Genetic Analysis of the Liverwort Marchantia Polymorpha Reveals That

Research

Genetic analysis of the liverwort Marchantia polymorpha reveals that R2R3MYB activation of ﬂavonoid production in response to abiotic stress is an ancient character in land plants

Nick W. Albert1 , Amali H. Thrimawithana2 , Tony K. McGhie1 , William A. Clayton1,3, Simon C. Deroles1 , Kathy E. Schwinn1 , John L. Bowman4 , Brian R. Jordan3 and Kevin M. Davies1 1The New Zealand Institute for Plant & Food Research Limited, Private Bag 11600, Palmerston North, New Zealand; 2The New Zealand Institute for Plant & Food Research Limited, Private Bag 92169, Auckland Mail Centre, Auckland 1142, New Zealand; 3Faculty of Agriculture and Life Sciences, Lincoln University, Christchurch 7647, New Zealand; 4School of Biological Sciences, Monash University, Melbourne, Victoria 3800, Australia

Summary Author for correspondence: The flavonoid pathway is hypothesized to have evolved during land colonization by plants Kevin M. Davies c. 450 Myr ago for protection against abiotic stresses. In angiosperms, R2R3MYB transcription Tel: +64 6 3556123 factors are key for environmental regulation of flavonoid production. However, angiosperm Email: [email protected] R2R3MYB gene families are larger than those of basal plants, and it is not known whether the Received: 29 October 2017 regulatory system is conserved across land plants. We examined whether R2R3MYBs regulate Accepted: 19 December 2017 the flavonoid pathway in liverworts, one of the earliest diverging land plant lineages. We characterized MpMyb14 from the liverwort Marchantia polymorpha using genetic New Phytologist (2018) 218: 554–566 mutagenesis, transgenic overexpression, gene promoter analysis, and transcriptomic and doi: 10.1111/nph.15002 chemical analysis. MpMyb14 is phylogenetically basal to characterized angiosperm R2R3MYB flavonoid regulators. Key words: CRISPR/Cas9, evolution, Mpmyb14 knockout lines lost all red pigmentation from the flavonoid riccionidin A, flavonoid, liverwort, MYB, stress. whereas overexpression conferred production of large amounts of flavones and riccionidin A, activation of associated biosynthetic genes, and constitutive red pigmentation. MpMyb14 expression and flavonoid pigmentation were induced by light- and nutrient-deprivation stress in M. polymorpha as for anthocyanins in angiosperms. MpMyb14 regulates stress-induced flavonoid production in M. polymorpha, and is essential for red pigmentation. This suggests that R2R3MYB regulated flavonoid production is a conserved character across land plants which arose early during land colonization.

Introduction Many flowering plants (angiosperms) produce colourless UV- B absorbing flavone and flavonol glycosides in response to UV-B Flavonoid biosynthesis is a central component of plant secondary exposure, as well as following other abiotic stresses. They also metabolism that is thought to have evolved during land coloniza- produce red/purple anthocyanin flavonoid pigments in response tion, c. 450–500 Myr ago (Ma) (Vogt, 2010; Weng, 2013; de to nutrient deprivation, drought, wounding, cold and high light, Vries et al., 2017). Derived from the larger phenylpropanoid as well as for signalling to animals for pollination and seed disper- pathway, the acquisition of flavonoids is hypothesized to have sal (Allan et al., 2008; Agati & Tattini, 2010; Davies et al., been a key adaptation for coping with the additional abiotic 2012a; Cheynier et al., 2013). The flavones/flavonols and antho- stresses faced with the transition to a nonaquatic lifestyle. These cyanins can directly screen excess UV-B or white light energy, included drought, extreme temperature fluctuations and, in par- respectively, but also are thought to provide broad stress tolerance ticular, increased exposure to UV-B light (Bj€orn et al., 2002). through scavenging of reactive oxygen species (ROS; Agati & UV-B has severe detrimental effects on plant cells, and was at Tattini, 2010; Agati et al., 2012; Landi et al., 2015). particularly high intensities during the period of land coloniza- In all angiosperm species examined to date, R2R3MYB tran- tion because the ozone layer was not fully developed. Flavonoids scription factors are essential for environmental and developmen- are ubiquitous in extant land plants, including the liverworts, tal regulation of flavonoid biosynthesis (Albert et al., 2011; Feller which may be the closest living relatives of the first land plants. et al., 2011; Cheynier et al., 2013; Xu et al., 2015; Lloyd et al., However, they are absent from all algae, including Charophytes, 2017). For activation of the anthocyanin or proanthocyanidin which are thought to be most similar to the aquatic ancestors of pigment pathways, R2R3MYBs of subgroups (SGs) 5 or 6 act as land plants (de Vries et al., 2017). part of an R2R3MYB-bHLH-WDR (MBW) transcriptional

554 New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust New Phytologist Research 555 complex. Additionally, members of SG4 can have a repressive acting redundantly in regulation of flavone biosynthesis. role on genes of the phenylpropanoid pathway as part of the Riccionidin A production is induced in liverworts by similar MBW complex and contain transcriptional repressor domains in abiotic stress factors as previously characterized for anthocyanin their C-terminal (Matsui et al., 2008; Albert et al., 2014; production in angiosperms. It is notable that structurally dis- Cavallini et al., 2015). For activation of flavonol biosynthesis tinct red flavonoid pigments are produced in marchantia and R2R3MYBs of SG7 act independently of the MBW complex. angiosperms, although some of the same flavonoid biosynthetic Although the MBW complex is often regarded as universal for genes are activated in each plant group. The results suggest that anthocyanin regulation, it is actually an open question as to what R2R3MYB proteins were co-opted for regulation of the aspects of the system are conserved across the major land plant flavonoid pathway from the early stages of land plant evolution, groups (Albert et al., 2014; Liu et al., 2015; Xu et al., 2015; Lloyd and supports the hypothesis that an inducible flavonoid path- et al., 2017). The recent finding of a conserved MBW complex way was an early evolutionary adaption to the abiotic stresses of able to activate flavonoid biosynthesis in a gymnosperm species life on land. (Nemesio-Gorriz et al., 2017) suggests an ancient origin in the plant lineage, as angiosperms and gymnosperms are thought to Materials and Methods have diverged c. 300 Ma (Lu et al., 2014). However, genomes of the basal plant groups liverworts (Bowman et al., 2017) and Plant lines and growth conditions mosses (Rensing et al., 2008) contain markedly fewer MYB and bHLH genes than those of angiosperms and gymnosperms, sug- Marchantia polymorpha L. accessions Sey-1 (male) and Aud-2 (fe- gesting a more limited range of regulatory activities. male) were grown on Jiffy-7 peat pellets (Egmont Seeds, New Liverworts together with mosses and hornworts form the three Plymouth, New Zealand), and were induced to produce sexual main lineages of the basal land plant group, the bryophytes. structures with supplemental far-red light (Chiyoda et al., 2008) Marchantia polymorpha (hereafter, marchantia) has been devel- and crossed to generate spores for transformations. oped as a model species for liverwort studies. Marchantia has a Plant lines were maintained asexually through the propagation dominant haploid gametophytic generation, rapid growth rate, is of gemmae, either plated directly on agar medium (0.5 9 Gam- dioecious and can reproduce asexually in large numbers through borg’s B5 medium (Duchefa Biochemie, Haarlem, the Nether- single-cell derived clonal gemmae (Bowman et al., 2016). The lands) 1% (w/v) sucrose, 1.2% (w/v) agar) or onto a sterile filter last common ancestor of liverworts and angiosperms is thought paper disc covering the medium. Standard culture conditions to have existed c. 450 Ma, so comparative studies of the two plant were 25°C, 16 h photoperiod and 30 lmol m 2 s 1 light inten- groups can help to inform us about systems that have an early sity provided by cool fluorescent tubes. evolutionary origin. Liverworts lack some of the typically For nutrient deprivation experiments, gemmae (three per tub/ flavonoid pigmented tissues of angiosperms, such as flowers, biological replicate; three biological replicates per treatment) were fruits and seed coats, and knowledge about flavonoid production plated onto filter papers on complete media (0.5 9 B5, 1% in liverworts is comparatively limited. They produce a range of sucrose, 1.2% agar) and grown for 4 wk. After this time, plants vacuolar-located flavone O-orC-glycosides similar to those of were transferred on the filter paper to fresh complete media or angiosperms (Asakawa et al., 2013). Additionally, although liver- onto minimal media (1% sucrose, 1.2% agar) and grown for a worts lack the diversity of anthocyanins reported for further 13 d. Tissue was collected and immediately frozen in liq- angiosperms, they do produce a cell wall-localized red flavonoid uid nitrogen; aliquots were used for RNA extraction or freeze- pigment that has been identified as the aglycone anthocyanidin dried for metabolite analysis. riccionidin A, including from marchantia (Kunz et al., 1993). For light induction experiments, gemmae (three to four per The analysis of Bowman et al. (2017) identified only 21 tub/biological replicate; three biological replicates per treatment) R2R3MYB genes in the marchantia genome, as compared for were plated onto complete media (0.5 9 B5, 1% sucrose, 1.2% example to the estimated 138 in Arabidopsis, 157–188 in agar) and grown for 4 wk under high or low light conditions. maize (Zea mays) and 244–288 in soybean (Glycine max) Illumination was provided by a combination of warm and cool (Stracke et al., 2001; Feller et al., 2011; Liu et al., 2015). Often white LEDs (high light, 200 lmol m 2 s 1). The low light treat- the marchantia genome contains one or two genes that are basal ment (50 lmol m 2 s 1) was performed using the same light to large clades containing multiple angiosperm subgroups with source, except that plants were shielded with shade cloth. b differing but related functions. There are no clear orthologues MYB14pro:GUS (GUS, -glucuronidase) plants were grown in marchantia for genes of the phenylpropanoid-related SGs 4, under the same high/low light conditions, except that they were 5, 6 or 7, but MpMyb02 and MpMyb14 are basal to a clade sampled at either 16 or 20 d. containing 19 Arabidopsis R2R3MYBs comprising SGs 4, 5, 6, 7, 15 and 44 (Bowman et al., 2017). Following the hypothesis Light microscopy that R2R3MYB gene(s) will regulate abiotic stress-responsive flavonoid production in liverworts, we characterized the func- Images of marchantia plants and hand sections were taken with a tion of the candidate gene MpMyb14. The findings show that Leica DF550 digital camera mounted on a Leica M205FA dis- MpMyb14 activates both flavone glycoside and riccionidin A secting microscope. Z-stack images were generated with Leica biosynthesis, being essential for riccionidin A production but Application Suite X software.

