A Root Tip-Specific Expressing Anthocyanin Marker for Direct Identification of Transgenic Tissues by the Naked Eye in Symbiotic Studies

The Agrobacterium rhizogenes hairy root transformation system is widely used in symbiotic studies of model legumes. It typically relies on fluorescent reporters, such as DsRed, for identification of transgenic roots. The MtLAP1 transcription factor has been utilized as a reporter system in Medicago truncatula based on production of anthocyanin pigment. Here, we describe a version of this reporter driven by a root-cap specific promoter for direct observation of anthocyanin accumulation in root tips, which allows the identification of transgenic hairy roots by the naked eye. Results from our analysis suggest that the reporter had no significant effects on nodulation of M. truncatula. This approach, by virtue of its strong and specific expression in root cap cells, greatly reduces false positives and false negatives, and its use of an easily scored visible pigment should allow greater versatility and efficiency in root biology studies.


Introduction
The model legume species Medicago truncatula has been used extensively to study interactions with nitrogen-fixing rhizobia and arbuscular mycorrhizal fungi [1][2][3]. Both interactions occur within plant roots, and M. truncatula has also been used for studies on root biology, such as in hormone regulation, nutrient response, water stress, root architecture, etc. [4,5]. A common approach to study root biology at the molecular level in plants other than arabidopsis is to generate composite plants with transgenic roots using the Agrobacterium rhizogenes-based 'hairy root' transformation system. Hairy roots can be both nodulated and colonized by arbuscular mycorrhiza, and are thus very useful for symbiotic studies, particularly since transgenic roots can be obtained within several weeks while conventional regeneration of transgenic plants roots takes 3 to 6 months [6]. However, the efficiency of hairy root transformation is not 100%, and it is therefore essential to distinguish transgenic and non-transgenic roots. This is most commonly achieved using β-glucuronidase/GUS staining and fluorescent proteins (GFP, YFP, DsRed, etc.) encoded in plasmids being transformed, but these identification methods have many limitations [7,8]. For example, GUS staining assay is destructive, preventing further study of the material [9]. Observation of fluorescent markers needs a specially equipped microscope, and there may be problems with background autofluorescence depending on the plants or tissues being studied and fluorescent proteins being used. Another issue is that use of a fluorescent reporter protein often precludes its use in protein fusions, and so reduces the options available to the researcher.
In recent years, several studies have indicated the feasibility of using anthocyanins as a visible marker to avoid using expensive equipment and tedious histochemical procedures. Zhang et al. designed a set of binary transformation vectors (pPurpleRoot) using the MtLAP1 (Legume Anthocyanin production 1) gene, and demonstrated its use for direct identification of transformed plants in M. truncatula [10]. Fan et al. also developed an anthocyanin reporter for several legume species including Glycine max, Lotus japonicus, L. corniculatus, and M. truncatula by using the AtMyb75/PAP1 (PRODUCTION OF ANTHO-CYANIN PIGMENTS 1) gene [11]. Both markers are suitable for non-destructive and direct visual identification of transgenic hairy roots.
MtLAP1 is a MYB transcription factor that induces the production and accumulation of anthocyanin in legumes [12]. Anthocyanins are pigmented secondary metabolites with important physiological functions in plants like UV-shielding in leaves and coloration of flowers and fruits [13]. They are a type of flavonoid and share the same precursors, such as tetrahydroxychalcone and naringenin, as other important flavonoids like flavones, flavonols and isoflavonoids, which play essential roles in the nodulation process [14,15]. The accumulation of anthocyanins can be easily affected by sucrose and light treatments. It has been shown that flavonoid and anthocyanin biosynthetic pathways are strongly up-regulated following sucrose treatment while light signaling also induces anthocyanin biosynthesis via the AN3 and COP1 network [16,17]. Thus, these biochemical features of anthocyanins need to be taken into consideration when they are used as the transgenic reporter and the transformation system may also require optimization.
Here, we describe a new binary vector (pMtRC-MtLAP1) that carries a reporter gene, MtLAP1, under a root-cap specific promoter. Anthocyanin accumulation can be seen in the root tip of transgenic roots by the naked eye, which allows the easy visual screening of transgenic roots without the use of histochemical staining or fluorescent microscopy. In addition, we show that the MtLAP1 marker does not affect M. truncatula nodulation. We also investigate the endogenous production of anthocyanins in roots responding to light and sucrose treatments and their effects on identification of transgenic roots and compare transformation efficiency between pMtRC-MtLAP1 and pAtUbi10-DsRed, which is the most commonly used marker for M. truncatula hairy root transformation.

