The Establishment of Two Efficient Transformation Systems in Quinoa


 Quinoa (Chenopodium quinoa Willd.) provides a gluten-free food with abundant nutrition and is potential to become a major crop in future. Quinoa has a lot of unique features including the high tolerance to multiple stresses, of which the underlying mechanisms may help improve other crops. Genetic manipulation will provide powerful tools to investigate the function of key genes in regulation of quinoa development and stress responses, which will further improve quinoa planting in the fields. However, the efficient transformation system for quinoa has not been well developed yet. Here, we established two rapid and efficient transformation systems for quinoa by using hairy root and leaf agroinfiltration, which provide useful tools for quick analyses of genes function in quinoa. Hairy roots were obtained from three types of explants, including cotyledon-nod with hypocotyl, cotyledon itself, and hypocotyl pieces. Interestingly, explants of cotyledon-nod with hypocotyl showed the highest transformation efficiency at 67.9%, and cotyledon displayed medium efficiency at 42.2%, while hypocotyl with the lowest at 31.6%. We also obtained transgenic quinoa roots successfully in-vivo, which showed low efficiency, but this provides a potential method to test gene function in live plants. By using young leaves for agroinfiltration, direct injection showed a better transgenic effect compared with vacuum penetration. Overall, the transformation systems using both hairy root and leaf agroinfiltration provide efficient and convenient ways to manipulate and analyse gene functions in quinoa.


Introduction
Quinoa, which belongs to the Chenopodiaceae family, contains abundant nutrition and is used as the staple food for the indigenous inhabitants of South America, including the Andean regions for more than 7000 years (Wilson 1990;Jacobsen et al. 2003). Quinoa, considered as a potential future major crop, has attracted intense public attention since the Food and Agriculture Organization of the United Nations announced "International Year of Quinoa" in 2013 (Bazile et al. 2016). Previous analyses showed that quinoa constitutes all of the essential amino acids required for human health, and contains more protein content than rice, barley, corn, rye and sorghum (R. et al. 2003;Bodner-Montville et al. 2006;Abugoch et al. 2010;Gonzalez et al. 2012;Escuredo et al. 2014;Aloisi et al. 2016;Filho et al. 2015). Moreover, quinoa can sustainably endure well in the extreme growth conditions, including drought, salinity and frost (Ruiz et al. 2014;Ruiz et al. 2016). For instance, some varieties of quinoa can tolerate the salinity around 700 mM NaCl, which is higher than the salinity of seawater (Hariadi et al. 2011;Orsini et al. 2011;Adolf et al. 2012). These unique properties of quinoa, including stress tolerance and high nutrition, result in its considerable roles in global food security due to climate change paradigm shifts. Recently, there is a surge in the functional analysis of quinoa genes own to the availability of the genome draft and sequence ( (Zou et al. 2017;Jarvis et al. 2017). However, all the functional veri cations were only through the comparison of gene expression pro les based on the relative phenotypes. Moreover, it is crucial to develop a method to study the gene function through genetic manipulation.
At present, Agrobacteriumtumefaciens-mediated stable transformation is a prominent method that is used to analyze gene function regulating plant development or environmental adaptation. Although the stable transformation in quinoa has been studied since last century, it is still unsuccessful (Komari 1990;López-Marqués et al. 2020). The function analyses are alternatively done in model plant systems such as Arabidopsis thaliana or Nicotiana benthamiana (Imamura et al. 2019). However, there is still a signi cant barrier referring to unique genes and pathways of quinoa. Thus, both transient and stable transformation systems are critically needed to be established to investigate the function of valuable genes in quinoa.
Agroin ltration is A. tumefaciens-mediated instant transformation method for gene function related studies, which is fast, real-time and economical. Agroin ltration is a convenient tool in plant gene function research without resorting to stable transformation (Del Toro et al. 2014). Through the injection of A. tumefacien, extraneous genes are transiently expressed in leaves, fruits, owers, petals, roots, needles and other tissues (Ma et al. 2020). Compared with stable transformation that produces minimal transgenic lines and needed to consider the effect of T-DNA insert position, agroin ltration infects a large number of cells with high e ciency in a short time and dose not required T-DNA integration (Yang et al. 2000;Gelvin and Stanton 2010).
Agrobacterium rhizogenes-mediated transformation system has many advantages compared with A. tumefaciens-mediated transformation which is time-consuming, genotype dependent and laborious. A. rhizogenes can potentially infect numerous varieties of dicotyledonous plants and subsequently induce transgenic hairy root lines (Gomes et al. 2019;Lam et al. 1984). The hairy root system can excellently characterize root development and related gene functions, which has been extensively applied in metabolic engineering, rhizosphere physiology, environmental stress response, etc. However, when the target genes or biological pathways don't work in the plant roots only, the hairy root system will not be applicable. The composite plant consisting of the wild-type shoots and transgenic hairy roots can be used to study root-shoot interactions and resistance against the various environmental and abiotic stresses, especially when roots are in stress conditions, such as salinity, drought or heavy metal (Mellor et al. 2012;Xue et al. 2017;Kajikawa et al. 2010;Cho et al. 2000;Cai and D. 1997;Savka 1990). Besides, the composite plants were employed in previous studies of hormone transport, symbiosis, secondary metabolites and root nutrient uptake. Up to date, various plant species have been established ex-vitro composite plant transformation system, such as pea (Clemow et al. 2011), soybean , tea (Alagarsamy et al. 2018) and peanut (Guimaraes et al. 2017). Both hairy roots in-vitro and composite plants could be derived in a cost-effective, fast and e cient way, which were considered as a practical protocol, especially for gene function study in roots. In this study, we developed the hairy root system and leaf agroin ltration system, which were used to analyse gene function rapidly in quinoa.
Healthy seeds were carefully selected and then surface sterilized with chlorine gas (produced by mixing sodium hypochlorite and HCl) in a desiccator for 4-6 h. Sterilized seeds were then moved to a clean bench for air blowing for one day and germinated in half strength Murashige and Skoog medium (MS) at 50 μmol/s/m 2 light exposure 16 h light/8 h dark at 25±1 ℃ for 5-6 days to obtain suitable explants from seedlings for hairy roots induction.
For in-vivo seedlings, healthy seeds were carefully selected and sowed in a mixture of soil: vermiculite: perlite (5:3:2) in the greenhouse at 540 μmol/s/m 2 light exposure 16 h light/8 h dark, two-week-old plants with only primary leaves spread out were used for agroin ltration, two-month-old plants were used for invivo hairy roots induction.

