In this paper, detailed meiotic analysis was investigated in seven species of Curcuma (Linnaeus, 1753) which can contribute significantly to our understanding about species inter-relationship, speciation and evolution. The species were divided into two groups viz., Group I having 2n = 42 (C. comosa Roxburgh, 1810, C. haritha Mangaly & M.Sabu, 1993, C. mangga Valeton & Zijp, 1917, and C. motana Roxburgh, 1800) and Group II with 2n = 63 (C. caesia Roxburgh, 1810, C. longa Linnaeus, 1753 and C. sylvatica Valeton, 1918). Both groups display varying degree of chromosome associations. Group I species showed the prevalence of bivalents, however occasional quadrivalents besides univalents were also encountered. About 48% of the PMCs analyzed in C. mangga showed 21 bivalents (II) meiotic configurations, 32% in C. comosa and 16% in C. haritha. Group II species as expected showed the presence of trivalents besides bivalents, univalents and quadrivalents. About 32% of the PMCs analyzed at MI in C. sylvatica showed 21 trivalents (III) meiotic configurations, 24% in C. longa and 8% in C. caesia. Overall, low frequency of multivalent associations as compared to bivalents indicates that Curcuma is an allopolyploid complex. Moreover, x = 21 is too high a basic number, therefore, we suggest that the genus Curcuma has evolved by hybridization of species with different chromosome numbers of 2n = 24 and 18, resulting in a dibasic amphidiploid species.

Polyploidy, amphidiploid, inter-specific crosses, diversification

The genus Curcuma Linn. belonging to the tribe Zingibereae of the family Zingiberaceae consists of about 120 species and is pan-tropical in distribution (Kress et al. 2002, Škorničková et al. 2007, Záveská et al. 2012). It contains many taxa with multifaceted uses and quite a few species of Curcuma (e.g. C. amada, C. caesia, C. longa, etc.) are used as spice, dye, medicine, cosmetics, ornamental and as a source for starch (Sasikumar 2005, Velayudhan 2012).

Curcuma, a rhizomatous, perennial and herbaceous group of plant displays a great degree of diversity in ploidy levels which is evident from earlier cytogenetical studies wherein various chromosome numbers of 2n = 22, 42, 63, 77, 105, etc., have been reported. Moreover, continuous dispute concerning the basic chromosome number in Curcuma (x = 7, 8, 16 and 21) has been highlighted in early cytological studies of Raghavan and Venkatasubban (1943), Sharma and Bhattacharya (1959), Ramachandran (1961), Islam (2004), Škorničková et al. (2007). Whilst a lot of information on the somatic chromosome number is available for the genus Curcuma, essential information about the homology among the chromosome complements and level of polyploidy has yet to be investigated.

Meiosis, a highly conserved and specialized process in eukaryotes, not only generates genetic variability but also ensures gamete viability and constancy of ploidy levels (Pagliarini 2000, Hamant et al. 2006, Kumar and Singhal 2011, Brownfield and Köhler 2011). However, disruption of meiosis as well as pre- and post- meiotic events can have a severe effect on the genetic stability and viability of the gametes (Brownfield and Köhler 2011). Moreover, the degree of association and behaviour of chromosome pairing, chiasma distribution and its frequencies, disjunction of chromosomes in anaphase I/II can also provide significant insight on speciation and structural details of genomic organization and species inter-relationships (Sharma et al. 2011). Chromosome pairing, an important feature of meiosis, has often been used to infer genome relationship in hybrids and polyploid species (Grandont et al. 2013). Such studies might also contribute to the better understanding of cytological evolution of species which can be utilized for future genetic improvement and conservation of the genetic resources (Kumar and Singhal 2011). However, detailed studies on male meiosis are very much limited in the genus Curcuma except for a few reports of Ramachandran (1961), Nambiar (1979) and Puangpairote et al. (2016). The possible reason may be due to rare flowering of the plants under non-optimal environment and factors like inherent difficulty in obtaining good analyzable cytological preparations, small chromosome size and stainability (Puangpairote et al. 2016).

In this context, seven species of Curcuma: C. comosa Roxburgh, 1810, C. haritha Mangaly & M.Sabu, 1993, C. mangga Valeton & Zijp, 1917, C. montana Roxburgh, 1800, C. caesia Roxburgh, 1810, C. longa Linnaeus, 1753 and C. sylvatica Valeton, 1918, were taken up for the present investigation for analysis of meiotic pairing behaviour in order to find evidence on species inter-relationship, speciation and evolution. From our previous investigations on chromosome count, the somatic chromosome number in C. comosa, C. haritha, C. mangga and C. montana was observed to be 2n = 42 while 2n = 63 was recorded in C. caesia, C. longa and C. sylvatica (Lamo and Rao 2014, 2017).

For the present investigation, Curcuma germplasm along with their specimen voucher numbers were obtained from Indian Institute of Spices Research, Kozhikode. Flower buds were obtained from the plants growing in polyhouse conditions at the Department of Biotechnology and Bioinformatics, North-Eastern Hill University, Shillong.

Flower buds of appropriate size were fixed in freshly prepared Carnoy’s solution (1:3 glacial acetic acid: 95% ethanol) for 4 days at room temperature and stored in 70% ethanol at 4ºC. Anthers were squashed in 2% aceto-carmine solution and in some cases ferric chloride solution was used as mordant. The slides were examined and photographed using Leica DM 4000 B microscope attached to Leica CCD camera at ×1000 magnification. For meiotic analysis each preparation was determined by microscopy as well as photomicrographs. On an average 25 PMCs/species were used for detailed analysis at diplotene, diakinesis and/or metaphase I.