Ó 2018 Plant & Food Research New Phytologist (2018) 218: 554–566 New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com New 556 Research Phytologist

® The UHPLC system used was a Dionex Ultimate 3000 Rapid Transformation Separation LC system equipped with a binary pump (HPR- Overexpression constructs with the CaMV35S promoter were gen- 3400RS), autosampler (WPS-3000RS), column compartment erated by LR recombination of the MpMyb14 coding sequence (TCC-3000RS) and a diode array detector (DAD-3000RS). The (synthesized by Genscript, Piscataway, NJ, USA) into pHARTII, analytical column used was a Kinetex XB-C18 50 mm 9 3mm, which confers Hygromycin resistance (10 mg l 1). CRISPR/Cas9 2.6 lm (Phenomenex, Torrance, CA, USA), maintained at 35°C. constructs were generated by annealing oligonucleotides corre- A binary solvent program was used with Solvent A (formic sponding to the variable regions of the guide RNAs, ligation into acid : MQ water, 1 : 99) and Solvent B (acetonitrile) at a flow of pMpU6pro/pENTR and LR recombination into the binary desti- 1mlmin 1. The initial solvent composition was 90%A 10%B nation vector pMpGE010 (Sugano et al., 2014). Marchantia until 0.5 min, then changed to 50%A 50%B at 2.5 min, and 5%A sporelings were transformed by inoculation with Agrobacterium 95%B at 3.5 min. After a 1-min hold at 5%A 95%B, the compo- tumefaciens (GV3101) strains harbouring binary vectors, essen- sition was returned to 90%A 10%B ready for the next injection. tially as described in Tsuboyama & Kodama (2014) with selection Total UHPLC analysis time was 6 min per sample. All solvent gra- on hygromycin (10 mg l 1) and ticarcillin (500 mg l 1). dients were linear. The injection volume was 3 ll. Spectral data (260–600 nm) were collected for the entire analysis. Flavones were quantified by integrating peak areas from Screening CRISPR mutations chromatograms at 340 nm, and external calibration curves were Genomic DNA was isolated from plants harbouring the constructed for luteolin 7-O-glucoside (Extrasynthese, Genay, CRISPR/Cas9 construct by CTAB/chloroform extraction (Ste- France). Flavone concentrations were calculated as luteolin 7-O- wart & Via, 1993), followed by PCR amplification of the target glucoside equivalents. Individual flavones were identified based region within MpMyb14 (NA322 50GCATTCTCTCCGAC on absorption spectra and LC-HRAM-MS/MS fragmentation ACGAGA30, NA460 50GATCCACCTCAGTCTACAGCT30). data (described below), and order of elution reported in PCR reactions were prepared for DNA sequencing by Exonucle- Markham et al. (1998). Riccionidin metabolites were quantified aseI/Shrimp Alkaline Phosphatase treatment, and sequenced from chromatograms extracted at 484 nm and reported as peak directly using BigDye v.3.1 chemistry. Genomic DNA samples areas, because no standard is available for these compounds. that contained multiple MpMyb14 alleles (chimeras) were readily Chromatographic data were collected and manipulated using the ® identified by mixed signals on the chromatogram where the Chromeleon Chromatography Management System v.7.2 sequences differed. Once plants reached a stage where the tissue (Dionex Corp., Sunnyvale, CA, USA). culture media became exhausted, mutants for MpMyb14 could The LC-HRAM-MS/MS system was composed of a Dionex ® readily be identified because of their lack of red pigmentation Ultimate 3000 Rapid Separation LC and a micrOTOF QII and genetic chimeras were easily detected visually. These mutants high resolution mass spectrometer (Bruker Daltonics, Bremen, were confirmed by DNA sequencing. Gemmae were propagated Germany) fitted with an electrospray ion source. The LC column (G1 generation) from mutant tissue and re-sequenced to ensure was a Hypersil Gold 200 9 2.1 mm, 1.9 lm (Thermo Scientific, that pure mutant lines were obtained. Plants used for experiments Auckland, New Zealand) and was maintained at 40°C. The flow were G2 generation or later. was 400 ll min 1. The solvents were A = 0.2% formic acid and B = 100% acetonitrile. The solvent gradient was: 10%A 90% B until 0.5 min; linear gradient to 100% A 0.5 to 22 min; composi- Promoter:GUS tion held at 100% A 22 to 40 min; linear gradient to 10% A MpMyb14 promoter:GUS fusions were made by amplifying 2 kb 90% B 40 to 40.2 min; to return to the initial conditions before of sequence upstream of the initiating ATG (Bowman et al., another sample injection at 45 min. The injection volume for 2017) and ligating this in front of an intron-containing GUS samples and standards was 1 ll. The micrOTOF QII parameters ° reporter gene (pDAH2, Davies et al., 2012b). The NotI cassette for flavonoid analysis were: temperature 225 C; drying N2 flow 1 was ligated into the binary vector pKART, conferring resistance 6 l min ; nebulizer N2 1.5 bar, endplate offset 500 V, mass to geneticin/G418 (7.5 mg l 1). Histochemical localization of range 100–1500 Da, data were acquired at 2 scans s 1. Negative GUS activity was performed as described in Albert et al. (2011). ion electrospray was used with a capillary voltage of 3500 V. Post-acquisition internal mass calibration used sodium formate clusters with the sodium formate delivered by a syringe pump at Flavonoid analysis the start of each chromatographic analysis. Flavonoids were Flavonoids were extracted from 10 mg freeze-dried ground thal- quantified using QUANTANALYSIS (Bruker Daltonics, Bremen, lus tissue with 1 ml methanol : water : formic acid (80 : 19 : 1). Germany) software. Ultra High Performance Liquid Chromatography (UHPLC) was used to separate and quantify the flavonoids present in extracts of RNA-seq and DNA sequence analysis thallus tissue. Liquid chromatography – High Resolution Accu- – rate Mass mass spectrometry (LC-HRAM-MS) and LC- Total RNA was extracted from control (35Spro:GFP)or35Spro: HRAM-MS/MS were used to confirm the identity of the com- MYB14 (line #8) thallus tissue with acid guanidinium/phenol pounds. extraction (Chomczynski & Sacchi, 1987) followed by

