Auxin Produced by the Indole-3-Pyruvic Acid Pathway Regulates Development and Gemmae Dormancy in the Liverwort Marchantia polymorpha

Auxin, which is synthesized in the liverwort Marchantia polymorpha via the smallest known toolkit among land plants, plays critical roles in development and dormancy of the gametophyte generation.

Abstract

The plant hormone auxin (indole-3-acetic acid [IAA]) has previously been suggested to regulate diverse forms of dormancy in both seed plants and liverworts. Here, we use loss- and gain-of-function alleles for auxin synthesis- and signaling-related genes, as well as pharmacological approaches, to study how auxin regulates development and dormancy in the gametophyte generation of the liverwort Marchantia polymorpha. We found that M. polymorpha possess the smallest known toolkit for the indole-3-pyruvic acid (IPyA) pathway in any land plant and that this auxin synthesis pathway mainly is active in meristematic regions of the thallus. Previously a Trp-independent auxin synthesis pathway has been suggested to produce a majority of IAA in bryophytes. Our results indicate that the Trp-dependent IPyA pathway produces IAA that is essential for proper development of the gametophyte thallus of M. polymorpha. Furthermore, we show that dormancy of gemmae is positively regulated by auxin synthesized by the IPyA pathway in the apex of the thallus. Our results indicate that auxin synthesis, transport, and signaling, in addition to its role in growth and development, have a critical role in regulation of gemmae dormancy in M. polymorpha.

INTRODUCTION

Tissues and cells with the ability to initiate and release dormancy can be found in all extant land plants from the basal-most bryophytes to flowering plants (Schröder, 1886; Patterson and Baber, 1961; Mapes et al., 1989). Because of the sessile nature of the lifestyle of land plants, they are required to successfully cope with a changing environment at their specific location. Thus, the place and timing for spore or seed germination as well as the ability for plants to adapt their growth pattern to their ecological context and to various environmental factors are of utmost importance. Plants have evolved diverse sets of dormancy to cope with the consequences of a sedentary lifestyle. Seeds or spores become dormant to prevent germination, e.g., inside the fruit or sporangia (reviewed in Linkies et al., 2010). The seed or spore maintains dormancy to delay germination until favorable environmental conditions are perceived. This delay also allows for widespread dispersal of offspring before germination. Another type of dormancy, referred to as apical dominance, occurs when lateral or axillary shoots remain inactive to allow the plant to focus its growth and energy on only one or a few apical meristems (reviewed in Leyser, 2009). The lateral shoots initiate growth if the primary meristem is lost, e.g., to a grazing herbivore, and can thus be seen as a backup apical meristem.

Different forms of dormancy are to a large extent regulated by different combinations of hormones. Seed dormancy is antagonistically regulated by abscisic acid (ABA) and gibberellic acid (reviewed in Graeber et al., 2012), and recently it was shown that auxin (indole-3-acetic acid [IAA]) transcriptional responses regulate seed dormancy by interacting with ABA-INSENSITIVE3 in the ABA signaling pathway (Liu et al., 2013). Liu et al. (2013) observed that Arabidopsis thaliana seeds expressing the auxin synthesis gene iaaM showed delayed release of dormancy, while seeds having loss-of-function alleles of auxin synthesis genes of the YUCCA (YUC) family or auxin transcriptional response genes of the AUXIN RESPONSE FACTOR (ARF) family exhibited reduced dormancy. Similarly, apical dominance in vascular plants is believed to be positively regulated by auxin transport from the source in the apical region of the shoot. This auxin transport signal has been postulated to inhibit bud growth by regulating cytokinin and strigolactone synthesis pathways (Thimann and Skoog, 1933, 1934; Nordström et al., 2004; Tanaka et al., 2006; Hayward et al., 2009; de Jong et al., 2014; reviewed in Müller and Leyser, 2011). Apical dominance has also been observed in basal lineages of land plants such as liverworts, and here also, an apical auxin source is believed to be responsible for the apical dominance of shoots during thallus growth (Maravolo, 1976).

In some complex thalloid liverworts, such as Marchantia polymorpha and Lunularia cruciata, disc-formed propagules called gemmae are formed in cup-like structures, called gemma cups, on the dorsal side of the dorsiventral haploid thallus (Barnes and Land, 1908). Mature gemmae are a means of asexual reproduction and remain dormant in the cup until dispersed (Molish, 1922), which in natural conditions generally is a result of rainfall (Brodie, 1951). Once the gemmae has been splashed out of the gemma cup, dormancy is released and initiation of growth occurs within hours, leading to the hypothesis that the parent plant secretes a gemma growth inhibitor, much in the same way as the fruit was thought to secrete a seed germination inhibitor (Molish, 1922). This hypothesis was supported by the observation that desiccation-tolerant gemmae from dried (dead) M. polymorpha thalli will commence growth on the remains of the parent plant once rehydrated (Schröder, 1886).

M. polymorpha gemmae will release dormancy and grow on crushed gemma cup tissue just as well as gemmae placed on nutrient solution, suggesting the dormancy signal requires intact cups or that it might not originate in the cup (Oppenheimer, 1922). However, gemmae in gemma cups only germinated partially even if the cups had been isolated from the surrounding thallus—only when the cups began to disintegrate was clear gemmae germination observed (Tarén, 1958). When gemmae were removed from the cups of intact plants and later replaced, they continued to germinate within the cup, but did not cause release of dormancy in other gemmae inside the cup, suggesting growing gemmalings do not stimulate their neighbors to germinate and once germination has commenced the inhibitor is ineffective (Oppenheimer, 1922; Tarén, 1958). LaRue and Narayanaswami (1957) demonstrated that removal of the apical notch of L. cruciata thalli resulted in release of dormancy of gemmae in cups directly posterior to the excised notch, leading to a hypothesis that the gemma dormancy signal in liverworts originates in the thallus apical meristem. Similar experiments by the same authors in M. polymorpha were deemed inconclusive, leading to the authors suggesting the quick regeneration of the apical meristem in M. polymorpha but not L. cruciata prevents complete release of gemma dormancy (LaRue and Narayanaswami, 1957).

When applying auxin to the cut end of the basal L. cruciata thallus fragment containing the gemma cup, gemmae germination was not observed (LaRue and Narayanaswami, 1957). As endogenous auxin was detected in extracts from apical regions of L. cruciata thalli, the authors suggested that auxin produced in the apex acts to inhibit gemmae germination within the cup (LaRue and Narayanaswami, 1957). Somewhat counterintuitively, Tarén (1958) observed addition of auxin directly into the gemma cup can stimulate rhizoid emergence and gemma growth, suggesting auxin can overcome the inhibition imposed by residence in the gemma cup, and proposed a cup-localized antiauxin also contributes to dormancy. The available data are consistent with the hypothesis that there are two inhibitory substances: one that emanates from the apical meristem and acts broadly throughout the thallus and another one localized to the cups (Tarén, 1958).

In angiosperms, a significant amount of IAA required for growth and development of the sporophyte is produced from Trp via indole-3-pyruvic acid (IPyA) in a two-step pathway (Mashiguchi et al., 2011; Stepanova et al., 2011; Won et al., 2011). The enzymes catalyzing the conversion of Trp to IPyA and from IPyA to IAA are members of the TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS (TAA) and YUC families, respectively (Stepanova et al., 2008; Tao et al., 2008; Mashiguchi et al., 2011; Dai et al., 2013). There are also other Trp-dependent and -independent pathways proposed for angiosperms (reviewed in Tivendale et al., 2014), but their contribution to the total IAA pool and their biological significance remain to be determined (Normanly et al., 1993; Ouyang et al., 2000; Pollmann et al., 2003, Brumos et al., 2014). In liverworts and other non-seed plants, it has been proposed that a majority of IAA is produced from a Trp-independent pathway (Sztein et al., 2000; reviewed in Cooke et al., 2002). However, growing M. polymorpha on Trp-supplemented media resulted in a distinct phenotype similar to that observed for gemmalings grown on auxin (Dunham and Bryan, 1968), and Jayaswal and Johri (1985) demonstrated that cultures of the moss Funaria hygrometrica could use exogenous Trp to synthesize IAA via IPyA. Homologs of both TAA and YUC genes were identified in the genome of the moss Physcomitrella patens (Rensing et al., 2008). These results suggest the Trp-dependent IPyA pathway is evolutionary conserved among land plants and thus could contribute to the total IAA pool also in bryophytes.

With the advent of the liverwort M. polymorpha as a model genetic system (Ishizaki et al., 2008, 2012, 2013; Kubota et al., 2013; Sugano et al., 2014; Althoff et al., 2014; Flores-Sandoval et al., 2015a), we are now in a position to begin elucidating the molecular genetic mechanisms regulating dormancy in the basal-most land plants. We have recently characterized mutants perturbing the M. polymorpha auxin transcriptional response (Flores-Sandoval et al., 2015b; Kato et al., 2015), noting that gemmae dormancy is affected in plants with altered auxin homeostasis or signaling. In this study we examine the role of auxin biosynthesis and response pathways on M. polymorpha gemmae dormancy, clarifying the role for auxin as a regulator of dormancy in basal land plants.

RESULTS

Dormancy Regulation by the Apical Thallus Notch Is Conserved among Liverworts

Because previous studies resulted in inconclusive data regarding differences in regulation of gemmae dormancy between M. polymorpha and L. cruciata (LaRue and Narayanaswami, 1957; Tarén, 1958), we examined whether the apical meristem acts to impose dormancy on M. polymorpha gemmae in our growth conditions. Pfeffer (1871) and Fitting (1939) previously determined that the first sign of a gemma breaking dormancy is the emergence of rhizoids. This is followed by apical cell divisions leading to gemmaling expansion (Pfeffer, 1871; Fitting, 1939). To measure dormancy, we scored gemma cups for the presence of rhizoids and/or elongating gemmalings (nondormant) or their absence (dormant). We removed the apical meristem from wild-type thalli, and 5 d following removal of the apical part of mature thalli, we saw the first signs of growth of gemmae in posterior gemma cups (Supplemental Figures 1A and 1C). Seven days after removing the apex, almost 40% of gemma cups displayed one or more nondormant gemmae compared with 0% in controls with intact apex (Supplemental Figure 1B). No gemma cups showed more than five nondormant gemmae, suggesting release of dormancy was weak. At 7 d, a regenerated thallus apex was clearly visible in most cut thalli. From 7 to 12 d after removing the apical notch, no change in the dormancy status of proximal gemma cups was found and the nondormant gemmae grew slower than gemmae displaced from the cup, suggesting suppression of gemmaling growth. Our observations suggest the regenerated apex may quickly reestablish a dormancy signal to proximal gemma cups, as previously proposed by LaRue and Narayanaswami (1957). Thus, our data suggest the apical signal regulating gemmae dormancy is conserved between M. polymorpha and L. cruciata but also support the notion that M. polymorpha is less sensitive than L. cruciata to this treatment (LaRue and Narayanaswami, 1957).

