Colchicine-Enabled Genomic Doubling In Oil Palm ( Elaeis Guineensis Jacq . )

Wonder Nunekpeku (  nuswonder1@gmail.com ) Biotechnology and Nuclear Agriculture Research Institute https://orcid.org/0000-0002-9179-5421 Daniel Nyadanu Cocoa Research Institute of Ghana Richard Akromah Kwame Nkrumah University of Science and Technology College of Agriculture and Natural Resources Harry Mensah Amoatey Biotechnology and Nuclear Agriculture Research Institute Alexander Wireko Kena Kwame Nkrumah University of Science and Technology College of Agriculture and Natural Resources Abdelbagi Mukhtar Ali-Ghanim International Atomic Energy Agency Daniel Agyei-Dwarko Oil Palm Research Institute, CSIR King Manteaw Odoi Biotechnology and Nuclear Agriculture Research Institute Kenneth Ellis Danso Biotechnology and Nuclear Agriculture Research Institute

subjected to colchicine treatment at concentrations of 0, 1.26, 3.76, 6.26 and 8.76 mM for 0, 3, 6, 12 or 24 hours. The colchicine treatment produced variable physio-morphologies in the treated seedlings. Principal component analysis and agglomerative hierarchical clustering of genotype dependent treatments were carried out to provide an ordination guide. The response of oil palm seedlings to colchicine treatment showed genotypic, concentration of colchicine applied and treatment duration dependency. Ploidy level analysis of the colchicine treated seedlings revealed that polyploidization in oil palm increased signi cantly (p ≤ 0.05) with increasing colchicine concentration. Seedlings of the genotype G-131 responded more favourably to the colchicine treatment than seedlings of G-132. Optimal colchicine concentration for genome doubling in G-131 (28.9 %) and G-132 (17.8 %) was 8.76 mM applied for a period of 24 h. A population of 154 tetraploids out of 1800 seedlings of the two genotypes mutagenized has been identi ed and being raised for further evaluation and genetic exploitation.

Background
The African oil palm (Elaeis guineensis. Jacq.) is an oleaginous, erect and single-stemmed monoecious tree belonging to the family Arecaceae. The oil palm trunk is cylindrical with dense adventitious roots at the base occurring in the upper 40-60 cm depth of the soil and spread out to a radius of 3-5 m wide. The crop is naturally a diploid species (2n = 2x = 32) with a relatively small genome size of approximately 1.8 Gb (Singh et al, 2013). It is the most important oil bearing perennial tree crop cultivated in most tropical regions worldwide. In spite of its huge economic importance, it has a very narrow genetic base making its improvement using conventional techniques very di cult. All the commercial planting materials are hybrids from predominantly heterozygous pisifera and dura parents selected through reciprocal recurrent and family-individual selection approaches. (Barcelos et al, 2015;Corley and Tinker, 2003). The breeding populations used by breeders in developing these parental lines often referred to as breeding populations of restricted origins (BPROs) could be identi ed by speci c traits traceable to usually unique and small clusters of wild or unimproved ancestral palms (Barcelos et al., 2015). The genetic implications of this dependence on such breeding materials for oil palm improvement are limited available genetic diversity and a reduced rate of progress in terms of overall genetic improvement of the crop. The imperativeness of generating adequate novel materials with su cient genetic diversity in order to increase progress in oil palm improvement is therefore, very essential.. Genomic manipulations through chromosome doubling may present opportunities to broaden the genetic base and increase the diversity within the commercial oil palm species.
Though chromosome doubling can occur spontaneously, there are a number of bottle necks to its wide spread application. Firstly, the event is rare in nature and its frequencies are often extremely low. For instance, in citrus, spontaneous autotetraploids occur at a frequency of between 1-7% in seedlings of haploid apomictic seedlings (Saleh, et al., 2008). The process of chromosome doubling can come about as a result of nuclear fusion following cell division and therefore culminating in genome doubling by somatic diploidization and meiotic or gametic un-reduction leading to the formation of 2n gametes and polyploid individuals (Castillo et al., 2009). Secondly, to date in oil palm, the use of naturally occurring polyploids for the breeding of the crop has not been reported. Thus the use of induced genome doubling may be an essential and unavoidable process in the exploitation of haploids, doubled haploids or polyploids with higher ploidy levels for improvement of the crop. E cient genome doubling protocols using anti-mitotic agents are therefore, pre-requisite to successful deployment of doubled haploids and other genomic variants or polyploids in oil palm breeding programmes. Antimitotic agents are chemicals that inhibit the formation of microtubules and spindle bres during cell division at early anaphase. By their action, though duplication of chromosomes may occur during mitosis further cell division is Due to the di culty in obtaining haploids in such su cient quantity to enable practical replicated trials genome doubling protocols established using their diploids counterparts may also be useful in doubled haploid development programmes. In this manner valuable haploid materials which are di cult regenerate are not wasted and therefore preserved for subsequent doubling to produce doubled haploids. In chemical mutagenic treatment the concentration, as well as the duration of exposure or treatment are very essential in determining optimum survival and mutation frequency (Jancowicz-Cieslak and Till (2016). Developing an e cient protocol for genome doubling using colchicine is therefore essential and will complement haploid production efforts and help raise large populations of doubled haploids and also allow for development, manipulation and exploitation of polyploid genotypes for genetic improvement of oil palm. The objectives of this study were, therefore, to optimize genomic doubling in oil palm by determining optimum colchicine concentration for chromosome doubling and also outline procedures for screening for putative polyploids developed.