New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust New Phytologist Research 557

® purification with a Nucleospin RNA clean-up column lineage, and examplar non-phenylpropanoid R2R3MYBs (Sup- (Macherey-Nagel, Duren,€ Germany). RNA sequencing was on porting Information Fig. S1). MpMyb14 was basal to a clade Illumina TruSeq Stranded mRNA libraries by Illumina HiSeq, containing the angiosperm SG4, 5, 6 and 7 sequences, which and was conducted by the Australian Genome Research Facility also contained representatives from all the major groups of land (Melbourne, Australia). Bioinformatics analysis used DEseq2 plants examined. MpMyb02 was basal to the SG5/6/7 sub-clade (Love et al., 2014) with comparison of reads to a de novo transcript within this. A further sub-clade of SGs 5 and 6 was present con- assembly and the published marchantia genome sequence and taining lycophyte (Selaginella moellendorffii), gymnosperm and transcript assemblies (Bowman et al., 2017). The deduced amino angiosperm sequences but no bryophyte sequences. We chose to acid sequences of the R2R3MYB domain were aligned using functionally characterize MpMyb14, as it may relate to the sister CLUSTALWintheGENEIOUS suite of software (Kearse et al., 2012). ancestral sequence for the proposed phenylpropanoid clade. Additionally, 35SCaMV:MYB02 plants showed no change in flavonoid content based on UHPLC/LC-HRAM-MS analysis Quantitative RT-PCR (present authors’ unpublished data). Total RNA was isolated from frozen ground thallus tissue using the SpectrumTM Plant Total RNA Kit (Sigma-Aldrich) according MpMYB14 can activate flavonoid production in transgenic to the manufacturer’s instructions. First strand cDNA was pre- marchantia plants pared from DNAseI-treated RNA, and used for quantitative reverse transcription polymerase chain reaction (qRT-PCR) Transgenic marchantia lines were produced by overexpressing essentially as described in Albert (2015). Unless otherwise stated, MpMyb14 under the control of the 35SCaMV (35S) promoter, three biological replicates were used. Transcript abundance was which confers moderate constitutive gene expression in marchan- determined relative to the reference gene ACTIN (Pfaffl, 2001). tia, but is not strongly expressed in meristems (Althoff et al., Primers used were 2014). The 35Spro:MYB14 lines showed orange/red colour from MpACTIN NA336 50GATCTTGCTGGACGTGATCTT30, when transgenic cell-clusters first began to regenerate, and thalli NA337 50GCTTCTCCTTCATGTCTCTCAC30; of fully differentiated plants were pigmented orange/red except MpMyb14 transgene NA534 50GTCATCCCGCTGCAATC for around the meristem. Plants grown from gemmae of these T30, NA535 50GATCACCTTATTCTGAGCACCA30; lines had red pigmentation from a young age and adult plants MpMyb14 NA320 50GTCCGACAGATTCCTGTGTAAA30, had the same intense pigmentation as the clonal parents, includ- NA321 50GCCGCAAACAAACTTGTAAGA30; ing red pigmentation of developing gemmae (Figs 1a, S2). The MpCHS NA513 50GAGTTGGAATCAGAGTGGGTATG30, red pigment was localized within cell walls (Fig. S2b), which NA514 50TTCAGGTTGCAGAGCAGTTTA30; matched early observations of red pigmentation in liverwort cell 0 0 MpCHI NA378 5 AGCGCCTGTGGACAAATTA3 ,NA379 walls (Nagai, 1915; Herzfelder, 1921). Control plants (35Spro: 50TCAAGCTCATCTCCGTTGTG30; GFP) grown under the same conditions were predominantly MpCHI-Like NA515 50ACTCTCCACCTCTCAAATTTCT green, with red pigmentation restricted to a small region of the 0 0 0 C3 , NA516 5 TAATGTCCTCACTGGCATACAC3 . ventral thallus midrib (Fig. 1a). Analysis of the 35Spro:MYB14 lines in comparison to controls using UHPLC found a five-fold increase in amounts of flavone O-glucuronides (Fig. 1b,d) and a Statistical analyses 50-fold increase in amounts of an orange-red pigment (lambda The data were analysed by ANOVA using GENSTAT v.18 (VSN max c. 486 nm, Fig. 1c,e), identified as riccionidin A based on International, Hemel Hempstead, UK). The data were log- the UV-Vis absorption spectrum and accurate mass data from transformed as necessary, to equalize variances before performing LC-HRAM-MS. Riccionidin A is a cell wall located anthocyani- ANOVA. Post hoc comparisons among treatment means were din comprising four rings rather than the usual three of antho- made using Fisher’s Least Significant Difference (LSD) values at cyanins, based on the MS/NMR defined structure, and is the a = 0.05. only red pigment identified from marchantia (Kunz et al., 1993). The major ion of riccionidin A (compound 2 in Fig. 1) had an Results accurate mass at m/z 283.0230 (consistent with the published C15H7O6 structure, mass difference 1.8 mDa, isotope ratio 22.8 mSigma), and this was detected in both control and transgenic Marchantia R2R3MYB gene MpMyb14 is basal to a clade lines. However, there was a second compound (compound 1 in of R2R3MYB phenylpropanoid pathway regulatory genes Fig. 1) that was abundant in the transgenic plants but below the The analysis of Bowman et al. (2017) found that MpMyb02 and accurate detection limit in control plants. LC-HRAM-MS gave a MpMyb14 were basal to a clade containing Arabidopsis SGs 4, major ion for compound 1 with an accurate mass at m/z 5, 6 and 7. We generated a phylogenetic tree using the deduced 591.1340, consistent with a formula of C27H27O15 (mass differ- MYB domain amino acid sequences of MpMyb02 and ence 1.6 mDa, isotope ratio 5.4 mSigma). MS/MS fragmentation MpMyb14 aligned against the MYB domain sequences of charac- of this m/z 591.1340 ion produced a fragment ion at m/z terized phenylpropanoid regulators from a range of gymnosperm 283.0232, suggesting that compound 1 is a previously unidenti- and angiosperm species, BLAST hits from across the land plant fied glycoside of riccionidin A.

Ó 2018 Plant & Food Research New Phytologist (2018) 218: 554–566 New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com New 558 Research Phytologist

(a) (d)

(e)

(b) (c)

Fig. 1 Phenotypes of 35Spro: MYB14 transgenic and control Marchantia polymorpha plants. (a) Two examplar transgenic lines having medium (#13) or strong (#4) intensity visual phenotypes of red pigmentation are shown. The left panel shows a control transgenic line (35Spro:GFP) of the same age for comparison. The top panels show whole thallus branches (bars, 2 mm), whereas the lower panels show close-up images of gemmae cups containing

mature gemmae (bars, 0.5 mm). (b, c) Flavone and riccionidin content of control and 35Spro:MYB14 plants. Means SEM are shown; control (n = 4), medium (n = 3), strong (n = 4). Signiﬁcant differences (LSD, P = 0.05) between means are indicated where letters above bars differ. (d) UHPLC

chromatogram of ﬂavone compounds (340 nm) present in control of 35Spro:MYB14 plants. Peaks were identiﬁed based on LC-High Resolution Accurate Mass (HRAM)-MS ion masses and comparison to Markham et al. (1998). (1) Luteolin-7,30 di-O-glucuronide; (2) Luteolin-7-O-glucuronide; (3) Apigenin- 7-O-glucuronide and (4) Luteolin-40-O-glucuronide. (e) UHPLC chromatograms (484 nm) of riccionidin metabolites. Based on LCMS and MS/MS data, peak (1) corresponds to a glycosylated form of riccionidin A and (2) the riccionidin A aglycone. Absorption spectra for the two major peaks are shown.