Endogenous Auxin Produced by the IPyA Pathway Is a Positive Regulator of Gemmae Dormancy in M. polymorpha

Consistent with auxin imposing dormancy on gemmae within the cup, we noticed that inhibition of TAA activity by the alternate TAA substrate Kyn induced premature release of dormancy in M. polymorpha gemmae (Figure 1; He et al., 2011). We previously demonstrated that Kyn treatment of M. polymorpha results in developmental defects opposite to those of exogenous auxin, e.g., hyponastic versus epinastic thallus curvature and decreased versus increased rhizoid number (Flores-Sandoval et al., 2015b). To assess effects of global reduction of TAA activity, 14-d-old wild-type gemmalings were transferred to plates supplemented with 75 or 125 μM Kyn and grown for an additional 3 weeks. These plants exhibited the expected developmental defects, e.g., reduced and hyponastic-like growth and delayed gemmae initiation (Flores-Sandoval et al., 2015b). Plants on 125 μM Kyn had a somewhat reduced growth, but the number of gemma cups was not significantly altered (Supplemental Figure 2A), and most gemma cups developed in a normal fashion (Figures 1A and 1B). We observed that the youngest cups of Kyn-treated plants had less developed gemmae than in the mock-treated plants (Figures 1C and 1D). However, in contrast to mock-treated plants where mature gemmae were mainly dormant, Kyn-treated plants had gemmalings developing inside a majority of their cups, indicating reduced dormancy (Figures 1C and 1D). A dose-dependent loss of dormancy was found when quantifying the ratio between gemma cups with both dormant and nondormant gemmae and cups with only dormant gemmae (Figure 1E).

Figure 1.

Global Reduction of TAA Activity Negatively Affects Dormancy of Gemmae in Gemma Cups.

(A) Fourteen-day-old wild type transferred to mock and grown for an additional 3 weeks.

(B) Fourteen-day-old wild type transferred to 125 μM Kyn and grown for an additional 3 weeks.

(C) Close-up of dashed region in (A).

(D) Close-up of dashed region in (B).

(E) Graph shows the percentage of gemma cups that contain nondormant gemmae.

Dormancy of gemmae in gemma cups was scored in 5-week-old plants (shown in [A] and [B]) grown on supplemented media for the last 3 weeks. Bars show the average of three biological replicates. Error bars indicate se. P values from two-tailed t test are 0.0011 (mock/75) and 0.0002 (mock/125). Numbers in (C) and (D) represent the position and relative age of the gemma cups, with 1 being the gemma cup closest to the apical notch. Bars = 1 cm in (A) and (B).

This encouraged us to do a more detailed investigation of the role for auxin in gemmae dormancy regulation. We first set out to characterize the IPyA auxin synthesis pathway previously described in seed plants (Mashiguchi et al., 2011; Stepanova et al., 2011; Won et al., 2011) and in the M. polymorpha thallus; thereafter we used the obtained tools for manipulation of auxin synthesis, homeostasis, and signaling to study the effect of auxin on dormancy of M. polymorpha gemmae.

The M. polymorpha IPyA Auxin Synthesis Pathway

To verify the presence of the IPyA pathway in liverworts, and to better understand the evolution of genes regulating auxin synthesis, we made an inventory of TAA and YUC family members in the liverwort M. polymorpha. We used known TAA and YUC family members as queries in BLAST searches against M. polymorpha genome (female accession BC4-1C) and transcriptome sequence databases. These searches revealed two TAA/TAR-like genes and several FLAVIN MONOOXYGENASE (FMO)-like genes, of which only two were highly similar to known YUC family members from other species.

To determine orthology, we used PhyML3.0 (Guindon et al., 2010) to construct maximum likelihood (ML) phylogenetic trees from the best scoring BLAST hits aligned with homologous regions of amino acid sequences from the moss P. patens, the lycophyte Selaginella moellendorffii, and angiosperms. These ML analyses revealed that M. polymorpha has a single autosomal Arabidopsis TAA/TAR1/TAR2 ortholog, putatively involved in the production of IPyA from Trp, as well as one At-TAR3/TAR4 ortholog with unknown function (Supplemental Figure 3 and Supplemental Data Set 1).

In male individuals, the autosomal Mp-TAA gene is accompanied by dozens of Mp-TAA copies (position −482 to +6607 relative to the ATG initiator codon) in the heterochromatic region of the M. polymorpha male-specific Y chromosome (annotated as M2D3.1; Ishizaki et al., 2002; Yamato et al., 2007). The putative coding sequence of the Y chromosomal Mp-TAA genes only differs in four nucleotide positions relative to the coding sequence of the autosomal Mp-TAA. The female X chromosome does not appear to encode a homolog (Yamato et al., 2007; Joint Genome Initiative M. polymorpha genome sequencing project). Given that Mp-TAA expression levels did not significantly differ between male and female lines (Ishizaki et al. 2002), and given that no Mp-TAA sequences with the four Y chromosomal-specific polymorphisms could be detected in RNA-seq libraries from either male and female gametophytic tissues, the male-specific copies of Mp-TAA were believed to be inactive.

ML analysis showed that the M. polymorpha genome contains two YUC paralogs, duplicated within the liverwort lineage, putatively catalyzing the conversion of IPyA to IAA (Supplemental Figure 4 and Supplemental Data Set 2). The best M. polymorpha non-YUC plant FMO (locus 10699) resided in a clade of non-YUC FMOs of other land plants with high statistical support (Supplemental Figure 5 and Supplemental Data Set 3). Hence, our analysis shows that the last common ancestor of extant land plants had one TAA and one YUC gene and that M. polymorpha harbors the smallest IPyA-dependent auxin synthesis toolkit identified in any land plant. M. polymorpha TAA and YUC genes were named TAA, YUC1, and YUC2. The M. polymorpha At-TAR3/4 ortholog was named TAR, and we refer to the Y-chromosomal Mp-TAA genes collectively as TAA ON Y-CHROMOSOME (TAY).

Constitutive YUC or TAA Expression Mimics the Effects of Exogenous Auxin Application

To confirm the function of the IPyA-dependent auxin synthesis pathway components in liverworts, we cloned and expressed all three genes, Mp-TAA, Mp-YUC1, and Mp-YUC2, using the constitutive Mp-ELONGATION FACTOR1-α promoter (_proEF1; Ishizaki et al., 2012; Althoff et al., 2014). Exogenous auxin has previously been shown to reduce growth, increase rhizoid formation, and affect dorsiventral patterning, gemma cup development, and air pore/chamber development, as well as to affect curvature of the thallus (Fitting, 1939; Halbsguth and Kohlenbach, 1953; Rousseau, 1950, 1951a, 1951b, 1952, 1953a, 1953b; Kaul et al., 1962; Maravolo and Voth, 1966; Maravolo, 1980; Ishizaki et al., 2012). Constitutive expression of Mp-YUC1 and Mp-YUC2 gave identical phenotypes, highly similar to plants treated with high levels of exogenous auxin (Figures 2A to 2D). Constitutive Mp-TAA expression gave a similar, although considerably milder, phenotype of increased ventral rhizoid initiation and length, as well as slight growth inhibition (Figures 2E to 2H). This phenotype is similar to the effects of low to moderate levels of exogenous auxin (Ishizaki et al., 2012; Flores-Sandoval et al., 2015b). RT-PCR indicated that Mp-TAA transcript levels in the transgenic lines positively correlated with the severity of the phenotype (Figure 2I).

Figure 2.

Mp-YUC and Mp-TAA Overexpression Mimics the Effects of Exogenous Auxin Application.

(A) Wild type (6 weeks old).

(B) _proEF1:YUC2#4 (6 weeks old). Intermediate phenotype.

(C) _proEF1:YUC2#6 (8 weeks old) T1 generation. Strong phenotype.

(D) _proEF1:YUC1#1 (6 weeks old) T1 generation. Intermediate phenotype. Photograph taken from a 45° angle to illustrate the compact growth. Arrow indicates epinastic curvature of thallus.

(E) Wild type (4 weeks old).

(F) _proEF1:TAA#11 (4 weeks old).

(G) Wild type. Close-up of dashed square in (E).

(H) _proEF1:TAA#11. Close-up of dashed square in (F). Light setting in (G) and (H) was optimized to more clearly see the rhizoids.

(I) Eight-day-old gemmalings on standard growth media. Wild type, left panel; _proEF1:TAA#1, middle panel; _proEF1:TAA#18, right panel. Mp-TAA RT-PCR on cDNA from 8-d-old gemmalings is shown in the lower panel. Mp-EF1 was used for normalization.

(J) Nine-day-old gemmalings. Genotype/treatment as indicated in the figure. Mp-YUC2 RT-PCR on cDNA from 9-d-old gemmalings. Mp-EF1 was used for normalization.

(K) Gemma cup from the wild type.

(L) Gemma cup from _proEF1:YUC2#1.

(M) _proEF1:YUC2#9 grown on 250 μM Kyn.

(N) _proEF1:YUC2 _proEF1:TAA double transformant (T1 generation) grown on 250 μM Kyn.

(O) IAA and OxIAA measurements in the wild type and _proEF1:YUC1#1.

Graph shows the average of four replicates. Error bars show se. Two-tailed t test gave P values of 0.04 (IAA) and 0.003 (OxIAA). (A) to (C) are the same scale. Bars = 1 cm in (A), (D) to (F), and (M), 5 mm in (I), (J), and (N), and 1 mm in (K) and (L).