Results
Variables distribution and clustering of the genotypecolchicine treatment associations Normal probability plot correlation coe cients of the variables used to carry out ordination and cluster analysis indicates that the data showed normal distribution ( Fig. 2a) with very high correlation coe cients ranging from 0.863 (Surv-16) to 0.9963 (CC) (Fig. 2b). Cluster analysis performed on the genotype-colchicine treatment associations also revealed ve distinct clusters with a cophenetic correlation coe cient of 0.7341 (Fig. 3.) based on the physio-morphological traits in response to the mutagenic treatment. The rst cluster (cluster I) consisted of 3 colchicine-induced morphotypes (CIMs) which are morpho-variants of G-132 resulting from colchicine treatment of G-132 at low concentrations of 1.26 and 3.76 mM and for long and short durations of 12, 24 and 3 hours respectively. Individuals in this cluster included V2C1-12, V2C1-24 and V2C2-3.
The fourth cluster (cluster IV) comprised solely of individuals derived from the genotype G-132 treated with only distilled water as controls for durations ranging from 3-24 hours. Individuals in the fourth cluster were V2C0-3, V2C0-6, V2C0-12 and V2C0-24.
Variables with the widest range and variability and therefore contributed signi cantly to variations observed among the colchicine-mutagenized oil palm populations in order of decreasing strength include stomatal conductance (SC), chlorophyll content (CC), leaf area (LA) and plant height (PH) (Fig. 4 ).
Frequency of aberrancy in the colchicine-mutagenized oil palm population.
Seedlings with aberrant (off-type) phenotypes or morphologies such as thick and reduced leaf sizes and plant height of approximately 50% of the control were also studied and statistically signi cant, (p ≤ 0.05) variation was found between the oil palm genotypes (Fig. 5). Genotype x concentration, concentration x treatment duration and genotype x treatment duration interactions were also statistically signi cant (p ≤ 0.05). At higher colchicine concentrations (6.26 and 8.76 mM) the genotype G-131 produced signi cantly higher proportions of aberrant seedling types than G-132. No aberrant seedlings were recorded among the control seedlings and seedlings treated with 1.26 mM colchicine for 3, 6, 12 and 24 hours respectively in both genotypes. Furthermore, G-132 did not record any aberrant seedling at a lower concentration of 3.76 mM applied to the seedlings for 3 hours.