was specific to individual genes in the large marchantia PAL and MpMYB14 activates flavonoid biosynthetic gene expres- CHS gene families, which contain 10 and 24 members, respec- sion tively. The activation of CHS (family member Mapoly0021s0159) In order to compare the regulatory action of MpMyb14 with the and CHI-L by MpMyb14 overexpression was confirmed by qRT- published data on angiosperm flavonoid-related R2R3MYB acti- PCR in independent transgenic lines (Fig. S2d). vators, RNAseq was used to identify changes in transcript abun- Biosynthetic routes to flavone O-glucuronides and riccionidin dance in the 35Spro:MYB14 transgenics. The transgene induced A after the chalcone isomerase/CHI-L steps have not been deter- changes in transcript abundance for a large number of genes, with mined for liverworts, and there could be flavonoid biosynthetic c. 600 genes having a greater than two-fold change and 223 having genes upregulated in the 35Spro:MYB14 transgenics that have yet an adjusted P-value of < 1 9 10 5 (Table S1). These values are in to be functionally annotated. Two ABCG-type transporter genes a similar range to the numbers of genes showing at least a two-fold were significantly upregulated (Mapoly0015s0086 and change in response to overexpression of flavonoid-related Mapoly0027s0141). ABC proteins transport anthocyanins, R2R3MYBs in angiosperm transgenics (Zhang et al.,2015). flavones or other flavonoids into the vacuole in species such as Transcripts for three phenylpropanoid biosynthetic enzymes – grape (Vitis vinifera), and Medicago truncatula (Francisco et al., phenylalanine ammonia lyase (PAL), chalcone synthase (CHS) 2013; Hwang et al., 2016; Biala et al., 2017), and also transport and chalcone isomerase-like (CHI-L) – were among the top 25 lignin precursors to the cell wall (Alejandro et al., 2012). changes ranked by either log2FoldChange or adjusted P-value There are alternative routes to flavone production in (Tables S1–S3). These were the only three enzymes whose genes angiosperms involving either cytochrome P450 mono-oxygenase were upregulated by at least two-fold from among the 61 genes (CytP450) or 2-oxoglutarate dependent-dependent dioxygenase annotated (Bowman et al., 2017) for the shikimate and phenyl- (2OGD) genes, with examples from both enzyme groups for propanoid pathways of marchantia (Table S3). This suggests that flavone synthase (FNSI and FNSII) and flavanone 2-hydroxylase MpMyb14 specifically promotes the channeling of substrate into (F2H) (Bredebach et al., 2011; Farrow & Facchini, 2014; Du the phenylpropanoid section of the shikimate pathway, particu- et al., 2016). The 2OGD and CytP450 groups are large gene larly the flavonoid branch. Moreover, the activation by MpMyb14 families in plants, including in marchantia (Bowman et al.,

New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust New Phytologist Research 559

2017). There are five CytP450 genes among the 100 most highly produced using CRISPR/Cas9 mutagenesis. Agrobacterium- upregulated genes in the RNA-seq data (Table S2), but no mediated transformation of spores was conducted with constructs 2OGD genes. The only liverwort FNS candidate sequence func- encoding two guide RNAs that targeted the MpMyb14 gene, and tionally characterized to date, from Plagiochasma appendiculatum, > 30 independent transgenic events were obtained. These plants is a 2OGD with both FNSI and F2H activity (Han et al., 2014). were screened for mutations within MpMyb14 by PCR amplifi- There is one strong marchantia BLAST hit to the cation and DNA sequencing; 14 lines were identified with muta- P. appendiculatum sequence, Mapoly0002s0224 with 67% amino tions ranging from single nucleotide deletions or insertions to acid identity, and this is the only marchantia gene that clades in deletions of up to 592 nucleotides (Fig. 2a). the 2OGD sub-group (DOXC28) that contains FNSI and F2H Wild-type (WT) marchantia plants generally lacked red pig- of angiosperms. However, Mapoly0002s0224 is not significantly mentation in the thallus, but developed red pigmentation in upregulated in the 35Spro:MYB14 transgenics, suggesting that older thallus tissues when tissue culture media become exhausted. CytP450s may be involved in flavone biosynthesis in marchantia. The myb14 mutants never developed red pigmentation, even In addition to activation of the flavonoid biosynthetic path- when grown to an age and stage when WT plants developed way, many of the other strongly induced genes in 35Spro:MYB14 strong red colour in the central part of the plant (Fig. S3). This plants have predicted roles in abiotic stress or defence responses was well illustrated by examining plant lines that were chimeric (Tables S1, S2). Analysis of the Gene Ontology terms (where for the myb14 mutation, as somatic events were easily visualized available from BLAST matches) for transcripts with at least a two- in the haploid background (Fig. 2b). Sharply divided green and fold change found that the two largest categories for either up- or red sectors were present, and in some examples presumed myb14 downregulated genes, were ‘oxidation-reduction process’ (22% CRISPR events could be seen as green sectors within the red tis- and 11% of genes, respectively) and ‘membrane’ (15% and 14% sues. Genotyping tissue from green or red thallus regions of the of genes, respectively). MpMyb14 may thus regulate other stress same plant confirmed their chimeric nature. Except for loss of responsive pathways in addition to flavonoids. Alternatively, red pigmentation, the myb14 mutants were phenotypically like these changes could reflect downstream responses to the produc- the control lines, indicating a specific action of MpMyb14 tion of flavonoids, as it is suggested that flavonoids can modify (Figs 2b, S3). general redox signalling pathways (Agati & Tattini, 2010). Light- or nutrient-deprivation stress induces MpMyb14 MpMYB14 is essential for production of the red flavonoid expression and flavonoid production in marchantia pigment riccionidin A in marchantia Although the physiological roles of the anthocyanins during abi- In order to confirm the functionality of MpMyb14 in activation otic stress is uncertain, it is generally accepted that they can of the flavonoid pathway and to examine whether there is redun- provide screening against damaging intensities of white light and dancy of pathway regulation, knockout mutant lines were it is thought that they also function as antioxidants under

(a)

Fig. 2 CRISPR/Cas9-generated myb14 Marchantia polymorpha mutants. (a) DNA sequence of CRISPR-induced mutations in (b) the MpMyb14 gene (Mapoly0073s0038). Green text corresponds to the guide RNA sequence, blue text is the PAM (Protospacer adjacent motif), and red text highlights the mutations generated. (b) Example of a primary transgenic plant (line #31) that is chimeric for a CRISPR-induced mutation of MpMyb14. Boxed area marked in the left hand image is shown enlarged on the right, with an arrow indicating a small sector from an independent CRISPR event.

Ó 2018 Plant & Food Research New Phytologist (2018) 218: 554–566 New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com New 560 Research Phytologist

conditions that promote oxygen free radical stress. Little is quantities of flavone O-glucuronides that were equivalent to WT known about functions of nonphotosynthetic pigments in (Figs 3b,c, S3). Transcripts for MpMyb14, determined by qRT- bryophytes such as marchantia. However, it has long been noted PCR, were induced by the nutrient stress treatment in both WT that red pigmentation in different liverwort species becomes and the myb14-1 mutant (Fig. 3d). These findings confirmed more intense in high light conditions (Kny, 1890; Nagai, 1915; MpMyb14 as a major activator of the flavonoid pathway with no Herzfelder, 1921), and riccionidin A production is induced by redundant regulation for riccionidin A production. However, the UV-B in Antarctic liverworts (Newsham, 2010). We thus exam- presence of flavones in the myb14 mutant lines showed that there ined stress induction of riccionidin A in marchantia and whether was some MpMyb14 independent activation of the core flavonoid this was mediated by MpMyb14, by exposing WT, 35Spro: biosynthetic pathway. MYB14, and myb14 mutant lines to a variety of nutrient and Wild-type and myb14 mutants were grown under low or high light stresses that are known to induce anthocyanin and/or intensity white light conditions, to determine if light intensity flavonol/flavone production in angiosperms. alters flavonoid accumulation in liverworts as it does in Transfer of WT marchantia from nutrient-rich medium to angiosperms. Additionally, because in angiosperms the minimal medium induced production of red pigmentation R2R3MYB genes determine the spatial and temporal regulation within c. 7 d, which intensified with increasing time (Fig. 3a). of flavonoids (Schwinn et al., 2006; Stracke et al., 2010a; Davies The myb14 mutant lines completely lacked red pigmentation, et al., 2012a), marchantia plants transformed with a MYB14 and UHPLC analysis found no detectable riccionidin A but promoter-GUS reporter construct were used to examine whether

(a)

(b) (c) (d) Fig. 3 Nutrient deprivation induces flavone and riccionidin accumulation. Wild-type (WT) and myb14-1 mutant Marchantia polymorpha plants grown on complete medium or transferred to minimal medium. (a) Pigmentation phenotypes (bars, 2 mm). (b) Quantification of flavones and (c) riccionidin compounds. ND, not detected. (d) Relative transcript abundance of MpMyb14. Means SE, n = 3 biological replicates are shown; significant differences (LSD, P = 0.05) between means are indicated where letters above bars differ.

