New Evidence for Ancient Origins of Bowman-Birk Inhibitors from … · 2017. 3. 14. · 1 RESEARCH...

1

RESEARCH ARTICLE

Evidence for Ancient Origins of Bowman-Birk Inhibitors from Selaginellamoellendorffii

Amy M. James1,2, Achala S. Jayasena1,2,, Jingjing Zhang1,2, Oliver Berkowitz3, David Secco2, Gavin J. Knott1, James Whelan3, Charles S. Bond1, Joshua S. Mylne1,2*

The University of Western Australia, 1 School of Molecular Sciences & 2 The ARC Centre of Excellence in Plant Energy Biology, 35 Stirling Highway, Crawley, Perth 6009, Australia 3La Trobe University, School of Life Sciences, ARC Centre of Excellence in Plant Energy Biology, Melbourne 3086, Australia * Corresponding Author: [email protected]

Short title: Exploring Bowman-Birk Inhibitor Evolution

One-sentence summary: A sequence from spikemoss uncovers the true extent of the Bowman-Birk Inhibitor family to be wider than cereals and legumes, solving a long-standing mystery about their origins.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantcell.org) is: Joshua S. Mylne (to [email protected])

ABSTRACT

Bowman-Birk Inhibitors (BBIs) are a well-known family of plant protease inhibitors first described 70 years ago. BBIs are known only in the legume (Fabaceae) and cereal (Poaceae) families, but peptides that mimic their trypsin-inhibitory loops exist in sunflowers and frogs. The disparate biosynthetic origins and distant phylogenetic distribution implies these loops evolved independently, but their structural similarity suggests a common ancestor. Targeted bioinformatic searches for the BBI inhibitory loop discovered highly divergent BBI-like sequences in the seedless, vascular spikemoss Selaginella moellendorffii. Using de novo transcriptomics, we confirmed expression of five transcripts in S. moellendorffii whose encoded proteins share homology with BBI inhibitory loops. The most highly expressed, BBI3, encodes a protein that inhibits trypsin. We needed to mutate two lysine residues to abolish trypsin inhibition, suggesting BBI3’s mechanism of double-headed inhibition is shared with BBIs from angiosperms. As Selaginella belongs to the lycopod plant lineage, which diverged ~200-230 million years before the common ancestor of angiosperms, its BBI-like proteins imply there was a common ancestor for legume and cereal BBIs. Indeed, we discovered BBI sequences in six angiosperm families outside the Fabaceae and Poaceae. These findings provide the evolutionary missing links between the well-known legume and cereal BBI gene families.

Plant Cell Advance Publication. Published on March 14, 2017, doi:10.1105/tpc.16.00831

©2017 American Society of Plant Biologists. All Rights Reserved

2

INTRODUCTION 1

Bowman-Birk Inhibitors (BBIs) were first described 70 years ago and were the subject of 2

classical experiments in biochemistry (Bowman, 1946). BBIs are one of the many 3

different families of plant protease inhibitors reviewed by Habib & Fazili (2007), who 4

used the structural classifications provided by the MEROPS peptidase database 5

(Rawlings et al., 2016). BBIs originally characterized from soybean (Glycine max) are 6

dual inhibitors of both trypsin and chymotrypsin (Birk, 1961; Birk et al., 1963); this 7

duality was later found to be the result of two separate inhibitory loops, each formed by 8

a disulfide bridge (Odani and Ikenaka, 1973). This presence of two spatially separated 9

loops is referred to as a ‘double-headed’ structure (Figure 1). This double-headed 10

structure is predicted to have arisen from an internal gene duplication of a single-11

headed ancestral inhibitor (Mello et al., 2003). The inhibitory loops are bound by 12

proteases in a substrate-like conformation, a standard mechanism for protease 13

inhibition. A double-headed BBI can bind two proteases simultaneously (Song et al., 14

1999). BBIs also typically possess many internal disulfide bonds between conserved 15

Cys residues that provide structural stability and maintain the active conformation of the 16

inhibitory loops. Highlighting this, a recent study showed that mutating a single Cys 17

residue of a typical ‘double-headed’ BBI abolished the activity attributed to both loops of 18

the mutated BBI in seed extracts (Clemente et al., 2015). 19

Since their discovery, BBIs have been shown to be widely distributed in the legume 20

(Fabaceae) and cereal (Poaceae) families (Mello et al., 2003). Although these two plant 21

groups are distantly related (Figure 1b), the high sequence and structural similarity 22

between their BBIs suggests they share a common ancestor (Mello et al., 2003). Cereal 23

BBIs lack the double inhibitory-loop structure common in legume BBIs. This is due to a 24

lack of inhibitory-loop-forming Cys residues in the region where the second loop is 25

located in legumes, resulting in the second loop becoming non-functional (Park et al., 26

2004; Qi et al., 2005). Cereal BBIs differ in size from 8 kDa to 20 kDa, whereas legume 27

BBIs are consistently around 8 kDa in size. The wide size range for cereal BBIs was 28

caused by internal gene duplication events. Although cereal BBIs lack the Cys residues 29

required for a functional second inhibitory loop, the internal duplication events have 30

3

resulted in multiple inhibitory loops for a single protein (Prakash et al., 1996; Song et al., 31

1999) (Figure 1). 32

The inhibitory loops of BBIs can function independently from the rest of the BBI protein, 33

as demonstrated by analysis of synthetic peptides made for the nine-residue loop alone 34

(Nishino et al., 1977). These small, synthetic peptides were also used to determine the 35

essential residues necessary for specificity to trypsin or chymotrypsin (Terada et al., 36

1978). The sequences of the synthetic nonapeptides analysed were CTKSNPPQC and 37

CALSTPAQC, which correspond to the soybean BBI trypsin inhibitory and chymotrypsin 38

inhibitory loops, respectively (Terada et al., 1978). By mutating the P1 residue 39

(underlined; the first amino acid in the N-terminal direction from the cleaved bond 40

following the notation described by (Schechter and Berger, 1967)), they showed that 41

Lys or Arg at the P1 position confers specificity for trypsin, whereas a Tyr or Leu will 42

confer specificity for chymotrypsin (Terada et al., 1978). 43

Since the first studies on synthetic BBI-loop peptides, every residue in the nonapeptide 44

except the P5' residue, which shows the weakest conservation, has been tested for its 45

functional importance (McBride et al., 2002). Apart from the obviously critical P1 residue 46

and loop forming Cys residues (P3, P6'), the remaining residues also influence BBI 47

structure and activity. For example, the hydroxyl and methyl group of the conserved Thr 48

at position P2 directly interacts with trypsin to stabilize the active site, resulting in 49

reduced hydrolysis (McBride et al., 1998). An Ile at P2' is optimal for inhibition of trypsin 50

(Gariani et al., 1999). The P3' and P4' Pro residues maintain a polyproline II 51

conformation important for structure (Park et al., 2004). Furthermore, the P3' Pro 52

residue must be in a cis conformation for correct positioning of the P1 residue (Brauer et 53

al., 2002). The P1' position is a highly conserved Ser, however, substitution with Ala 54

only slightly reduced BBI inhibitory activity (Brauer and Leatherbarrow, 2003). The low 55

tolerance for substitution in these studies illustrates that the high conservation observed 56

within BBI inhibitory loops is required for the optimal inhibition of trypsin. 57

Two decades after the first analysis of synthetic BBI inhibitory loops, a similar and 58

naturally occurring peptide was discovered in sunflower seeds (Luckett et al., 1999). 59

Sunflower Trypsin Inhibitor-1 (SFTI-1) shares the same function, a nearly identical 60

4

amino acid sequence, and three-dimensional structural conformation with the trypsin 61

inhibitory loop of BBIs; however, SFTI-1 is composed of only 14 residues and is a head-62

to-tail macrocycle (Figure 1). Originally thought to be the smallest member of the BBI 63

family, SFTI-1 has an altogether different biosynthetic and evolutionary origin. SFTI-1 64

emerged from a protein that is also a precursor for a napin-type, seed storage albumin 65

(Mylne et al., 2011). Evidence suggests SFTI-1 evolved de novo, in a stepwise manner 66

within a standard preproalbumin gene, independently of BBIs (Elliott et al., 2014). A less 67

potent, but similar, inhibitory loop also exists in ORB proteins from frogs (Li et al., 2007), 68

implying BBI-like inhibitory loops have evolved independently multiple times for this 69

function. 70

The existence of a highly conserved sequence in disparate proteins and protein 71

contexts within distinct phylogenetic groups suggests the BBI inhibitory loop is a product 72

of convergent evolution. Here we investigated whether this trypsin inhibitory loop has 73

appeared in additional protein contexts. We searched databases using the query 74

sequence CTKSIPPIC, the sequence shared by SFTI-1 and BBIs, and found an 75

identical match to a predicted protein from the ancient seedless, vascular plant 76