The terminalization coefficient was calculated using the following formula:

In the present study, seven species of Curcuma showed varying degree of chromosome association(s) viz. bivalents, multivalents and univalents. Group I species showed the prevalence of bivalent associations besides univalents and occasional quadrivalents with a near- normal meiotic behaviour. On the other hand Group II species as expected showed trivalent associations besides bivalents, univalents and quadrivalents. Similar observations were also reported by Ramachandran 1961, Nambiar 1979 and Puangpairote et al. 2016 in C. aromatica, C. decipens, C. longa, C. comosa and C. latifolia. It is interesting to note that univalents in C. montana lie in close proximity to each other at diplotene suggesting a residual attraction between homologues and their recent separation (Ghosh et al. 2016). However, in the remaining six species, the occurrence of univalents cannot be deciphered whether it is a consequence of synaptic variation or precocious separation of the chromosomes.

The present study strongly support that Curcuma is an allopolyploid complex which is evident from the low frequency of multivalent associations and in view of the fact that chromosome associations at the first meiotic division are the usual source of information concerning the type of polyploidy in a given plant (Swaminathan 1953). Allopolyploidization mechanisms involving interspecific and intergeneric hybridization, followed by chromosome doubling for obtaining a stable allopolyploid lineage, plays a pivotal role in the plant evolution (Stebbins 1971, Feldman and Levy 2005, Ozkan and Feldman 2009, De Strome and Mason 2014). Allopolyploids are characterized by a diploid-like meiotic behaviour. Male meiotic events in Curcuma species clearly signify that species differentiation is helped by polyploid events and the resultant products are yet to be stabilized in nature.

Members of the zingiber family viz. Zingiber and Mantisia exhibit varying degree of meiotic irregularities have contributed to reduce fertility and poor seed set (Ramachandran 1969, Sharma et al. 2012). This might be the probable reason for vegetative propagation by means of bulbils and rhizomes (Puangpairote et al. 2016). Likewise, Curcuma species have also adopted vegetative mode of propagation which apparently help to overcome meiotic disturbances. Furthermore, polyploidy has offered a strong evolutionary advantage to adapt to a wide range of ecological niche and better survivability than their diploid counterpart (Stebbins 1971, Grant 1971, Feldman and Levy 2005). Several studies have reported that Curcuma species with 2n = 63 (probable triploids) are geographically widespread (Ramachandran 1961, Škorničková et al. 2007) and have been slightly successful in cultivation, mainly for their productive rhizomes and competitive ability in natural environment (Puangpairote et al. 2016).

From comprehensive male meiotic investigation in seven species of Curcuma, we speculate that the speciation in Curcuma might have been affected by inter-specific crosses. We hypothesize that Curcuma species with 2n = 24 (e.g. Curcuma gracillima, etc.) might involved in hybridization events with species of related taxa belonging to the order Zingiberales having 2n = 18 (e.g. Costus speciosus) resulting in F₁ progeny with 2n = 21 (Fig. 3). Such hybridization events might be followed by natural and expected chromosome doubling giving rise to amphidiploids with 2n = 42, a somatic number more common in the genus Curcuma e.g. C. aromatica, C. mangga, C. decipens, etc. In the course of subsequent evolution, these amphidiploid species might have underwent yet another round of chromosome doubling resulting in species derivatives with 2n = 84, a presumed octoploid viz. Curcuma attenuata. Few probable triploid species of Curcuma such as C. caesia, C. longa, C. sylvatica, etc., could be possible due to inter-specific hybridization at heteroploid levels involving amphidiploids (e.g. C. aromatica, C. comosa, C. mangga, etc.) and inter-specific octoploid (e.g. C. attenuata). Our hypothesis amply gains support from cytogenetical investigation carried out in the present study, wherein male meiotic analysis of amphidiploid species viz. C. mangga showed the presence of more bivalents (94.02%) as compared to univalent or any other type of associations. On the other hand, triploid (presumed) species like C. longa showed the presence of significant number of trivalents (35.86%), a hallmark feature of triploids. However, detail meiotic data from species with 2n = 84, like C. attenuata (presumed naturally occurring octoploid) needs to be further investigated for approval of the hypothesis proposed.

Figure 3.

Proposed scheme for Curcuma speciation and diversification.

Besides the reason for considering Costus speciosus as a putative diploid parent is that there is no published literature on chromosome counts with 2n = 18 in any of the species belonging to Zingiberaceae, Hedychieae and Globba, the closely related tribes of the order Zingiberales. Moreover, Costaceae showed a close relationship with Zingiberaceae and was even previously placed as a subfamily within the family Zingiberaceae and immensely shared broad similarities in inflorescence and floral traits (Specht and Stevenson 2006). Futhermore, x = 21 is too high a basic number to be considered (Škorničková et al. 2007), therefore, we suggest that the genus Curcuma has evolved by hybridization of species with different chromosome numbers of 2n = 24 and 18, resulting in a dibasic amphidiploid species which is in complete support of Ramachandran (1961, 1969) and Nambiar (1979) findings with regard to speciation of the genus Curcuma.

The authors are thankful to the Head of the Department for providing necessary facilities. The authors are also grateful to the Director, Indian Institute of Spices Research (IISR), Kozhikode for providing the germplasm; Department of Biotechnology, Government of India for financial assistance (Ref. No. DBT-JRF/2010-11/582 dated 27^th January, 2011); DST-FIST (16-7/BTBI/FIST/10) and UGC-SAP (17-5/BTBI/UGC-SAP/2015). Sincere thanks are also due to the members of Plant Biotechnology Laboratory, Department of Biotechnology and Bioinformatics, North-Eastern Hill University, Shillong for their constant help and encouragement.