Anthocyanin Markers Allow the Easy Visual Screening of Transgenic Roots
To confirm that the expression of MtLAP1 in roots of M. truncatula can induce anthocyanin accumulation, an AtE47 promoter-driven MtLAP1 construct, named pAtE47-MtLAP1 ( Figure S1a), was transformed into the M. truncatula ecotype A17 by A. rhizogenesmediated hairy root transformation. As reported by Zhang et al., we observed roots with strong anthocyanin coloration (Figure 1a-c) [10]. The anthocyanins were deposited mainly at the endodermis and pericycle due to the specificity of the AtE47 promoter as previously reported (Figure 1c) [10].
To construct a binary vector using a root tip-specific expressing promoter to drive MtLAP1, we took advantage of the Golden Gate cloning system, which can easily assemble multiple DNA fragments in a desired order [18]. The resulted vector was named pMtRC-MtLAP1 (Figure S1b), which was transformed into A. rhizogenes strain ARqua1 for generating transgenic hairy roots of M. truncatula cv. Jemalong A17 and R108 plants. The anthocyanin specifically accumulated in the tips of transformed roots in both A17 and R108 plants (Figure 1d,f; data not shown for R108). Upon closer inspection, we observed that the anthocyanins accumulated specifically in vacuoles of root cap cells and border cells, often forming dark purple precipitates ( Figure S2a). The marker was also scorable in older nodulated plants that were grown in substrate ( Figure S2b). In contrast, anthocyanins were absent in the non-transgenic roots (Figure 1d). Plants transformed with the control vector pAtUbi10-DsRed did not accumulate anthocyanins in their root tips (Figure 1g,h). The marker was also functional in tomato (Solanum lycopersicum), although transformation frequencies were very low ( Figure S3). control vector pAtUbi10-DsRed did not accumulate anthocyanins in their root tips ( Figure  1g,h). The marker was also functional in tomato (Solanum lycopersicum), although transformation frequencies were very low ( Figure S3).

Light and Sucrose Induce Endogenous Anthocyanin Production in Roots but Do Not Affect Visual Screening of Transgenic Roots Expressing pMtRC-MtLAP1
During our experiments, we occasionally observed purple sections along the roots transformed with either pAtUbi10-DsRed or pMtRC-MtLAP1 (Figure 2a). In order to determine whether it was caused by the transgenes and whether it could affect identification of transgenic roots expressing pMtRC-MtLAP1, we tested both transgenic and wild type A17 and R108 plants under different conditions. Sucrose treatment and light exposure are known to influence anthocyanin accumulation, so we tested their influence on the system. To limit light exposure, we wrapped the lower part of the plates with aluminum foil to cover the roots and shielded them from direct light, while others remained uncovered. To test the effect of sucrose, we added 10 g/L sucrose into the media. Interestingly, R108 plants did not develop any purple sections on roots of untransformed seedlings roots or transgenic roots expressing the pAtUbi10-DsRed in any of the treatments (Figure S4e-h; Figure S5e-h). However, in all A17 plants, the roots exposed to light exhibited clear purple