Selaginella moellendorffii. S. moellendorffii belongs to the lycopod lineage, which is 77

estimated to have diverged 400 million years ago, an estimated 200-230 million years 78

before the evolution of angiosperms (Banks, 2009; Bell et al., 2010). Closer analysis 79

revealed that rather than being a new protein context, the Selaginella BBI-like protein 80

represents an ancient member of the BBI protein family, as it shares conserved Cys 81

residues outside the conserved inhibitory loops. Furthermore, a recombinant S. 82

moellendorffii BBI inhibited trypsin, and mutating predicted P1 residues removed its 83

trypsin inhibitory activity, suggesting it uses the same mechanism of inhibition as BBIs 84

in seed plants. Consistent with the hypothesis that the Selaginella BBI-like proteins 85

share a common ancestor with legume and cereal BBIs, we discovered BBIs in six 86

angiosperm families outside the Fabaceae and Poaceae plant families. 87

5

RESULTS 88

Identification of BBI-like genes in S. moellendorffii 89

Between SFTI-1 and BBIs, only the inhibitory loop is shared. As this loop is sufficient for 90

inhibitory activity, we performed BLAST searches using the sequence CTKSIPPIC as 91

bait to identify other proteins that possess this functional motif. We identified two 92

sequences derived from the S. moellendorffii genome that contained highly similar 93

motifs: BBI1 (NCBI accession: XM_002980953) and BBI2 (NCBI accession: 94

XM_002980955). 95

The proteins encoded by BBI1 and BBI2 showed conservation within the inhibitory 96

motifs, but less sequence similarity outside of these motifs (Figure 1). The translated 97

sequences possessed a high number of Cys residues (13.7% of the complete primary 98

amino acid sequence), which is a characteristic of BBIs. The top plant BLASTp hit in 99

NCBI (NCBI accession: ADV40041) was a predicted BBI from the legume Lathyrus 100

sativus that shares 25.4% identity and 44.9% similarity with BBI1, as calculated in a 101

global alignment. Similarly, the most similar BBI sequence to BBI2 by BLASTp (NCBI 102

accession: NP_001236539) was a BBI from the legume Glycine max that shares 29.7% 103

identity and 50.7% similarity in a global alignment. 104

A S. moellendorffii transcriptome identifies five BBI-like sequences 105

Two BBI-like sequences were found through mining the published S. moellendorffii 106

genome. To determine if these genes are expressed and to determine if BBI-like 107

sequences are conserved within Selaginella, the transcriptomes of S. moellendorffii, 108

S. kraussiana, and S. martensii were assembled from RNA-seq data. Searches within 109

the S. martensii and S. kraussiana transcriptomes found no BBI-like sequences. A 110

search of publicly available RNA-seq data for S. stauntoniana identified a sequence with 111

two CTKSIPPIC-like motifs (CTMSYPPSC and CTKAPPNC). This suggests there are 112

BBI-like sequences present in Selaginella species other than S. moellendorffii, despite 113

being undetectable in S. kraussiana and S. martensii RNA-seq datasets. 114

In total, five BBI-like genes were identified as being expressed in S. moellendorffii 115

(Figure 2; Supplemental Figure 1). All five genes were sequenced from both cDNA 116

6

and genomic DNA and further validated by mapping RNA-seq reads to the complete 117

open reading frames. 118

Using SignalP 4.1 (Petersen et al., 2011), we found that all encoded BBIs except BBI4 119

had a predicted endoplasmic reticulum (ER) localization signal (Figure 2). BBI4 is 39 120

residues shorter and shows the greatest sequence divergence (38-43% similarity) from 121

the other four BBI-like sequences, which share >89% similarity. The shorter length was 122

verified by 5' RACE. Altogether, these data imply that BBI4, although expressed, 123

probably does not code for a functional BBI and is a pseudo-gene. 124

S. moellendorffii BBI3 is a functional trypsin inhibitor 125

To test if the Selaginella proteins share a similar function with characterized BBIs, we 126

produced recombinant BBI3 in E. coli and purified it for trypsin inhibition assays 127

(Supplemental Figure 2). BBI3 was chosen because it had the highest number of 128

reads mapped to its open reading frame (ORF), suggesting it is the most highly 129

expressed of the five BBI transcripts (Table 1). Trypsin inhibitory ability was determined 130

by end-point analysis after 45 minutes, assuming that the reaction had reached 131

equilibrium at this stage. Therefore, the results will not indicate any intermediate levels 132

of inhibition due to the inhibitor being in molar excess. Comparing BBI3 to soybean BBI 133

in trypsin inhibition assays showed that both were effective trypsin inhibitors, reducing 134

activity by 99.0% and 99.5%, respectively (Supplemental Figure 3). A protein model of 135

BBI3 predicted a double-headed structure similar to that of legume BBIs (Figure 2). 136

Based on the protein model generated, Lys67 and Lys105 were identified as the 137

putative P1 Lys residues (Figure 2). However, a third region within the BBI3 sequence 138

shares similarity with the BBI inhibitory motif and included a third potential inhibitory P1 139

residue, Lys77. In order to provide support to our protein model and to test which 140

predicted P1 residues contribute to the inhibition by BBI3, we generated a series of 141

BBI3 mutants (Figure 3). Mutating Lys67, Lys77 or Lys105 to Ala in isolation had no 142

effect on the ability of BBI3 to inhibit trypsin. Similarly, a double mutant containing 143

Lys77Ala and Lys105Ala did not reduce inhibition of trypsin activity compared to the 144

wildtype control. The inhibitory activity of the Lys67Ala/Lys105Ala BBI3 double mutant, 145

however, was completely abolished (Figure 3). Furthermore, the triple mutant showed 146

7

no difference in inhibition compared to the Lys67Ala/Lys105Ala double mutant (Figure 147

3). Taken together, these data suggest that a combination of Lys67Ala and Lys105Ala 148

mutations are sufficient to completely abolish inhibitory activity to a level equivalent to 149

that of a negative control and that Lys77 is not an inhibitory residue. 150

BBIs are widely distributed in the plant kingdom 151

The discovery of BBI-like sequences in S. moellendorffii suggests there was an ancient 152

common ancestor for BBIs in legumes and cereals. If so, BBIs should be more widely 153

distributed in the plant kingdom. Comprehensive BLAST searches in published plant 154

genomes identified BBI-like sequences in plant lineages outside legumes and cereals. 155

BBI-like sequences were not identified in the genomes of the two bryophyte species 156

searched. At the time the BLAST searches were completed, there were no published 157

genomes for monilophytes (ferns and horsetails), which diverged after the evolution of 158

lycopods and prior to the evolution of angiosperms. We did not find any sequence in the 159

genome of loblolly pine (Pinus taeda), the only gymnosperm genome searched. All plant 160

genomes searched are included in Figure 4 along with the number of BBIs identified in 161

each genome. 162

To provide experimental support of the publicly available data, we chose to confirm the 163

presence of three of the BBI-like sequences from banana (Musa acuminata), which 164

does not belong to the legume or cereal families. Sequences that matched the publicly 165

available sequences were cloned from banana leaf genomic DNA (Figure 5). This 166

supports the bioinformatics finding that angiosperm-type BBIs are found in plant 167

lineages outside legumes and cereals. As we did not identify BBI-like sequences in the 168

other Selaginella species, S. kraussiana and S. martensii, we assembled a 169

transcriptome of an available lycopod, Isoetes drummondii, and found a BBI-like 170

sequence. The I. drummondii BBI-like sequence was verified by cloning and sequencing 171

the full-length cDNA (Figure 5). Therefore, despite not being expressed in 172

S. kraussiana and S. martensii, BBIs likely are present in the common ancestor of 173

Selaginella and Isoetes. 174

The number of BBIs identified in each plant genome searched was mapped onto a 175

phylogenetic reconstruction derived from trees from Cogepedia and Angiosperm 176

8

Phylogeny Group III (Figure 4) 177

(http://genomevolution.org/wiki/index.php/Sequenced_plant_genomes;(Angiosperm 178