Mp-EF1-regulated Mp-YUC1 and Mp-YUC2 expression resulted in a much more severe phenotype compared with that seen for the _proEF1:TAA lines (Figures 2B to 2D). In young _proEF1:YUC gemmalings, growth inhibition, developmental defects, and rhizoid overproduction were observed, similar to that of auxin-treated wild type (Figure 2J). As in _proEF1:TAA lines, the strength of the phenotype was positively correlated with Mp-YUC2 transcript levels, as estimated by RT-PCR (Figure 2J). Adult _proEF1:YUC1 and _proEF1:YUC2 thalli were smaller and appeared darker than the wild type (Figure 2C). All lines also had epinastic curvature of thalli, making the plant appear to be more compact and higher, relative to its width, than the wild type (Figures 2A and 2D). Ectopic Mp-YUC1 or Mp-YUC2 expression also led to irregular gemma cup formation with many gemma cups developing as narrow, elongate tubes (Figures 2K and 2L). The older plants, as well as younger plants displaying the strongest phenotypes, showed signs of hyperpigmentation, as previously seen in P. patens gametophytic tissues with increased endogenous IAA levels (Figure 2C; Eklund et al., 2010). All developmental abnormalities described above for the _proEF1:YUC lines were also seen in wild-type M. polymorpha treated with moderate to high levels of exogenous auxin, at various stages of development (Fitting, 1939; Halbsguth and Kohlenbach, 1953; Rousseau, 1950, 1951a, 1951b, 1952, 1953a, 1953b; Kaul et al., 1962; Maravolo and Voth, 1966; Maravolo, 1980; Ishizaki et al., 2012).

To maximize the IAA production from the IPyA pathway, we transformed wild-type sporelings with both _proEF1:YUC2 and _proEF1:TAA constructs. However, a majority of these double transformants did not survive when plated on standard selection medium. Only when plating sporelings were on 250 μM Kyn were we able to isolate 11 highly pigmented double transformants (Figures 2M and 2N). These plants did not survive, ceasing growth and dying after a few weeks on Kyn, and even more rapidly when transferred to standard plates. Because application of exogenous auxin severely inhibits growth of gemmalings and causes wilting at high concentrations (Ishizaki et al., 2012), our observation suggests that the high endogenous auxin synthesis triggered by Mp-TAA and Mp-YUC overexpression was severely inhibiting growth and causing lethality.

To confirm a connection between phenotype and an increased IAA production, we measured the levels of IAA and the irreversibly oxidized degradation product 2-oxindole-3-acetic acid (OxIAA; Reinecke and Bandurski, 1983; Pencík et al., 2013; Tanaka et al., 2014) in the _proEF1:YUC1#1 line displaying a severe phenotype (Figure 2D). The _proEF1:YUC1#1 line had significantly increased levels of both IAA and OxIAA compared with the wild type (Figure 2O), just as previously described for Arabidopsis lines overexpressing At-YUC1 from the 35S promoter (Novák et al., 2012). Our metabolite measurements thus confirm overexpression of Mp-YUC genes leads to increased IAA synthesis in the M. polymorpha thallus.

We hypothesized the phenotype caused by constitutive Mp-TAA activity, i.e., reduced growth and increased rhizoid formation, to be the result of an Mp-TAA-dependent increase of IAA production. To test this, we transferred 7-d-old wild type, and several independent _proEF1:TAA lines, from standard growth media to media supplemented with 100 and 500 μM Kyn. All transgenic lines were essentially insensitive to 500 μM Kyn (Supplemental Figure 6A), whereas the wild type responded as expected by slightly decreased growth, gemma cup fusions, and upward bending of the thallus, resembling hyponastic curvature of leaves on 100 μM Kyn and by even more extreme growth reduction, curvature, and reduced and abnormal gemma cup formation on 500 μM Kyn.

If IPyA levels become rate limiting when Mp-YUC1 or Mp-YUC2 is ectopically expressed, then Kyn treatment of _proEF1:YUC2 lines should result in plants gradually becoming more similar to the wild type when subjected to increasing levels of Kyn. Already at 100 μM Kyn, both mild and severe _proEF1:YUC2 lines showed a less dramatic auxin overproduction phenotype, e.g., more wild type-like growth and branching pattern and by regularly producing functional gemma cups. At 250 μM Kyn the transgenic lines behaved almost like the wild type, suggesting IPyA availability (TAA function) becoming rate limiting for IAA production (Supplemental Figure 6B).

Sites of IAA Synthesis in Thalli of M. polymorpha

To characterize the spatial expression domains of the two Mp-YUC genes and Mp-TAA, we performed RT-PCRs on cDNA from thallus tissue or developing and mature sporophytes. The RT-PCRs showed that Mp-TAA and Mp-YUC2 transcripts could be detected in both samples, while Mp-YUC1 transcripts only could be detected in sporophytes (Figure 3A). This experiment suggests Mp-YUC1 is sporophyte specific, while Mp-YUC2 has a broader expression pattern.

Figure 3.

Mp-TAA and Mp-YUC2 Have Partially Overlapping Expression Domains in the Thallus.

(A) Mp-TAA, Mp-YUC1, and Mp-YUC2 RT-PCRs on cDNA from thallus (Th) and sporophyte (Sp) tissues. Mp-EF1 was used for normalization.

(B) _proTAA:GUS#5 expression in adult thallus.

(C) to (F), (H) to (K), and (N) _proYUC2:GUS#6 expression.

(C) One-day-old gemmaling. Arrowheads indicate the two apical notches. Asterisk indicates the position the connecting stalk cell had before release of the gemmae.

(D) Two-day-old gemmaling. Asterisk indicates the position the connecting stalk cell had before release of the gemmae.

(E) Seven-day-old gemmaling with four apical notches.

(F) Adult thallus, dorsal side. Arrowheads indicate two developing gemma cups.

(G) Section through air chamber of _proYUC2:GUS#3.

(H) Young developing gemma cup. Close up of dashed section in (F).

(I) Mature gemma cup wall.

(J) Adult thallus, ventral side. Arrowheads indicate apical notches.

(K) Close-up of apical notch in adult plant, seen from ventral side.

(L) Thallus margins of _proYUC2:GUS#9. Two partial gemma cups are seen on the dorsal side along the midrib (lower side of panel).

(M) Developing gemma of line _proYUC2:GUS#9, still attached to gemma cup when stained.

(N) Dormant gemmae of line _proYUC2:GUS#6. Arrowheads indicate three notches from two different dormant clonal gemmae.

(O) IAA levels in different tissues of 3-week-old Tak-1 (Japanese wild-type strain).

Graph shows the average of three biological replicates. Error bars show sd. Bars = 1.5 mm in (B), 0.5 mm in (C), (D), and (K), 1 mm in (E), 2 mm in (F), (J), and (L), 100 μm in (G), and 0.25 mm in (M) and (N).

Between the ATG of Mp-YUC1 and the end of the next upstream transcript there is 3.1 kb of noncoding sequence, of which at least 117 bp is the 5′ untranslated region (UTR) of Mp-YUC1. To further characterize the expression pattern of Mp-YUC1, we cloned a 2.9-kb fragment including the full 5′UTR of Mp-YUC1 and transcriptionally fused it with the coding sequences of the GUS reporter gene encoding a β-glucuronidase (Jefferson et al., 1987). As expected, all transgenic lines analyzed completely lacked GUS signal in the thallus and male and female gametangiophores. Even though our results do not completely rule out Mp-YUC1 function outside the sporophyte, these results indicate that the gene determining the location of gametophytic IPyA-dependent IAA synthesis in M. polymorpha is Mp-YUC2. Our expression data thus suggest there are single genes for both TAA and YUC families that are active during M. polymorpha thallus development.

Previous studies have suggested that IAA is produced in the apical region of liverwort and fern thalli and is subsequently transported through the thallus to other tissues and organs in an apical-basal manner (Albaum, 1938; LaRue and Narayanaswami, 1957; Maravolo, 1976). To determine the spatial localization of auxin synthesis during gametophyte development, we made _proTAA:GUS and _proYUC2:GUS constructs. The genomic fragments used included a 5.0-kb genomic region containing the promoter (4.1 kb), the putative 5′UTR (0.5 kb), and a part of the first exon (0.4 kb) of Mp-TAA, as well as a 4.8-kb promoter region and the putative 5′UTR (0.7 kb), directly upstream of the ATG of Mp-YUC2. We analyzed more than 40 independent transgenic lines, of each construct, of which a majority showed the same spatial expression domains, with subtle differences only in signal strength.

_proTAA:GUS was mainly detected in apical notches (meristematic regions) and at the bases of gemma cups, with much weaker signal along the midrib (Figure 3B). The potential length of the Mp-TAA promoter region (14 kb to the next upstream predicted open reading frame) and the spatially restricted _proTAA:GUS pattern in the thallus prompted us to analyze _proTAA:GUS expression also in other tissues and organs where we would expect strong signal, such as the sporophyte (Figure 3A), to verify that our construct revealed all expected expression domains. _proTAA:GUS gave signal in antheridiophores and archegoniophores (male and female gametangiophores, respectively) and the signal was high in the foot of developing sporophytes, similar to the expression pattern of the Mp-IAA auxin signaling repressor (Supplemental Figure 7; Kato et al., 2015). This strengthens the notion that the GUS signal reflects the endogenous expression domains of Mp-TAA in the thallus.

In young gemmalings, _proYUC2:GUS expression was detected mainly in the apex, but weaker expression was found in the rhizoid initiation zone (Figures 3C and 3D) and in cells adjacent to the stalk cell previously connecting the gemma to the maternal tissue in the base of the gemma cup (Figure 3C). As a gemmaling grows, air chambers develop on the dorsal surface and a midrib is evident behind the apical meristem. Along the midrib, water-conducting smooth rhizoids develop on the ventral side (Bowen, 1935; McConaha, 1939, 1941). In 7-d-old gemmalings, Mp-YUC2 expression had increased in the rhizoid initiation zone stretching along the midrib, and the apical signal had also increased in intensity compared with earlier stages (Figures 3C to 3E). On the dorsal surface of mature plants Mp-YUC2 expression was mainly found in the apex, along the midrib, in the air chambers/pores, and to some extent in thallus margins (Figure 3F). A clear signal was also detected in cells constituting the wall of the air chamber and air pore and in cells of the assimilatory filaments inside the air chambers (Figure 3G; Supplemental Figures 8A and 8B). A strong signal was detected throughout newly formed and developing gemma cups (Figure 3H), but as gemma cups matured the signal weakened and was spatially restricted toward the base, where continued meristematic activity is located (Douin, 1923; Figure 3I). On the ventral side, no expression was detected in rhizoids or scales, but the apical signal remained strong and a weak signal was detected in cells of the midrib connecting to rhizoid cells (Figures 3J and 3K; Supplemental Figure 8B). In adult thalli, a region directly encircling the apical notch did not show any staining, except for in developing gemma cups within this region (Figures 3H and 3I). A weak signal was sometimes detected in older parts of the thallus margins (Figure 3L).

