Prospects for improvement of Plectranthus edulis (Vatke) Agnew: A high potential food security crop

Plectranthus edulis (Vatke) Agnew is a tuber crop originated and culti­ vated in Ethiopia for food and medicinal uses. It has higher quality nutrient composition than other tuber crops. The cultivation of P. edulis is under threat because of introduction of exotic species, land use change, habitat destruction, population pressure, selection of only few local varieties by farmers and clima­ te change. Although P. edulis is a high potential food security crop, there is no any improved variety as there has been very little research on the crop. This lack of research attention including absence of improved varieties and lack of conservation strategy may lead to significant genetic erosion of this crop. Previous studies of P. edulis include documentation of indigenous knowledge, analysis of nutritional composition and the use of its starch in drug formulation, traditional cultivation practices, micropropagation, in vitro regeneration, and morphological and molecular genetic diversity. However, more extensive research in all aspects is required for its improvement, use and conservation. Therefore, it is imperative to review the research results of this crop for the research community. This article reviewed research findings of this crop and presents it in a comprehensive way so that the readers will get ample informa­ tion and can refer original research findings for the details.


Introduction
Plectranthus edulis (Vatke) Agnew (synonym. Coleus edulis) is an ancient indigenous tuber crop that occurs both as cultivated and wild spe cies in Ethiopia. Although it is known by different local names depending on the place where it is cultivated, Oromo dinich is the most popular local name. The cultivation of P. edulis is restricted to Ethiopia. It is mainly cul tivated in the south and south western parts of the country at altitudes ranging between 1880 and 2200 m a.s.l. as food, and sometimes medicine (Demissie, 1988;Taye et al., 2007;Megersa, 2010). P. edulis is a dicotyle donous plant that grows up to a height of 1.5 m (Taye, 2008).
P. edulis is a high energy food crop and the tubers contain substantial amounts of micro and macronutrients. When compared to Irish potato (Solanum tuberosum), P. edulis has higher food energy and contains twice the amount of fat and cal cium (EHNRI, 1997). It contains a similar amount of protein to that of Irish potato and almost twice that of sweet potato (Ipomoea batatas) after cooking (EHNRI, 1997).
The cultivation of P. edulis is under threat because of a decline in diversified local food production systems as a result of introduction of exotic species, land use change, habitat destruction, population pressure, selection of only a few local varieties by far mers, farming system intensification and climate change (Smolders, 2006;FAO, 2010). The decline in cultivation of the crop may result in erosion of the genetic base preventing the use of distinctive useful traits that are used for crop adaptation and improve ment (Padulosi et al., 2002;IBC, 2005).
There is enormous untapped potential in Ethiopia to exploit the rich and diverse plant genetic resour ces of underutilized root and tuber crops including P. edulis. In spite of this potential, very little research has been done to improve the productivity of P. edulis (IBC, 2007). Currently, however, P. edulis is reali zed as a high potential food security crop that is attracting research attention.
The poor storability of tubers is found to be the major production constraint of P. edulis. It is propa gated by vegetative methods so that it is susceptible to diseases that are transmitted from one generation to the next through infected planting material (Taye, 2008). This problem can be solved by integration of new technologies including biotechnological approa ches. Documentation of information on indigenous knowledge, uses, and distribution of this crop is essential for its conservation, further utilization, and improvement (Mathenge, 1995). It is also important to initiate research on seed production of P. edulis. So far, there is no any improved variety of P. edulis. The objective of this article is to review some aspects of this crop, with particular regard to biotechnologi cal approaches and present it in a comprehensive way so that the readers will get summarized informa tion to design research or policy on conservation and improvement of the crop.