Phylogeny Group, 2016). BBIs appear to be maintained throughout monocot evolution, 179

with examples of BBI-like sequences found in all monocot species genomes searched 180

with the exception of duckweed (Spirodela polyrhiza). In dicots, the presence and 181

absence of genes indicates a complex pattern of gene loss. The identification of a BBI-182

like sequence in the basal eudicot, the Colorado blue columbine (Aquilegia coerulea), 183

suggests that dicot BBIs share a common ancestor followed by several independent 184

loss events. BBI-like sequences were not found in any asterid species searched, 185

suggesting there was a loss event prior to their divergence. This is true for Malphigiales 186

and rosid clade II as well. The pattern of loss events in rosid clade II is more complex, 187

with BBI-like sequences found in Fabaceae and Rhamnaceae but apparently lost in all 188

other rosid clade II species searched. 189

All dicot sequences analysed had the typical double-headed BBI motif with the 190

exception of one sequence from jujube (Ziziphus jujuba) (XP_015882735.1), which has 191

lost two inhibitory loop forming Cys residues and another that appears to have lost one 192

of the two loop forming residues (XP_015882689.1). As predicted, most monocot 193

sequences lack the conserved Cys residues forming the second inhibitory loop. Three 194

sequences from the monocot species, Musa acuminata, have two inhibitory loops 195

similar to dicot sequences, two of which were confirmed by sequences as well as a third 196

that lacks the two Cys residues similarly to other monocot sequences (Figure 5). This 197

suggests these sequences could represent the ancestral form of monocot BBIs prior to 198

the predicted loss of the two Cys residues of the second BBI loop. One of the BBI-like 199

sequences from Amborella trichopoda is predicted by sequence homology to possess 200

the same double-headed motif as dicots, suggesting this is likely the ancestral structure 201

of angiosperm BBIs. However, four of the five A. trichopoda BBI-like sequences have an 202

additional Cys residue within their first BBI-loop, suggesting they may no longer function 203

as a protease inhibitors and have potentially evolved a new function. 204

BBI-like sequences identified in angiosperms as well as those found in lycopods were 205

used in phylogenetic analysis (Figure 6). The phylogenetic tree of full-length sequences 206

9

was divided into six clades based on the branching pattern of the tree. The lycopod 207

clade forms a species-specific clade with 100% bootstrap support, strongly suggesting 208

that they share a single common ancestor. All angiosperm BBIs are predicted to have 209

evolved from a single ancestral BBI sequence as they together form a single clade 210

separate from lycopod BBIs. The dicot sequences all fall within a single clade. Aquilegia 211

coerulea and Ziziphus jujuba each form species-specific nodes with strong (≥88.6%) 212

bootstrap support, suggesting the sequences belonging to these nodes probably 213

duplicated following speciation. A duplication event occurring prior to the divergence of 214

Glycine max and Phaseolus vulgaris likely resulted in two paralagous sequences, with 215

one paralog being lost in Phaseolus vulgaris, as all Fabaceae sequences fall into two 216

distinct bootstrap supported (≥63.3%) nodes. 217

Nearly all monocot sequences fall within three clades designated the monocot clade, 218

Poaceae clade I, and Poaceae clade II. The nodes leading to these three clades have 219

weak bootstrap support and therefore conclusions on the relationships between these 220

groups cannot be made. Both Poaceae clades contain nodes with bootstrap support 221

that contain sequences from more than one species, suggesting potential independent 222

duplication events resulting in multiple paralogous sequences within Poaceae. The 223

African oil palm (Elaeis guineensis) and pineapple (Ananas comosus), both sister to 224

Poaceae, each form species-specific nodes with bootstrap support (≥73.5%) within the 225

monocot clade. This suggests that the BBI sequences belonging to each species 226

evolved from a single common ancestral sequence, so all monocot sequences share a 227

common ancestral sequence. 228

Musa acuminata BBI-like sequences showed the highest sequence divergence from 229

other BBIs, with the five sequences analysed falling into three distinct nodes. 230

Potentially, species-specific duplication events occurred followed by sequence 231

divergence. Only one M. acuminata sequence falls into the monocot clade and the other 232

sequences form two distinct nodes from other monocot sequences. Sequences from 233

other species do fall within these nodes, but with very low bootstrap support. Possibly, 234

duplication events leading to the divergent sequences are independent of other 235

monocots and could therefore be a result of the three whole genome duplication events 236

10

in the Musa lineage (D'Hont et al., 2012). Alternatively, these divergent M. acuminata 237

sequences could be the result of a duplication event in the ancestor of monocots and 238

therefore could be paralogs of sequences found in the monocot clade and Poaceae 239

clade II. Any definitive conclusions cannot be made due to the lack of bootstrap support 240

for the branches leading to these sequences. 241

DISCUSSION 242

BBIs were first studied 70 years ago and since then their structure and function have 243

been well defined (Bowman, 1946). Here we report five BBI genes from the seedless, 244

vascular plant S. moellendorffii. The genes identified are highly divergent from those in 245

legumes and cereals but possess the conserved trypsin inhibitory motif. They also 246

possess a conserved primary protein architecture, sharing a conserved pattern of Cys 247

residues around the inhibitory loop motifs, and all but one S. moellendorffii BBI 248

possesses a predicted ER signal. A predicted structure of a S. moellendorffii BBI 249

modelled against other known BBIs suggests a similar double-headed structure. 250

Together, these results support the notion that the BBI-like genes identified here are 251

related to BBIs in angiosperms. Selaginella represents the most ancient extant lineage 252

of vascular plants and therefore the identification of these BBIs suggests an ancient 253

ancestral origin of BBIs in angiosperms. 254

We have determined that at least one S. moellendorffii BBI protein, BBI3, is a functional 255

trypsin inhibitor. The structure of BBI3 is predicted to be double-headed based on 256

homology-based modelling. However, due to high sequence divergence outside of the 257

predicted inhibitory sequence motifs and the presence of a third potential inhibitory Lys 258

residue, we further tested the predicted double-headed structure by analysing modified 259

proteins. By mutagenesis, we demonstrated that Ala substitutions at both K67 and K105 260

were required to abolish BBI3 inhibitor activity, suggesting that there are two 261

independent sites that bind to and inhibit trypsin. Consistent with this, single Ala 262

substitutions at K67 and K105 did not affect BBI3 inhibition. The residues K67 and K105 263

reside within the short BBI3 sequences that displayed high similarity to the angiosperm 264

BBI inhibitory motifs and are predicted to share a similar loop structure from our protein 265

model. Mutation of a third Lys residue that also fell within a short sequence that shared 266

11

similarity to BBI inhibitory motifs, but was not predicted to be part of an inhibitory loop 267

from our protein model, had no effect on BBI3 inhibition. Studies of short peptide mimics 268

of the BBI inhibitory loop showed similar results when the P1 residue was mutated 269

(Terada et al., 1978; Domingo et al., 1995). 270

Since BBIs had thus far only been described in legumes and cereals (Mello et al., 2003; 271

Qi et al., 2005), and these two plant families are distantly related, we reasoned that if 272

the BBIs in these two families share an ancient ancestral origin, they are likely more 273

widely distributed in angiosperms. We demonstrated that BBIs are indeed found in 274

angiosperm lineages outside the legumes and grasses. Supporting a hypothesis for a 275

common ancestral origin for angiosperm BBIs, we identified BBI-like sequences in what 276

is considered to be the most basal living representative of angiosperms, Amborella 277

trichocarpa (Soltis et al., 2008). 278

BBIs appear to be widespread throughout the monocot lineage, being absent from only 279

one of the genomes searched. In the evolutionary scheme proposed by Mello et 280

al.(2003), the loss of two Cys residues in the monocot lineage results in the loss of 281

functionality of the second inhibitory loop and the double-headed structure characteristic 282

of dicot BBIs. Our findings support this model, as we observed a loss of Cys residues in 283

monocot BBI-like sequences; however, we also demonstrated that the loss of the two 284

inhibitory loop-forming Cys residues potentially occurred following the divergence of 285

monocots from dicots, as we found double-headed BBI-like sequences in the monocot 286

M. acuminata. 287

Given our identification of the Selaginella BBIs and the now apparent widespread 288

distribution in angiosperms, we expected to find BBI-like sequences in gymnosperms, 289

which diverged after lycopods, but prior to the evolution of angiosperms. However, 290

searches failed to identify any BBI-like sequences in P. taeda. This could be due to the 291

lack of genetic data for other gymnosperm species, or potentially the BBIs might have 292

been lost during evolution. Alternatively, BBI-like proteins might have evolved 293

independently in Selaginella and are not related to those found in angiosperms. 294

However, this is unlikely given the high sequence similarity of the inhibitory motif and 295

shared inhibitory function. 296

12

BBIs appear to have been lost in several angiosperm lineages. The main mechanisms 297

for gene loss, as reviewed by Albalat and Canestro (2016), are through unequal 298

crossing over during meiosis, the mobilization of a transposable element leading to 299

physical loss of the gene from the genome, or through the introduction of iterative 300

mutations resulting in pseudogenization or new functionality. Examples of BBI gene loss 301

have been observed in pea (Pisum sativum) germplasm collections, with evidence of 302

partial BBI sequences resulting from premature stop codons (Clemente et al., 2015). 303