Principal component analysis
The principal components determined on the 12 physio-morphological traits in the oil palm genotypes subjected to the colchicine mutagenic treatment together with the eigen values and the proportion of the variance explained by each of the principal components are presented in Table 1. The analysis revealed 12 principal axes but only 4 which had proportion of the eigen values greater than 1% were retained. The scree plot (Fig. 6a) also showed that after the rst 4 principal components the percentage contribution of the remaining axes were negligible. Together the 4 principal components explained 99.95% of the total variation observed in the colchicine-induced response variables. The principal component 1 (PC1) explained 65.89% of the total variation while the second, third and fourth principal components, PC2, PC3 and PC4 accounted for 27.40% 5.59% and 1.09% of the total variation observed, respectively. The most variable parameters on the rst two principal axes PC1 and PC2 were SC, LA, CC and PH (Fig. 6b).
The loadings derived to determine the most relevant traits, which in uenced the 4 principal component axes are shown in Table 2. Principal axis 1 (PC1) correlated positively with stomatal conductance and chlorophyll content and negatively with leaf area. On PC2, the most relevant traits were, number of leaves per plant, leaf emission rate, plant height, proportion of aberrant seedlings, leaf area ratio, and plant height ratio. PC3 was de ned mainly by chlorophyll content, leaf area, stem girth, plant height and leaf area ratio. PC4 correlated with leaf thickness, survival of seedlings/ and frequency of seedlings with aberrant morphologies  Ploidy level determination by ow cytometry in the colchicine-mutagenized population.
Variations in chromosome doubling frequency in the oil palm seedlings treated with colchicine are presented in Fig. 7. In determining the ploidy level of the genotypes, the commercial tenera (control, untreated) with a diploid ploidy status (Figs. 8 and 9) was used as an internal reference standard. Variation observed in the chromosome doubling frequency between the genotypes was found to be statistically signi cant (p ≤ 0.05). Concentration, treatment duration as well as genotype x concentration, genotype x duration and concentration x duration interactions were also statistically signi cant (p ≤ 0.05). Genotype x concentration x duration interaction was, however, not statistically signi cant. The untreated seedlings of both genotypes, G-131 and G-132 did not record any doubling of chromosome at all durations of soaking in sterile distilled water. Nevertheless, in both genotypes, the colchicine treatment induced several morphological variations (Fig. 10) and genome doubling frequency generally increased signi cantly (p ≤ 0.05) with increasing colchicine concentration and duration of treatment. In G-131, doubling frequency for treated seedlings ranged from 0-6.7, 6.7-22.2, 8.9-28.3 and 13.6-28.9% for treatment durations of 3, 6, 12 and 24 hours respectively. In G-132 however, between 0-6.7, 0-8.9, 0-13.3 and 6.7-17.8% chromosome doubling frequency were observed for treatment durations of 3, 6, 12 and 24 hours respectively. At a relatively mild colchicine concentration of 1.26 mM, G-131 and G-132 began to experience doubling in ploidy level but at a lower frequency of 6.7% at treatment durations of 6 and 24 hours respectively. Seedlings treated with 8.76 m M colchicine for 24 hours recorded the highest doubling frequency in ploidy level in G-131 (28.9%) and G-132 (17.8%).

Discussion
The oil palm genotypes exhibited differential responses to the colchicine treatment. In both genotypes seedling survival decreased signi cantly (p ≤ 0.05) with increasing colchicine concentration. However, higher colchicine concentrations applied for longer durations exceeding 12 hours had favourable effects on chromosome doubling in the oil palm genotypes. Genome doubling is very critical for deployment of haploids in breeding programmes. Though chromosome doubling can occur spontaneously, its occurrence is rare. This makes induced genome doubling very essential in doubled haploid breeding programmes (Nelson et al., 2010). Haploids in most regeneration systems are obtained in scanty numbers and this precludes their practical use in replicated trials for establishment of optimal genome doubling rates. However, protocols developed using their diploid counter parts may have relevance in application to doubled haploid production.
The signi cant (p ≤ 0.05) variation observed in the morphology of colchicine-treated oil palm seedlings provides an indication of the effectiveness of the colchicine treatment in inducing genetic variations in oil palm. Colchicine is an anti-mitotic agent used in the induction of polyploidy in plants (Stanys et al., 2004). It acts by preventing formation of microtubules and spindle bres during cell division at anaphase (Petersen et al., 2003). Chromosomes at this stage of cell division get duplicated but because further cell division is inhibited the cell content becomes doubled resulting in polyploidy. These polyploids generally have larger cells, tissues and organs than their diploid relatives (Vainola, 2000). Contrarily, in the present study, the oil palm polyploids identi ed among the aberrant seedling types had smaller tissues and organs than their diploid counterparts (Fig. 3). Though preliminary visual selection of aberrant types based on previous description of oil palm polyploid morphology, (Madon et al., 2005) resulted in large number of off-type individuals actual polyploid frequencies determined by ow cytometry were lower. Ploidy level and genome size determinations using ow cytometry provides a more reliable and accurate estimates than any other conventional technique. This suggests that the use of morphological attributes is only useful in pre-screening and in assigning tentative polyploid status in oil palm. In a similar investigation, Nelson et al. (2010) suggested use of ordination approaches to improve the e ciency of using morphological characteristics in classifying off-types and normal genotypes of oil palm. Cluster analysis performed on the colchicine treatments-, durations-and genotype-associations revealed 5 classes with a cophenetic coe cient 0.7341 indicating high quality and reliability of using colchicineinduced morpo-variables in pre-screening.
In the present study, not only did the colchicine treatment of the oil palm genotypes provide a guide to ploidy level and genome doubling in oil palm but also produced an array of tenera tetraploids (104 from G-131 and 50 from G-132) which in themselves can be evaluated for genetic variations in important agronomic characteristics or represent a valuable genetic stock for oil palm improvement. For instance, sel ng of the tenera tetraploids will lead to generation of homozygous duras (ShShShSh) and pisiferas (shshshsh) and wide range of teneras (ShShShsh, ShShshsh and Shshshsh) with variable shell-thickness characteristics which can be exploited in breeding programmes for oil palm genetic improvement.