Production practices
Farmers select seed tubers from the previous har vest or they buy from the market. Most farmers plow the land more than three times before planting and mix farmyard manure with the soil (Mekbib and Weibull, 2012). Farmers do not use chemical fertili zers and other commercial inputs for production of P. edulis. The seed tubers are planted from March to April as there is brief rain during this time. Farmers commonly slice a single tuber into pieces and plant the pieces because they claim planting tuber pieces increases tuber yield (Taye et al., 2007;Mekbib and Weibull, 2012). Research findings also showed that planting tuber pieces results in higher yields than whole tubers (Taye, 2008).
In southern Ethiopia, almost 50% of the farmers plant the tubers with their sprouts whereas the others plant after removing the sprouts (Taye, 2008). Most P. edulis growing farmers remove the shoot apex with one or two pairs of leaves from the main stem and branches to increase stem number.
Research results show that this practice significantly increases the number of stems and enhances faster stolen formation. Most farmers commonly build up the soil around the stem to enhance plant growth and increase tuber yield. The absence of positive cor relation between frequency of removing the shoot apex and amount of the soil built up around the stem was reported, although both activities serve the same purpose (Taye, 2008).
Weeding is carried out at least three times at dif ferent growing stages of P. edulis. Following harvest of P. edulis, farmers plant other crops, and the yields of crops that follow P. edulis in the rotation are high. The reason for increase in yield could be due to the slowly released nutrients into the soil and improved soil structures as the result of the organic manure applied in previous cropping seasons ( Mekbib and Weibull, 2012). Depending on the type of variety, P. edulis is harvested from six to eight months after planting. Although the cultivation is mainly perfor med by male farmers, female farmers also play important roles in selecting tubers for different end use criteria related to the household food require ments (Mekbib and Weibull, 2012).
Farmers used to maintain several local varieties of P. edulis for different purposes. However, the num ber of local varieties preserved by farmers has been declining. So far, the total number of local varieties all over the country is not known. However, there are at least six local varieties of P. edulis that grow in spe cific areas as study results on three kebeles (the smallest administrative units in Ethiopia), namely Deleboatewaro, Delebowogene and Koketemare chare of Sodo Zuria district, South Ethiopia have shown (Mekbib and Weibull, 2012). Growth habits, taste, tuber skin color, days to maturity, drought tolerance, storage period and marketability are the main characteristics of the plant used by farmers to identify local varieties (Mekbib and Weibull, 2012). Distribution and seed exchange of local varieties by farmers play key roles in conservation and utilization of local varieties (de Boef, 2008). Younger farmers mostly purchase the planting material from markets whereas older farmers (older than fifty years) use their own saved seed tubers from the previous har vest. Older farmers have much richer indigenous knowledge than younger farmers about the cultural value of the crop, indicating a limitation in the docu mentation of knowledge on the management. These knowledge gaps may lead to cultivation of only few local varieties (Negri, 2003;Mekbib and Weibull, 2012). This calls for more research on documentation of local knowledge and transferring this knowledge to the young farmers.
One of the major challenges of P. edulis produc tion is seed tuber storage. The tuber is stored in the field where it is produced until used for the required purpose. This storage method results in significant tuber yield loss due to excessive heat especially during dry seasons. The land where the tubers are stored remains idle until all the tubers are removed. Therefore, as farmers do not want their land to be idle, most of them often obtain seed tubers from other sources during planting season (Mekbib and Weibull, 2012). Farmers cover P. edulis fields with mulching materials to protect the tubers from direct sunlight until the next planting season. After the pre paration of the land for planting, tubers are transfer red to the pits for a day to facilitate the germination and performance of seedlings. Most seed tubers are obtained from farmers who have enough land to pro duce P. edulis as they produce more than a hou sehold requirement and save the surplus for sale (Negri, 2003). Although different cultural practices have been designed to overcome the production con straints of P. edulis, very little research has been done to tackle the problems. Another important con straint is shortage of land for planting (Mekbib and Weibull, 2012). This can be solved by amending land usage policy.
Some efforts have been made to evaluate perfor mance of P. edulis with regard to yield and yield rela ted characters and reports show different perfor mance of the crop at different environments (Taye et al., 2013). Tuber fresh weights of 45564933 g m 2 , or 4549 Mg ha 1 were obtained at experimental sites, which is higher than yields produced by farmers (7401480 g m 2 ). This higher experimental yields compared to farmers' yields is because farmers har vest the tubers early Late emergence, slow develop ment of the canopy, full coverage of the ground in a short period of time and decline in soil cover during senescence caused poor radiation interception. Radiation interception of a crop can be enhanced by increasing ground cover by the canopy. This could be achieved by planting large seed tuber pieces (Wiersema and Cabello, 1986;Lommen and Struik, 1994), by planting more tuber pieces per area (Spitters, 1990) or by using traditional practices that enhance the canopy development (Taye et al., 2012).