However, we failed to identify any sequences, including partial sequences, showing 304

similarity to BBIs in whole plant lineages including rosid class II and the asterids, 305

suggesting gene loss occurred by complete gene loss or through the introduction of 306

mutations resulting in the sequence becoming unrecognizable as a BBI. The loss of 307

BBIs is likely prevalent through angiosperm evolution as a result of functional 308

redundancy, as there are several families of protease inhibitors in plants that could 309

compensate for the loss of BBIs (Habib and Fazili, 2007). For example, the serpin family 310

of protease inhibitors is widespread throughout angiosperms, including species lacking 311

BBIs such as Arabidopsis and cucumber (Roberts and Hejgaard, 2008). Species-312

specific gene loss has been observed for other protein-based plant defences such as 313

polyphenol oxidases (Tran et al., 2012). Gene loss is common and occurs more 314

frequently for genes coding for non-essential proteins that have minor influences on 315

plant fitness (Hirsh and Fraser, 2001; Krylov et al., 2003). 316

The evolutionary split between Isoetes and Selaginella is predicted to have occurred 317

370 million years ago (Arrigo et al., 2013). Therefore, the common ancestor of the BBI-318

like sequences we have found in Isoetes and Selaginella is at least this old and pre-319

dates the evolution of angiosperms, which are predicted to have evolved between 167-320

199 million years ago (Bell et al., 2010). Both the Isoetes and Selaginella BBI-like 321

sequences have a double-headed motif that is observed in legume BBIs and therefore it 322

can be predicted that the ancestor of both lycopod and angiosperm BBIs had a double-323

headed motif. This is supported by the identification of BBI-like sequences in the basal 324

angiosperm species A. trichopoda. Although the double-headed motif likely arose from 325

an internal gene duplication of a single inhibitory loop, we found no current day 326

examples of a single-headed BBI other than those found in cereals, whose sequences 327

13

suggest they lost two Cys residues resulting in the second loop becoming non-328

functional. 329

In conclusion, we have shown that the highly conserved BBI inhibitory motif has been 330

maintained through vascular plant evolution since at least prior to the divergence of 331

Isoetes and Selaginella. This supports the many studies on synthetic peptide mimics of 332

this inhibitory loop showing that the CT[K/R]SIPPXC motif is optimal for trypsin 333

inhibition. Having identified BBIs throughout the angiosperm lineage, it is clear that BBIs 334

in angiosperms share a common ancestor. We propose that the lycopod BBIs identified 335

here share an ancient ancestral origin with the well-characterized BBIs in angiosperms 336

given the sequence similarity at the conserved loops, a shared protein architecture, and 337

the common mode of action. 338

METHODS 339

BLAST searches using the CTKSIPPIC motif as bait 340

To determine whether the sequence CTKSIPPIC existed in different protein contexts 341

other than SFTI-1 and the legume and cereal BBIs, we performed a BLASTp analysis in 342

the BLAST portal of Phytozome v11.0 excluding the genus Helianthus (sunflower SFTI-343

1) and Fabaceae or Poaceae families (well-known BBI-containing families). The top 344

matches in this analysis were two unknown proteins from Selaginella moellendorffii 345

(BBI1 GenBank: XP_002980999, BBI2 GenBank: XP_002981001). To assess the 346

similarity to other BBIs, the BBI-like protein sequences were used as bait sequences in 347

a BLASTp search in the NCBI non-redundant protein database restricted to Fabaceae 348

and Poaceae sequences. The sequence similarity between the S. moellendorffii BBI-like 349

genes and their corresponding BLASTp hits were calculated in a global alignment using 350

the default settings of the ExPASy LALIGN program. 351

Plant tissue 352

Cuttings were taken from S. kraussiana and S. martensii growing at Hilltop Nursery 353

(Menzies Creek, Victoria, Australia), frozen on dry ice, and shipped to Western Australia 354

(entry into WA under Quarantine Inspector’s Direction No. KH090713). Cuttings were 355

taken from S. moellendorffii (Cat. #3228, Plant Delights Nursery Inc., Raleigh, NC USA) 356

14

growing at the Joint BioEnergy Institute. Plant Delights Nursery was the source of the 357

S. moellendorffii used to generate the published genome sequence (Banks et al., 2011). 358

The S. moellendorffii cuttings were preserved in RNAlater RNA Stabilization Reagent 359

(Ambion) and shipped on ice packs to Western Australia. 360

RNA extraction from S. moellendorffii, S. kraussiana, and S. martensii 361

Tissue from each species was ground with glass beads to a fine powder under liquid 362

nitrogen. Some of the thicker S. martensii stems were removed during the grinding. 363

Total RNA from approximately 0.3 mL of frozen tissue powder was extracted as 364

previously described (Mylne et al., 2012) using phenol:chloroform and a 2 M lithium 365

chloride RNA-selective precipitation. Contaminating genomic DNA was removed by 366

digesting total RNA with DNase and subsequent purification with a NucleoSpin RNA 367

Clean-up kit (Macherey-Nagel). The resulting total RNA was analysed on a NanoDrop 368

Spectrophotometer for its A260/280 and A260/230 ratios, and all three samples 369

displayed distinct banding when run on a 1% agarose gel, implying the mRNA was 370

intact. 371

Transcriptome sequencing, assembly, mining data for BBI-like sequences 372

Sequencing libraries were generated using the TruSeq® Stranded Total RNA LT with 373

Ribo-Zero Plant kit (Illumina) with 300-1000 ng of purified total RNA according to the 374

manufacturer’s instructions. Sequencing was then performed on an Illumina HiSeq 1500 375

instrument as 101 bp single read runs or 2 x 101 bp paired-end read runs. 376

The de novo transcriptome assemblies were done as described (Jayasena et al., 2014). 377

Raw reads were inspected for quality using FastQC 378

(www.bioinformatics.babraham.ac.uk/projects/fastqc/). Quality trimming and filtering 379

was done using the FASTX toolkit (hannonlab.cshl.edu/fastx_toolkit/). Raw reads were 380

trimmed to maintain a phred score of 30, which sets the base call accuracy to be 99.9%, 381

and the minimum length after trimming was set at 50. Trimmed reads were filtered with 382

a quality threshold of 22 and the percentage of bases that match the quality threshold 383

was set to 90. 384

15

De novo transcriptomes were assembled using CLC Genomics Workbench 7.0 385

(CLCBio). RNA from S. moellendorffii was sequenced only as single reads. S. martensii 386

and S. kraussiana were sequenced as paired-end reads. Each dataset was assembled 387

initially with the default CLC settings (i.e. word size: 23, bubble size: 50) and 388

subsequently with three different word sizes (i.e. 30, 40, and 60), which determines the 389

size the reads are fragmented into before being assembled, keeping all other 390

parameters at default settings. In each case, reads were mapped back to contigs 391

keeping the mapping parameters as follows: mismatch cost: 2, insertion cost: 3, deletion 392

cost: 3, length fraction: 0.8, similarity fraction: 0.8, update contigs: yes. Relevant 393

statistics are presented in Table 2. In this way, four transcriptomes for each of the three 394

species were assembled. 395

Each assembly was queried using tBLASTn for the two BBI-like genes identified from 396

the S. moellendorffii genome. 397

The sequence referred to from S. stauntoniana was found within the NCBI short read 398

archive (Run ERR364347) with the specific sequence ID within the dataset being 399

ERR364347.14864137.1. The pertinent region of the translated sequence encoded 400

TVCTMSYPPSCFCTKAPPNCGVGSSCCSD, in which the CTKSIPPIC-like motifs are 401

underlined. 402

Cloning of S. moellendorffii BBI-like sequences 403

5' and 3' RACE-ready cDNA was generated using the SMARTer RACE cDNA 404

Amplification Kit (Clontech). Primers for 3' and 5' RACE were designed against 405

sequences from the assembled transcriptome (for BBI3 and BBI4) and from the 406

published genome (for BBI1 and BBI2). All primers are included in Supplemental Table 407

1. Sequences were amplified using Taq DNA polymerase. Purified 3' and 5' RACE PCR 408

products were cloned into pGEM-T Easy vector (Promega) and sequenced. Primers 409

designed based on the RACE results were used to clone full-length sequences with Taq 410

DNA polymerase. For each PCR-amplified product, at least three independent clones 411

were sequenced to account for errors induced by Taq DNA polymerase. Because of the 412

high sequence similarity between BBI2 and BBI3, a single primer was designed directly 413

16

upstream of the polyA tail that matched both the BBI2 and BBI3 3' UTR sequences. 414

BBI2 and BBI3 full-length sequences were obtained with the same primer pair. 415

Sequencing and validation of BBI genes 416

Searches of the S. moellendorffii transcriptome identified two BBI-like sequences (BBI3; 417