New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust New Phytologist Research 561

(a) (b)

Fig. 4 MpMyb14 promoter activity and light- (c) responsive riccionidin accumulation. (a) A mature wild-type (WT) Marchantia polymorpha plant showing production of red riccionidin A pigment on the ventral midrib (with tissue cleared of chlorophyll shown in the right panel). (b) Activity of the MpMyb14 promoter in young plants, as measured by GUS, b-glucuronidase (GUS) activity staining of MYB14pro:GUS line #2. GUS activity was restricted to the ventral midrib in a similar pattern to that in mature plants, although no riccionidin A was (d) (e) apparent at this growth stage. (c) GUS (f) activity pattern in MYB14pro:GUS plants grown under low (50 lmol m 2 s 1) or high (200 lmol m 2 s 1) intensity white light. Bars, 2 mm. Quantification of flavones (d) and riccionidin (e) compounds in WT and myb14-1 mutant plants grown under high or low light conditions. (f) Relative transcript abundance of MpMyb14. Means SE, n = 3 biological replicates are shown; significant differences (LSD, P = 0.05) between means are indicated where letters above differ.

MpMyb14 also contributes to spatial and temporal regulation flavonols (SG7) do not (Feller et al., 2011). The conserved motif patterning. Older WT plants accumulate red riccionidin pigmen- [DE]Lx2[RK]x3Lx6Lx3R has been identified as facilitating the tation on the ventral midrib, particularly within tissues that con- MYB-bHLH interaction. Changes in any one of the conserved tact soil (Fig. 4a). The activity of the MpMyb14 promoter, as amino acids can reduce or abolish the bHLH interaction for Ara- measured by GUS activity, was restricted to the ventral midrib in bidopsis proteins, with amino acids L and R that start and end a similar pattern (Figs 4b, S4a), although the young plants the Lx6Lx3R sequence of particular importance (Zimmermann assayed in tissue culture at this stage lacked red pigmentation. In et al., 2004; Dai et al., 2016). MpMYB14 has only a partial con- WT plants, high light induced the production of flavones (two- served motif (Fig. 5), including lacking two of the conserved fold) and riccionidin A (seven-fold) compared with the low light residues [DE] and [RK]. MpMYB02 also has only a partial motif. treatment (Fig. 4). The myb14 mutants also had increased flavone However, greater conservation of the motif is found in content in response to the high light treatment, but had no R2R3MYBs from the moss Physcomitrella patens, and is fully con- detectable riccionidin A under either light regime (Figs 4d,e, S4b, served in those from the gymnosperm Picea. c). MYB14pro:GUS plants grown under high intensity white light had an overall increase in amount and a spreading of the pro- Discussion moter activity region compared with those grown under low light conditions (Figs 4c, S4a). Transcript abundance for MpMyb14, The findings on marchantia (Marchantia polymorpha)MpMyb14 determined by qRT-PCR, confirmed that MpMyb14 is transcrip- show that R2R3MYBs are conserved as key activators of abiotic tionally regulated in response to high light intensity (Fig. 4f). stress-responsive flavonoid production between angiosperms and their most distant land plant relatives, the liverworts. Thus, this may be a character trait that has been inherited from their com- The amino acid motif required for function within the mon early land plant ancestor which is conserved across all land MBW complex is only partially conserved in MpMYB14 plants. All of the major plant groups examined contained In angiosperms, the R2R3MYBs activators for anthocyanins sequences that formed a clear phylogenetic clade with MpMyb14 (SG5 or SG6) and proanthocyanidins (SG5) act as part of the at the base. As the R2R3MYB gene family has undergone less MBW complex, whereas the R2R3MYB activators for flavones/ diversification during liverwort evolution than has occurred in

Ó 2018 Plant & Food Research New Phytologist (2018) 218: 554–566 New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com New 562 Research Phytologist

Fig. 5 CLUSTALW alignment of 28 R2R3MYB sequences from flavonoid-related subgroups, showing only the conserved R2R3 domain. The conserved residues of the domain for interaction with bHLH partners are indicated below the sequence. Amino acid positions identified by Schwinn et al. (2016) as differing between SG5 and SG6 are indicated with asterisks. The Sequence logo of amino acid frequency at each position is shown at the top. In addition to Marchantia polymorpha MYB02 and MYB14, the sequences are from: the monocot species Allium cepa (MYB1, KX785130; MYB29, KX785133), Lilium hybrid (MYB12, BAJ05398), Oncidium Gower Ramsey (MYB1, ABS58501), Oryza sativa (C1, BAD04024; MYB3, BAA23339), Phalaenopsis schilleriana (UMYB8, ACH95789), Vanda hybrida (MYB6, ADQ57817), and Zea mays (C1, AAA33482; P1, AAC49394); the dicot species Antirrhinum majus (ROSEA1, ABB83826), Arabidopsis thaliana (MYB12, ABB03913; MYB75/PAP1, AAG42001; MYB123/TT2, CAC40021), Lotus japonicus (TT2a, BAG12893), Malus 9 domestica (MYB10, ACQ45201; MYB22, AAZ20438), Petunia 9 hybrida (AN2, AAF66727) and Vitis vinifera (MYBA1, BAD18977; MYBPA1, AM259485; MYBF1, FJ948477); the gymnosperms Ginkgo biloba (Gb_39081), Picea abies (MYB30, ANB66413) and Picea mariana (MYBF1, AAA82943); the lycophyte Selaginella moellendorffii (XP_002978781); and the moss Physcomitrella patens (XP_001752936). The amino acid sequences of the R2R3MYB domain were aligned in CLUSTALW.

the angiosperm linage, MpMyb14 may represent the ancestral and HY5 is central to regulating light-responsive changes in the gene for phenylpropanoid/flavonoid pathway regulators of transcription of AtMYB12 (SG7) and AtMYB75 (SG6) and plants. However, data are lacking from additional bryophytes, associated flavonoid biosynthetic gene expression in Arabidopsis including the other bryophyte linage with liverworts and mosses, (Stracke et al., 2010b; Shin et al., 2013). It is possible that HY5 the hornworts. There are no published genome sequences for could contribute to the redundancy for activation of flavone gly- hornworts to examine, although this may be resolved soon coside biosynthesis that we found in marchantia. The lack of (Sz€ovenyi, 2016). Additionally, functional data on phenyl- change in flavone or riccionidin A production in 35S:MYB02 propanoid pathway regulation is required from other plant plants indicates that MpMyb02 does not contribute to activation groups, because besides the marchantia data presented here, func- of these flavonoid pathway branches (present authors’ unpub- tional data are available only for gymnosperm and angiosperm lished data). sequences. This would also resolve which genes in the R2R3MYB The lack of a fully conserved bHLH interaction motif in clade regulate flavonoid biosynthesis specifically (such as MpMyb14 (Fig. 5) suggests that it is not forming part of an MpMyb14) and which may regulate other branches of the MBW complex. Thus, MpMyb14 may be more like ZmP1 phenylpropanoid pathway. (SG7) from maize, which acts independently of a bHLH partner The regulatory role of MpMyb14 in marchantia shows many (ZmP1 lacks three of the conserved motif residues; Hernandez commonalities with the characterized flavonoid-related et al., 2004) yet can activate production not only of flavone, but R2R3MYBs of angiosperms. This includes both the genes tar- also red phlobaphene flavonoid pigments (Feller et al., 2011). geted within the flavonoid pathway and the environmental trig- Phlobaphenes are 3-deoxyflavonoids formed from the oxidation gers that induce MpMyb14 expression. Moreover, the of colourless flavan-4-ol monomers or polymers, with the flavan- MYB14pro:GUS expression pattern suggests that, as with 4-ols being formed by a variant dihydroflavonol 4-reductase angiosperm R2R3MYB flavonoid regulators (Schwinn et al., using flavanones as substrates rather than dihydroflavonols 2006; Stracke et al., 2010a), MpMyb14 provides the patterning (Morohashi et al., 2012). The existence of the R2R3MYB- information for production of flavonoids, in particular riccioni- regulated phlobaphene pathway raises interesting questions din A. It is yet to be determined whether or not the upstream regarding riccionidin A biosynthesis. The biosynthetic enzymes signal transduction pathways that control MpMyb14 promoter that produce riccionidin A are still unknown. It may be that only activation are also conserved with those reported for angiosperms riccionidin A’s red colour and the structural similarity of the core (Feller et al., 2011; Liu et al., 2015; Xu et al., 2015). However, structure to that of anthocyanins has led to it being classified as marchantia contains a strongly conserved ELONGATED an anthocyanidin. It cannot be ruled out that it is produced by a HYPOCOTYL 5 (HY5) candidate gene (Bowman et al., 2017), yet to be described branch of the flavonoid pathway.