Nutritional content
P. edulis contains substantial amounts of nutrients. Moisture content of the tuber dry matter ranges from 14.1% to 17.5%, whereas the protein content ranges from 0.70% to 1.76% (Hellemans et al., 2017). These values are higher than cassava tuber, which contains 0.3% protein and 0.1% to 0.3% starch (Alvani, 2011;Waterschoot et al., 2015). Another study reported 11.2% moisture, 0.14% ash, 0.21% lipid, 0.43% protein, and 99.22% starch con tent on a dry weight basis (Assefa, 2015) (Table 1). Samples collected from different regions of Ethiopia showed variation in fiber content based on their geographic location. Starch extracted from tuber samples collected from Wolaita zone (Chenqoua, Inuka and Lofua areas) showed signifi cantly lower fiber content than tuber samples collec ted from Arjo white, Chencha and Jarmet. Moreover, the samples collected from Arjo white, Chencha and Jarmet contain higher fiber content than cassava or potato. Whether this variation is due to the environ ment or genetic variation needs further investigation. P. edulis contains amylase ranging from 14.2% to 23.9% (Waterschoot et al., 2015).
Tuber samples collected from Abay Chomen district of Oromia Region, Ethiopia, showed variation in mineral content may be due to genotype or envi ronmental variation. P. edulis is found to be rich in potassium followed by magnesium, calcium, zinc, manganese, and copper. Mineral content of P. edulis is presented in Table 2. The trace metals (chromium and nickel) and the heavy metals (lead and cadmium) were not detected, indicating P. edulis is safe with regard to the toxic heavy metal health concerns (Lema, 2016). The phosphorus content of most P. edulis samples is twice as much as in Irish potato. High phosphorus content is linked to resistant starch, paste viscosity and gel strength (Lu et al., 2012). Higher phosphorus content is responsible for the gra nular structure, and especially for the crystalline region to be less rigid and therefore enabling absorp tion of more water (Lin et al., 2013). High phospho rus content was also found to impart high viscosity to the starch and increase gel strength. This indicates P. edulis starch can be used in food processing that requires high gel strength and can provide resistant starch in functional food preparation (Hellemans et al., 2017). However, further research should be con ducted to confirm this finding. P. edulis is found to be richer in most of the mineral elements than Irish potato.

Starch physicochemical properties
Scanning Electron Microscopy (SEM) of the starch granules showed elliptical and some oval shapes. The starch showed a normal monomodal granule size distribution. XRay Diffraction pattern of the starch showed typical Btype with a distinctive peak (Assefa, 2015). On the contrary, other authors reported distinct bimodal distribution pattern of starch granu lar characteristics, which indicates the presence of small (Btype) and large (Atype) granules (Hellemans et al., 2017). The shape of Btype granules is spheri cal whereas that of Atype is elliptical when visuali zed by the cryoSEM. For starch extracted from P. edulis tuber, such bimodal distribution is not com mon. Starch extracted from Irish potato and cassava shows normal distribution pattern (Waterschoot et al., 2015). Therefore, P. edulis can be used in food industry and as an alternative for potato starch, which is less suitable to be used in pasteurized foods.
Typical water absorption (WA) pattern, swelling power and relative humidity of P. edulis is reported to be higher than Irish potato starch but its solubility at all studied temperatures is lower than Irish potato (Assefa, 2015). The physicochemical analysis of the starch showed 1.08 g/g WA of P. edulis (Hellemans et al., 2017). This means that the starch granules absorb on average of 1.08 times their own weight after it is completely hydrated. Only a small difference was observed in WA among P. edulis starch samples of dif ferent accessions. The high water binding ability of the starch could reduce the stickiness of dough, increase moistness, improve handling and soften the texture of baked products, which makes P. edulis starches more suitable for use in baked goods (Taggart, 2004).
Differential Scanning Calorimeter (DSC) thermo grams of P. edulis starch showed higher onset tempe rature (To) of 69.2°C, peak temperature (Tp) of 74.3°C and ending temperature (Te) of 83.3°C than the starch of Irish potato (Assefa, 2015). Similarly, the pasting properties showed P. edulis starch has  (Bryan, 1983). These problems might be solved by using the benefits of modern plant biotechnology such as plant tissue culture that can supply a large number of diseasefree planting materials to growers who continually acquire it.