In developing gemmae still attached to the gemma cup, strong signal for _proYUC2:GUS was detected in the stalk cell (Figure 3M). Weaker signal was, to a varying degree, found along the margins and cells adjacent to the stalk cell. No signal specific for the apical notch could be detected (Figure 3M). In mature gemmae collected from gemma cups with no visible rhizoids or elongating gemmae, there was no detectable _proYUC2:GUS signal, except for a very weak signal rarely found in a few cells in the apical notch and in cells adjacent to the stalk cell (Figure 3N). Developing gemmae of _proTAA:GUS lines gave an almost identical signal to that of _proYUC2:GUS, with the majority of signal found in margins close to the base and the stalk cell (Supplemental Figure 9). No _proTAA:GUS signal could be detected in dormant gemmae, except for the very weak signal sometimes found in cells adjacent to the stalk cell (Supplemental Figure 9), suggesting that dormant gemmae produce no or very little IAA via the IPyA pathway.

To ascertain whether _proYUC2:GUS activity depicts sites of relatively high auxin concentrations, we performed measurements of IAA levels in 3-week-old wild-type thalli dissected according to the _proYUC2:GUS pattern into four samples: apical notch, midrib, gemma cup, and thallus margin. We found that IAA levels in wild-type tissues correlated with the signal intensity and pattern of _proYUC2:GUS expression, being significantly highest in the apical notch (Figure 3O). There were no significant differences between IAA levels in gemma cups and midrib, but margins had significantly lower IAA levels.

Loss of the IPyA Pathway Results in Loss of Cell and Tissue Differentiation

To obtain tools to reduce auxin synthesis in the thallus, we used several different strategies in an attempt to either completely eliminate, or reduce, the contribution of the IPyA pathway to the total IAA pool in the gametophyte of M. polymorpha. To create a complete knockout of the IPyA pathway, we employed gene disruption/replacement by homologous recombination of Mp-TAA. Because we speculated a complete loss of the IPyA pathway would be lethal, we also attempted to knock down the IPyA pathway throughout the thallus by employing two artificial microRNAs (amiRs) targeting either Mp-TAA or Mp-YUC2 and regulated by the Mp-EF1 promoter.

By homologous recombination, deleting a 378-bp fragment of Mp-TAA exon 2, and disrupting the remaining coding region by insertion of a 3.6-kb hygromycin resistance cassette (Supplemental Figure 10), we obtained four stable lines with severe and indistinguishable defects in growth and development (Figures 4A and 4D). All four lines were molecularly verified to have predicted 5′ and 3′ recombination events (Supplemental Figure 10). We therefore regarded the four lines as independent taa knockout lines. Treating the taa lines with 0.1 μM IAA resulted in partial rescue of the phenotype (Figure 4A). When one line (taa#190) was transformed with a genomic Mp-TAA fragment, the aberrant phenotype was complemented. The complete recovery of the obtained taa#190/TAA transformants, and the partial rescue of taa on exogenous IAA, demonstrated that the taa phenotype is due to loss of Mp-TAA function (Figures 4A and 4B).

Figure 4.

Knockout of the Mp-TAA Gene.

(A) Tak-1 (wild type) grown from gemmae and Mp-TAA knockout line taa#190, T1 generation. Plants were grown for 2 weeks on media supplemented with mock or 0.1 μM IAA as indicated in the figure.

(B) Two independent lines from a transformation event where taa#190 was transformed with a genomic Mp-TAA fragment. Plants were grown for 2 weeks on mock media as in (A).

(C) Wild-type gemma grown on 500 μM Kyn for 5 weeks. Asterisks indicate the growing apical regions. Circle outlines the size of the dormant gemma before breaking dormancy. Inset is a scanning electron micrograph of the apical region in the dashed square.

(D) and (E) Scanning electron micrographs of 3-week-old plants. taa#190 (D) and dorsal side of a wild-type thallus (E).

(F) _proEF1:iaaL#102. Side view.

(G) _proEF1:iaaL#103 seen from above.

(H) IAA and OxIAA levels in Tak-1, the taa#190 mutant, and the complemented mutant taa#190/TAA#1.

Graph shows the average of four replicates. Error bars show se. Two-tailed t test gave a P value of 0.0005 (OxIAA, wild type versus taa#190). Bars = 5 mm in (A) and (B), 1 mm in (C), (F), and (G), and 0.5 mm in (D) and (E).

All taa lines grew very slowly and had no distinguishable dorsal or ventral structures, and, although some rhizoids were formed, the plants consisted mainly of a smooth undifferentiated cell mass (Figures 4A and 4D). We observed similar tissues in apical regions of wild-type gemmae placed directly on very high concentrations (500 μM) of the TAA inhibitor Kyn (Figure 4C). In Kyn-treated wild-type gemmae, we observed a similarly slow growth: 5-week-old Kyn-treated plants had a diameter of 5 mm instead of the 10 cm observed for mock-treated wild type. Treatment with 500 μM Kyn also led to a severe reduction of both dorsal structures and rhizoids (Figure 4C; Flores-Sandoval et al., 2015b). The taa plants appeared almost spherical with small and seemingly unorganized protrusions, growing equally in all directions. We have never observed such a strong phenotype in the Kyn-treated wild type, but strong _proEF1:iaaL lines displayed a similar, albeit less severe, phenotype (Figures 4F and 4G; Flores-Sandoval et al., 2015b). These results are in line with our previous observations suggesting auxin promotes differentiation in liverworts (Flores-Sandoval et al., 2015b; Kato et al., 2015).

The M. polymorpha Y chromosome has a region of repeated Mp-TAA copies (TAY). While we initially speculated these loci to be completely inactive, they could potentially provide residual TAA function in the event of a knockout of the autosomal Mp-TAA gene. In an attempt to resolve if the obtained taa lines were independent of any Y chromosomal activity, we determined the sex of the taa lines. All four taa lines obtained were male, suggesting that our taa lines may depend on residual TAA function provided by TAY.

We measured IAA and OxIAA levels in the taa#190 mutant and the complemented mutant (taa#190/TAA#1; Figure 4H). taa#190 contained the same IAA levels, per gram of tissue, as those of the wild type and the complemented mutant. As expected, no significant differences in OxIAA levels were detected between the wild type and the complemented mutant. However, OxIAA levels were significantly increased in the taa#190 mutant compared with the other lines.

An amiR targeting Mp-TAA, using the Mp-miR160 stem-loop backbone (Flores-Sandoval et al., 2015a), was designed and constitutively expressed throughout the thallus (_proEF1:amiR-TAA^miR160; Supplemental Figure 11A). From several independent transformation events, more than 20 lines with conspicuous and similar phenotypes, but with variations in severity, were selected for analysis. 5′ RNA ligase-mediated rapid amplification of cDNA ends (RLM-RACE) on two independent lines established that the expected microRNA-specific degradation product could be detected (Supplemental Figure 11B; Alvarez et al., 2006). IAA and OxIAA measurements confirmed that the _proEF1:amiR-TAA^miR160#1 line had a profile identical to the taa#190 mutant, further verifying that the same process has been targeted in these different lines (Supplemental Figure 12). The most severe _proEF1:amiR-TAA^miR160 lines were clearly different from the wild type and similar to strong _proEF1:iaaL lines or taa knockouts; they displayed a spherical shape due to hyperbranching and/or callus-like growth, complete reduction of air pore/chamber development, and inability to produce gemma cups or scales (Figures 5A, 5B, and 5D to 5G). No clear dorsal or ventral structures were observed, and although rhizoids were formed, they were less numerous than on the ventral side of the wild type and the rhizoids were evenly distributed over _proEF1:amiR-TAA^miR160 plant surfaces and not concentrated to a ventral midrib as in the wild type (Figures 5D to 5G). The less severe lines were much smaller than the wild type but had clear thalloid growth and thus grew in a more wild type-like manner than the most severe _proEF1:amiR-TAA^miR160 lines (Figures 5B, 5C, 5H, and 5I); they displayed increased branching with narrow thalli, reduced air pore development, and almost complete reduction of air chamber development including the complete loss of assimilatory filaments (Figures 5B, 5C, 5H, 5I, 5L, and 5M). The few gemma cups that were formed did not elongate, but produced misshapen gemmae that appeared to lack dormancy and instead developed into gemmalings inside the gemma cup (Figures 5I, 5K, and 5L). Also, rhizoid development appeared to be reduced along the midrib of the ventral side (Figure 5J).

Figure 5.

Constitutive Expression of a miR Targeting Mp-TAA/TAY Phenocopies the taa Mutant.

(A) Wild type (3 weeks old).

(B) _proEF1:amiR-TAA^miR160#1 and _proEF1:amiR-TAA^miR160#101 (inset). Both plants are 2 months old and are shown in same scale.

(C) Thalli of _proEF1:amiR-TAA^miR160#2.

(D) to (M) Scanning electron micrographs.

(D) _proEF1:amiR-TAA^miR160#101.

(E) Close up of dashed square in (D).

(F) _proEF1:amiR-TAA^miR160#5 (2-month-old plant).

(G) Close-up of dashed square in (F).

(H) Thallus from _proEF1:amiR-TAA^miR160#1.

(I) and (J) Thallus of _proEF1:amiR-TAA^miR160#2. Dorsal side (I) and ventral side (J).

(K) Close-up of a developing gemma cup shown in (I).

(L) Wild-type gemma cup and air pores with protruding assimilatory filaments (arrowheads).

(M) Air pores from the old part of _proEF1:amiR-TAA^miR160#2 thallus, without visible assimilatory filaments inside air pores (arrowheads).

(N) Mp-TAA and TAY RT-PCR on cDNA from male wild type and three independent _proEF1:amiR-TAA^miR160 lines. Mp-EF1 was used for normalization.