Light and Sucrose Induce Endogenous Anthocyanin Production in Roots but Do Not Affect Visual Screening of Transgenic Roots Expressing pMtRC-MtLAP1
During our experiments, we occasionally observed purple sections along the roots transformed with either pAtUbi10-DsRed or pMtRC-MtLAP1 ( Figure 2a). In order to determine whether it was caused by the transgenes and whether it could affect identification of transgenic roots expressing pMtRC-MtLAP1, we tested both transgenic and wild type A17 and R108 plants under different conditions. Sucrose treatment and light exposure are known to influence anthocyanin accumulation, so we tested their influence on the system. To limit light exposure, we wrapped the lower part of the plates with aluminum foil to cover the roots and shielded them from direct light, while others remained uncovered. To test the effect of sucrose, we added 10 g/L sucrose into the media. Interestingly, R108 plants did not develop any purple sections on roots of untransformed seedlings roots or transgenic roots expressing the pAtUbi10-DsRed in any of the treatments (Figure S4e-h; Figure S5e-h). However, in all A17 plants, the roots exposed to light exhibited clear purple sections, and sucrose and hairy root transformation enhanced the purple pigments on the roots (Figures S4a-d and S5a-d). Nonetheless, in both A17 and R108 plants, the root tips of all wild type and transgenic roots expressing pAtUbi10-DsRed under all treatments were white ( Figures S4 and S5), while the ones expressing pMtRC-MtLAP1 were purple (Figure 2c,d). Based on this, we conclude that neither light treatment nor sucrose supplement affect the accumulation of anthocyanin pigments in the root tip of either genotype, although they both induce endogenous anthocyanin production in A17 root segments above the root tip.
sections, and sucrose and hairy root transformation enhanced the purple pigments on the roots (Figures S4a-d and S5a-d). Nonetheless, in both A17 and R108 plants, the root tips of all wild type and transgenic roots expressing pAtUbi10-DsRed under all treatments were white ( Figures S4 and S5), while the ones expressing pMtRC-MtLAP1 were purple (Figure 2c,d). Based on this, we conclude that neither light treatment nor sucrose supplement affect the accumulation of anthocyanin pigments in the root tip of either genotype, although they both induce endogenous anthocyanin production in A17 root segments above the root tip.

The Apparent Transformation Efficiency of pMtRC-MtLAP1 Is Higher than pAtUbi10-DsRed
After ensuring the feasibility of the pMtRC-MtLAP1 visual screening system, we investigated the transformation efficiency and compared it with the pAtUbi10-DsRed control vector. We then used these markers to test the percentage of plants that developed at least one transgenic hairy root. In both A17 and R108, over 50% of the plants (71-78% for A17, and 52-64% for R108) showed transformed hairy roots with either DsRed fluorescence or pMtRC-MtLAP1. Considering just the transformed plants, the frequency of transformed hairy roots for pAtUbi10-DsRed ranged from 36 to 38%, while for pMtRC-MtLAP1, it ranged from 48 to 51% (Figure 3a). We noted that DsRed-positive roots showed large fluctuations in intensity, which may result in false negative observations, while the pigment intensity of MtLAP1 positive roots was quite consistent. To test if the low transformation rates observed for DsRed were due to false negatives or were instead due to unforeseen effects of the transgenes on the transformation process, we constructed a vector that contained both markers (pAtUbi10-DsRed pMtRC-MtLAP1; Figure S1c). Upon transformation into M. truncatula, we found that almost every DsRed-positive root had a purple root tip, while many purple tipped roots did not show any DsRed fluorescence (Figure 3c,d). Thus, as observed when analyzing the markers independently, higher rates of transformation were estimated using the anthocyanin marker in both A17 and R108 backgrounds when the markers were present on the same vector (Figure 3b), confirming that pAtUbi10-DsRed produces false negatives. Together, these data indicate that pMtRC-MtLAP1 is a better reporter for transformation of M. truncatula.