Micropropagation and in vitro shoot regeneration
Increased use of rapid multiplication techniques enables production of large amounts of pathogen free planting material. Most rapid multiplication techniques in root and tuber crops involve the use of aerial portions of the plant so that contact with soil and tuber is broken and most nonsystemic patho gens can be eliminated (Bryan, 1983). Meristem cul ture is the most popular tissue culture technique for production of disease free planting materials. It can be used alone or in combination with thermotherapy to eliminate a number of viruses and bacteria (Kartha and Gamborg, 1975). Meristem culture technique eli minates systemic pathogens found in the mother stock plant.
The first micropropagation protocol of P. edulis was developed from meristem culture (Tsegaw and Feyissa, 2014) followed by in vitro shoot regenera tion from leaf derived callus (Aschale and Feyissa, 2019). Shoots were initiated from meristem explants using Murashige and Skoog (MS) (Murashige and Skoog, 1962) medium. The MS medium was supple mented with 1.0 mg/l gibberellin (GA 3 ) and 0.1 mg/l anaphthalene acetic acid (NAA) in combination with different concentrations of benzylaminopurine (BAP) (Tsegaw and Feyissa, 2014). According to the authors, MS medium containing the aforementioned hormones was proved to be the best in terms of shoot initiation from meristem for mother plants col lected from two locations, Holeta and Wolaita. The shoots were transferred to multiplication medium and the highest mean shoot number per explant (7.2) was obtained on medium containing 1.0 mg/l kinetin in combination with 0.1 mg/l NAA for shoots derived from mother plants collected from Holeta while, 6.2 shoots per explant was obtained on medium containing 3.0 mg/l kinetin in combination with 0.05 mg/l NAA for shoots derived from mother plants collected from Wolaita. The effect of solid and liquid media on in vitro axillary shoot proliferation was investigated by culturing shoots on MS medium containing different concentrations of BAP in combi nation with thidiazuron (TDZ) (Yimam, 2013). The highest mean shoot number per explant, 5.85 and 6.07, were obtained on solid and liquid media, higher pasting temperature than Irish potato (Hellemans et al., 2017). A rapid increase in viscosity occurred at Tp range of 70.7 to 74.0°C compared to Irish potato starch (52.2°C).
The disintegrating abilities of starch of P. edulis and Irish potato were compared in paracetamol tablet formulations prepared by wet granulation method (Assefa, 2015). The granules were prepared with the starches at different disintegrant concentra tions and characterized for particle size distribution and flow properties. The prepared tablets showed excellent flow property as manifested by the weight uniformity. After studying the crushing strength, fria bility, disintegration time, and dissolution rate of the tablets using standard methods, the results showed that P. edulis starch favorably competed with potato starch as a disintegrant in the paracetamol tablet for mulations.
In other recent study, carboxymethylated P. edulis starch was evaluated as a suspending agent in metro nidazole benzoate suspensions by comparing with sodium carboxymethyl cellulose (NaCMC) (Brhane, 2020). The author reported viscosities of the formu lations prepared with carboxymethylated P. edulis starch were significantly lower than that of NaCMC. The flowability of the suspensions was in the order of carboxymethylated P. edulis starch (CMPS) greater than NaCMC. At 1% concentration, carboxymethyla ted P. edulis starch resulted in significantly higher sedimentation volume than NaCMC. Potassium dihy drogen phosphate (KH 2 PO4), which is employed as a flocculating agent, significantly increased the sedi mentation volume of the suspensions prepared with carboxymethylated P. edulis starch and NaCMC. The redispersibilities of CMPS were found to be better than NaCMC. The author pointed out that all suspen sions resulted in a release of greater than 85% of drug within 1 h. All suspension formulations were found to be stable (Brhane, 2020). The author con cluded carboxymethylated P. edulis starch can be used as alternative suspending agent.

