Enhanced production of falcarinol-type polyacetylenes in hairy roots of cultivated carrot (Daucus carota subsp. sativus)

Polyacetylenes (PAs) are a large group of bioactive phytochemicals, which are primarily produced by higher plants of the families Apiaceae and Araliaceae. Especially aliphatic C 17 -polyacetylenes of the falcarinol-type such as falcarinol (FaOH) and falcarindiol (FaDOH) are known for their numerous biological functions including positive effects on human health. In this study we investigate the potential of carrot hairy root (HR) cultures for accumulation of PAs in vitro. Three individual plants of seven differently colored carrot cultivars were used for the development of HR cultures by transformation of root discs with the wild-type Rhizobium rhizogenes strain 15834. A total of 51 individual HR lines were established and quantitatively analysed by HPLC/DAD together with root, petiole and leaf tissue samples for their FaOH and FaDOH levels. Among the ve tissues sampled from the donor plants, root periderm samples generally exhibited the highest PA levels with FaDOH as prevailing PA and large differences between cultivars. Compared with the periderm, FaOH levels were highly increased in HRs of all investigated carrot cultivars, whereas the FaDOH levels were not signicantly increased. Considering the low to moderate PA concentration in root and leaf tissues of the orange cultivars there was an up to 30-fold increase of the FaOH concentration in HRs derived from orange genotypes. In this study, we demonstrate, that carrot HRs are able to produce large amounts of PAs. Hence, we suggest that HRs might be a suited tool to assess the function of candidate genes involved in the biosynthesis of major PAs.


Introduction
Polyacetylenes (PAs) are a large group of bioactive phytochemicals, which are primarily produced by higher plants of the families Apiaceae and Araliaceae both belonging to the order Apiales. They are also common in Asteraceae family members (Dawid et al. 2015). Among the more than 1.400 known PAs (Christensen and Brandt 2006), a subset of 12 structurally related PAs were isolated from D. carota (Dawid et al. 2015). These PAs are found in common vegetables and herbs of the Apiaceae family such as carrots, parsnip, fennel, celery, and parsley, with the C 17 -PAs falcarinol (FaOH) and falcarindiol (FaDOH) as the main PAs Hoffmann 2003, Kidmose et al. 2004). Both FaOH and FaDOH are known to have a major impact on bitter off-taste in carrots, which is often the reason for consumer rejection of carrots or carrot products (Schmiech et al. 2008;Kreutzmann et al. 2008). On the other hand, several in vitro and in vivo studies showed that in particular the aliphatic C 17 -PAs of the falcarinol-type exhibit potent anti-microbial, anti-in ammatory, and anti-cancer effects (Zidorn et al. 2005;Christensen and Brandt 2006;Christensen 2011;Dawid et al. 2015;Christensen 2020). For instance, falcarinol-type PAs may function as inhibitors of the breast cancer resistance protein BCRP/ABCG2, indicating their prospective use as agents for cancer chemotherapy (Tan et al. 2014). Rat feeding experiments with puri ed FaOH and FaDOH suggest preventing effects of these PAs on the development of colorectal cancer (Kobaek-Larsen et al. 2019). Despite their numerous biological functions the biosynthesis of PAs and its genetic basis are poorly understood. An e cient test system based on fast growing transformed root cultures might support the functional characterization of gene candidates involved in PA production.
Hairy root (HR) cultures obtained after transformation with the soil bacterium Rhizobium rhizogenes (Rrhi), formerly known as Agrobacterium rhizogenes, have often been used for tissue culturebased production of specialized plant metabolites (Malik 2018). Especially medicinal plants were transformed in the past with wild-type Rrhi strains to produce pharmaceutically relevant phytochemicals (Guillon et al. 2008). Since the advent of genome editing techniques such as CRISPR/Cas9 HR-based transformation systems are evolving to an e cient research tool to elucidate physiological processes and biosynthetic pathways (Shi et al. 2021). Examples for genome editing approaches based on HRs were reported for several plant species such as Salvia (Li et al. 2017), Eucalyptus (Dai et al. 2020), and Medicago (Zhang et al. 2020).