New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust New Phytologist Research 563

Angiosperms have at least four separate R2R3MYB subgroups light screening ability rather than ROS scavenging may be the (SGs 4, 5, 6 and 7) known to be involved in regulation of the main function in liverworts, and may have been the principal flavonoid pathway (Feller et al., 2011; Cheynier et al., 2013), and function in the early land plant ancestor. In this case, there would each of these is typically represented as a small multigene family be an evolutionary drive towards the red colour itself rather than (SG4, 6 and 7 each have four members in Arabidopsis). By con- the specific structure generating it. A previously unreported gly- trast, marchantia has only two putative R2R3MYB orthologues to coside of riccionidin A was detected in the 35Spro:MYB14 trans- these (MpMyb02 and MpMyb14). For the residues within the genics, and at trace amounts in wild-type (WT) plants. MYB domain identified as differing between SG6 and the other Compared to the amount of cell wall-bound riccionidin, the subgroups (Schwinn et al., 2016), in most cases MpMYB02 and amount of soluble riccionidin glycoside in WT plants was very MpMYB14 are more closely aligned to SG5/7 and not SG6. This low, probably too low to be significant in stress tolerance. It is is supported by the phylogenetics, as no SG6 sequences have been possible that it could represent a transport intermediate between identified from nonangiosperm species (Du et al., 2015; and the site of synthesis and the cell wall. Fig. S1). The marchantia genes could reflect the more basal evolu- Anthocyanins and riccionidin A may have arisen indepen- tionary state, with the other multigene flavonoid-related R2R3MYB dently during evolution since the last common ancestor of subgroups being more recent developments through gene duplica- angiosperms and liverworts. Yet as the biosynthesis of these struc- tion and sub-functionalization following the acquisition of flowers, turally distinct pigments is controlled by putatively orthologous fruit and seed. The SG4 of transcriptional repressors may have genes it raises the question of whether one ancestral compound arisen after the separation of the bryophyte lineage, as no bryophyte or both compounds were present in the common ancestor of all sequences corresponding to SG4 were apparent, and SG6 may have land plants. Notably, there is one report of riccionidin A in an arisen after the separation of the angiosperm lineage. Alternatively, angiosperm species, although there is no information on the rela- the progenitor genes corresponding to these subgroups may have tionship of the biosynthetic pathway to that of liverworts been lost in bryophytes during evolution since the separation from (Taniguchi et al., 2000). If both types of pigment were present in the last common ancestor with other plant groups. a common ancestor then retention of anthocyanins in The RNA-seq data from the 35Spro:MYB14 transgenics con- angiosperms may have been favoured as they can generate a wider tained candidate genes for the yet to be identified biosynthetic range of colours and have ROS scavenging activities. Interest- routes to flavones and riccionidin A in marchantia. This included ingly, it is thought that when the ancestor of the Caryophyllales two ABCG transporter genes. ABCG proteins transport the fla- species Beta vulgaris (beet) switched from producing antho- vanone liquiritigenin to the vacuole in M. truncatula (Biala et al., cyanins to betacyanins it retained the SG6 R2R3MYB for path- 2017) and monolignols to the cell wall in Arabidopsis (Alejandro way activation (Hatlestad et al., 2015; Lloyd et al., 2017). et al., 2012; Hwang et al., 2016). In grape, an ABCC subgroup Analysis of the beet betacyanin-related BvMYB1 found that it protein transports anthocyanins to the vacuole (Francisco et al., had amino acid changes that caused it to lose the ability to inter- 2013). Marchantia has 95 annotated ABC transporter genes, with act with bHLH partners and that it could activate the betalain 23 assigned as ABCG and 15 as ABCC (Bowman et al., 2017). biosynthetic genes independent of the MBW complex, thus Assignment of function based solely on sequence similarity showing similarities to the action of MpMYB14. between the marchantia and angiosperm sequences is difficult, We show here that the R2R3MYB gene MpMyb14 is a key reg- and functional studies will be required to determine the roles of ulator of flavonoid production in the liverwort M. polymorpha. the candidate genes. This is also the case for the flavone synthase Knockout lines of myb14 lose all red flavonoid pigmentation gene candidates. The liverwort P. appendiculatum FNSI/F2H is a from riccionidin A, whereas overexpression confers production of 2OGD (Han et al., 2014), as is the FNSI of the horsetail species large amounts of flavonoids, activation of the associated biosyn- Equisetum arvense, which phylogenetically is positioned between thetic genes, and constitutive red pigmentation. MpMyb14 bryophytes and angiosperms (Bredebach et al., 2011). However, expression and riccionidin A production are induced by the same only CytP450 genes were strongly upregulated by MpMyb14. light- and nutrient-deprivation stress triggers in marchantia as for The major non-carotenoid red pigments of angiosperms are anthocyanins in angiosperms, despite riccionidin A being cell the mutually exclusive anthocyanins and betacyanins. Beta- wall-located and anthocyanins vacuolar. The results suggest that cyanins replace anthocyanins in some families of the plant order R2R3MYB-activation of flavonoid production is a conserved Caryophyllales (Jain & Gould, 2015; Lloyd et al., 2017). character across land plants that may have arose early during land Although structurally distinct and the products of different colonization. The regulatory system in marchantia could reflect a biosynthetic pathways, betacyanins and anthocyanins are induced basal state that has subsequently been expanded through duplica- by the same set of abiotic stresses. They are both vacuole-located tion and sub-functionalization of R2R3MYB genes during evolu- glycosides and it has been suggested that they serve similar function, facilitating the diversity of pigmentation and flavonoid tions, by screening excess light but also through scavenging reac- functions observed in angiosperms. tive oxygen species (ROS) (Jain & Gould, 2015). The riccionidin pigments of liverworts are another group of struc- Acknowledgements turally distinct red pigments also induced by a similar variety of abiotic stresses. However, they are principally cell wall-bound This work was supported by The Marsden Fund of New aglycones, which would limit antioxidant function. Thus, the Zealand. We thank: Cathie Martin for her helpful discussions;

Ó 2018 Plant & Food Research New Phytologist (2018) 218: 554–566 New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com New 564 Research Phytologist