BBI4) (Figure 2). These sequences were not initially found in the published genome 418

using BLASTp searches. Using the RNA sequences in BLASTn searches, BBI3 and 419

BBI4 were found in the genome v1.0 (BBI3, gene locus: scaffold 54: 291,062…291,531; 420

BBI4, gene locus: scaffold 53: 389,315...389,727). To verify the sequences identified in 421

both published genome and our transcriptome assembly, the sequences were amplified 422

by 5' and 3' RACE, cloned and sequenced. All sequences were further verified by 423

amplification of an identical sequence from genomic DNA, which also demonstrated the 424

genes for these transcripts are intronless. 425

We cloned a fifth BBI-like sequence (BBI5) using primers originally designed for BBI3 426

(Figure 2). We found BBI5 through searches in the genome v1.0 427

(scaffold 47: 346,127…346,596). Following identification of five BBI-like sequences, the 428

RNA-seq reads were mapped to all five BBI-like sequences (Table 1; Supplemental 429

Figure 1). Reads supporting each sequence indicated all were expressed. In the 430

current model for BBI2 (GenBank: XM_002980955.1), an intron is predicted. However, 431

upon cloning the full length sequence from cDNA we found no intron in the sequence. 432

We verified this by mapping the RNA-seq reads onto the genomic sequence with and 433

without the predicted intron present (Supplemental Figure 4). RNA-seq reads mapped 434

to the predicted intron sequence, but not across the exon-exon junction with the 435

predicted intron manually removed, confirming that BBI2 is actually intronless. 436

S. moellendorffii BBI mRNA abundance 437

To estimate the relative abundance of each BBI gene at the mRNA level, clean 438

RNA-seq reads from S. moellendorffii mRNA (NCBI SRA BioSample Accession: 439

SAMN05958544) were mapped onto the ORF of each gene using CLC Genomics 440

Workbench 6.5.1 (CLC bio). To avoid non-specific reads mapping due to the high 441

sequence similarity between the BBI-like genes, stringent parameters were used. 442

17

Mapping parameters were set to a length fraction of 1.0 and similarity fraction of 1.0 to 443

allow only perfect matches to map to each gene. 444

To determine if BBI2 was intronless as we predicted, reads were mapped to genomic 445

DNA with and without the currently annotated (GenBank: XP_002981001) intron using 446

the same parameters above (Supplemental Figure 4). 447

Homology modelling of S. moellendorffii BBI3 448

A protein model of BBI3 was generated using EasyModeller 4.0, a graphical user 449

interface for MODELLER. Only the BBI-like domain (BBI346-116) was used for the protein 450

model. The model was built against the following templates: soybean BBI (1BBI), the 451

snail medick (Medicago scutellata) BBI (1MVZ), and cowpea (Vigna unguiculata) BBI 452

(2R33), which show 28%, 31% and 28% sequence similarity with the BBI3 BBI domain, 453

respectively. 454

Recombinant protein expression, purification and mutagenesis 455

The protein encoded by BBI3 was chosen for expression in E. coli as it had the highest 456

number of reads map to its sequence, indicating that it is the most abundant BBI (Table 457

2; Supplemental Figure 1). A synthetic BBI3 ORF that included an N-terminal six-His 458

tag and a TEV protease cleavage site in lieu of its ER signal was designed with optimal 459

codon usage for E. coli (GENEART) (Supplemental Figure 2). The sequence was sub-460

cloned into the BamHI and SalI sites of pQE30 (Qiagen). The pQE30-BBI3 construct 461

and the suppressor plasmid pREP4 (Qiagen) were co-transformed into the E.coli strain 462

Shuffle Express (New England Biolabs, Cat. # C3028H) for protein production. 463

To test the predicted P1 inhibitory residues of BBI3, a series of mutant sequences were 464

generated by site-directed mutagenesis (Zheng et al., 2004). Three BBI3 motifs are 465

similar to the conserved BBI inhibitory motif. In all three cases, the putative P1 Lys was 466

mutated to Ala by site-directed PCR mutagenesis using primers listed in Supplemental 467

Table 1. 468

E. coli expressing BBI3 or BBI3 mutants were grown in 750 mL LB in 2 L flasks at 30°C 469

to an OD600 of 0.8, induced by isopropyl β-D-1-thiogalactopyranoside addition to 1 mM, 470

and incubated overnight at 16°C. Bacterial pellets (~5 mL) were resuspended in 25 mL 471

18

lysis buffer (50 mM Tris-HCl pH 8.0, 1 M sodium chloride, 30 mM imidazole) and lysed 472

by sonication. After centrifugation, the cleared lysate was incubated with Ni-NTA resin 473

(BioRad) at 4°C overnight with mild agitation. The resin was washed with lysis buffer 474

and eluted in elution buffer (50 mM Tris-HCl pH 8.0, 100 mM sodium chloride, 300 mM 475

imidazole). The protein was further purified by gel-filtration chromatography (BioLogic 476

DuoFlow, Bio-Rad) on a 10/300 Superdex™ 200 Increase column (GE Healthcare) or 477

10/300 Superdex™ 75 GL column (GE Healthcare) in gel filtration buffer (20 mM Tris-478

HCl pH 8.0, 50 mM sodium chloride). 479

Trypsin inhibition assay 480

Inhibition was determined as described (Prasad et al., 2010). Protein concentration was 481

determined with a Pierce™ BCA Protein Assay Kit (Thermo Scientific) using BSA as a 482

standard. BBI from soybean was used as a positive control (Sigma Aldrich, Cat. # 483

T9777). The inhibitors were diluted to 100 µg/mL in gel filtration buffer and 5 µL was 484

added to 20 µL of 25 µg/mL trypsin from bovine pancreas (Sigma Aldrich) dissolved in 485

50 mM Tris-Cl, 20 mM calcium chloride, pH 8.0 and incubated for 15 min at 37°C. 486

Residual trypsin activity was determined by the addition of 125 µL of 1 mM 487

N-α-benzoyl-DL-arginine-p-nitroanilide (BAPNA) substrate (Sigma Aldrich) dissolved in488

99% 50 mM Tris-HCl, 20 mM calcium chloride, pH 8.0, 1% (v/v) dimethyl sulfoxide and 489

incubation for 45 min at 37 °C. The reaction was stopped with the addition of 25 µL of 490

30% (v/v) acetic acid. The absorbance was measured at 410 nm. 491

Comprehensive BLAST searches and phylogenetic analysis 492

BLAST searches to identify BBI-like sequences from other species were performed 493

against the NCBI and phytozome genome databases using BBI sequences from 494

legumes and cereals as well as the Selaginella BBI sequences as baits. A list of 495

genome version numbers used for tBLASTn searches is included in Supplemental 496

Data Set 1. Searches were completed by tBLASTn using an E-value threshold of 50. 497

Sequences that shared similarity around the conserved motif were collected and pooled 498

into a single file. Only full-length sequences were used for analysis. Accession numbers 499

for all sequences are included in Supplemental Data Set 1. These sequences were 500

19

analysed in a PFAM batch sequence search and each sequence was found to be have 501

homology with the conserved BBI domain. 502

An alignment using the predicted amino acid sequences from all BBI-like sequences 503

identified through tBLASTn searches, as well as those from Selaginella moellendorffii 504

and Isoetes drumondii, was generated with ClustalW (2.0.12) followed by manual 505

editing with BioEdit Sequence Alignment Editor v7.2.5 (Hall, 1999). Other than manually 506

adjusting the alignment, no deletions were made to avoid reducing the accuracy of the 507

analysis. According to a comprehensive study of filtering methods by (Tan et al., 2015), 508

all current filtering methods reduce accuracy of the resulting phylogenetic analysis. A 509

phylogenetic tree was generated from a protein distance matrix using neighbour joining 510

methods using the Jones-Taylor-Thornton model in the Phylip package (v 3.67). Highly 511

similar sequences were not included. Sequences that resulted in very long branch 512

lengths were removed as potential gene model errors. Following the removal of these 513

sequences, the alignment and phylogenetic analysis was repeated with a 105 BBI-like 514

sequences (Supplemental File 1). Bootstrap values were calculated from 1,000 515

phylogenetic constructions using a distance neighbour joining analysis and mapped 516

back onto the original tree. The phylogenetic tree graphical representation was 517

generated with FigTree (v 1.4). 518

Sequencing BBI-like genes from banana and quillwort 519

Leaf tissue was collected from a Cavendish banana (Musa acuminata) tree sourced 520

from Bunnings Warehouse. Live Isoetes drummondii plants were collected by Kingsley 521