Bars = 1 cm in (A) and (B), 5 mm in (C), 1.75 mm in (D), 250 μm in (E), 1.5 mm in (F) and (H), 0.5 mm in (G) and (M), 0.85 mm in (I), 0.75 mm in (J), 175 μm in (K), and 0.6 mm in (L).

In an attempt to bring clarity to whether reduced Mp-TAA activity could induce TAY activity, thereby facilitating survival of male Mp-TAA loss-of-function lines, we performed RT-PCRs using primers with bias for either Mp-TAA or TAY on male wild type and three independent _proEF1:amiR-TAA^miR160 lines, of which two had the intermediate phenotype (Figures 5B and 5H) and one the strong phenotype (Figures 5B and 5D). Because of the similarities between Mp-TAA and TAY, only the 3′-most nucleotide of the primers differed between Mp-TAA and TAY biased primers (Supplemental Table 1). The amplicons spanned two additional nucleotide differences, as well as the amiR-TAA binding site and several intron splice sites. All RT-PCRs were run on a gel (Figure 5N) and were also directly sequenced (Supplemental Figure 11D). RT-PCR on male wild type with Mp-TAA biased primers resulted in the amplification of only Mp-TAA according to the sequencing chromatogram (Supplemental Figure 11D), while TAY biased primers clearly amplified both Mp-TAA and TAY (Supplemental Figure 11D). We also performed PCRs on male wild type using unbiased primers. Chromatograms from direct sequencing of this PCR product showed no TAY (Supplemental Figure 11D), suggesting TAY is expressed at negligible levels in the male wild type. In the three _proEF1:amiR-TAA^miR160 lines, Mp-TAA biased primers amplified only Mp-TAA, and TAY biased primers amplified only TAY (Supplemental Figure 11D). The RT-PCRs thus indicate that Mp-TAA is downregulated in _proEF1:amiR-TAA^miR160 lines, while TAY is upregulated in the two less severe lines but is found at similar levels as the wild type in the severe line analyzed (Figure 5N). The amiR-TAA binding site was placed over an Mp-TAA/TAY nucleotide difference, allowing us to detect degraded TAY in the RACE PCR. As seen in Supplemental Figure 11B, degraded Mmp-TAA and TAY could be detected in the chromatogram from direct sequencing of the 5′ RLM-RACE product. Our data thus show that amiR-TAA^miR160 can degrade both Mp-TAA and TAY transcripts, although the efficiency for TAY might be lower than for Mp-TAA.

Auxin synthesis was also knocked down by constitutively expressing an amiR targeting Mp-YUC2 (_proEF1:amiR-YUC2^miR160; Supplemental Figure 11A). Thirteen independent transformants with three categories of phenotypes were found and examined. Differences in Mp-YUC2 transcript levels between strong, intermediate, and weak lines were detected using RT-PCR, suggesting a correlation between phenotype and amiR activity (Figure 6A). Among _proEF1:amiR-YUC2^miR160 lines, one was very similar to strong _proEF1:amiR-TAA^miR160 or _proEF1:iaaL lines. Two transformants were almost phenotypically wild type, and the remaining 10 had an intermediate phenotype (Figures 6B to 6D). The intermediate phenotype included increased branching, reduced air pore/air chamber formation, hyponastic-like curvature, upward folding of thallus margins, and abnormalities in gemma cup and scale development (Figures 6F to 6K). A few _proEF1:amiR-YUC2^miR160 lines managed to produce gemma cups that failed to develop normally. These gemma cups did not elongate but developed serrated marginal tissue, normally found on the distal gemma cup rim (Figures 6F and 6I). Inside these cups gemmae developed like those in cups of _proEF1:amiR-TAA^miR160 lines.

Figure 6.

Global Reduction of Mp-YUC2 Transcripts by an Artificial miR.

(A) Mp-YUC2 RT-PCR on cDNA from three independent _proEF1:amiR-YUC2^miR160 lines. Mp-EF1 was used for normalization.

(B) _proEF1:amiR-YUC2^miR160#6. Intermediate phenotype.

(C) _proEF1:amiR-YUC2^miR160#10. Strong phenotype.

(D) _proEF1:amiR-YUC2^miR160#12. Weak phenotype.

(E) Wild-type thallus. Dorsal side.

(F) to (J) Scanning electron micrographs.

(F) Wild-type thallus. Dorsal side. Apical region with gemma cup.

(G) and (H) Dorsal side of _proEF1:amiR-YUC2^miR160#6 thalli.

(G) Young thallus.

(H) Old thallus. Arrowhead indicates a gemmae cup.

(I) Close-up of gemma cup shown in (H). Arrowheads indicate tissue normally found in the gemma cups rim (compared with [F]). Arrows indicate gemmae at an early stage of development.

(J) Wild-type thallus. Ventral side. Apical region. Rhizoids were deliberately removed to better visualize the scales.

(K) Ventral side of _proEF1:amiR-YUC2^miR160#6. Rhizoids were not removed.

Bars = 7 mm in (B), 2 mm in (C) to (E), 1 mm in (F), (J), and (H), 1.75 mm in (G) and (K), and 150 μm (I).

Because Mp-YUC2 is highly expressed in the meristematic regions of thalli, we used the apical notch-specific Mp-SHORT INTERNODES (SHI) promoter (Flores-Sandoval et al., 2015b) to study the effects of reduction of Mp-YUC2 levels in the meristem. As with Mp-EF1-regulated expression of amiR-YUC2^miR160, _proSHI:amiR-YUC2^miR160 transformants displayed a range of phenotypes suggesting variations in reduction of notch specific auxin synthesis. The three strongest lines were similar to the strong _proEF1:amiR-YUC2^miR160 line and also resembled strong or intermediate _proEF1:amiR-TAA^miR160 lines or _proEF1:iaaL (Supplemental Figures 13A and 13B). However, more than 20 weak and intermediate lines were more similar to the wild type, but still with a distinct phenotype different from weak _proEF1:amiR-YUC2^miR160 lines (Supplemental Figures 13A and 13B). Weak and intermediate _proSHI:amiR-YUC2^miR160 lines developed air chambers, gemma cups, and rhizoids similar to the wild type, but as with intermediate _proSHI:iaaL lines, many apical notches of intermediate or weak _proSHI:amiR-YUC2^miR160 lines became pigmented and ceased growth (Supplemental Figure 13C; Flores-Sandoval et al., 2015b). 5′ RLM-RACE preformed on _proSHI:amiR-YUC2^miR160#3 (intermediate line) produced the expected microRNA-specific degradation product (Supplemental Figure 11C).

Reduced Auxin Synthesis or Levels in the Apical Notch Result in Loss of Gemmae Dormancy

In all gemma cup-producing transgenic lines with reduced expression of genes encoding IPyA-mediated auxin biosynthesis enzymes, or ectopic iaaL expression, described above, we also found another striking phenotype: Gemmae in gemma cups did not properly induce and/or maintain dormancy, just as previously described for the wild type treated with Kyn (Figure 1). The similarity between the transgenics and our pharmacological results encouraged us to perform a more detailed examination of dormancy in transgenic lines expressing iaaL (_proEF1:iaaL and _proSHI:iaaL) and in those where Mp-YUC2 transcript levels were altered in the apical meristem (_proSHI:amiR-YUC2^miR160).

Intermediate _proSHI:amiR-YUC2^miR160 lines resembled weak _proEF1:amiR-YUC2^miR160 lines (Figures 6D, 7C, and 7D) and both produced gemma cups. The most severe _proEF1:iaaL, _proSHI:iaaL, _proSHI:amiR-YUC2^miR160, and all taa, _proEF1:amiR-TAA^miR160, and _proEF1:amiR-YUC2^miR160 lines previously described by us (this article; Flores-Sandoval et al., 2015b) did not produce any, or only rarely produced, gemmae or gemma cups. Thus, we focused on _proEF1:iaaL, _proSHI:iaaL, and _proSHI:amiR-YUC2^miR160 lines with weaker phenotypes, all of which grew slowly, producing fewer gemma cups than the wild type, although development of initiated gemma cups proceeded as in the wild type (Supplemental Figure 2B). Examination of the youngest gemma cups in 7-week-old plants revealed that gemmae dormancy in all transgenic lines was affected in a manner similar to that of Kyn-treated wild type (Figures 1C, 1D, and 7A to 7D). Quantification of the ratio between nondormant and dormant gemma cups revealed significantly reduced dormancy for all lines analyzed (Figure 7E). These results suggest that reduction of auxin levels in the apex of the thallus is sufficient to reduce gemmae dormancy in proximal gemma cups.

Figure 7.

IAA Produced in the Thallus Apex Regulates Dormancy of Gemmae in Gemma Cups.

(A) Wild-type thalli.

(B) Apical region of the wild type.

(C) Thalli of _proEF1:amiR-YUC2^miR160#4.

(D) Close-up of dashed region in (C).

(E) Graph shows the percentage of gemma cups that contain nondormant gemmae.

Dormancy of gemmae in gemma cups was scored in 5-week-old plants. Bars show the average of three biological replicates. Error bars indicate se. P values from two-tailed t test are <0.001 for all comparisons with the wild type. Numbers in (A) to (D) represent the position and relative age of the gemma cups, with 1 being the gemma cup closest to the apical notch. Asterisks indicate the position of apical notches. Bars = 5 mm in (A) and (C).

ARF-Mediated Auxin Signaling Has a Positive Effect on Gemmae Dormancy

To examine whether prolonged dormancy could be obtained by positively manipulating auxin signaling, we constitutively expressed a truncated version of the single M. polymorpha class A ARF, Mp-ARF1, as well as two different Mp-miR160-based amiRs both targeting the single Mp-IAA gene (amiR-IAA⁷ and amiR-IAA⁹; Flores-Sandoval et al., 2015b). Mp-ARF1 is an ortholog of At-ARF5/MONOPTEROS and acts as the sole transcriptional activator ARF in M. polymorpha (Flores-Sandoval et al., 2015b; Kato et al., 2015). ARF proteins have domains (three and four) that interact with homologous domains in AUX/IAAs, linking the DNA binding ARF to the TOPLESS transcriptional repressor (Szemenyei et al., 2008). Since a truncated ARF1 (Mp-ARF1ΔD34) cannot interact effectively with AUX/IAA (Krogan and Berleth, 2012), _proEF1:ARF1ΔD34 lines should be insensitive to reduced auxin levels; accordingly, their growth was unaffected when treated with Kyn (Figures 8A and 8B; Flores-Sandoval et al., 2015b). Although the morphology and development of _proEF1:ARF1ΔD34 lines recently have been described by us in detail (Flores-Sandoval et al., 2015b), we have previously not reported on any dormancy effects. All _proEF1:ARF1ΔD34 plants were small and highly branched with slightly increased gemma cup production compared with wild-type plants (Flores-Sandoval et al., 2015b; Supplemental Figure 2C). All analyzed _proEF1:ARF1ΔD34 lines maintained gemmae dormancy significantly longer than the wild type grown under identical conditions (Figures 8A, 8C, and 8D). The prolonged dormancy of _proEF1:ARF1ΔD34 was also maintained in the presence of 125 μM Kyn (Figures 8B and 8D), suggesting the modified class A ARF encoded by the transgene was responsible for the increased dormancy in these lines.