The cultivated carrot (Daucus carota ssp. sativus) is one of the most important root vegetable crops grown worldwide. Carrots provide not only one of the richest sources of the vitamin A precursor βcarotene, but contain many other bioactive and putatively health-relevant phytochemicals including anthocyanins, phenylpropanoids, terpenoids, and the above mentioned PAs. Several Apiacae species were investigated in the past for production of plant natural products in HR culture systems (Baranski 2008). Starting with pioneer work in early 1980s (Chilton et al. 1982;Willmitzer et al. 1982) several HRbased studies were reported for D. carota (Kim and Yoo 1996;Ridgway et al. 2004;Baranski et al. 2006). Recently, HR cultures derived from dark-purple carrots were shown to be a suited platform for research on the in vitro production of anthocyanins and phenolic compounds (Barba-Espin et al. 2020).
To our knowledge, in carrots the production of PAs in HR cultures has not been described yet.
In this study we use a Rrhi-based transformation system for generation of HRs derived from orange, yellow and purple carrots and elucidate their potential to produce su cient amounts of two major PAs, FaOH and FaDOH. We show that especially FaOH is strongly increased in HRs of orange carrots compared to peridermal root tissue and suggest that carrot HR cultures are well suited for functional studies of genes involved in carrot PA biosynthesis.
Afterwards, roots were latitudinally halved and an 8 mm thick slice was removed from each root half. Using a scalpel phloem and xylem samples were prepared from both slices. The remaining root material was sterilized and used for transformation. Petiole samples were obtained by cutting a 5-cm section from the middle of three petioles. Leaf tissue was sampled from younger lea ets. Plant material was shock frozen in liquid nitrogen immediately after harvest and stored at -80°C until freeze-drying.

Transformation of carrot with Rhizobium rhizogenes
Pieces of tap roots remaining after PA sampling were surface-sterilized via immersion in 70% ethanol for 5 min followed by a 20 min treatment with 4% sodium hypochlorite supplemented with 0.1% Tween 20.
Each carrot root was sliced into 0.5 -0.8 cm thick discs and dead tissue of the outer edge of the discs was removed. The apical (shoot-averted) side was labeled with a sterile toothpick, and the carrot discs were placed (apical side up) onto water agar (2% agar in water; Duchefa, Haarlem, Netherlands). Wildtype Rrhi strain ATCC 15834 (Lippincott et al. 1973) was used to grow bacterial inocula in dark conditions at 28°C for about 24 h in liquid CPY medium (bacto-yeast extract 0.1%, peptone 0.5%, saccharose 0.5%, bacto-agar 20 g L -1 , MgSO 4 x7H 2 O 200 mM) on a rotary shaker (180 rpm). About 100 µl of the bacterial suspension (OD 600 of 0.6 -0.8) was spread along the cambial ring on the apical side of the root discs. After co-cultivation for seven days on 2% water agar medium the carrot discs were placed for up to ve weeks on 2% water agar supplemented with 200 mg L -1 cefotaxime (Duchefa, Haarlem, Netherlands) and 100 mg L -1 carbenicillin (Roth, Karlsruhe, Germany) to eliminate the bacteria. HRs started to regenerate generally after two weeks and were harvested weekly alongside with a small piece of original root tissue. Regenerated HRs were placed in small petridishes (6 cm diameter) on hormonefree half-concentrated MS (Murashige & Skoog) medium with 200 mg L -1 cefotaxime and 100 mg L -1 carbenicillin (½ MS-CC) and cultured in the dark at 22 o C. HR cultures developing from single excised roots that originated unambigously from different regions of the root disc were considered to have an independent origin of regeneration. They were placed in a small single petridish and were designated as 'hairy-root line' (HR line). About four weeks later the propagating HRs were subcultured on the same culture medium. After further one to two subcultures in small plates, the HRs were placed on lter paper located on the surface of a ½ MS culture medium without antibiotics in 9 cm petri dishes. Four weeks later the HR lines were harvested and frozen immediately at -80 o C until chemical analysis.