The Apparent Transformation Efficiency of pMtRC-MtLAP1 Is Higher than pAtUbi10-DsRed
After ensuring the feasibility of the pMtRC-MtLAP1 visual screening system, we investigated the transformation efficiency and compared it with the pAtUbi10-DsRed control vector. We then used these markers to test the percentage of plants that developed at least one transgenic hairy root. In both A17 and R108, over 50% of the plants (71-78% for A17, and 52-64% for R108) showed transformed hairy roots with either DsRed fluorescence or pMtRC-MtLAP1. Considering just the transformed plants, the frequency of transformed hairy roots for pAtUbi10-DsRed ranged from 36 to 38%, while for pMtRC-MtLAP1, it ranged from 48 to 51% (Figure 3a). We noted that DsRed-positive roots showed large fluctuations in intensity, which may result in false negative observations, while the pigment intensity of MtLAP1 positive roots was quite consistent. To test if the low transformation rates observed for DsRed were due to false negatives or were instead due to unforeseen effects of the transgenes on the transformation process, we constructed a vector that contained both markers (pAtUbi10-DsRed pMtRC-MtLAP1; Figure S1c). Upon transformation into M. truncatula, we found that almost every DsRed-positive root had a purple root tip, while many purple tipped roots did not show any DsRed fluorescence (Figure 3c,d). Thus, as observed when analyzing the markers independently, higher rates of transformation were estimated using the anthocyanin marker in both A17 and R108 backgrounds when the markers were present on the same vector (Figure 3b), confirming that pAtUbi10-DsRed produces false negatives. Together, these data indicate that pMtRC-MtLAP1 is a better reporter for transformation of M. truncatula.

The MtLAP1 Screening Marker Does Not Affect Nodulation
Hairy root transformation has been widely used for nodulation studies [19,20]. To test if the MtLAP1 reporter system is suitable for use in these studies, we assessed the number and nitrogen fixation potential of nodules in pMtRC-MtLAP1-transformed M. truncatula roots. The transgenic roots were inoculated with Sinorhizobium meliloti 2011 HemA::LacZ, and nodule number was calculated and the acetylene reduction assay was carried out 21 days later. Our results show that acetylene reduction activity and nodule number are not significantly affected in pMtRC-MtLAP1-transformed roots (Figure 4), suggesting that it is a useful marker for nodulation studies.

The MtLAP1 Screening Marker Does Not Affect Nodulation
Hairy root transformation has been widely used for nodulation studies [19,20]. To test if the MtLAP1 reporter system is suitable for use in these studies, we assessed the number and nitrogen fixation potential of nodules in pMtRC-MtLAP1-transformed M. truncatula roots. The transgenic roots were inoculated with Sinorhizobium meliloti 2011 HemA::LacZ, and nodule number was calculated and the acetylene reduction assay was carried out 21 days later. Our results show that acetylene reduction activity and nodule number are not significantly affected in pMtRC-MtLAP1-transformed roots (Figure 4), suggesting that it is a useful marker for nodulation studies.

The MtLAP1 Screening Marker Does Not Affect Nodulation
Hairy root transformation has been widely used for nodulation studies [19,20]. To test if the MtLAP1 reporter system is suitable for use in these studies, we assessed the number and nitrogen fixation potential of nodules in pMtRC-MtLAP1-transformed M. truncatula roots. The transgenic roots were inoculated with Sinorhizobium meliloti 2011 HemA::LacZ, and nodule number was calculated and the acetylene reduction assay was carried out 21 days later. Our results show that acetylene reduction activity and nodule number are not significantly affected in pMtRC-MtLAP1-transformed roots (Figure 4), suggesting that it is a useful marker for nodulation studies.