Conclusions
The colchicine treatment produced variable physio-morphologies in the treated seedlings. Polyploidization was dependent upon genotype, concentration of colchicine applied and treatment duration. Consequently, 28.9 5% polyploidy was observed in G-131 which was higher than 17.8% recorded for G-132. Polyploidization in oil palm increased with increasing colchicine concentration. Autopolyploidy in oil palm can therefore, be induced with colchicine treatment at optimum concentration (8.76 µM) for longer durations (≥ 24 hrs).

Methods
At the end of the treatment period, the mini sacks containing the seedlings were carefully removed from the colchicine solution using an attached metal rod into a plastic basket placed in a plastic basin. The basket and basin were immediately placed under running tap for 30 min to wash off excess colchicine from the seedlings. After washing, the mini sacks were removed, air-dried on a concrete platform at 27-

Data collection and statistical analysis
All morphometric data were taken on seedlings surviving at the nursery up to or at sixteen (16) weeks after treatment (Surv-16) with the colchicine mutagen. Ploidy level analysis was carried out at fty-four (54) weeks after treatment. Chlorophyll content (CC) was measured using the chlorophyll content meter, CCM 300 plus (Opti-Sciences, USA) and expressed in mg/g. Stomatal conductance (SC) was was determined using the SC-1 leaf porometer (Decagon Devices, Inc. USA). Leaf thickness (LT) was determined on rst fully expanded leaves using Vernier calipers. Stem girth (SG) was measured at the widest portion close to the base of seedlings using a pair of Vernier callipers and expressed in centimetres. Plant height (PH) was determined as the total length of seedling as measured from the base of the seedling to the longest leaf and expressed in centimetres. Plant height ratio was computed as ratio of seedling height of the colchicine treated seedlings to that of the control seedlings. Number of leaves (NL) referred to the total number of leaves developed by the seedling over the nursery period. Leaf emission rate (LER) referred to the number of leaves produce per unit time interval expressed in months. For each leaf, the length and the broadest width were measured. Leaf area (LA) was calculated as; where; A = leaf area (cm 2 ); L = length of leaf (cm); B = broadest width of leaf (cm); Cf = leaf area correction factor of 0.57 for single undivided leaves and 0.5 for divided leaves .
Leaf area ratio was calculated as the ratio of leave area of colchicine treated seedlings to that of the control seedlings. Seedlings with aberrant morphology were determined by visual observation of the colchicine treated seedlings and comparing them to the control seedlings. Seedling survival rate referred to the total number of surviving seedlings at 16 weeks expressed as percentage of the control. The Paleontological Statistics (PAST3) software (version 3.21) was used to carry out principal component analysis and also perform agglomerative hierarchical clustering of genotype dependent treatment units by Unweighted Paired Group Method with Arithmetic mean (UPGMA). Euclidean distance was used to estimate the similarity index.

Ploidy Level Analysis
Ploidy level of colchicine treated seedlings were analysed using the CyStain UV Precise P DNA test kit, 05-5002 (Sysmex-Partec, Germany) containing 4′,6-diamidino-2-phenylindole (DAPI) uorescent stain. Approximately 5 mg fresh leaf samples were taken from the rst fully expanded leaves and placed in a plastic petri dish (15 mm x 60 mm) containing 0.5 ml of nuclei extraction buffer and chopped with a sharp razer blade to release the DNA of samples into the solution. A 1.5 ml of staining buffer was added and mixed thoroughly with a 3.5 ml plastic dropper. The mixture was sieved through a 30 µm sieve into a 3.5 ml sample tube and incubated for 30 min in the dark. At the end of the incubation period, the C-DNA contents of the samples were analysed on a Cy owSpace ow cytometer (Partec, Germany) operating with the FloMax Software (version 2.4 ) and tted with UV light source with an internal band pass (IBP) of 455 nm. The gain value of the instrument was set at 505. Lower and upper limits were set at 80 and 999.9 respectively. The samples were run on channel FL3 UV at a speed of 0.5 µls − 1 and an acquisition time of 60 seconds (s). The commercial oil palm variety, tenera was used as an internal reference standard.  Effect of colchicine on induction of aberrant (off-type) seedlings in two oil palm genotypes G-131 and G-132.

Figure 6
Eigen values of variables used to determine the oil palm traits that in uenced the major principal components (A), and biplot of the variables the determined the rst two principal components (B).

Figure 7
Variation in frequency of chromosome doubling in two colchicine-treated oil palm genotypes. Figure 10