Prospects of biotechnology for improvement of P. edulis
In addition to the shortage of seed tubers and the poor storability of the tubers, systemic diseases, viru ses, viroids and mycoplasma as well as several patho genic bacteria are the most devastating root and tuber crops including P. edulis, in terms of yield loss liquid medium containing 0.1 mg/l IAA. The highest mean root number per shoot of 10.15 on MS medium supplemented with 2.0 mg/l IBA was also reported (Aschale and Feyissa, 2019).
All in vitro regenerated plants of Holeta origin and 96% of Wolaita origin survived after one month of acclimatization in a greenhouse. Similar study repor ted 100% and 82.85% survival of plantlets derived from solid and liquid media, respectively, after accli matization in greenhouse (Yimam, 2013). All in vitro rooted and 78 % ex vitro rooted microshoots survi ved in the greenhouse (Aschale and Feyissa, 2019). The development of these protocols for micropropa gation and in vitro regeneration of shoots from callus is the prerequisite for further advanced biotechno logy research such as genetic transformation and genome editing of this crop.

Genetic diversity of P. edulis
For improvement of any crop, the existence of adequate genetic diversity is imperative. This could be achieved if and only if there is appropriate conser vation strategy. Selection and cultivation of only few local varieties of P. edulis and its current replacement by other tuber crops has been causing serious gene tic erosion of the local gene pool of this crop.

Genetic diversity analyses using morphological traits
There are several studies on morphological gene tic diversity of P. edulis. Genetic diversity of 36 acces sions of P. edulis using 16 morphological traits was assessed and highly significant variation among the accessions for all the analyzed traits except length of tuber was reported (Garedew et al., 2013). In another study, 20 P. edulis accessions that were assessed using 29 morphological traits showed 75.7% variance among the accessions (Mekbib, 2007). This study showed that there are accessions collected from different regions of the country but showed genetic similarity, and some of those collected from the same geographic area showed different genetic background. This indicates geographic diversity does not necessarily represent genetic diversity. Similarly, genetic diversity analysis of 20 accessions using 13 traits revealed significant variation of all traits except flower length and leaf width (Soresa, 2017). Recent extensive genetic diversity study of 174 accessions at three environments based on 12 qualitative and 16 quantitative traits also showed a highly significant respectively, containing 1.5 mg/l BAP in combination with 0.5 mg/l TDZ.
Developing in vitro regeneration protocol through callus phase or somatic embryogenesis is a prerequi site to genetically improve a crop through biotechno logical approaches including genetic engineering, genome editing and creating somaclonal variants. Recently, in vitro regeneration protocol of P. edulis have been developed by using leaf explants (Aschale and Feyissa, 2019). The highest percentage of callus induction (100%) was obtained on MS medium sup plemented with 1.5 mg/l NAA in combination with 1.0 mg/l BAP and 2.0 mg/l NAA in combination with 0.5 mg/l BAP. The highest shoot regeneration per centage (46.6%) from callus was obtained on MS medium containing 0.1 mg/l BAP in combination with 0.5 mg/l TDZ. The highest mean shoot number per callus (1.66) was obtained on medium containing 1.5 mg/l BAP in combination with 1.0 mg/l TDZ. The highest mean shoot length (0.63 cm) was obtained on medium supplemented with 0.5 mg/l BAP in com bination with 0.1 mg/l TDZ. The highest mean shoot number per explant (58.68) was obtained on medium supplemented with 0.5 mg/l BAP in combination with 0.4 mg/l GA3.

