > ca 1.2, the grains are termed prolate, if P/E < ca. 0.8, the grains are oblate. Moreover, the subcategories prolatespheroidal and oblate-spheroidal were used for the grains that are slightly prolate or oblate, respectively. Pollen sculpturing is an external feature of the pollen wall. The pollen grains with reticulate sculpturing pattern present elements termed a murus, and space between muri is termed a lumen.

Statistical analysis
We applied one-way ANOVA test to assess the palynological characteristics and Pearson's coefficient of correlation to determine the relationships between the pollen features using SPSS ver. 15. The data for principal coordinates analysis (PCOA) and correspondence analysis (CA) were standardized. We performed the CA-joined plot represented pollen features similarity relationships among the species, linking the palynological traits using MVSP ver. 3.2 (Podani, 2000).

Statistical analysis
ANOVA test showed significant differences (p ≤ 0.05) among all quantitative palynological traits, except colpus width, polar axis length, polar/equatorial length ratio, and mesocolpium (Table 3). Principal component analysis proved that five quantitative variables, namely colpus length, width and length/width ratio, polar axis length, and equatorial axis length, comprise 77.79 % of total variations (Table 4).  We registered significant positive/negative correlations between some quantitative palynological characteristics. For example, the colpus length had significant positive correlations (p ≤ 0.01) with equatorial (r = 0.54) and polar (r = 0.65) axes length, but it had significant negative correlation with the pori diameter (p ≤ 0.05, r = -0.33). We also revealed positive correlation (p < 0.01, r = 0.45) between the colpus width with mesocolpium. Polar axis length had positive correlation (p < 0.01, r = 0.76) with equatorial axis length, we also recorded neg-ative correlation (p < 0.05, r = -0.32) between the pori diameter and equatorial axis length.
We separated the studied taxa and their populations by PCOA plot of palynological data. The populations of the same species placed considerably far from each other, moreover, the species of the same group did not cluster closely, while the species of different groups were clustered together. We revealed (Figure 8) that the species plotting did not coincide with the sectional classifications proposed by Jamzad et al. (2013).    Table 1.
CA-joined plot showed that some of the studied taxa and their populations were characterized by special trait(s), which could be useful in species identification. For example, the polar/equatorial (P/E) ratio was a reliable trait for the identification of N. glomerosa (Jiroft population). Polar and equatorial axe lengths should be considered as prompt diagnostic features for N. cataria (Khorasan population) and N. cataria (Mazandaran population), respectively. N. fissa (Albourz population) and N. meyeri (Lasem population) had different colpus width, whereas the apocolpium diameter was a specific characteristic for N. sessilifolia (Alvand population). N. fissa (Albourz population) was separated from other species by mesocolpium width value and N. crassifolia was identified by lumina specific shape (Figure 9). Figure 9. CA-joint plot of the studied taxa with their pollen morphological variables (see Table 1 for taxa numbers).

Discussion
Nepeta belongs to the subfamily Nepetoideae with hexacolpate pollen grains (Harley et al., 1992;Abu-Asab & Cantino, 1994;Razavi et al., 2017). We found that all the examined taxa present 6-zonocolpate monad pollen grain, this was also reported by Celenk et al. (2008a). Most of the species and their populations had hexacolpate pollen grains, whereas the tetracolpate and abnormal pollen grains have been rarely described before.
Variations in pollen size among the populations of the same species were observed in various Lamiaceae genera such as Mentha L. (Gocmen et al., 1997;Celenk et al., 2008b) and Lycopus (Moon & Hong, 2003;Trudel & Morton, 1992) and our results confirmed this. The pollen size cannot be used as a diagnostic character for the studied Nepeta species because the pollen size greatly varied among the populations (Table 5). We observed different shapes of pollen grains derived from P/E ratio among the studied taxa, though the more frequent type was subprolate. N. kotschyi var. persica (Kashmar population) was the only species with the perprolate pollen grains. It broadly agrees with the results of early studies of N. cataria, N. fissa, N. heliotropifolia, N. meyeri, N. racemosa, N. crassifolia, N. hausskneckhtii, and N. sessilifolia (Azizian et al., 2001;Celenk et al., 2008a;Razavi et al., 2017).
N. sessilifolia (Arak population) and N. fissa (Albourz population) had completely different shape than the reported one in previous research (Jamzad et al., 2000). It seems that these variations could be related to the creation of infraspecific taxonomical ranks. Both species have a wide distribution in Iran. Previous studies on various populations of these species (Talebi et al., 2018(Talebi et al., , 2019 have revealed that environmental factors should determine the essential oil compositions and anatomical characteristics, and substantially lead to the infraspecific variations. We believed that these variations cause the creation of new infraspecific taxonomical ranks, which should be not morphologically identified. Nevertheless, the identification of these hide ranks needs an additional comprehensive molecular and cytological studies. Pollen grain shape is a nearly stable trait among the studied species and should be a valid taxonomical characteristic.
We compared the exine ornamentations between the populations of the same species and compared them with results of previous palynological studies. We observed infraspecific exine ornamentation variations for populations of N. kotschyi var. persica and N. cataria; moreover, the variations in exine ornamentation were found in different palynological study on the same species. For example, Celenk et al. (2008a) recognized the bireticulate exine in N. fissa, but Razavi et al. (2017) found that these species has microreticulate ornamentation.
The polygonal lumina shape was present in all most studied taxa. We registered such variations in lumina shape within the genus and populations. For example, two different shapes of lumina were registered in N. meyeri and N. fissa (Celenk et al., 2008a;Razavi et al., 2017). We observed some infraspecific variations in lumina shapes between populations of the same species; therefore it seems to be a semi-plastic character.
Therefore, in the current study, pollen grain morphology was investigated to examine the value of pollen features in the infrageneric taxonomy. We compared our results with sectional classification patterns proposed by Jamzad (2012) and revealed that group/section species do not occur together towards the species of different sections.

Conclusion
The pollen morphological features of Nepeta kotschyi var. persica, N. lasiocephala, N. menthoides, N. mirzayanii, N. pogonosperma, N. saccharata, N. satureioides, and N. wettsteinii were described for the first time. The size of the pollen grains of the studied taxa varied greatly among the different species. The length of the polar and equatorial axes varied significantly between the populations of the same species. Nevertheless, the equatorial pollen view is a specific feature, while the polar view was similar in all species and cannot be used as a diagnostic trait.
We have found two types of exine sculpture in the Nepeta species and determined significant infraspecific variability. The analysis of variance (ANOVA) indicated that there was a significant difference in some quantitative characteristics, namely apocolpium diameter and pori diameter. Some of the studied taxa were characterized by special variable(s) such as P/E ratio by CAjoined plot. Finally, we conclude that the taxa arrangement in PCOA plot does not correspond with previously proposed sectional classification.