Preparation of the Agrobacterium strains for infection
A. rhizogenes K599 carrying the binary vector PTF102 harbouring the GUS gene was used to induce positive hairy roots. The hairy roots of negative control were induced using the same vector but without the GUS gene. The K599 was streaked on solidi ed YEP medium (10 g/Ltryptone, 5 g/L yeast extract, 5 g/L NaCl, 15 g/L agar, pH 6.8) with a disposable plastic inoculation loop. After being incubated at 28 ℃ for two days, a half loop of the cultures were collected and restreaked on a new YEP medium then incubated for another one day. The A. rhizogenes cultures were collected and suspended to OD 660 of 0.8 with infection medium (MS medium, Acetosyringone (AS 19.6 mg/L), 2-(4-Morpholino) ethanesulfonic acid (MES 500 mg/L), 3% Sucrose, pH 5.4). A. tumefaciens EHA105 carrying the binary vector pBI121 harbouring the GUS gene was used for agroin ltration of leaves. EHA105 was prepared in the same way as K599. The cultures were collected and suspended to OD 660 of 1.0 with infection medium(1/10 MS medium, 19.6 mg/L AS, 100 mg/L MES, 2% Sucrose, pH 5.6).
The vector pX11 harbouring CYP76AD6, CYP76AD1, BvDODA1 and cDOPA5GT (kindly provided by Dr. Asaph Aharoni, Weizmann Institute of Science, Israel) was also used to test and verify the hairy root system. The transgenic hairy roots could produce red pigment betacyanins, which showed red-violet colour and could be visually observed (Polturak et al. 2016).