Figure 8.

Mp-ARF1-Mediated Auxin Signaling Positively Affects Gemmae Dormancy.

(A) and (B) Fourteen-day-old _proEF1:ARF1D34#1 transferred to mock (A) and 125 μM Kyn (B) and grown for an additional 3 weeks.

(C) Close-up of dashed region in (B). Numbers indicates relative gemma cups age, where 1 is the most apical (youngest) gemma cup.

(D) to (F) Graphs show the percentage of gemma cups in transgenic lines that contain nondormant gemmae. Bars show the average of three or more biological replicates. All error bars show se.

(D) Five-week-old wild-type and _proEF1:ARF1ΔD34#3 plants grown on supplemented media for the last 3 weeks. P values from two-tailed t test are 0.00006 (wild type mock/wild type Kyn), 0.005 (wild type mock/Mp-ARF1 mock), and 0.93 (Mp-ARF1 mock/Mp-ARF1 Kyn).

(E) Six-week-old wild-type, _proEF1:amiR-IAA⁷#1, and _proEF1:amiR-IAA⁹#2 plants grown on standard growth media. P values from two-tailed t test are 0.038 (wild type/⁹#2) and 0.010 (wild type/⁷#1).

(F) Five-week-old wild-type and _proEF1:amiR-ARF1^miR160 plants grown on standard growth media. P value from two-tailed t test is <0.001.

Plants in (A) and (B) can be directly compared with the wild type shown in Figures 1A and 1B. Bars = 1 cm in (A) and (B).

Increased auxin sensitivity has previously also been observed for plants constitutively expressing _proEF1:amiR-IAA⁷ and _proEF1:amiR-IAA⁹, targeting the single Mp-IAA gene. These lines are phenotypically similar to the wild type under standard growth conditions but are auxin hypersensitive (Flores-Sandoval et al., 2015b). Line _proEF1:amiR-IAA⁹#2 has a stronger auxin hypersensitivity phenotype than line _proEF1:amiR-IAA⁷#1, an observation that held true also for gemmae dormancy as we could detect slightly increased dormancy in line _proEF1:amiR-IAA⁷#1 compared with the wild type, and to a higher degree in _proEF1:amiR-IAA⁹#2 (Figure 8E). No significant difference in gemma cup production was observed between wild-type and _proEF1:amiR-IAA lines (Supplemental Figure 2D).

To confirm that manipulation of auxin signaling in the thallus notch or developing gemmae did not result in reduced or increased dormancy of mature gemmae, we also expressed ARF1ΔD34 using the Mp-SHI promoter. Six independent _proSHI:ARF1ΔD34 lines with slightly increased branching were found to have wild type-like Kyn sensitivity, suggesting that ARF1ΔD34 expression in the cup is necessary to enhance dormancy (Supplemental Figure 14).

To analyze if reduced Mp-ARF1 activity leads to reduced gemmae dormancy, we used a previously designed amiR that specifically targets the transcript of Mp-ARF1 (Flores-Sandoval et al., 2015a). Reduced dormancy was observed in all analyzed _proEF1:amiR-ARF1^miR160 lines, with >80% of gemma cups in 5-week-old plants containing nondormant gemmae (Figure 8F; Supplemental Figure 2E). Taken together, our results demonstrate that the auxin-regulated transcriptional activator Mp-ARF1 acts as a positive regulator of gemmae dormancy.

DISCUSSION

The M. polymorpha Genome Contains the Smallest Known Toolkit for the IPyA Auxin Synthesis Pathway in a Land Plant

In our inventory of auxin synthesis components in the liverwort M. polymorpha, we found only three autosomal genes, one TAA and two YUC, with functions in the IPyA pathway. However, this low number is not surprising considering that very small gene family sizes appears to be the norm for genes regulating development in M. polymorpha (Kanazawa et al., 2013; Ueda et al., 2013; Komatsu et al., 2014; Kubota et al., 2014). Mp-TAA, Mp-YUC1, and Mp-YUC2 are structurally typical TAA and YUC family members. Our functional studies studying loss- and gain-of-function alleles leave little doubt their biochemical functions in the IPyA pathway are conserved between liverworts and angiosperms, as phenotypes of transgenic plants mimic those obtained by exogenous auxin treatment or global reduction of auxin levels. As expected, gain-of-function Mp-YUC alleles resulted in increased IAA and OxIAA levels in the M. polymorpha thallus, supporting a role for Mp-YUCs in a rate-limiting step in the IPyA pathway. The notion of YUCs being rate-limiting in the gametophyte is further supported by the weak auxin-related phenotype seen in _proEF1:TAA lines, suggesting IPyA overproduction has no toxic effects. Given that the effect of Mp-TAA overexpression was much milder than Mp-YUC overexpression, conversion from IPyA to IAA by Mp-YUC2 seems to be the rate-limiting step for this pathway during normal M. polymorpha thallus development.

We have shown that M. polymorpha has the smallest known toolkit for the IPyA auxin synthesis pathway in any land plant genome. The presence of so few paralogs in M. polymorpha, together with the many recently obtained tools such as Agrobacterium tumefaciens-mediated transformation, sexual and asexual propagation, and the use of CRISPR/Cas9, homologous recombination-mediated gene deletion/disruption, and amiRNAs to create loss-of-function alleles (Ishizaki et al., 2013; Sugano et al., 2014), makes this species the ideal system to investigate the effects of auxin synthesis on development and growth in basal land plants.

The IPyA Pathway Is Essential for Basal Land Plants

Previous studies have shown that genes encoding IPyA pathway components exist also in the genome of the moss P. patens (Rensing et al., 2008), and our results reveal the functional significance of the IPyA-dependent auxin synthesis pathway in a bryophyte. Previous reports of IAA metabolism in several major divisions of land plants suggested that IAA synthesis in a selection of non-seed plants is performed through a Trp-independent pathway (Sztein et al., 2000). The authors presented data supporting the hypothesis that IAA is synthesized in the apex of charophytes and liverworts independently of Trp. However, our functional studies of IPyA pathway components in M. polymorpha show that manipulation of the IPyA synthesis pathway dramatically impairs plant development, mimicking the effects of increased irreversible IAA conjugation (_proEF1:iaaL). This conclusively demonstrates the conserved and critical role of the Trp-dependent IPyA pathway in tissue and organ development of the gametophyte in the complex thalloid liverwort M. polymorpha and suggests the role of the IPyA pathway is conserved throughout the land plant lineage.

Considering TAY was upregulated in _proEF1:amiR-TAA^miR160 lines with reduced Mp-TAA levels, it may be that TAA function can be supplied from the male sex chromosome. While all obtained Mp-TAA knockout lines were male, we only produced four independent lines. Hence, whether TAA function originating from either the autosome or male sex chromosome is essential for survival of M. polymorpha sporelings requires further investigation. The phenotypes of taa, _proEF1:amiR-TAA^miR160, _proEF1:amiR-YUC2^miR160, and _proEF1:iaaL lines indicate a significant amount of IAA produced in M. polymorpha in the examined tissues and developmental stages originates from the IPyA pathway. Furthermore, the dramatic phenotypes also indicate no alternative Trp-dependent or Trp-independent IAA synthesis pathways can compensate for loss of the IPyA pathway during the early sporeling stage of development, making this pathway an evolutionarily conserved essential regulator of plant development and growth.

We speculate IAA synthesis is severely decreased in taa and _proEF1:amiR-TAA^miR160 lines. Because auxin is critical for the growth of M. polymorpha, our transgenic lines likely grow in proportion to the amount of auxin they can produce, with the growth of taa much slower than the wild type. This could account for the lack of apparent difference in IAA level per fresh weight between taa, _proEF1:amiR-TAA^miR160, and the wild type. Cellular IAA levels are likely maintained at a certain level as a result of complex and strict homeostasis regulation. This regulation would lead to reduced IAA catabolism in the taa and _proEF1:amiR-TAA^miR160 lines, resulting in slower turnover of IAA catabolites in transgenic lines than in the wild type. This would explain the apparent accumulation of OxIAA in taa and _proEF1:amiR-TAA^miR160 lines relative to the wild type. The wild type-like IAA levels in taa might suggest the existence of an alternative IAA synthesis pathway; however, residual TAA activity from TAY may also account for observed IAA levels. The increased activity of TAY in Mp-TAA loss-of-function lines could contribute to the survival and maintenance of the slow-growing and highly abnormal thallus but fail to provide levels of IAA in the appropriate spatio-temporal domains to facilitate normal growth of the male M. polymorpha thallus (relative to the identified Mp-TAA regulatory sequences, that of TAY is significantly truncated and contains only 482 bp of the 5′UTR and no upstream promoter sequence). Regardless, the severe phenotypes of taa, _proEF1:amiR-TAA^miR160, and _proEF1:iaaL demonstrate the significance of endogenous IPyA-mediated auxin synthesis in the development of the M. polymorpha gametophyte. The significance of additional Trp-dependent or independent IAA biosynthesis pathways in M. polymorpha remains to be resolved.