Rooting and acclimatization
Shoots cultured on half salt strength MS medium rooted better than those cultured on full salt strength (Tsegaw and Feyissa, 2014). Shoots derived from explants collected from Holeta resulted in the highest percentage of rooting (100%) and 3.12 mean root number per shoot whereas shoots whose explants were collected from Wolaita resulted in 76.7% rooting and 2.2 mean root number per explant. Microshoots were also directly planted in a greenhouse for rooting and acclimatization by bypas sing the in vitro rooting stage and the overall best result in rooting was obtained in ex vitro rooting con dition. It was promising in cost reduction as it bypas ses the in vitro rooting stage. Reducing the cost of in vitro production is a key for increasing the applica tion of the method. Rooting of microshoots in vitro is expensive and can even double the price of the plan tlets (Zimmerman, 1988;De Klerk, 2002). In another study, shoots were cultured on half strength MS medium containing different concentrations of indole acetic acid (IAA) and indole butryic acid (IBA) for roo ting (Yimam, 2013). The highest mean root number per shoot (10.55) was obtained on solid medium and 13.11 mean root number per shoot was obtained in variation among the populations and nonsignificant environmentpopulation interaction for most of the quantitative traits (Gadissa et al., 2020). The first six principal axes of principal components analysis accounted for 77% of the total variation. The popula tions were grouped into four clusters but there was no strong groupings based on geographical locations from where the genotypes were collected. This indi cates the historical or contemporary gene flow, parti cularly tuber exchange among the different geo graphical regions. Although there are reports of rela tively high genetic diversity, still there is a need for more collections and evaluation over multiple loca tions and seasons for wellrefined genetic diversity estimation. However, as the analyzed traits are affec ted by the environment, the morphological genetic diversity results should be complemented by molecu lar genetic diversity studies.

Genetic diversity analyses at molecular level
Assessment of molecular genetic diversity of P. edulis was conducted by using inter simple sequence repeat (ISSR) markers for the first time in 2015 (Shiferaw, 2015). The Shannon information index ranged from 0.25 to 0.39 with overall index of 0.48. Nei's gene diversity ranged from 0.18 to 0.30 with overall mean diversity of 0.33, and 74.23% within population variation. Analysis of Molecular Variance (AMOVA) showed within population diversity range of 41.30% to 65.22%. Another study using the same marker exhibited overall percent polymorphism of 95%, 0.62 Shannon information index and 0.40 Nei's gene diversity (Gebrehiwet et al., 2019). More exten sive study of genetic diversity assessment was con ducted by using expressed sequence tag simple sequence repeats (ESTSSRs) derived from P. barbatus (Gadissa et al., 2018). The author identified 128 alleles in 12 populations that consisted of 287 indivi dual plants. The results showed gene diversity index range of 0.31 to 0.39 with overall mean of 0.35. Among the 12 populations, Wenbera, Awi and Wolaita populations showed the highest genetic diversity, and hence these populations can be consi dered for in situ conservation and identification of genotypes with traits of interest that can be used in breeding programs. AMOVA showed low population differentiation with only 3% of the total variation accounting for variation among populations. STRUC TURE and cluster analyses did not group the popula tions into distinct clusters, which may be attributed to historical and contemporary gene flow and also the reproductive biology of the crop.

Conclusions
P. edulis is a high potential food security crop con taining substantial amounts of nutrients. However, there is no any improved variety of the crop, and cur rently its cultivation is declining. Shortage of tuber seed is the major problem of this crop. Recently, the farmers are turning to produce other cash crops that generate more income. With advances in plant research, it is important to integrate both conventio nal and modern technologies to improve this crop. Unless urgent attention is given to such neglected and underutilized crops, these crops are extremely subjected to genetic erosion as the farmers shy away from producing them. Using biotechnological resear ch tools such as plant tissue culture, genetic enginee ring, genome editing and plant molecular breeding might be a future hope for improvement of this crop, although the international funding opportunities for such orphan crops are highly limited. In vitro propa gation and in vitro regeneration protocols from diffe rent explants have already been developed as reviewed in this article and these protocols can be used for improvement of P. edulis. In addition, infor mation from the genetic diversity results reported so far could be used as a baseline for more research in the area and improve this crop using genomic tools including marker assisted breeding.