In order to verify the transgenic character of the HRs and to prove that they were free of contaminating Rrhi bacteria we performed PCR analyses. For DNA isolation, 50 -100 mg fresh hairy roots were snap frozen in liquid N 2 and grinded using a swing mill to extract total genomic DNA with the innuPREP Plant DNA kit (Analytik Jena, Jena, Germany). The DNA content was measured via NanoDrop device (Thermo Fisher). To check the transgenic status of the HR lines, we used PCR primer for the rolC gene from the Ri plasmid of Rrhi, which are diagnostic for the integration of the T-DNA into the plant genome (for primer information and PCR conditions, see Medina-Bolivar et al. 2007). To prove that the presence of the rolC gene is due to its integration into the genome of D. carota and not due to contamination with Rrhi we performed PCR analysis with primer for the Rhizobium virD2 gene which is not transmitted into the plant genome (Medina-Bolivar et al. 2007).

Results And Discussion
Establishment of carrot hairy root lines Hairy root induction in seven differently coloured D. carota cultivars was accomplished by co-culture of carrot root disks and Rrhi wild strain 15834. Two weeks after inoculation rst HRs protruded from the upper surface of the root disks. HR lines were established from single excised roots and, after selection for su cient growth, used for PA analysis. To prove the presence of the T-region of the Ri plasmid and the integration of the rol (root loci) genes into the genome of D. carota we performed PCR analysis with rolC primer. The transgenic character of the HR lines was veri ed, and the virD2 fragment indicating bacterial contamination was not detected in any HR line analyzed (PCR results not shown). A single transformation experiment based on three individual tap roots of each carrot cultivar used for PA analysis resulted in a total number of 51 well growing HR lines at the end of the propagation process (Table 1). Cultivars DP, YS, and LR totally yielded the desired three lines per plant (totally 9), and for the other cultivars the number of HR lines varied from 8 in RO to 4 in AN (Table 1). For graphical presentation and statistical evaluation of PA data from different plant tissues and HRs (Fig. 1) only the cultivars DP, YS, LR and RO were considered, for which HRs from all three original plants were available.
Rrhi strain ATCC 15834 proved to be highly effective for transformation of D. carota. This wild strain, used rst for carrot transformation by Willmitzer et al. (1982), has shown strong virulence in many plant studies. Among the about 100 plant species, which have been reported for a subsequent plant regeneration from HRs carrying a wild type Ri plasmid, about 20 were transformed by the 15834 strain (Desmet et al. 2020). John et al. (2017) examined four different Rrhi strains, of which strain 15834 showed signi cantly higher transformation e ciency and was capable of inducing HRs from different explants of Achyranthes aspera. Transformation of 'Indian ginseng' (Withania somnifera) was only successful with strain 15834 (Saravanakumar et al. 2012).