Dixon from Alison Baird Reserve, Kenwick in Western Australia. 522

I. drummondii was sequenced by Illumina NextSeq500 as 2×151 pair-end read length523

runs. The transcriptome was assembled as described for Selaginella species. Relevant 524

statistics are presented in Table 2. 525

To confirm the presence of BBI-like sequences in Musa acuminata, genomic DNA was 526

extracted using the DNeasy Plant Minikit (Qiagen). RNA was extracted from 527

I. drummondii as described for Selaginella species. Reverse transcription of this RNA528

was performed using a ProtoScript® II First Strand cDNA Synthesis Kit (NEB). Primers 529

20

designed to amplify the complete ORF from cDNA or genomic DNA are included in 530

Supplemental Table 1. Purified PCR products were cloned into pGEM-T Easy 531

(Promega) and sequenced. At least two independent clones were used to account for 532

errors introduced by Taq DNA polymerase. 533

Accession numbers 534

Sequence data generated for this article can be found in the NCBI short read archive 535

and GenBank under accession numbers SRP092379 and KY069178-KY069186, 536

respectively. Protein Data Bank codes for data used in figures are: SFTI-1 1JBL; Gmaa 537

1BBI; HvuBBIa 2FJ8; MscBBI 1MVZ. GenBank accession numbers for sequences used 538

in figures are Gmab: ACU13240; HvuBBIb: BAJ91702. 539

Supplemental Data 540

Supplemental Figure 1. Relative BBI transcript abundance in S. moellendorffii tissue 541

shown as the number of clean RNAseq reads that mapped to each open reading frame. 542

Supplemental Figure 2. BBI3 synthetic protein sequence used for recombinant protein 543

production and SDS-PAGE gel of expressed protein before and after TEV cleavage to 544

remove the 6-His tag. 545

Supplemental Figure 3. Trypsin inhibition by S. moellendorffii BBI3 and Glycine max 546

BBI. 547

Supplemental Figure 4. Mapping of RNA-seq reads to genomic DNA sequence shows 548

that BBI2 is intronless. 549

Supplemental Table 1. All primers used in this study. 550

Supplemental Data Set 1. Accession numbers and sequence information for 551

sequences used to generate the phylogeny presented in Figure 5. Sequences used in 552

the phylogeny presented in Figure 5 have a corresponding ID. Assembly version 553

numbers are given for genomes accessed through the Phytozome BLAST portal. 554

Assembly accession numbers are given for genomes accessed through the NCBI 555

21

BLAST portal. Genomes for which no BBI-like sequences were found are shown in red 556

text. 557

Supplemental File 1. Sequence alignment of full-length BBI sequences used to 558

generate the BBI phylogeny shown in Figure 6. 559

ACKNOWLEDGMENTS 560

The authors thank J. Heazlewood for S. moellendorffii cuttings and staff at Hilltop 561

Nursery for cuttings from S. kraussiana and S. martensii. A.M.J. was supported by a 562

University International Stipend and a Scholarship for International Research Fees from 563

UWA; A.J. was supported by an International Postgraduate Research Scholarship and 564

an Australian Postgraduate Award from UWA; J.Z. was supported by an International 565

Postgraduate Research Scholarship and an Australian Postgraduate Award from UWA; 566

G.J.K. was supported by a Hackett Postgraduate Scholarship from UWA; J.W. was 567

supported by an Australian Research Council (ARC) Centre of Excellence grant 568

CE140100008; J.S.M. was supported by an ARC Future Fellowship (FT120100013). 569

This work was supported by ARC grant DP130101191. 570

AUTHOR CONTRIBUTIONS 571

J.S.M. conceived the study. A.M.J. designed and performed all experimentation with the 572

exception of the following: J.S.M. performed RNA extractions; A.S.J. and J.Z. 573

assembled transcriptomes; D.S., O.B., and J.W. prepared libraries and performed 574

sequencing; all authors analyzed the data; G.J.K. and C.B. optimized methods for 575

protein purification. A.M.J. and J.S.M. wrote the manuscript with contributions from all 576

authors. 577

Table 1. SmoBBI mRNAexpression. 578

The number of RNA-seq clean reads mapping to each SmoBBI open reading frame and 579

the average coverage, which is indicative of expression level. Average coverage 580

represents the average number of reads mapping to each position of the reference 581

gene. Mapped reads are shown as a graphical representation in Supplemental Figure 582

1. 583

22

Gene Mapped reads

Average Coverage

SmoBBI1 77 20. 7SmoBBI2 1,539 413.2SmoBBI3 2,383 639.7SmoBBI4 30 11.7SmoBBI5 589 158.1

Table 2. De novo transcriptome assembly statistics. 584

The number of contigs, number of raw and clean reads, the word size used to generate 585

each assembly and the N50 statistic (summary of the lengths of the longest contigs until 586

50% of the total contig length is reached) are shown. 587

Species Raw reads Clean reads Word size Contigs N50 S. moellendorffii 63,151,678 56,711,459 40 (Single) 44,564 584 S. martensii 90,146,302 82,241,287 40 (Paired) 51,848 557 S. kraussiana 145,585,252 135,198,965 60 (Paired) 44,996 780 I. drummondii 103,198,884 68,341,138 60 (Paired) 264,900 427 588

589

References 590

Albalat, R., and Canestro, C. (2016). Evolution by gene loss. Nat. Rev. Genet. 17, 379-391. 591 Angiosperm Phylogeny Group. (2016). An update of the Angiosperm Phylogeny Group classification for 592

the orders and families of flowering plants: APG IV. Bot. J. Linn. Soc. 181, 1-20. 593 Arrigo, N., Therrien, J., Anderson, C.L., Windham, M.D., Haufler, C.H., and Barker, M.S. (2013). A total 594

evidence approach to understanding phylogenetic relationships and ecological diversity in 595 Selaginella subg. Tetragonostachys. Am. J. Bot. 100, 1672-1682. 596

Banks, J.A. (2009). Selaginella and 400 million years of separation. Annu. Rev. Plant Biol. 60, 223-238. 597 Banks, J.A., Nishiyama, T., Hasebe, M., Bowman, J.L., Gribskov, M., dePamphilis, C., Albert, V.A., Aono, 598

N., Aoyama, T., Ambrose, B.A., Ashton, N.W., Axtell, M.J., Barker, E., Barker, M.S., Bennetzen, 599 J.L., Bonawitz, N.D., Chapple, C., Cheng, C., Correa, L.G.G., Dacre, M., DeBarry, J., Dreyer, I., 600 Elias, M., Engstrom, E.M., Estelle, M., Feng, L., Finet, C., Floyd, S.K., Frommer, W.B., Fujita, T., 601 Gramzow, L., Gutensohn, M., Harholt, J., Hattori, M., Heyl, A., Hirai, T., Hiwatashi, Y., 602 Ishikawa, M., Iwata, M., Karol, K.G., Koehler, B., Kolukisaoglu, U., Kubo, M., Kurata, T., 603 Lalonde, S., Li, K., Li, Y., Litt, A., Lyons, E., Manning, G., Maruyama, T., Michael, T.P., Mikami, 604 K., Miyazaki, S., Morinaga, S.i., Murata, T., Mueller‐Roeber, B., Nelson, D.R., Obara, M., Oguri, 605 Y., Olmstead, R.G., Onodera, N., Petersen, B.L., Pils, B., Prigge, M., Rensing, S.A., Riaño‐606 Pachón, D.M., Roberts, A.W., Sato, Y., Scheller, H.V., Schulz, B., Schulz, C., Shakirov, E.V., 607 Shibagaki, N., Shinohara, N., Shippen, D.E., Sørensen, I., Sotooka, R., Sugimoto, N., Sugita, M., 608

23

Sumikawa, N., Tanurdzic, M., Theißen, G., Ulvskov, P., Wakazuki, S., Weng, J.K., Willats, 609 W.W.G.T., Wipf, D., Wolf, P.G., Yang, L., Zimmer, A.D., Zhu, Q., Mitros, T., Hellsten, U., Loqué, 610 D., Otillar, R., Salamov, A., Schmutz, J., Shapiro, H., Lindquist, E., Lucas, S., Rokhsar, D., and 611 Grigoriev, I.V. (2011). The Selaginella genome identifies genetic changes associated with the 612 evolution of vascular plants. Science 332, 960-963. 613

Bell, C.D., Soltis, D.E., and Soltis, P.S. (2010). The age and diversification of the angiosperms re-614 revisited. Am. J. Bot. 97, 1296-1303. 615

Birk, Y. (1961). Purification and some properties of a highly active inhibitor of trypsin and α-616 chymotrypsin from soybeans. Biochim. Biophys. Acta 54, 378-381. 617

Birk, Y., Gertler, A., and Khalef, S. (1963). A pure trypsin inhibitor from soya beans. Biochem. J. 87, 281. 618 Bowman, D.E. (1946). Differentiation of soy bean antitryptic factors. Proc. Soc. Exp. Biol. Med. 63, 547-619