Hairy roots induction from three types of explants
Three types of explants were obtained from ve-day-old aseptic seedlings ( Fig. 1-A) (MS509/N, QN108B, 141R, 20ALC), including the cotyledons with hypocotyl explant (CHE), were excised approximately 1cm below the node of the cotyledons; the cotyledon explants (CE) excised from the seedlings; the hypocotyl explants (HE) were cut into a length of 1-1.5 cm.
CHEs were gently pierced near the cotyledonary node with a syringe tip for two to three times, CEs were pierced in the middle of the abaxial side ( Fig. 1-B), while the HEs were obtained by just cutting off the hypocotyls into pieces. The explants were immediately incubated in the infection medium for 30 min at 25 ℃ ( Fig. 1-C), then transferred to the co-culture medium at 22 ℃ in the dark ( Fig. 1-D). Two days later, the explants were transferred to hairy roots induction medium (Gamborg B-5 basal medium, 500 mg/L MES, 3% Sucrose, pH 5.6) at 25 ℃ in the dark for two weeks.

Hairy roots inductionin-vivo
In-vivo hairy roots were induced by injecting the Agrobacterium culture into the taproots of two-month-old seedlings (MS509/N) (Fig. 3-A). Part of the soil was moved away to expose the taproot. The A. rhizogenes K599 infection culture of OD 660 of 1.0 was slowly injected into the taproots with a 2.5 mL syringe ( Fig. 3-B). After injection, the taproot was covered with wet vermiculite and a plastic lm to maintain humidity ( Fig. 3-C). The seedlings were grown at 540 μmol/s/m 2 light exposure 16 h light/8 h dark at 25℃ until the hairy roots emerged.

Agroin ltration of quinoa leaves
Two-week-old seedlings (MS509/N) with only primary leaves spread out were chosen for agroin ltration (Fig. 4-A). Two methods were developed for agroin ltration of quinoa leaves, one was syringe injection, and the other was vacuum penetration.
The cotyledons and one of the primary leaves were removed. The A. tumefaciens EHA105 infection culture was injected into the other primary leaf remained on the seedlings with a 2.5 mL needle-free syringe (Fig. 4-B). The leaf in ltrated with EHA105 carrying pBI121 without GUS gene was used as the negative control. After injection, the treated seedlings were covered with a plastic jar and grew at 100 μmol/s/m 2 light intensity with the photoperiod of 16 h light/8 h dark at 22 ℃ for 4 days ( Fig. 4-C).
In vacuum penetration, the remaining primary leaf was gently damaged by pressing a wire brush on the abaxial surface while placing a rubber cork wood on the adaxial surface ( Fig. 4-D). The damaged leaf was immediately incubated in Agrobacterium culture and vacuumed at 0.07 MPa for 1 min in a tightly sealed desiccator (Fig. 4-E). After in ltration, the seedlings were treated for recovery in the same way as injection method.

Histochemical staining assay of GUS activity
The hairy roots produced from three types of explants and in-vivo injection were collected for GUS staining assay. The leaves after agroin ltration were subjected to GUS assay then soaked in 100% ethanol to remove and destain the chlorophyll.

Polymerase chain reaction (PCR) veri cation of GUS gene
When hairy roots grew out from the explants, they were collected for PCR analysis of GUS gene. The plasmid of PTF102 was used as a positive control. To con rm the insertion of GUS gene, PCR was performed with a pair of primers speci c to GUS gene (5'-GCAGGTCACTGGATTTTGGT-3'; 5'-ACGGCAGAGAAGGTACTGGA-3'). The ampli cation conditions were: a 3-min melting step at 95 ℃, followed by 30 cycles of a 30-s melting step at 95 ℃, a 30-s annealing step at 55 ℃, and a 25-s elongation step at 72 ℃ and a 5-min elongation step at 72 ℃ after cycles. PCR products were analyzed by electrophoretic separation on 1% agarose gels.