Discussion
In this study, we investigated whether anthocyanins produced exclusively in the root tip could be used as a screening marker for M. truncatula hairy root transformation. The MtLAP1 expression was driven by the MtRC promoter, which is highly expressed in root tips. The expression of MtLAP1 led to the accumulation of red/purple-colored anthocyanins in the root cap cells of Medicago and tomato, which can be seen without a microscope. The pMtRC-LAP1 marker was easily scored in both commonly used Medicago ecotypes, R108 and A17. Our analysis also showed that the effective transformation efficiency of the pMtRC-MtLAP marker compared favorably with that of DsRed, the latter being the most commonly used marker. We found this difference can be attributed to the large variation in DsRed expression in transgenic roots that results in a high rate of false negatives. Unlike DsRed, which is expected to have little or no impact on the cells in which it is expressed, MtLAP1 induces an entire suite of genes for production of anthocyanins, which could influence other metabolic pathways, for instance, flavonoid biosynthesis [12,21]. The anthocyanins could themselves impact on cell biology by blocking light or exerting effects as antioxidants [22,23]. Our use of a highly tissue-specific expression promoter helps decrease the risk of these effects on nodulation, and no obvious effects were observed on nodule numbers or N 2 -fixation capacity. Our results showed that using a sugar-free medium and shielding the roots from light can minimize anthocyanin production in roots for the A17 genotype, while neither sugar nor light-exposure induced anthocyanin accumulation in the R018 background. Importantly, neither sugar nor light exposure caused anthocyanin accumulation in roots tips of either genotype, and so, did not interfere with the pMtRC-MtLAP1 screening system. In conclusion, it is clear that anthocyanin production can be used as an effective marker for convenient visual screening for hairy root transformation to study the legume-rhizobia symbiosis.
This MtLAP1 anthocyanin marker, by replacing the common fluorescence indicator DsRed, makes DsRed or other fluorescent proteins with overlapping emission spectra available for use in protein fusions. The MtLAP1 system offers obvious advantages in efficiency over GUS staining. Another key advantage is the 'live-staining' aspect of the MtLAP1 system, allowing identification and monitoring of transgenic roots with minimal interference. We also successfully deployed pRC-MtLAP1 in hairy root transformation of tomato ( Figure S3). Although its transformation efficiency was lower than Medicago, this might be improved by the use of different Agrobacterium strains. In addition, the anthocyanin accumulation appeared to be strictly limited to the tissues where MtLAP1 was expressed, presumably due to glycosylation and vacuolization of the produced flavonoids [12]. This suggests that MtLAP1 can be developed as a marker to monitor gene expression in different tissues, particularly in the R108 background, a possibility that is currently being investigated.

A. Rhizogenes-Mediated Hairy Root Transformation
M. truncatula plants were germinated on water-agar plates at 22 • C in dark after being scarified with sandpaper, sterilized with bleach, washed with autoclaved dH 2 O, and imbibed for 3 h and vernalized at 4 • C in the dark for 4-7 days. An overnight liquid culture of A. rhizogenes strain ARqua1 containing the binary vector was grown. The seedling radicle was cut approximately 3 mm from the root tip after seedlings radicle length reached about 1 cm, and the wounded seedling was then dipped into the bacterial solution for about 10-15 min and then placed on a FP medium for 7 days at 22 • C (16 h photoperiod). The non-transgenic roots that grew before inoculation were then removed, and the seedling was transferred onto MFP medium. About 3 weeks after transformation, the number of plants with transgenic hairy roots was calculated for each treatment, plasmid construct, and plant genotype. The seedlings with at least one transgenic root were scored for transformation efficiency (percentage of transformed hairy roots on a given plant). The seedlings with transgenic roots were then moved into soil consisting of vermiculite and perlite (1:1) and grown in a greenhouse at 22 • C with 16/8 h light/dark cycle. One week later, the plants were inoculated with the rhizobia strain S. meliloti 2011 HemA::LacZ and nodulation phenotypes were scored 3 weeks after inoculation.

The Acetylene Reduction Assay
Nitrogenase not only catalyzes the reduction of atmospheric N 2 to NH 3 but can also reduce acetylene to ethylene, therefore reduction of ethylene by nitrogenase can be used to estimate N 2 -fixation capacity of nodules. For determination of acetylene-reduction activity (ARA), plants were either assayed 21 dpi with S. meliloti 2011 HemA::LacZ with tap water. For the AR assay, 3 to 7 biological replicates were used. Each replicate comprised of at least 7 nodulated roots for each genotype. The nodulated roots were introduced into 20 mL glass vials sealed with rubber stoppers. After injection of 1 mL acetylene, they were subsequently incubated for 2 h at room temperature. Samples of 200 µL of gas from each bottle were used to measure ethylene production using a gas chromatograph (7820A, Agilent Technology, Shanghai, China) and the roots were weighed. The nodules were then excised and weighed and the total nitrogenase activity was calculated as nanomoles of ethylene per gram of nodules, per nodule and per transgenic plant.