Eduardo Flores and Bernardo Pollak for advice on marchantia medicarpin biosynthetic pathway. Journal of Experimental Botany 68: 3231– 3241. transformation; Huaibi Zhang for advice on the chemical analy- € sis; Steve Arathoon, Andrew Mullen, Ian King and Julie Ryan for Bjorn LO, Widell S, Wang T. 2002. Evolution of UV-B regulation and protection in plants. Advances in Space Research 30: 1557–1562. technical assistance; and Andrew McLachlan for statistical analy- Bowman JL, Araki T, Kohchi T. 2016. Marchantia: past, present and future. ses. The authors acknowledge the services of the Australian Plant and Cell Physiology 57: 205–209. Genome Research Facility. Bowman JL, Kohchi T, Yamato KT, Jenkins J, Shu S, Ishizaki K, Yamaoka S, Nishihama R, Nakamura Y, Berger F et al. 2017. Insights into land plant evolution garnered from the Marchantia polymorpha genome. Cell 171: Author contributions 287–304. Bredebach M, Matern U, Martens S. 2011. Three 2-oxoglutarate-dependent K.M.D., S.C.D., K.E.S., N.W.A. and B.R.J. conceived the pro- dioxygenase activities of Equisetum arvense L. forming flavone and flavonol – ject; N.W.A. performed experiments; T.K.M. conducted from (2S)-naringenin. Phytochemistry 72: 557 563. Cavallini E, Matus JT, Finezzo L, Zenoni S, Loyola R, Guzzo F, Schlechter R, metabolite analyses; A.H.T., K.M.D., N.W.A. and J.L.B. con- Ageorges A, Arce-Johnson P, Tornielli GB. 2015. The phenylpropanoid ducted bioinformatic analysis; N.W.A., W.C., S.C.D., K.E.S., pathway is controlled at different branches by a set of R2R3-MYB C2 B.R.J. and K.M.D. contributed to experimental design and inter- repressors in grapevine. Plant Physiology 167: 1448–1552. pretation of data; and K.M.D. and N.W.A. wrote the Cheynier V, Comte G, Davies KM, Lattanzio V, Martens S. 2013. Plant manuscript, and all authors contributed to manuscript editing. phenolics: recent advances on their biosynthesis, genetics, and ecophysiology. Plant Physiology and Biochemistry 72:1–20. Chiyoda S, Ishizaki K, Kataoka H, Yamato KT, Kohchi T. 2008. Direct ORCID transformation of the liverwort Marchantia polymorpha L. by particle bombardment using immature thalli developing from spores. Plant Cell Reports Nick W. Albert X http://orcid.org/0000-0002-8579-529X 27: 1467–1473. Amali H. Thrimawithana X http://orcid.org/0000-0001-6325- Chomczynski P, Sacchi N. 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Analytical Biochemistry 2334 162: 156–159. Tony K. McGhie X http://orcid.org/0000-0002-8016-0244 Dai X, Zhou L, Zhang W, Cai L, Guo H, Tian H, Schiefelbein J, Wang S. Simon C. Deroles X http://orcid.org/0000-0001-8154-1922 2016. A single amino acid substitution in the R3 domain of GLABRA1 leads Kathy E. Schwinn X http://orcid.org/0000-0002-6337-3627 to inhibition of trichome formation in Arabidopsis without affecting its – John L. Bowman X http://orcid.org/0000-0001-7347-3691 interaction with GLABRA3. Plant, Cell & Environment 39: 897 907. Davies KM, Albert NW, Schwinn KE. 2012a. From landing lights to mimicry: Kevin M. Davies X http://orcid.org/0000-0001-5652-5015 the molecular regulation of flower colouration and mechanisms for pigmentation patterning. Functional Plant Biology 39: 619–638. Davies KM, Deroles SC, Boase MR, Hunter DA, Schwinn KE. 2012b. Biolistics-based gene silencing in plants. In: Sudowe S, Reske-Kunz A, eds. References Methods in molecular biology. Heidelberg, Germany: Springer, 63–74. Agati G, Azzarello E, Pollastri S, Tattini M. 2012. Flavonoids as antioxidants in Du H, Liang Z, Zhao S, Nan MG, Tran LS, Lu K, Huang YB, Li JN. 2015. The plants: location and functional significance. Plant Science 196:67–76. evolutionary history of R2R3-MYB proteins across 50 eukaryotes: new insights Agati G, Tattini M. 2010. Multiple functional roles of flavonoids in into subfamily classification and expansion. Scientific Reports 5: 11037. photoprotection. New Phytologist 186: 786–793. Du H, Ran F, Dong H-L, Wen J, Li J-N, Liang Z. 2016. Genome-wide analysis, Albert NW. 2015. Subspecialization of R2R3-MYB repressors for anthocyanin classification, evolution, and expression analysis of the Cytochrome P450 93 and proanthocyanidin regulation in forage legumes. Frontiers in Plant Science 6: family in land plants. PLoS ONE 11: e0165020. 1165. Farrow SC, Facchini PJ. 2014. Functional diversity of 2-oxoglutarate/Fe(II)- Albert NW, Davies KM, Lewis DH, Zhang H, Montefiori M, Brendolise C, dependent dioxygenases in plant metabolism. Frontiers in Plant Science 5: 524. Boase MR, Ngo H, Jameson PE, Schwinn KE. 2014. A conserved network of Feller A, Machemer K, Braun EL, Grotewold E. 2011. Evolutionary and transcriptional activators and repressors regulates anthocyanin pigmentation in comparative analysis of MYB and bHLH plant transcription factors. Plant – eudicots. Plant Cell 26: 962–980. Journal 66:94 116. Albert NW, Lewis DH, Zhang H, Schwinn KE, Jameson PE, Davies KM. 2011. Francisco RM, Regalado A, Ageorges A, Burla BJ, Bassin B, Eisenach C, et al. Members of an R2R3-MYB transcription factor family in Petunia are Zarrouk O, Vialet S, Marlin T, Chaves MM 2013. ABCC1, an ATP developmentally and environmentally regulated to control complex floral and binding cassette protein from grape berry, transports anthocyanidin 3-O- – vegetative pigmentation patterning. Plant Journal 65: 771–784. glucosides. Plant Cell 25: 1840 1854. Alejandro S, Lee Y, Tohge T, Sudre D, Osorio S, Park J, Bovet L, Lee Y, Han X-J, Wu Y-F, Gao S, Yu H-N, Xu R-X, Lou H-X, Cheng A-X. 2014. Geldner N, Fernie AR et al. 2012. AtABCG29 is a monolignol transporter Functional characterization of a Plagiochasma appendiculatum flavone involved in lignin biosynthesis. Current Biology 22: 1207–1212. synthase I showing flavanone 2-hydroxylase activity. FEBS Letters 588: – Allan AC, Hellens RP, Laing WA. 2008. MYB transcription factors that colour 2307 2314. our fruit. Trends in Plant Science 13:99–102. Hatlestad GJ, Akhavan NA, Sunnadeniya RM, Elam L, Cargile S, Hembd A, Althoff F, Kopischke S, Zobell O, Ide K, Ishizaki K, Kohchi T, Zachgo S. 2014. Gonzalez A, McGrath JM, Lloyd AM. 2015. The beet Y locus encodes an Comparison of the MpEF1a and CaMV35 promoters for application in anthocyanin MYB-like protein that activates the betalain red pigment pathway. – Marchantia polymorpha overexpression studies. Transgenic Research 23:235–244. Nature Genetics 47:92 96. Asakawa Y, Ludwiczuk A, Nagashima F. 2013. Chemical constituents of Hernandez JM, Heine GF, Irani NG, Feller A, Kim M, Matulnik T, Chandler bryophytes; bio- and chemical diversity, biological activity, and VL, Grotewold E. 2004. Different mechanisms participate in the R-dependent chemosystematics. In: Kinghorn AD, Falk H, Kobayashi J, eds. Progress in the activity of the R2R3 MYB transcription factor C1. Journal of Biological – chemistry of organic natural products, vol. 95. Vienna, Austria: Springer, 1–796. Chemistry 279: 48 205 48 213. € € Biala W, Banasiak J, Jarzyniak K, Pawela A, Jasinski M. 2017. Medicago Herzfelder H. 1921. Beitrage zur Frage der Moosfarbungen. Beihefte zum – truncatula ABCG10 is a transporter of 4-coumarate and liquiritigenin in the Botanischen Centralblatt 38: 355 400.

New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust New Phytologist Research 565