Hairy roots induction in-vitro
We tested hairy roots induction in four quinoa cultivars, and found all four cultivars were able to produce hairy roots. We mainly used MS509/N for hairy roots induction in this study. All three types of explants, including CHE ( Fig. 1-E), CE ( Fig. 1-F) and HE ( Fig. 1-G), produced hairy roots after two-week induction.
GUS staining assay showed that CHE got the highest transformation e ciency at 67.9% (Fig. 2-A) , followed by CH at 42.2% (Fig. 2-B) and the lowest was HE at 31.6% (Fig. 2-C). Hairy roots from the negative control didn't show GUS expression (Fig. 2-D). PCR analysis also con rmed the GUS gene expressed successfully in hairy roots (Fig. 2-I). The transformation e ciency was calculated by the ratio of the number of explants produced positive hairy roots to the total explants number ( Table 1). The transformation e ciency differed among these explants, CHE and CE showed relatively higher transformation e ciency, which suggested they may be suitable explant types for transformation.
In order to test whether this method can be used for gene function analysis in quinoa, we transformed pX11 containing betacyanins biosynthesis genes to hairy roots, and found the accumulation of red colour in the hairy roots ( Fig. 2-E, F, G, H). This suggested hairy root might be a good system to test gene function in quinoa.

Hairy roots induction in-vivo
Many hairy roots initiated from the injection sites grew out from soil two weeks after the injection (Fig. 3-D). By using GUS reporter assay, although most of the roots showed negative results in GUS staining, we still could get positive ones (Fig. 3-E). Taproot injection was a tedious, time-consuming method with lower transformation e ciency compared with in-vitro system, however, this provided potential to study some genes expressed in root and regulated the shoot development.

Leaf Agroin ltration
For leaf agroin ltration, we injected agrobacterium carrying pBI121 plasmids with GUS gene, the whole leaf showed blue colour in GUS staining assay (Fig. 4-G), while no blue colour appeared in the negative control ( Fig. 4-F). We also tested the vacuum penetration system, and found only partial wounded spots appeared (Fig. 4-H). Considering the homogeneity of leaf age and Agrobacterium concentration (OD 660 of 1.0), the result indicated that vacuum penetration had much less throughput compared with syringe injection. For the cultivar MS509/N in this study, the primary spread leaves from two-week-old seedlings were the most suitable for agroin ltration. Leaves from younger seedlings were too small and delicate to carry out the syringe technique and thus resulted in severe damages from the injection leading to necrosis, while older leaves had lower transformation e ciency.

Discussion
The hairy roots induction in-vitro and in-vivo Hairy roots could be induced from different organs of quinoa, such as cotyledonary node, cotyledon, hypocotyl and in-vivo taproot in soil. Both CHE and CE had cotyledons and higher transformation e ciency. It was inferred that cotyledons provided more nutrition for hairy roots induction and subsequent elongation. CHE was considered to be the optimal explant type for its highest transformation e ciency according to our experiments.
For the in-vivo system, the composite plant with transgenic roots is an excellent research material to study genes expressed in roots and regulated the upper part of the plant. In this study, it was quite di cult to get positive hairy roots by injecting the taproot. It is necessary to optimize this protocol to obtain more composite plants. Many factors should be further explored such as the seedling age, the temperature and the soil humidity after injection (Alagarsamy et al. 2018;Guimaraes et al. 2017;Cao et al. 2009).
Hairy roots grow fast even without hormonal supplements, which can be used as bioreactors to produce bioactive substances more e ciently and economically. Quinoa contains abundant bioactive substances including saponins, polyphenols and avonoids (Alvarez-Jubete et al. 2010). These secondary metabolites accumulate richly in roots (Lim et al. 2020). In our research, transgenic hairy roots produced betacyanins due to the expression of the genes from pX11, which will be ideal approach for commercially production of nutritious betacyanins from quinoa hairy roots.
For quinoa, A. tumefaciens-mediated stable transformation has not been developted yet. Hairy root is originated from a single cell and usually non-chimeric, which can be chosen as explant to regenerate stable transgenic plantlets (Chen et al. 2018). Up to date, some species have regenerated successfully from hairy root cultures, such as tobacco (Gurusamy et al. 2017), alfalfa (Jin et al. 2003), maize (Xu et al. 2006), potato (Butler et al. 2020). Especially for spinach, which is closely related to quinoa, the transgenic shoots have been regenerated from the hairy roots (Ishizaki et al. 2002). Therefore, the hairy root system here provides a new way to explore quinoa genetic transformation.