550. 620 Brauer, A.B.E., and Leatherbarrow, R.J. (2003). The conserved P1' Ser of Bowman-Birk-type proteinase 621

inhibitors is not essential for the integrity of the reactive site loop. Biochem. Biophys. Res. 622 Commun. 308, 300-305. 623

Brauer, A.B.E., Domingo, G.J., Cooke, R.M., Matthews, S.J., and Leatherbarrow, R.J. (2002). A 624 conserved cis peptide bond is necessary for the activity of Bowman-Birk inhibitor protein. 625 Biochemistry 41, 10608-10615. 626

Clemente, A., Arques, M.C., Dalmais, M., Le Signor, C., Chinoy, C., Olias, R., Rayner, T., Isaac, P.G., 627 Lawson, D.M., Bendahmane, A., and Domoney, C. (2015). Eliminating anti-nutritional plant 628 food proteins: the case of seed protease inhibitors in pea. PLoS One 10, e0134634. 629

D'Hont, A., Denoeud, F., Aury, J.M., Baurens, F.C., Carreel, F., Garsmeur, O., Noel, B., Bocs, S., Droc, G., 630 Rouard, M., Da Silva, C., Jabbari, K., Cardi, C., Poulain, J., Souquet, M., Labadie, K., Jourda, C., 631 Lengelle, J., Rodier‐Goud, M., Alberti, A., Bernard, M., Correa, M., Ayyampalayam, S., McKain, 632 M.R., Leebens‐Mack, J., Burgess, D., Freeling, M., Mbeguie, A.M.D., Chabannes, M., Wicker, T., 633 Panaud, O., Barbosa, J., Hribova, E., Heslop‐Harrison, P., Habas, R., Rivallan, R., Francois, P., 634 Poiron, C., Kilian, A., Burthia, D., Jenny, C., Bakry, F., Brown, S., Guignon, V., Kema, G., Dita, 635 M., Waalwijk, C., Joseph, S., Dievart, A., Jaillon, O., Leclercq, J., Argout, X., Lyons, E., Almeida, 636 A., Jeridi, M., Dolezel, J., Roux, N., Risterucci, A.M., Weissenbach, J., Ruiz, M., Glaszmann, J.C., 637 Quetier, F., Yahiaoui, N., and Wincker, P. (2012). The banana (Musa acuminata) genome and 638 the evolution of monocotyledonous plants. Nature 488, 213-217. 639

Domingo, G.J., Leatherbarrow, R.J., Freeman, N., Patel, S., and Weir, M. (1995). Synthesis of a mixture 640 of cyclic peptides based on the Bowman-Birk reactive site loop to screen for serine protease 641 inhibitors. Int. J. Pept. Protein Res. 46, 79-87. 642

Elliott, A.G., Delay, C., Liu, H., Phua, Z., Rosengren, K.J., Benfield, A.H., Panero, J.L., Colgrave, M.L., 643 Jayasena, A.S., Dunse, K.M., Anderson, M.A., Schilling, E.E., Ortiz‐Barrientos, D., Craik, D.J., 644 and Mylne, J.S. (2014). Evolutionary origins of a bioactive peptide buried within preproalbumin. 645 Plant Cell 26, 981-995. 646

Engelken, J., Brinkmann, H., and Adamska, I. (2010). Taxonomic distribution and origins of the extended 647 LHC (light-harvesting complex) antenna protein superfamily. BMC Evol. Biol. 10, 233. 648

Gariani, T., McBride, J.D., and Leatherbarrow, R.J. (1999). The role of the P2' position of Bowman-Birk 649 proteinase inhibitor in the inhibition of trypsin. Studies on P2' variation in cyclic peptides 650 encompassing the reactive site loop. Biochim. Biophys. Acta 1431, 232-237. 651

Habib, H., and Fazili, K.M. (2007). Plant protease inhibitors: a defense strategy in plants. Biotechnol. 652 Mol. Biol. Rev. 2, 68-85. 653

Hall, T.A. (1999). BioEdit: a user-friendly biological sequence alignment editor and analysis program for 654 Windows 95/98/NT. Nucleic Acids Symp. S. 41, 95-98. 655

24

Hazkani‐Covo, E., Altman, N., Horowitz, M., and Graur, D. (2002). The Evolutionary History of 656 Prosaposin: Two Successive Tandem-Duplication Events Gave Rise to the Four Saposin Domains 657 in Vertebrates. J. Mol. Evol. 54, 30-34. 658

Hirsh, A.E., and Fraser, H.B. (2001). Protein dispensability and rate of evolution. Nature 411, 1046-1049. 659 Jayasena, A.S., Secco, D., Bernath‐Levin, K., Berkowitz, O., Whelan, J., and Mylne, J.S. (2014). Next 660

generation sequencing and de novo transcriptomics to study gene evolution. Plant Methods 10, 661 34. 662

Krylov, D.M., Wolf, Y.I., Rogozin, I.B., and Koonin, E.V. (2003). Gene loss, protein sequence divergence, 663 gene dispensability, expression level, and interactivity are correlated in eukaryotic evolution. 664 Genome Res. 13, 2229-2235. 665

Li, J., Zhang, C., Xu, X., Wang, J., Yu, H., Lai, R., and Gong, W. (2007). Trypsin inhibitory loop is an 666 excellent lead structure to design serine protease inhibitors and antimicrobial peptides. FASEB J. 667 21, 2466-2473. 668

Luckett, S., Garcia, R.S., Barker, J.J., Konarev, A.V., Shewry, P.R., Clarke, A.R., and Brady, R.L. (1999). 669 High-resolution structure of a potent, cyclic proteinase inhibitor from sunflower seeds. J. Mol. 670 Biol. 290, 525-533. 671

McBride, J.D., Brauer, A.B.E., Nievo, M., and Leatherbarrow, R.J. (1998). The role of threonine in the P2 672 position of Bowman-Birk proteinase inhibitors: studies on P2 variation in cyclic peptides 673 encompassing the reactive site loop. J. Mol. Biol. 282, 447-457. 674

McBride, J.D., Watson, E.M., Brauer, A.B., Jaulent, A.M., and Leatherbarrow, R.J. (2002). Peptide 675 mimics of the Bowman-Birk inhibitor reactive site loop. Biopolymers 66, 79-92. 676

Mello, M.O., Tanaka, A.S., and Silva‐Filho, M.C. (2003). Molecular evolution of Bowman–Birk type 677 proteinase inhibitors in flowering plants. Mol. Phylogenet. Evol. 27, 103-112. 678

Mylne, J.S., Colgrave, M.L., Daly, N.L., Chanson, A.H., Elliott, A.G., McCallum, E.J., Jones, A., and Craik, 679 D.J. (2011). Albumins and their processing machinery are hijacked for cyclic peptides in 680 sunflower. Nat. Chem. Biol. 7, 257-259. 681

Mylne, J.S., Chan, L.Y., Chanson, A.H., Daly, N.L., Schaefer, H., Bailey, T.L., Nguyencong, P., Cascales, L., 682 and Craik, D.J. (2012). Cyclic peptides arising by evolutionary parallelism via asparaginyl-683 endopeptidase-mediated biosynthesis. Plant Cell 24, 2765-2778. 684

Nishino, N., Aoyagi, H., Kato, T., and Izumiya, N. (1977). Studies on the synthesis of proteinase 685 inhibitors. I. Synthesis and activity of nonapeptide fragments of soybean Bowman-Birk inhibitor. 686 J. Biochem. 82, 901-909. 687

Odani, S., and Ikenaka, T. (1973). Studies on soybean trypsin inhibitors VIII. Disulfide bridges in soybean 688 Bowman-Birk proteinase inhibitor. J. Biochem. 74, 697-715. 689

Park, E.Y., Kim, J.A., Kim, H.W., Kim, Y.S., and Song, H.K. (2004). Crystal structure of the Bowman-Birk 690 inhibitor from barley seeds in ternary complex with porcine trypsin. J. Mol. Biol. 343, 173-186. 691

Petersen, T.N., Brunak, S., von Heijne, G., and Nielsen, H. (2011). SignalP 4.0: discriminating signal 692 peptides from transmembrane regions. Nat. Meth. 8, 785-786. 693

Prakash, B., Selvaraj, S., Murthy, M.R., Sreerama, Y.N., Rao, D.R., and Gowda, L.R. (1996). Analysis of 694 the amino acid sequences of plant Bowman-Birk inhibitors. J. Mol. Evol. 42, 560-569. 695

Prasad, E.R., Merzendorfer, H., Madhurarekha, C., Dutta‐Gupta, A., and Padmasree, K. (2010). 696 Bowman−Birk Proteinase inhibitor from Cajanus cajan seeds: purification, characterization, and 697 insecticidal properties. J. Agric. Food Chem. 58, 2838-2847. 698