Hwang J-U, Song W-Y, Hong D, Ko D, Yamaoka Y, Jang S, Yim S, Lee E, Stracke R, Favory JJ, Gruber H, Bartelniewoehner L, Bartels S, Binkert M, Khare D, Kim K et al. 2016. Plant ABC transporters enable many unique Funk M, Weisshaar B, Ulm R. 2010b. The Arabidopsis bZIP aspects of a terrestrial plant’s lifestyle. Molecular Plant 9: 338–355. transcription factor HY5 regulates expression of the PFG1/MYB12 gene in Jain G, Gould KS. 2015. Are betalain pigments the functional homologues response to light and ultraviolet-B radiation. Plant Cell & Environment 33: of anthocyanins in plants? Environmental and Experimental Botany 119: 88–103. 48–53. Stracke R, Jahns O, Keck M, Tohge T, Niehaus K, Fernie AR, Weisshaar B. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton 2010a. Analysis of PRODUCTION OF FLAVONOL GLYCOSIDES- S, Cooper A, Markowitz S, Duran C et al. 2012. Geneious Basic: an integrated dependent flavonol glycoside accumulation in Arabidopsis thaliana plants and extendable desktop software platform for the organization and analysis of reveals MYB11-, MYB12- and MYB111-independent flavonol glycoside sequence data. Bioinformatics 28: 1647–1649. accumulation. New Phytologist 188: 985–1000. Kny L. 1890. Bau und Entwickelung von Marchantia polymorpha In: Botanische Stracke R, Werber M, Weisshaar B. 2001. The R2R3-MYB gene family in Wandtafeln, vol VIII Abtheilung. Berlin, Germany: Paul Parey, 364–401. Arabidopsis thaliana. Current Opinion in Plant Biology 4: 447–456. Kunz S, Burkhardt G, Becker H. 1993. Riccionidins A and B, anthocyanidins Sugano SS, Shirakawa M, Takagi J, Matsuda Y, Shimada T, Hara- from the cell walls of the liverwort Ricciocarpos natans. Phytochemistry 35: 233– Nishimura I, Kohchi T. 2014. CRISPR/Cas9-mediated targeted 235. mutagenesis in the liverwort Marchantia polymorpha L. Plant and Cell Landi M, Tattini M, Gould KS. 2015. Multiple functional roles of anthocyanins Physiology 55: 475–481. in plant-environment interactions. Environmental and Experimental Botany Sz€ovenyi P. 2016. The genome of the model species Anthoceros agrestis. Advances 119:4–17. in Botanical Research 78: 189–211. Liu J, Osbourn A, Ma P. 2015. MYB transcription factors as regulators of Taniguchi S, Yazaki K, Yabu-uchi R, Kawakami K, Ito H, Hatano T, Yoshida phenylpropanoid metabolism in plants. Molecular Plant 8: 689–708. T. 2000. Galloylglucoses and riccionidin A in Rhus javanica adventitious root Lloyd A, Brockman A, Aguirre L, Campbell A, Bean A, Cantero A, Gonzalez A. cultures. Phytochemistry 53: 357–363. 2017. Advances in the MYB-bHLH-WD Repeat (MBW) pigment regulatory Tsuboyama S, Kodama Y. 2014. AgarTrap: a simplified Agrobacterium-mediated model: addition of a WRKY factor and co-option of an anthocyanin MYB for transformation method for sporelings of the liverwort Marchantia polymorpha betalain regulation. Plant and Cell Physiology 58: 1431–1441. L. Plant and Cell Physiology 55: 229–236. Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and Vogt T. 2010. Phenylpropanoid biosynthesis. Molecular Plant 3:2–20. dispersion of RNA-seq data with DESeq2. Genome Biology 15: 550. de Vries J, de Vries S, Slamovits CH, Rose LE, Archibald JM. 2017. How Lu Y, Ran JH, Guo DM, Yang ZY, Wang XQ. 2014. Phylogeny and divergence embryophytic is the biosynthesis of phenylpropanoids and their derivatives in times of gymnosperms inferred from single-copy nuclear genes. PLoS ONE 9: streptophyte algae? Plant and Cell Physiology 58: 934–945. e107679. Weng JK. 2013. The evolutionary paths towards complexity: a metabolic Markham KR, Ryan KG, Bloor SJ, Mitchell KA. 1998. An increase in the perspective. New Phytologist 201: 1141–1149. luteolin: apigenin ratio in Marchantia polymorpha on UV-B enhancement. Xu W, Dubos C, Lepiniec L. 2015. Transcriptional control of flavonoid Phytochemistry 48: 791–794. biosynthesis by MYB–bHLH–WDR complexes. Trends in Plant Science 20: Matsui K, Umemura Y, Ohme-Takagi M. 2008. AtMYBL2, a protein with a 176–185. single MYB domain, acts as a negative regulator of anthocyanin biosynthesis in Zhang Y, Butelli E, Alseekh S, Tohge T, Rallapalli G, Luo J, Kawar PG, Hill L, Arabidopsis. Plant Journal 55: 954–967. Santino A, Fernie AR et al. 2015. Multi-level engineering facilitates the Morohashi K, Casas MI, Falcone Ferreyra ML, Mejıa-Guerra MK, Pourcel L, production of phenylpropanoid compounds in tomato. Nature Yilmaz A, Feller A, Carvalho B, Emiliani J, Rodriguez E et al. 2012. A Communications 6: 9635. genome-wide regulatory framework identifies maize Pericarp Color1 controlled Zimmermann IM, Heim MA, Weisshaar B, Uhrig JF. 2004. Comprehensive genes. Plant Cell 24: 2745–2764. identification of Arabidopsis thaliana MYB transcription factors interacting Nagai I. 1915. Ueber roten pigmentbildung beu einigen Marchantia-Arten. with R/B-like BHLH proteins. Plant Journal 40:22–34. Botanical Magazine 29:90–98. Nemesio-Gorriz M, Blair PB, Dalman K, Hammerbacher A, Arnerup J, Stenlid J, Mukhtar SM, Elfstrand M. 2017. Identification of Norway Spruce MYB- Supporting Information bHLH-WDR transcription factor complex members linked to regulation of the flavonoid pathway. Frontiers in Plant Science 8: 305. Additional Supporting Information may be found online in the Newsham KK. 2010. The biology and ecology of the liverwort Cephaloziella Supporting Information tab for this article: varians in Antarctica. Antarctic Science 22: 131–143. Pfaffl MW. 2001. A new mathematical model for relative quantification in real- time RT-PCR. Nucleic Acids Research 29: 2002–2007. Fig. S1 Phylogenetic relationships of Marchantia polymorpha Rensing SA, Lang D, Zimmer AD, Terry A, Salamov A, Shapiro H, Nishiyama MybB02 and MpMyb14 to phenylpropanoid-related R2R3MYBs T, Perroud PF, Lindquist EA, Kamisugi Y et al. 2008. The Physcomitrella of other species. genome reveals evolutionary insights into the conquest of land by plants. – Science 319:64 69. Fig. S2 MpMyb14 overexpression activates flavonoid synthesis in Schwinn KE, Ngo H, Kenel F, Brummell DA, Albert NW, McCallum JA, Pither-Joyce M, Crowhurst R, Eady C, Davies KM. 2016. The onion (Allium Marchantia polymorpha L. cepa L.) R2R3-MYB gene MYB1 regulates anthocyanin biosynthesis. Frontiers. Plant Science 7: 1865. Fig. S3 MpMyb14 is essential for nutrient stress-induced ric- Schwinn K, Venail J, Shang Y, Mackay S, Alm V, Butelli E, Oyama R, Bailey P, cionidin accumulation in Marchantia polymorpha. Davies K, Martin C. 2006. A small family of MYB-regulatory genes controls floral pigmentation intensity and patterning in the genus Antirrhinum. Plant Cell 18: 831–851. Fig. S4 High light induces MpMyb14 transcription and Shin DH, Choi M, Kim K, Bang G, Cho M, Choi SB, Choi G, Park YI. 2013. flavonoid accumulation in Marchantia polymorpha. HY5 regulates anthocyanin biosynthesis by inducing the transcriptional activation of the MYB75/PAP1 transcription factor in Arabidopsis. FEBS Table S1 Adjusted P-value, BaseMean, log FoldChange and – 2 Letters 587: 1543 1547. annotation for the 223 genes showing an adjusted P-value Stewart CN Jr, Via LE. 1993. A rapid CTAB DNA isolation technique useful for < 9 5 RAPD fingerprinting and other PCR applications. BioTechniques 14: 748–750. 1 10 for transcript abundance difference in RNAseq

Ó 2018 Plant & Food Research New Phytologist (2018) 218: 554–566 New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com New 566 Research Phytologist

analysis of Marchantia polymorpha 35Spro:MYB14 transgenics vs of Marchantia polymorpha 35Spro:MYB14 transgenics vs control control plants plants.

Table S2 BaseMean, log2FoldChange, adjusted P-value and Please note: Wiley Blackwell are not responsible for the content annotation for the 100 genes showing the largest fold increase in or functionality of any Supporting Information supplied by the transcript abundance in RNAseq analysis of Marchantia authors. Any queries (other than missing material) should be polymorpha 35Spro:MYB14 transgenics vs control plants directed to the New Phytologist Central Ofﬁce.

Table S3 BaseMean and log2FoldChange values for genes of the shikimate and phenylpropanoid pathway from RNAseq analysis

New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-proﬁt organization dedicated to the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews and Tansley insights.

Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged. We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average time to decision is <26 days. There are no page or colour charges and a PDF version will be provided for each article.

The journal is available online at Wiley Online Library. Visit www.newphytologist.com to search the articles and register for table of contents email alerts.

If you have any questions, do get in touch with Central Office (np-centraloffi[email protected]) or, if it is more convenient, our USA Office (np-usaoffi[email protected])

For submission instructions, subscription and all the latest information visit www.newphytologist.com

New Phytologist (2018) 218: 554–566 Ó 2018 Plant & Food Research www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust