Evidence for Exaptation of the Marchantia polymorpha M20D Peptidase MpILR1 into the Tracheophyte Auxin Regulatory Pathway^,

Characterization of the ancient auxin conjugate hydrolase MpILR1 from liverwort provides evidence of its exaptation in the evolution of tracheophytes.

Abstract

Auxin homeostasis is tightly regulated by several mechanisms, including conjugation of the hormone to specific moieties, such as amino acids or sugar. The inactive phytohormone conjugate is stored in large pools in plants and hydrolyzed to regain full activity. Many conjugate hydrolases (M20D metallopeptidases) have been identified and characterized throughout the plant kingdom. We have traced this regulatory gene family back to liverwort (Marchantia polymorpha), a member of the most ancient extant land plant lineage, which emerged approximately 475 million years ago. We have isolated and characterized a single hydrolase homologue, dubbed M. polymorpha IAA-Leucine Resistant1 (MpILR1), from liverwort. MpILR1 can hydrolyze two auxin (indole acetic acid [IAA]) substrates (IAA-Leucine and IPA-Alanine) at very low levels of activity, but it cannot hydrolyze the two native auxin conjugates of liverwort (IAA-Glycine and IAA-Valine). We conclude from these results that liverwort likely does not employ active auxin conjugate hydrolysis as a regulatory mechanism and that conjugate homeostasis likely takes place in liverwort by passive background degradation. Furthermore, we present evidence that MpILR1 was probably exapted by tracheophytes over evolutionary time into the auxin regulatory pathway.

Auxin is a ubiquitous phytohormone that can be found throughout the entire plant kingdom, from nonvascular plants through gymnosperms to angiosperms. The hormone, most commonly found in the form of indole acetic acid (IAA), is essential in the regulation of growth and development of stems, shoots, leaves, and roots.

The concentration of IAA in a plant is critical: low concentrations of IAA positively impact plant physiological processes, but high concentrations are inhibitory and toxic (Bandurski et al., 1995). In addition, precisely minute, spatiotemporally regulated quantities of the phytohormone are required for proper stem, shoot, leaf, and root development. For this reason, auxin metabolism is tightly regulated by four major processes: biosynthesis, transport, conjugation, and degradation (Normanly and Bartel, 1999; Ljung et al., 2002).

One common regulatory system for auxin metabolism is molecular conjugation, in which auxin is stored in conjugated forms that are generally thought to be inactive (Ludwig-Müller, 2011). There are two major types of conjugated molecules: an amide-linked form bound to one or more amino acids and an ester-linked form mostly bound to sugar(s). These two types of conjugates are found in varying concentrations among the diverse tissues of bryophytes, sporophytes, gymnosperms, and angiosperms (Sztein et al., 1999, 2000). Up to 90% of auxin in a plant is conjugated into these storage forms (Bandurski et al., 1995; Campanella et al., 1996), although these forms vary by species. Free, active auxin may be made available by hydrolyzing the conjugate molecule from the side chain.

The evolution of the auxin conjugation process can be traced back to the earliest terrestrial plants. The most ancient land plants, the bryophytes, have been studied for their auxin metabolism (Sztein et al., 1999, 2000), and all have been found to produce auxin conjugates. Most liverwort species (including Marchantia polymorpha, Pallavicinia lyellii, Reboulia hemisphaerica, and Sphaerocarpos texanus) possess an IAA conjugate distribution in which IAA-amide conjugates account for more than half the auxin in the cell (∼52%), IAA-ester conjugates for ∼20%, and free IAA for ∼28% (Sztein et al., 1999). In mosses, IAA-amide conjugates dominate the total IAA composition (∼85%), with only a minor amount of IAA-ester conjugates (∼5%) and free IAA (∼10%; Sztein et al., 1999).

However, our understanding of the evolution of the regulatory mechanism for active auxin conjugate hydrolysis is incomplete. Evidence suggests that auxin production and regulation have changed over evolutionary time. Charophytic algae appear to be the direct aquatic ancestor of terrestrial bryophytes (Hori et al., 2014). Charophytes have relatively high levels of free IAA, and conjugation occurs slowly, if at all. These algae employ an IAA biosynthesis and degradation strategy to regulate auxin metabolism (Sztein et al., 1999, 2000). The first terrestrial plants may have evolved few changes from the processes of auxin regulation found in algae (Hori et al., 2014).

Sztein et al. (2000) concluded that, although bryophytic liverworts contain IAA conjugates, hydrolysis occurs at a very slow rate and may not be as active a process as the one seen in tracheophytes. Ludwig-Müller et al. (2009) proposed that auxin conjugation may occur in moss (Physcomitrella patens) as an intermediate in auxin inactivation and that conjugate hydrolase genes evolved in tracheophytes. Furthermore, Ludwig-Müller et al. (2009) suggested that moss may be a dead end for active auxin hydrolysis and that no earlier species possessed this regulatory pathway.

Our previous research has demonstrated that auxin conjugate hydrolase homologs are functionally highly conserved among angiosperms to help regulate auxin activity (Campanella et al., 2003b). More recent studies demonstrate that the genomes of gymnosperms and even ferns have genes encoding regulatory hydrolases (Campanella et al., 2014). If tracheophytes have putative auxin conjugate hydrolases, then they must have evolved from ancestral species.

Motivated by an understanding that highly conserved genes must have an evolutionary source, we were prompted to search back to the liverwort M. polymorpha, a member of the most ancient extant land plant lineage, which emerged approximately 475 million years ago (Wellman et al., 2003). As in any plant, M. polymorpha is dependent upon auxin for growth and development. Eklund et al. (2015) demonstrated that auxin is essential for gametophyte development and the promotion of gemma dormancy, and Solly et al. (2017) found that auxin works as a growth-promoting signal in thallus development. Additionally, liverwort is just as sensitive to the toxic effects of high auxin concentrations as other plants; exogenous auxin at low concentrations stimulated thallus growth, but interference with liverwort auxin conjugation caused severe stunting and the loss of differentiated cell fates (Solly et al., 2017).

We searched the genome of M. polymorpha, employing the newest genomic sequence (version 3.1; Bowman et al., 2017), and were able to identify one IAA-Leucine Resistant1 (ILR1) hydrolase homolog. Here, we characterize this liverwort hydrolase for enzymatic activity and discuss its implications for the evolution of the pathway regulating auxin conjugate hydrolase in Plantae.

RESULTS

The Gene Structure of MpILR1 in M. polymorpha Differs from That of Hydrolases in Tracheophytes

Previous experience with prokaryotic contamination of the first moss genome database (version 1.1; Rensing et al., 2008) sensitized us to the possibility that our liverwort MpILR1 (OAE20874.1) hydrolase might not be a plant gene. However, we were able to confirm its Plantae origins in M. polymorpha based on several factors. First, based on a GenBank BLAST search, the DNA sequence of MpILR1 demonstrated an average 38.8% similarity to a spectrum of tracheophyte hydrolases, and its protein sequence bore a moderate similarity (average 60%; Fig. 1). Additionally, we performed two-dimensional principal coordinate analysis of codon usage (Fig. 2) and found that MpILR1 fell squarely amid eukaryotic hydrolases, lying far from prokaryotic or archaean homologues. Included in this analysis were prokaryotic hydrolases capable of breaking down IAA-Asp. The structure of the MpILR1 protein includes the five conserved amino acids that characterize eukaryotic M20D peptidase structures: Cys-137, His-139, Glu-173, His-197, and His-397 (Rawlings and Barrett, 1995; Bitto et al., 2009; Fig. 3). These residues are not conserved in bacterial species and indicate the Mn⁺² ion-binding site of an amido-auxin-type hydrolase (structural subcategory Cd08017).

Figure 1.

Amino acid sequence similarity matrix of ILR1 orthologs from various species. The matrix was generated with MatGAT version 1.1 by using the default values for protein analysis. The boldface values are those related directly to MpILR1 similarity to the homologs.

Figure 2.

Principal coordinate analysis of codon usage to determine if MpILR1 was a bacterial contaminant acquired during genomic sequencing. The two-dimensional plot was created using ggplot2 (Wickham, 2009). MpILR1 is clearly positioned in the eukaryotic section of the plot.

Figure 3.

Amino acid alignment of several orthologue amidohydrolases. Alignment was performed using ClustalX version 1.81. Arrows indicate conserved amino acids (Cys-137, His-139, Glu-173, His-197, and His-397). The star indicates where the Leu was substituted in MpILR1^L244S with a Ser residue.

Further study of the genomic structure of MpILR1 indicates that the gene includes five exons and four introns (Fig. 4). The total length of the genomic sequence is 2,554 bp. The length of the MpILR1 coding sequence is 1,404 bp. The study conducted by Davies et al. (1999) and later genomic work indicate that the intronic structure of the tracheophyte M20D family consists of four introns and five exons; thus, it appears that the original liverwort intron/exon structure has been conserved since terrestrial plants first evolved.

Figure 4.

Intron/exon structure of MpILR1. The intron positions for this figure were generated using ClustalX using default settings. The arrow indicates the position of the active site Leu244 in the MpILR1 sequence.

MpILR1 does not possess an endoplasmic reticulum (ER) localization signal at its C-terminal end, such as observed in many angiosperm hydrolases, including those of Arabidopsis (Arabidopsis thaliana; Bitto et al., 2009).

MpILR1 Phylogeny Supports Its Position as the Ancestor of All Tracheophyte Auxin Conjugate Hydrolases

The phylogenetic analysis performed on the amidohydrolase orthologs (Fig. 5) suggests that the outgroup M. polymorpha MpILR1 may be the ancestor of all tracheophyte hydrolases that evolved after liverwort. The cladogram employs amino acid sequences and manifests clear phylogeny leading to present-day amido-conjugate hydrolases (i.e. the liverwort hydrolase shares a common ancestry with fern hydrolases and, eventually, gymnosperm and angiosperm hydrolases). This observation is supported by bootstrap values, which are all over 500.

Figure 5.

Neighbor-joining phylogram of the ILR1 protein family members. ClustalX was employed for alignment, and bootstrapping was performed 1,000 times. The outgroup used in this phylogenetic analysis is the uncharacterized putative amidohydrolase from K. flaccidum. The two important roots are in boldface.

A single copy of the putative auxin amidohydrolase homologue is found in liverwort. Additionally, we detected a single homologue in Klebsormidium flaccidum, the direct algal progenitor of liverwort (Hori et al., 2014). Going back further in evolutionary time, a single hydrolase homologue, related to that of both algae and liverwort, can be detected in the cyanobacterium Hassallia byssoidea. However, moving forward in time after liverwort, an expansion of the amidohydrolase family has occurred over the last 475 million years since liverwort initially evolved (Table 1). The genetic redundancy of the M20D family can be seen as early as in the ferns with two isoforms and later in gymnosperms that have anywhere between two and four isoforms (Fig. 1; Table 1). In angiosperms, it becomes even more apparent, with the redundant isoforms of amidohydrolases reaching as many as 15 in grape (Vitis vinifera) and 12 in soybean (Glycine max). This trend continues in monocots, with 11 paralogues detected in corn (Zea mays).

Table 1. Copy-number polymorphisms in the auxin conjugate hydrolase family across Plantae.

The presence of homologs was determined by BLAST analyses on the GenBank, The Institute for Genomic Research, and Phytozome Web sites. Asterisks indicate hydrolases that have been identified but not isolated and characterized.

Division	Species	Paralogue Hydrolases
Monocot (angiosperm)	Zea mays	11*
	Oryza sativa	9*
	Hordeum vulgare	7*
	Triticum aestivum	3*
Eudicot (angiosperm)	Vitis vinifera	15*
	Glycine max	12*
	Solanum lycopersicum	7*
	Solanum tuberosum	7*
	Arabidopsis thaliana	7
	Medicago truncatula	5
Gymnosperm	Picea sitchensis	4
	Pinus taeda	4*
	Picea glauca	2*
Bryophyte	Physcomitrella patens	3
	Marchantia polymorpha	1
Charophyte	Klebsormidium flaccidum	1*
Cyanobacterium	Hassallia byssoidea	1*

Wild-Type MpILR1 Has Weak Hydrolytic Activity against Auxin Conjugates

The MpILR1 hydrolase recognized only a few auxin conjugate substrates (IAA-Leu and IPA-Ala) and, furthermore, had very low activity against those substrates (∼10.2 and ∼1.66 pmol auxin min⁻¹ mL⁻¹, respectively; Table 2). These levels of hydrolysis are lower than those of the tracheophyte auxin conjugate hydrolases that have been characterized, where hundreds or more picomoles of auxin have been observed to be released per minute with the appropriate substrate (Davies et al., 1999; Campanella et al., 2003c, 2004, 2008, 2014). It should be noted that the MpILR1 mRNA appears to be expressed in liverwort, since we initially detected its cDNA in the liverwort expression database generated by Bowman et al. (2017).

Table 2. Enzyme/substrate specificity of the MpILR1 and MpILR1^L244S auxin hydrolases.

All hydrolase values are expressed as picomoles of auxin released per minute per milliliter (averages from four to seven replicate experiments) plus or minus the se. Untransformed E. coli (NovaBlue) cells were employed as negative hydrolysis controls. Positive hydrolysis is indicated by boldface values. IBA and IPA are indole butyric acid and indole proprionic acid, respectively. The comparative data for Medicago, loblolly pine, and Arabidopsis are from Campanella et al. (2008, 2014) and LeClere et al. (2002), respectively. Note that the data from LeClere et al. (2002) were expressed as nanomoles of auxin released per minute per milligram of protein. N/A indicates the data is not available.

Substrate	MpILR1	MpILR1L244S	PtIAR31	MtIAR34	AtILL2
IAA-Asp	0.00 ± 0.00	0.00 ± 0.00	220 ± 50.0	500 ± 123	62
IAA-Ala	0.00 ± 0.00	0.00 ± 0.00	20 ± 3.0	50 ± 13	1,600
IAA-Ile	0.00 ± 0.00	16.53 ± 6.65	0.00 ± 0.00	0.00 ± 0.00	67
IAA-Leu	10.20 ± 5.10	16.23+6.29	0.00 ± 0.00	70 ± 18	120
IAA-Phe	0.00 ± 0.00	0.00 ± 0.00	70 ± 10.0	0.00 ± 0.00	170
IAA-Val	0.00 ± 0.00	0.00 ± 0.00	<1.0	20 ± 3	96
IAA-Gly	0.00 ± 0.00	0.00 ± 0.00	150 ± 30.1	30 ± 5	46
IBA-Ala	0.00 ± 0.00	744.58 ± 134.87	1,930 ± 10.0	270 ± 29	N/A
IPA-Ala	1.66 ± 1.65	18.32 ± 16.69	N/A	N/A	N/A

The Mutated MpILR1^L244S Enzyme Has Both Greater Activity and Recognition against Auxin Conjugates

The low levels of MpILR1 hydrolysis against the most common auxin conjugate substrates suggested that this liverwort enzyme might have a structure unlike that of tracheophyte auxin conjugate hydrolases that arose later during evolution. Prompted by observation of the minimal levels of hydrolysis in the wild-type MpILR1, our next goal was to determine what structural alterations to MpILR1 may have led to the tracheophyte evolution in which the M20D enzyme family recognized and became more active against those substrates.

The x-ray crystallographic analysis of AtILL2 by Bitto et al. (2009) suggested to us potential structural changes that could have arisen over evolutionary time in MpILR1 to make it the more active enzyme found in tracheophytes. A range of mutant hydrolases were characterized by Bitto et al. (2009), where single amino acid changes led to the loss of function in AtILL2. Only one of these alterations could be detected in the MpILR1sequence to account for its reduced function. The mutation was identified originally by Davies et al. (1999) in AtIAR3. The structural alteration is a substitution in which Ser-206 is mutated to Leu.

We located the analogous amino acid 206 in MpILR1 (Leu244) and hypothesized that this wild-type Leu residue may be inhibiting substrate recognition and activity and obstructing the active site. Based on this hypothesis, we constructed a mutant version of MpILR1 in which Leu244 was changed to the Ser conserved in later tracheophytes. The new MpILR1^L244S construct was transformed into Escherichia coli and expressed for enzymatic analysis.

The MpILR1^L244S protein shows both greater substrate recognition and a higher level of hydrolytic activity than the wild-type enzyme (Table 2). The reengineered, mutant enzyme recognized IBA-Ala and IAA-Ile, which the wild type did not hydrolyze at all. Hydrolytic activity did increase slightly against IAA-Leu, from ∼10.2 to ∼16.23 pmol auxin min⁻¹ mL⁻¹ hydrolyzed (Table 2), while hydrolysis against IPA-Ala seemed to increase by ∼11-fold. We still observed no hydrolysis of the native auxin conjugates (IAA-Val and IAA-Gly) of liverwort.

It is unlikely that MpILR1^L244S was more active than the wild-type MpILR1 because of differences in protein concentration. We performed a PAGE control to determine if L233S and MpILR1 were being expressed at different levels. They appeared to be alike in terms of expression when examined at 52 kD in the PAGE analysis (data not shown).

DISCUSSION

MpILR1 and Evolution in Plantae

Auxin conjugate amidohydrolases create an evolutionary paradox: if bryophytes do not have active hydrolytic regulation of auxin conjugates, then where did these hydrolases arise? Exaptation (Gould and Vrba, 1982) could explain the evolutionary source of these enzymes, given the evidence that we present here. Exaptation can be defined as the evolutionary coopting and strengthening of one trait for a different purpose or role.

Liverwort may provide one of the earliest examples of an enzyme that would eventually be exapted for use in tracheophytes. MpILR1 is an inefficient and low-functioning hydrolase with standard auxin conjugates tested as substrates (Table 2), compared with tracheophyte hydrolases. Additionally, MpILR1 does not appear to hydrolyze either of the two auxin conjugates found in M. polymorpha: IAA-Val and IAA-Gly (Sztein et al., 1999; Table 2). We conclude from these two observations that MpILR1 may have an unknown substrate and function entirely differently from what we see later in tracheophytes. The initial role of MpILR1 may have been unconnected with the active regulation of auxin. However, through a series of mutations and alterations over evolutionary time, MpILR1 became exapted into the auxin regulatory pathway because it had the potential for a different purpose/function.

We note that Sztein et al. (1999) do caution that their identification of individual auxin conjugates was tentative in bryophytes, since they used thin-layer chromatography separation after incubation with IAA and not mass spectrometry. It is still possible that MpILR1 contributes to M. polymorpha physiology by hydrolyzing IAA-Leu and IAA-Ala at low levels, and this possibility will have to be tested in future experiments.

Although the active site of MpILR1 has a moderate similarity (57%–60%) to corresponding sequences in modern Plantae, it is difficult to identify which of many structural alterations from evolutionary time were necessary to modify it into the enzyme we recognize now in tracheophytes. We have characterized at least one structural change that was clearly necessary, but not sufficient, for the full activity in tracheophytes. The conserved Ser206 (Bitto et al., 2009) that is found in the active site of M20D hydrolases is not present in MpILR1, where it is replaced at the corresponding position with Leu244.

Bitto et al. (2009) suggested that this substitution results in a loss of hydrogen bonds to His380 and Asn337. Additionally, the larger, hydrophobic residue Leu206 position may disrupt Phe381, which helps form the indole-binding hydrophobic pocket. Our exchange of Leu244 with Ser244 in MpILR1^L244S altered both substrate recognition and hydrolysis quite dramatically (Table 2). Therefore, we must conclude that this mutation artificially evolved MpILR1 toward what a common progenitor gene would eventually become in vascular plants. We are confident that we have uncovered one clue that helps to bridge the evolutionary gap between the bryophyte and tracheophyte hydrolases.

We can demonstrate a clear line of heredity in the homologues that constitute the M20D peptidase family. The charophytic alga K. flaccidum has been characterized as possessing many of the genes that would eventually be required for the evolution of terrestrial plants (Hori et al., 2014). Among these important terrestrial genes were the M20D hydrolases that would eventually help regulate auxins. We have found a single orthologue (GAQ79540.1) for MpILR1 in K. flaccidum (Figs. 1 and 2), displaying 59.5% similarity to the liverwort protein. We can track the source of this enzyme further back in time to cyanobacteria. The cyanobacterium H. byssoidea also carries a homologue (KIF32975.1) for MpILR1 (Figs. 1 and 2), with both sharing 47.6% similarity. More interestingly, both the K. flaccidum and H. byssoidea orthologues have the same conserved manganese metal ion-binding structure as the M20D peptidases found in all Plantae.

Based on the phylogenetic analysis (Figs. 1 and 5) and structural homology, we propose an evolutionary model (Fig. 6) for M20D amidohydrolases. We believe that the source of the M20D peptidase family of enzymes in tracheophytes can be traced directly back to liverwort and that we can track the enzyme back further to charophytes and cyanobacteria. MpILR1 is further supported as the antecedent to the M20D peptidase family because of its highly conserved intron/exon structure (Fig. 4), which is mirrored in evolution up through the angiosperms.

Figure 6.

Proposed model of the M20D peptidase molecular evolution from cyanobacterium up through vascular plants. Mosses appear with a question mark because, although they seem to have high levels of conjugation (Sztein et al., 1999), it is still unclear whether they possess homologous auxin amidohydrolases in this evolutionary path. This model was based partly on that suggested by Cooke et al. (2002).

It should be noted that MpILR1 lacks the ER localization signal (KDEL or HDEL) seen in the majority of angiosperm amidohydrolases. Campanella et al. (2014) observed that no characterized gymnosperm M20D peptidase possesses an ER localization signal. We think that this signal did not evolve until after flowering plants arose. Sanchez et al. (2016) confirmed that, in Arabidopsis, the ILR1-like family of hydrolases regulates whole-cell auxin homeostasis from the ER. The two uncharacterized auxin conjugate hydrolases from the early-diverging flowering plant Amborella trichopoda (XP_011626860.1 and XP_006853919.1) contain HDEL signals, so we can trace the evolutionary origin of this sequence to at least the first known eudicot. Since a large proportion of angiosperm M20D hydrolases possess an ER localization signal, we speculate that some aspect of flowering plants called for the selection of greater homeostatic regulation of auxin.

Auxin Amidohydrolase Redundancy in Plantae

Although more recently evolved tracheophytes have multiple paralogues of the auxin amidohydrolases (Table 1), liverwort contains just the single copy of MpILR1. Since there is little evidence that MpILR1 is an active part of auxin regulation in liverwort, this single-copy status probably provides more insight into the later evolution of this enzyme family than into liverwort evolution per se. We know that, by the time ferns evolved, those first tracheophytes had at least two copies of this gene family (Table 1). The number of isoforms expands as we move into gymnosperms and, eventually, flowering plants.

Auxin conjugate hydrolysis may have become so important in vascular plants as a regulatory activity that evolutionary selection began to favor those plants that possessed additional paralogues as protection against potential mutation and loss. Alternately, paralogous forms may have evolved to carry out differing regulatory functions in various tissues and/or developmental stages. This conjecture is supported by the work of Rampey et al. (2004) which demonstrated that paralogues were expressed at different levels in varying tissues of several species. This change in expression suggests that the paralogues perform tissue-specific functions.

CONCLUSION

It remains unclear whether auxin conjugates are the substrate of MpILR1. The activity of the wild-type liverwort enzyme against these substrates is ineffective; therefore, it seems possible that other substrates may be better candidates. We are quite interested in testing putative candidate substrates of MpILR1 to identify what native function this enzyme may have in M. polymorpha. Performing a knockout for MpILR1 would allow us to ascertain if the auxin regulatory pathway (percentage free versus percentage conjugated) is altered by its loss. Additionally, we want to examine whether the degradation of auxin conjugates really does play a major role in auxin homeostasis in liverwort.

Further proteomic studies with MpILR1 can determine what structural changes in this bryophyte hydrolase led to its evolution into the enzyme eventually seen in tracheophytes. In addition, more in-depth physical analyses are required to determine where alterations have appeared in the MpILR1 active site and the surrounding amino acid residues.

Further study on moss is required to revisit the process of auxin regulation in the plant. As stated earlier, previous versions of the published moss genome have been contaminated with DNA from soil bacteria (unpublished data J.J. Campanella, J. Ludwig-Mueller). The most recently published P. patens genome should be examined to determine if it has genes encoding auxin conjugate amidohydrolases. It is possible that the speculation of Ludwig-Müller et al. (2009) was correct and moss is a more recently evolved dead end for active hydrolysis, but we need to ascertain whether this is indeed the case.

Finally, the same functional analyses need to be conducted with fern hydrolases. Since ferns were the earliest tracheophytes, we need to clarify if their hydrolases functionally fall somewhere between the liverwort and the gymnosperm hydrolases that evolved later. In short, are fern hydrolases more like the hydrolases in tracheophytes or bryophytes in function and hydrolysis?

MATERIALS AND METHODS

Detection of a Liverwort Hydrolase Homolog

A BLAST analysis (Altschul et al., 1990) of the Marchantia polymorpha genome (version 3.1) was performed (https://phytozome.jgi.doe.gov/pz/portal.html). The gymnosperm ortholog PsIAR31 DNA sequence (Campanella et al., 2014) was used in this search analysis. One M. polymorpha hydrolase ortholog was identified in the database, MpILR1 (OAE20874.1). A subsequent BLAST search on GenBank with the amino acid sequence of MpILR1 indicated that the newly identified liverwort hydrolase had a greater sequence similarity to ILR-like hydrolases than to the IAR subfamily. All BLAST analyses were performed with default search parameters.

Cloning and Construction

The gene sequences for MpILR1 and MpILR1^L244S were synthesized by Synbio Technologies. MpILR1^L244S was altered at amino acid 244 to code for a Ser, whereas the wild-type liverwort sequence coded for a Leu (Supplemental Fig. S1). Both these generated sequences included a T7-inducible IPTG promoter and coded for a Shine-Dalgarno sequence in the 5′ untranslated region just upstream of the open reading frames of the hydrolases. These constructed inserts were ligated into the BamHI site of a pUC57 plasmid. We transformed these constructs directly into NovaBlue Escherichia coli cells using heat shock transformation according to the procedure described by Sambrook et al. (1989). Transformants were selected on Luria-Bertani (LB) medium containing 50 µg mL⁻¹ ampicillin using blue-white selection (Sambrook et al., 1989).

Plasmids were isolated from transformants by alkaline lysis (Sambrook et al., 1989). Insert orientation and size were determined by endonuclease digestion, electrophoretic analyses using 1% agarose gels (w/v), and DNA sequencing of the insert regions using BigDye Terminator version 3.0 (Applied Biosystems) according to the manufacturer’s directions on an ABI model 3730 DNA Analyzer.

Enzyme Preparation from E. coli

MpILR1 and MpILR1^L244S cultures were grown overnight in 5 mL of LB medium containing 100 µg mL⁻¹ ampicillin. From this culture, 2 mL was transferred to a flask containing 50 mL of LB medium containing 100 µg mL⁻¹ ampicillin and 1 mm IPTG for gene induction. Induction was performed for 4 h with continuous shaking of the cultures. Untransformed NovaBlue control cells were grown in the same manner as the induced cells, but without IPTG.

Enzyme preparation and enzymatic activity assays were conducted as in our previous studies (Campanella et al., 2003c, 2004, 2008). After collecting the bacterial cells by centrifugation for 10 min at 8,000g, the pellet was resuspended in 1 mL of lysozyme buffer per initial 1 mL of bacterial culture (30 mm Tris-HCl, pH 8, containing 1 mm EDTA, 20% Suc, and 1 mg mL⁻¹ lysozyme; Sigma-Aldrich), 1 mL of glycerol, 5 µL of DNase, and 5 µL of RNase. This mixture on ice was sonicated in three cycles using a Microson Ultrasonic Cell Disruptor (Misonix) at full power for 1 min, followed by a 1-min cool down. The extract (100-µL volume per assay) was then used directly for the enzyme assay.