Polyacetylene contents and distribution patterns in carrot donor plants
Two main PAs FaOH and FaDOH were quanti ed in carrot plant tissues and HRs (Fig. 1). Among the ve leaf and root tissues we detected the highest concentration of both compounds in the periderm samples and the lowest in the leaf and xylem tissues, but there were large differences among the cultivars. Peridermal tissue of the purple cultivar DP accumulated approximately 600 µg/g DW (dry weight) FaOH, but in the orange cultivars LR and RO signi cantly lower FaOH levels of approximately 50 and 100 µg/g DW were registered, respectively. In DP the peridermal PA levels were signi cantly enhanced in comparison to the other plant samples (Fig. 1). However, due to the large variation of the PA data among the samples from individual plants most other differences were not signi cant. With regard to FaDOH, the purple cultivar DP also showed highest concentration in the periderm (1200 µg/g) whereas the peridermal FaDOH levels of the orange cultivars LR and RO were in the range of 450 -650 µg/g DW (Fig. 1). The highest FaDOH contents in the leaf samples were measured in DP with about 100 µg/g and in YS with 240 µg/g. PA accumulation patterns are in uenced by numerous factors such as genotype, developmental stage, and abiotic and biotic stress factors (Dawid et al. 2015). Besides, the PA distribution among the different plant organs and within the different root segments can vary considerably. Since the major aim of this study was not the analytical dissection of carrot plants, but to determine the potential of PA production in HRs, we analysed three selected individuals of each cultivar for HR induction and used leaf and root samples taken from each donor plant before Rrhi transformation as reference. Nevertheless, it was evident, that in all cultivars the periderm samples displayed the highest FaDOH levels. According to Czepa and Hofmann (2004) the most abundant PA in cultivated orange carrots is FaDOH, and the highest total PA levels were found in the periderm tissue (Baranska and Schulz 2005;Baranska et al. 2005). Busta et al. (2018) analysed tissue-speci c accumulation of PAs in the orange cultivar 'Danvers' and also measured the highest total PA level in the peridermal tissues, with FaDOH as the dominating PA.
With regard to FaOH, the purple cultivars DP and AN (data nor shown for AN) appeared to accumulate considerably higher contents in the periderm compared with the other tissue samples. One reason for this nding might be the ancestry of purple carrots. It is generally recognized, that the origin of the cultivated carrot was in Central Asia (Afghanistan), where purple-rooted wild carrots were domesticated and spread westward more than 1.000 years ago (Simon et al. 2008). Orange carrots occurred in the 15 th and 16 th centuries in Europe (Stolarczyk and Janick 2011), which indicates a secondary domestication process. Purple and red carrots persisted in Asia and the Middle East, and it can be assumed that current purple carrot cultivars are genetically closer to early domesticated purple carrots than to the modern orange forms, which were selected for less bitterness and likely for less PA contents. The results of our study indicate, that eld-grown carrots with relatively high PA contents, such as for instance purple cultivars like DP, might be used in future for the large-scale production of pharmaceutically relevant falcarinol-type PAs. Carrot HR cultures might be a suited alternative plant source, and bioreactor techniques seem to be a possible way to produce large quantities of PAs. However, for latter approach further research on the consistency and kinetics of PA production in HRs would be necessary. HR cultures ave been used in numerous plant species for the synthesis of a wide spectrum of phytochemicals (Malik 2018;Gutierrez-Valdes et al. 2020), but few reports are available for falcarinol-type PA production in HRs. FaOH was among the main components of oil samples from HR cultures of the Apiaceae species Levisticum o cinale (Santos et al. 2005), and Washida et al. (2003) reported the production of FaOH and FaDOH in HRs of Panax hybrids. HRs of Panax ginseng were used to investigate the natural formation of FaOH and FaDOH by 13 C-labeling experiments (Knispel et al. 2013). A major bottleneck in functional gene studies in non-model plants is often the lack of an e cient transformation protocol for recalcitrant genotypes or whole species. Since it is mostly easier to induce HRs using R. rhizogenes than to generate transformants using A. tumefaciens, HR-based functional gene studies might be a practicable alternative, especially for research on root-derived natural compounds. Figure 1 Falcarinol FaOH) and falcarindiol (FaDOH) levels in different leaf and root tissues and hairy roots. Bars represent means and error bars standard errors of mean (leaf, petiole, periderm, phloem, xylem: n = 3; hairy root: n = 9 for DP, YS and LR; n = 8 for RO). Signi cance analyses were performed after logtransformation using Tukey´s HSD test. Means with the same latin or greek letter are not signi cantly different (P ≥ 0.05).