Agroin ltration of quinoa leaves
Agroin ltration is an e cient system during the research of gene expression, promoter activities, metabolism, and protein-protein interaction (Levy et al. 2005; Mooney and Graciet 2020). To our knowledge, this is the rst report of agroin ltration application in quinoa. The whole agroin ltration process could be done within one month and independent of regeneration, which was e cient compared with tedious and laborious stable transformation.
Based on our research, seedling stage, humidity and recovery time were crucial to the system. In our protocol, the rst pair of spread leaves from two-week-old seedlings had higher agroin ltration e ciency, the explant older than one-month could hardly be injected. It was similar to Nicotiana benthamiana, Arabidopsis thaliana, Brassica oleracea and Thellungiella salsuginea, in which the younger leaves proved to be better for agroin ltration (Zhang et al. 2020). Keeping high humidity was vital for quinoa seedlings to survive from agroin ltration damage, the treated leaves were easy to wither or even to die if plants were not covered with jars to keep humidity.
The duration of recovery after agroin ltration also in uenced the assay result. It was inferred that genes started to express a few hours after transformation (Jones et al. 1989). Many reports showed that in plant tissues, gene expression level usually reached a peak 2-4 days after agroin ltration (Lacroix and Citovsky 2013;Krenek et al. 2015;Marion et al. 2008). In GUS assay, light blue colour started to appeare on the 2nd day and the dark blue appeared on the 4th day, so four-day recovery was considered suitable for quinoa seedlings.

Conclusions
Two rapid and e cient systems for gene function analysis were established in this report: hairy roots system and leaf agroin ltration system. Hairy roots could be obtained from three types of explants in two weeks with different e ciencies. The CHE showed the highest e ciency at 67.9%, thus was the recommended explant type for hairy roots induction in quinoa. The rst pair of spread leaves of twoweek-old seedlings were the most suitable material for agroin ltration. The leaf injection displayed a much better result than vacuum penetration in transient expression assay.

Con icts of Interest
All authors read and approved the nal manuscript. The authors have no con ict of interest.

Availability of data and material
All data generated or analysed during this study are included in this published article.  Figure 1 Hairy roots induction in-vitro. 5-day-old seedlings were obtained from half-strength MS medium(a), cotyledon was pierced with a syringe tip at the abaxial side (b), three types of explants were immersed into K599 culture (c), explants were transferred to co-culture medium covered with sterile lter paper (d), hairy roots from CHE in hairy root induction medium after two weeks (e), hairy roots from CE (f), hairy roots from HE (g). Scale bar a, c-g: 1 cm; b: 0.15 cm Hairy roots induction in-vivo. 2-month-old seedling was used for injection (a), the taproot was injected with a syringe needle (b), covered with wet vermiculite and plastic lm to keep humid (c), hairy roots of quinoa injection grew out from quinoa's tap root after two weeks (d), hairy roots cut from the tap root and assay for β-glucuronidase activity (e). Scale bar a: 0.5 cm; b-c: 1 cm; d-e: 0.25 cm Figure 4 Agroin ltration of quinoa leaves. Leaf was used for agroin ltration from two-week-old seedling(a), leaf was injected at the abaxial side with a needleless syringe (b), the treated seedlings were covered with a plastic jar for recovery for 4 days (c), leaf was pressed at the abaxial side by the brush (d), leaf was incubated in Agrobacterium culture and vacuumed at 0.07 MPa for 1 min (e), GUS staining of leaf injected with 1/10 MS culture (negative control ) (f), GUS assay of leaf injection (g), GUS assay of leaf vacuum penetration (h). Scale bar a-e: 1cm; f-h: 0.2 cm