Hydrolase Enzyme Assays

The enzyme assay for the hydrolysis of auxin conjugates was performed in a 500-µL reaction mixture containing 395 µL of assay buffer, 100 µL of bacterial enzyme extract (corresponding to ∼2.5 mg of total protein), and 5 µL of a 10 mm stock solution (dissolved in a small volume of ethanol, then diluted with water) of each substrate (final concentration of 100 µm; ethanol concentration was always less than 0.1%; Campanella et al., 2008). The substrates used in this study were the IAA-amide conjugates IAA-Asp, IAA-Ala, IAA-Gly, IAA-Leu, IAA-Ile, IAA-Phe, and IAA-Val (all from Sigma-Aldrich) and the amide conjugates IBA (Sigma-Aldrich) and IPA (Campanella et al., 2008) with Ala. The assay buffer consisted of 100 mm Tris (pH 8), 10 mm MgCl₂, 100 µm MnCl₂, 50 mm KCl, 100 µm PMSF, 1 mm DTT, and 10% Suc (Ludwig-Müller et al., 1996), but for the assays with IAA-Asp as substrate, no DTT was added. The reaction was incubated for 1 h at 40°C and stopped by adding 100 µL of 1 n HCl, and the aqueous phase was then extracted with 600 µL of ethyl acetate. This was followed by a 1-min centrifugation at 11,700g. We then transferred the upper organic phase to a new microfuge tube and evaporated it using the SpeedVac (Savant SpeedVac SC110, Savant Refrigerated Condensation Trap RT100, and Savant Two Stage VP190; Thermo Fisher Scientific) on a medium setting for 20 min. The evaporated pellet was resuspended in 200 µL of the appropriate running buffer (50% methanol or 50% methanol and 1% acetic acid) and analyzed by HPLC according to the procedure described by Campanella et al. (2003c).

The experiments were repeated three to six times using different enzyme preparations. All results represent means of three to six independent experiments + se. The uninduced cultures were evaluated as controls. The enzymatic activity was calculated as picomoles of IAA, IBA, or IPA released from cultures induced with IPTG minus the values obtained in control nonexpressing cultures.

Codon Usage Analysis

Principal coordinate analysis based on codon usage was performed using the family of auxin conjugate hydrolase orthologs within archaea, eubacteria, and eukaryotes (algae, bryophytes, gymnosperms, and angiosperms). EMBOSS CUSP (Lü et al., 2005) was then employed to obtain the relative abundance values. These values were then input into Vegan (Oksanen et al., 2016) to obtain a dissimilarity matrix (Bray and Curtis, 1957). The data from the dissimilarity matrix were used to perform the principal coordinate analysis via classical multidimensional scaling (R Core Team, 2014; Gower, 2015). The two-dimensional plot was created using ggplot2 (Wickham, 2009).

Global Alignment and Phylogenetic Tree Construction

The generation of the similarity matrix by global alignment of the hydrolase orthologs was performed using MatGAT version 1.1 (Campanella et al., 2003a). The program was set to the default configuration for amino acid analysis.

Alignments of all hydrolase homologs were created with the ClustalX version 1.81 software, using its default configurations (Thompson et al., 1997). The alignment data obtained from the ClustalX analysis were employed to create phylogenetic trees via the neighbor-joining method set to perform 1,000 bootstraps (Felsenstein, 1985; Saitou and Nei, 1987) and then visualized using TreeView (Page, 1996). The outgroup ortholog employed in the analysis was the alga Klebsormidium flaccidum.

Accession Numbers

The National Center for Biotechnology Information GenBank accession number for the MpILR1 gene is OAE20874.1. The following genes were employed in the phylogenetic analyses: K. flaccidum (GAQ79540), Arabidopsis (NP_175587.1), H. byssoidea (KIF32975.1), fern (XP_002981614.1 and XP_002976283.1), spruce (EF085410.1 and ED678121.1), pine (TC75408 and CT68471), tomato (XP_004235624.1), tobacco (XP_019233578.1), M. truncatula (XP_003597644.1), potato (XP_006343020.1), barley (BAJ94305.1), rice (XP_015621783.1), wheat (AAU06081.1), and corn (NP_001142151.1).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Sequence of MpILR1 along with the promoter structure that was inserted into the BamHI site of the pUC57 construct.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• ILR1 Gramene: AT3G02875

• ILR1 Araport: AT3G02875

• amino CHEBI: CHEBI:46882

• auxin CHEBI: CHEBI:22676