Meristematic Auxin Synthesis and Transport Regulates M. polymorpha Development

We found that Mp-YUC2, but not Mp-YUC1, is expressed in the gametophytic thallus. We also observed distinct differences in the spatial expression patterns of Mp-TAA and Mp-YUC2 in the thallus. The only tissues where we observe coincident expression of the two genes is in shoot apical regions of thalli and in gemma cup bases, both being meristematic tissues. Within the shoot apical region resides the apical cell and its meristematic derivatives, and in gemma cup bases there is continuous initiation of gemmae production. Our expression studies of the M. polymorpha TAA and YUC genes, as well as IAA measurements, in gametophytic thallus tissues thus suggest auxin is mainly produced in the meristematic regions of the M. polymorpha thallus. However, Ishizaki et al. (2012) showed the soybean (Glycine max) _proGH3:GUS auxin signaling marker was not highly expressed in the apical meristem compared with, e.g., gemma cup bases. Ishizaki et al. (2012) suggested the lack of Gm-GH3 promoter activity in the M. polymorpha meristem might be the result of lack of positive ARF signaling output in this tissue, as the Gm-GH3 promoter does not measure IAA levels directly, but rather IAA-mediated transcriptional activation via ARFs. In angiosperms, an alternative ARF-independent auxin marker, AUX/IAA-DII:GFP, shows there is potentially plenty of IAA in the center of the angiosperm shoot apical meristem even though the Gm-GH3 promoter is not active in that tissue (Vernoux et al., 2011). Likewise, in M. polymorpha, there appears to be a situation with high auxin synthesis and levels, but low positive auxin signaling output, in the apical meristem.

If auxin is primarily synthesized in meristematic tissues, how does it exert its effect throughout the growing thallus? LaRue and Narayanaswami (1957) suggested that a substance produced in the thallus tips of the liverwort L. cruciata, presumably IAA, was transported toward the older parts of the thallus through the midrib. This transport was later supported by the findings of Gaal et al. (1982) and Maravolo (1976), who showed that ¹⁴C-IAA is transported in an apical-basal and dorsal-ventral manner in the M. polymorpha thallus. Considering M. polymorpha has at least four PIN-FORMED (PIN) genes (Bennett et al., 2014a), these findings suggest PIN-mediated polar auxin transport is an important mechanism for gametophytic M. polymorpha development, as recently shown in the moss P. patens (Bennett et al., 2014b; Viaene et al., 2014). Our IAA measurements of different thallus tissues suggested IAA levels correlate well with _proYUC2:GUS expression but less well with _proTAA:GUS, as Mp-TAA is highly expressed only in the meristematic region and the base of gemma cups, with much weaker signal in the midrib and no signal in thallus margins. Our gene activity data as well as IAA measurements would thus indicate active transport or diffusion of IAA or IPyA, with IPyA potentially actively transported or diffusing to sites with low Mp-TAA but high Mp-YUC2 activity.

Auxin Is a Positive Regulator of Dormancy in Liverworts

Liu et al. (2013) showed that increased auxin synthesis has a positive effect and prolongs Arabidopsis seed dormancy, while reduced auxin synthesis or signaling has a negative effect and disrupts seed dormancy. Our findings, together with pharmacological experiments preformed up to 60 years ago (LaRue and Narayanaswami, 1957; Maravolo and Voth, 1966; Stange, 1971), clearly show that endogenous auxin acts as a positive regulator of gemma dormancy in liverworts. However, exogenous auxin has also been reported to induce rhizoid formation and growth of gemmae in gemma cups (Tarén, 1958), suggesting regulation of dormancy, via auxin, is a complex process.

Our observations are consistent with previous hypotheses that auxin or IPyA synthesized in the apex is transported basipetally and, at least in part, imposes dormancy on gemmae within gemma cups (LaRue and Narayanaswami, 1957; Gaal et al., 1982; Maravolo, 1976). Because _proGH3:GUS is highly expressed in gemma cup bases, it is possible that IAA, or auxin derived from transported IPyA, from the thallus apex is perceived at the gemma cup base.

That exogenous auxin added to the gemma cup induces gemmae germination indicates that auxin is likely not acting directly on gemmae dormancy, but rather through a second signal that is transported or diffuses into the gemmae in the gemma cup, where it is perceived. The cup-limited signal must be able to diffuse or be transported between gemmae within the confines of the cup. It has previously been suggested that auxin indirectly and positively regulates seed dormancy in Arabidopsis together with ABA (Liu et al., 2013), making ABA an obvious candidate for the gemma cup dormancy signal. While neither spores nor gemmae normally germinate in the dark, exogenous auxin can induce their germination in the dark (Pfeffer, 1871; Fitting, 1939; Rousseau, 1954), suggesting auxin can substitute for light stimulation. Addition of auxin directly into gemma cups may also mimic endogenous physiology during light-stimulated germination, overcoming cup-imposed germination signals. In this scenario, auxin acts as both a dormancy effector and germination stimulator. Apically produced auxin imposes gemmae dormancy via a second cup-limited signal, and once gemmae are displaced from the cup, light stimulated auxin synthesis acts to induce rhizoid emergence and gemmae growth.

A common theme in the regulation of dormancy in land plants is that it is imposed via auxin signaling acting in conjunction with other hormonal pathways, often ABA, but also cytokinin and strigolactone in some cases as in lateral shoots of angiosperms. The tissues in which dormancy is imposed vary widely among land plants and are often not homologous, e.g., the gemmae described here in M. polymorpha are limited to the Marchantiopsida lineage of the liverworts. Thus, similar genetic components are used to impose dormancy in analogous, rather than homologous, structures—a variation of the concept of deep homology whereby homologous genetic regulatory networks are used to direct disparate developmental processes (Shubin et al., 2009).

METHODS

Plant Material and Cultivation

Throughout the text, “wild type” refers to the Australian wild-type liverwort (Marchantia polymorpha) accession isolated at field location near Melbourne, Victoria (Flores Sandoval et al., 2015b), if not indicated otherwise. Strain Takaragaike-1 (Tak-1; Japanese male wild-type line; Ishizaki et al. 2008) was used as indicated in the text. Plants were grown on Gamborg’s B5 medium (PhytoTechLab) as previously described (Flores Sandoval et al., 2015b). When possible, phenotypic and expression pattern analysis was performed on first or higher order generation gemmae from primary transformants.

IAA (PhytoTechLab) was dissolved in ethanol. Kyn (Sigma-Aldrich) was dissolved in DMSO. IAA or Kyn was added to standard growth medium (Gamborg’s B5) for pharmacological assays as indicated in the text.

Sequence Retrieval and Phylogenetic Analysis

Arabidopsis thaliana sequences were downloaded from TAIR (www.arabidopsis.org). Physcomitrella patens sequences were downloaded from Cosmoss (www.cosmoss.org). Other species were collected from Phytozome (http://www.phytozome.net/). M. polymorpha homologs were identified from transcriptome and genome databases from the on-going M. polymorpha Genome Project (U.S. Department of Energy Joint Genome Institute, Walnut Creek, CA; http://www.jgi.doe.gov/). The raw reads from the genome sequencing have been deposited in the NCBI Sequence Read Archive database (http://www.ncbi.nlm.nih.gov/bioproject/251267).

Amino acid sequences were aligned using the MUSCLE algorithm (Edgar, 2004) in SeaView v.4.4.3 (Gouy et al., 2010). Resulting alignments were edited manually. Based on these alignments, midpoint-rooted (Supplemental Figures 3 and 4) and unrooted (Supplemental Figure 5) phylogenetic trees were generated using the ML method implemented in PhyML 3.0 (http://www.atgc-montpellier.fr/phyml/; Guindon et al., 2010) using the default settings (LG substitution model, fixed proportion of invariable sites, four substitution rate categories, estimated gamma shape parameters, five random BIONJ starting trees with SPR and NNI improvements, and optimized topology and branch lengths). ML bootstrap support (1000 iterations) was generated by PhyML 3.0.

Construction of Plasmids

Primer names and sequences are listed in Supplemental Table 1. If not specified in the text, reagents and enzymes were bought from New England Biolabs. Mp-YUC1 was PCR amplified from cDNA made from RNA isolated from sporophytes using primers ME169 and ME170. Mp-YUC2 and Mp-TAA were PCR amplified from cDNA made from RNA isolated from thallus tissue using primers ME23 and ME24, and ME28 and ME29, respectively. The PCR products were cloned into pCRII-TOPO (Invitrogen), creating plasmids pME289, pME23, and pME19, and then sequenced. The genes were subsequently subcloned into the KpnI/HindIII (Mp-YUC1) or EcoRI/KpnI (Mp-YUC2, Mp-TAA) sites of pEF1pro v.2.0 (Flores-Sandoval et al., 2015b), creating plasmids pME290, pME33, and pME34. A NotI fragment from each plasmid was subcloned into pSKF HART and/or pSKF KART (Flores-Sandoval et al., 2015b), creating plasmids pME291, pME42/pME43, and pME52/pME53.

Regulatory regions for Mp-YUC1 and Mp-YUC2 were PCR amplified from genomic DNA using primers ME5 and ME6, and ME7 and ME8, respectively. Mp-YUC1 and Mp-YUC2 fragments were cloned into pCRII-TOPO (Invitrogen), creating plasmids pME14 and pME22, and then sequenced. Promoter fragments were subcloned into the XhoI/BamHI (Mp-YUC1) or NdeI/BamHI (Mp-YUC2) sites of pRITA (Flores-Sandoval et al., 2015b), creating plasmids pME26 and pME36, respectively. A NotI fragment from each plasmid was subcloned into pSKF HART and pSKF KART, creating plasmids pME30/pME28 and pME85/pME86. To construct _proTAA:GUS, a 4986-bp fragment of Mp-TAA was amplified from Tak-2 genomic DNA by PCR using primers TAA-proL and TAA-proR, and cloned into pENTR-D-TOPO (Life Technologies). The entry clone was used to generate binary plasmids harboring _proTAA:GUS.

To create the Mp-TAA gene-targeting construct, a 3748-bp Mp-TAA 5′ fragment was PCR amplified from genomic DNA (accession Tak-1) with primers TAA_5IF_L and TAA_5IF_R and inserted into the PmeI site of plasmid pJHY-TMp1 (Ishizaki et al., 2013) using the In-Fusion HD cloning kit (Clontech). Subsequently, a 5896-bp Mp-TAA 3′ fragment was amplified by PCR using primers TAA_3IF_L and TAA_3IF_R and inserted into the PacI site of the plasmid obtained above using In-Fusion cloning.

The Mp-TAA genomic fragment was amplified by PCR using the primers TAA_CPL_L and TAA_CPL_R. The obtained genomic fragment was subsequently cloned into pENTR/D-TOPO (Life Technologies) to generate a plasmid containing the Mp-TAA genomic fragment from 5.0 kb upstream of the ATG to the 3′UTR (total of 9.4 kb). The Mp-TAA cassette was cloned into a binary vector using LR Clonase II (Life Technologies) according to the manufacturer’s protocol.

amiR genes based on the Mp-miR160 stem loop and targeting Mp-YUC2 and Mp-TAA were ordered from GeneScript (Supplemental Figure 11). The two genes were subcloned into the KpnI/HindIII sites of pEF1pro v. 2.0, creating pME193 and pME194. The amiR-YUC2^miR160 gene was also cloned into the KpnI/HindIII sites of plasmid pME159 (Flores-Sandoval et al., 2015b) to obtain pME196 (_proSHI:amiR-YUC2^miR160). NotI fragments of pME193, 194, and 196 were subcloned into pSKF HART, creating plasmids pME203, pME204, and pME215.

Plant Transformation

Constructs were introduced into Agrobacterium tumefaciens strain GV3001. M. polymorpha sporelings or thalli were transformed by agrotransformation essentially as described previously (Ishizaki et al., 2008; Kubota et al., 2013). Transformed sporelings or thalli were plated on selective media: 0.5 or 1× B5 with 10 μM hygromycin, 10 μM chlorosulfuron, and/or 10 μM G418, plus 200 μM Timentin (Phyto Technology Labs).

5′ RLM-RACE

Total RNA from _proEF1:miR-TAA^miR160 and _proSHI:miR-YUC2^miR160 lines and the wild type was extracted using an RNeazy Plant mini kit (Qiagen), DNase treated (Qiagen), and subsequently ligated to a single-stranded RNA adaptor as previously described (Flores-Sandoval et al., 2015a). cDNA made from these RNAs was analyzed by nested PCR using primers ME226 and ME187, followed by ME227 and ME188 for Mp-TAA/TAY, and primers ME226 and ME272, followed by ME227 and ME273 for Mp-YUC2. PCR products of the correct sizes, only found in the transgenic lines, were extracted (QIAquick Gel extraction kit; Qiagen) and sequenced.

Dormancy Scoring

A gemma cup was scored as dormant if no rhizoids and/or expanding gemmae could be detected among the gemmae inside the gemma cup and as nondormant if rhizoids and/or expanding gemmae were found. Newly formed gemma cups were not scored if no mature gemmae could be observed inside. All gemma cups in all lines for an individual experiment were scored during the same day.

Expression Analysis

Total RNA was isolated from tissues as described in Results using the Qiagen RNeasy plant mini kit and the Qiagen RNase-Free DNase set. cDNA was made using BioScript reverse transcriptase (Bioline). For all RT-PCRs, we used two or three biological replicates and at least three variations in cycle number to verify reproducibility and amplification in the logarithmic phase, respectively. In gel photos shown in Results, cycle numbers were optimized to give very weak signal in the sample with least signal, except for Figure 3A, where 30 cycles were used. To estimate the amount of PCR product obtained, PCR reactions were run on ethidium bromide-containing agarose gels and subsequently analyzed in UV light. During gel analysis and data collection, volumes and settings were chosen to avoid saturation of fluorescence signals. ME246 and ME247 were used to amplify Mp-EF1. ME290 and ME291 were used to amplify full-length Mp-TAA/TAY (Figure 5). ME24 and ME25 were used to amplify full-length Mp-YUC2 (Figure 6). ME89 and ME90 were used to amplify Mp-TAA (Figures 2 and 3). ME87 and ME88 were used to amplify Mp-YUC1 (Figure 3). ME85 and ME86 were used to amplify Mp-YUC2 (Figures 2 and 3). ME286 and ME287 were used as Mp-TAA biased primers, and ME284 and ME285 were used as TAY biased primers (Figure 5). For the biased primer pairs, the annealing temperature was 72°C to increase specificity. PCR products were extracted (QIAquick gel extraction kit; Qiagen) and sequenced to verify the identity of the amplicons.

GUS assays were performed essentially as described previously (Jefferson et al., 1987). Plants were incubated in GUS staining solution (0.5 mM potassium ferrocyanide, 0.5 mM potassium ferricyanide, and 1 mM X-Gluc) at 37°C and later cleared with ethanol.

For sectioning, GUS stained samples were fixed in FAA solution (50% ethanol, 5% acetic acid, and 10% formaldehyde) and embedded into Technovit 7100 resin according to the manufacturer’s instructions (Heraeus Kulzer). Embedded samples were then sectioned into 4-μm-thick sections and counterstained with 0.01% neutral red.

Liquid Chromatography-Electrospray Ionization-Tandem Mass Spectrometry Analysis of IAA and OxIAA Levels

Liquid chromatography-electrospray ionization-tandem mass spectrometry (MS/MS) analysis of IAA and OxIAA was performed using the Agilent 6420 Triple Quad system (Agilent). Freeze-dried plant tissues (∼30 mg) were homogenized with zirconia beads (3 mm) in 0.3 mL of 80% acetonitrile/1% acetic acid/water containing [phenyl-¹³C₆]IAA and [phenyl-¹³C₆]OxIAA using tissue lyser (Qiagen) for 3 min. Extracts were centrifuged at 13,000g for 3 min under 4°C and the supernatant was collected. Extraction was repeated twice without ¹³C-labeled internal standards. Extracts were combined and the volume was reduced (<500 μL) by evaporation using a SpeedVac. The extract was loaded onto Oasis HLB column (1 mL; Waters) and washed with 1% acetic acid/water (1 mL). IAA and OxIAA were coeluted with 80% acetonitrile/1% acetic acid/water (2 mL), and the volume was reduced (<300 μL) by evaporation using a SpeedVac. The IAA and OxIAA coeluted fraction was loaded onto Oasis WAX column (1 mL). After washing with 1% acetic acid/water (1 mL) and 80% acetonitrile/water (2 mL), IAA and OxIAA were eluted with 80% acetonitrile/1% acetic acid/water (2 mL) and evaporated using a SpeedVac. The fraction was redissolved in 1% acetic acid/water (30 μL) and injected into an Agilent 6420 Triple Quad system with a Zorbax Eclipse XDB-C18 column (1.8 μm, 2.1 × 50 mm). HPLC was performed with 0.01% acetic acid/water (solvent A) and acetonitrile/0.05% acetic acid (solvent B) using 3% solvent B for 3 min and a gradient from 3 to 15% of solvent B over 20 min at a flow rate of 0.2 mL/min. The MS/MS analysis conditions for IAA and [phenyl-¹³C₆]IAA (positive ion mode) were as follows: capillary = 4000 V, fragmentor voltage = 95 V, collision energy = 14 V, dwell time = 250 ms, and MS/MS transition (m/z) = 176/130 for unlabeled IAA and 182/136 for [phenyl-¹³C₆]IAA. The MS/MS analysis conditions for OxIAA and [phenyl-¹³C₆]OxIAA (postive ion mode) were as follows: capillary = 4000 V, fragmentor voltage = 80 V, collision energy = 12 V, dwell time = 250 ms, and MS/MS transition (m/z) = 192/146 for unlabeled OxIAA and 198/152 for [phenyl-¹³C₆]OxIAA. Quantification was performed using the extracted ion chromatogram of IAA and [phenyl-¹³C₆]IAA or those of OxIAA and [phenyl-¹³C₆]OxIAA. Endogenous levels of IAA and OxIAA in plants were calculated using a standard curve.

Microscopy

Plants were photographed under a Lumar.V12 dissecting microscope (Zeiss) equipped with an AxioCam HRc (Zeiss). Samples for scanning electron microscopy were prepared in FAA (50% ethanol, 5% acetic acid, and 10% formaldehyde) overnight at 4°C. Samples were dehydrated in an ethanol series before critical point drying in a CPD 030 (Baltec). Samples were coated with gold using a SCD 005 (Baltec). Scanning electron microscopy was performed at Monash Micro Imaging facilities using an S-570 (Hitachi) at 10 kV.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL libraries under the following accession numbers: M. polymorpha TAA (KP877963), TAY (KP877962), TAR (KP877964), YUC1 (KP877966), and YUC2 (KP877965).

Supplemental Data

• Supplemental Figure 1. Removal of the apical region of thalli reduces dormancy of gemmae in proximal gemma cups.

• Supplemental Figure 2. Gemma cup number in transgenic lines.

• Supplemental Figure 3. M. polymorpha has a single TAA family member.

• Supplemental Figure 4. M. polymorpha has two paralogous YUC family members.

• Supplemental Figure 5. Only two of the best hits from the M. polymorpha BLAST search clusters with known YUC sequences from other species.

• Supplemental Figure 6. Inhibition of TAA enzyme activity in _proEF1:TAA and _proEF1:YUC2 transgenic lines.

• Supplemental Figure 7. _proTAA:GUS expression in gametangiophores and the sporophyte.

• Supplemental Figure 8. _proYUC2:GUS is expressed in assimilatory filaments of the air chambers and in ventral cells adjacent to rhizoids.

• Supplemental Figure 9. _proTAA:GUS expression in developing and dormant gemmae.

• Supplemental Figure 10. Mp-TAA knockout strategy and molecular verification of transgenics.

• Supplemental Figure 11. Artificial miRNA genes: structure of amiRs and molecular verification of transgenic lines.

• Supplemental Figure 12. IAA and OxIAA measurements in _proEF1:amiR-TAA^miR160.

• Supplemental Figure 13. _proSHI:amiR-YUC2^miR160 lines resemble _proSHI:iaaL lines.

• Supplemental Figure 14. _proSHI:ARF1ΔD34 lines are sensitive to Kyn-induced release of gemmae dormancy.

• Supplemental Table 1. DNA oligonucleotide sequences used in this study.

• Supplemental Data Set 1. TAA/TAR amino acid alignment for phylogenetic analysis.

• Supplemental Data Set 2. YUC amino acid alignment for phylogenetic analysis.

• Supplemental Data Set 3. YUC and other FMOs amino acid alignment for phylogenetic analysis.

Glossary

ABA

abscisic acid

IAA

indole-3-acetic acid

IPyA

indole-3-pyruvic acid

Kyn

l-Kynuranine

ML

maximum likelihood

OxIAA

2-oxindole-3-acetic acid

UTR

untranslated region

amiR

artificial microRNA

RLM-RACE

RNA ligase-mediated rapid amplification of cDNA ends

MS/MS

tandem mass spectrometry