Qi, R.F., Song, Z.W., and Chi, C.W. (2005). Structural features and molecular evolution of Bowman-Birk 699 protease inhibitors and their potential application. Acta Biochim. Biophys. Sin. 37, 283-292. 700

Rawlings, N.D., Barrett, A.J., and Finn, R. (2016). Twenty years of the MEROPS database of proteolytic 701 enzymes, their substrates and inhibitors. Nucleic Acids Res. 44, D343-D350. 702

25

Roberts, T.H., and Hejgaard, J. (2008). Serpins in plants and green algae. Funct. Integr. Genomics 8, 1-703 27. 704

Schechter, I., and Berger, A. (1967). On the size of the active site in proteases. I. Papain. Biochem. 705 Biophys. Res. Commun. 27, 157-162. 706

Soltis, D.E., Albert, V.A., Leebens‐Mack, J., Palmer, J.D., Wing, R.A., dePamphilis, C.W., Ma, H., 707 Carlson, J.E., Altman, N., Kim, S., Wall, P.K., Zuccolo, A., and Soltis, P.S. (2008). The Amborella 708 genome: an evolutionary reference for plant biology. Genome Biol 9, 402. 709

Song, H.K., Kim, Y.S., Yang, J.K., Moon, J., Lee, J.Y., and Suh, S.W. (1999). Crystal structure of a 16 kDa 710 double-headed Bowman-Birk trypsin inhibitor from barley seeds at 1.9 Å resolution. J. Mol. Biol. 711 293, 1133-1144. 712

Tan, G., Muffato, M., Ledergerber, C., Herrero, J., Goldman, N., Gil, M., and Dessimoz, C. (2015). 713 Current Methods for Automated Filtering of Multiple Sequence Alignments Frequently Worsen 714 Single-Gene Phylogenetic Inference. Syst. Biol. 64, 778-791. 715

Terada, S., Sato, K., Kato, T., and Izumiya, N. (1978). Inhibitory properties of nonapeptide loop 716 structures related to reactive sites of soybean Bowman-Birk inhibitor. FEBS Lett. 90, 89-92. 717

Tran, L.T., Taylor, J.S., and Constabel, C.P. (2012). The polyphenol oxidase gene family in land plants: 718 Lineage-specific duplication and expansion. BMC Genomics 13, 395. 719

Zheng, L., Baumann, U., and Reymond, J.L. (2004). An efficient one-step site-directed and site-720 saturation mutagenesis protocol. Nucl. Acids Res. 32, e115. 721

722

26

723

Figure Captions 724

Figure 1. The CTKSIPPIC motif: structures, sequences and contexts. 725

(A) Structural alignment of SFTI-1 (PDB code: 1JBL), a BBI from soybean (GmaBBIa; 726

PDB code: 1BBI) and a BBI from barley (HvuBBIa; PDB code: 2FJ8). The trypsin 727

inhibitory loop of all three sequences (boxed) share structural similarity. 728

(B) BBIs are common in the phylogenetically separate cereals and legumes marked on 729

an angiosperm phylogeny of rbcL sequences (Elliott et al., 2014). 730

(C) Sequence alignment of SFTI-1, GmaBBI and HvuBBI. The homologous inhibitory 731

‘heads’ are indicated with black brackets. Inhibitory loops are indicated with Roman 732

numerals and the corresponding loops are also labelled in panel A. The soybean 733

sequence exemplifies the “double-headed” structure, with two homologous inhibitory 734

motifs, the second of which has been lost in cereal BBIs as a result of the loss of two 735

Cys residues (inhibitory loop II). Many monocot BBIs have undergone internal gene 736

duplications resulting in multiple inhibitory loops (e.g. inhibitory loop III). 737

Figure 2. S. moellendorffii BBI3 is predicted to be a BBI based on sequence similarity at 738

the inhibitory motifs and shared primary protein architecture. 739

(A) BOXSHADE alignment of S. moellendorffii BBI-like (SmoBBI) protein sequences 740

with soybean BBI (GmaBBIb) and barley BBI (HvuBBIb) shows conservation of the 741

trypsin inhibitory loop (grey box). All proteins, with the exception of SmoBBI4, share a 742

similar protein architecture, with an ER signal (line above sequence) followed by Cys-743

rich sequence with two spatially distinct conserved inhibitory motifs. Cys residues that 744

are conserved in all sequences are indicated with asterisks. 745

(B) Protein homology model of SmoBBI346-116 aligned with Medicago scutellata BBI 746

(PDB code: 1MVZ) is shown along with the sequence alignment. The conserved 747

inhibitory loops are shown as opaque, while the remaining protein is translucent and the 748

corresponding inhibitory motifs are boxed in the alignment. The side chains of the 749

labelled P1 residues (R64, K67, R90, K105) and disulfide bonds are displayed in stick 750

format. 751

27

Figure 3. Trypsin inhibition by wild-type and mutant S. moellendorffii BBI3. 752

(A) Full-length wild-type BBI3 showing predicted inhibitory residues (bold). 753

(B) Partial BBI3 sequences showing mutations (black highlight). 754

(C) Inhibitors were incubated with trypsin and a colorogenic substrate of trypsin whose 755

activity is measurable as OD410. Error bars represent standard deviation for three 756

technical replicates in a single microtiter plate. Trypsin incubated with BAPNA substrate 757

but no inhibitor was used as a control (NIC). 758

Figure 4. Distribution of BBIs in land plants. 759

The number of BBI-like sequences identified in each plant genome searched. Accession 760

numbers for all sequences are included in Supplemental Data Set 1. Predicted gene 761

loss events are indicated with an X. The phylogeny of the plant genomes investigated is 762

derived from species trees available from Cogepedia and APG III. 763

Figure 5. BBI-like sequences identified from quillwort and banana. 764

Alignment of translated sequences of genes cloned from genomic DNA from banana 765

(Musa acuminata, Mac) and cDNA from quillwort (Isoetes drummondii, Idr). Sequences 766

are aligned with a BBI encoded by a gene from soybean (GmaBBIb) to illustrate 767

conserved regions including the inhibitory loops (grey boxes). Conserved Cys residues 768

are indicated by asterisks. 769

Figure 6. Phylogeny of BBI-like sequences in plants and study of sequence diversity 770

through phylogenetic analysis of BBI ‘heads’. 771

Neighbour-joining tree from a protein distance matrix of 104 BBI-like sequences. 772

Sequences were identified by searches within plant genome assemblies. Major clades 773

are indicated and defined by branching patterns. The five S. moellendorffii sequences 774

and the I. drummondii sequence were used to root the tree with the assumption they 775

were evolutionarily most distant from all other sequences. Percent bootstrap values 776

from 1000 replicates of a distance neighbour joining analysis are indicated at each node 777

(only nodes with >50% bootstrap support are labelled). Identifiers and accession 778

numbers are listed in Supplemental Data Set 1. The scale bar represents number of 779

substitutions per site. 780

AICTRSNPPTCRCVDEVKKCAPTCKTCLPSRSRPSRRVCIDSYFGPVPPRCTPR124

G1RCTKSIPPICFPD

14

D1DESSKPCCDQCACTKSNPPQCRCSDMRLNSCHSACKSCICALSYPAQCFCVDITDFCYEPCKPSEDDKEN

71

A1GKKRPWKCCDEAVCTRSIPPICTCMDEVFE-CPKTCKSCGPSMGDPSRRICQDQYVGDPGPICRPWECCDK

SFTI-1

GmaBBIa

HvuBBIa

HvuBBIa

III

II

I

III

III

A B

C

cereals

legumes

Figure 1. The CTKSIPPIC motif: structures, sequences and contexts.

(A) Structural alignment of SFTI-1 (PDB code: 1JBL), a BBI from soybean (GmaBBIa; PDB code: 1BBI) and a BBI from barley

(HvuBBIa; PDB code: 2FJ8). The trypsin inhibitory loop of all three sequences (boxed) share structural similarity.

(B) BBIs are common in the phylogenetically separate cereals and legumes marked on an angiosperm phylogeny of rbcL

sequences (Elliott et al., 2014).

(C) Sequence alignment of SFTI-1, GmaBBI and HvuBBI. The homologous inhibitory ‘heads’ are indicated with black brackets.

Inhibitory loops are indicated with Roman numerals and the corresponding loops are also labelled in panel A. The soybean

sequence exemplifies the “double-headed” structure with two homologous inhibitory motifs, the second of which has been lost in

cereal BBIs as a result of the loss of two Cys residues (inhibitory loop II). Many monocot BBIs have undergone internal gene

duplications resulting in multiple inhibitory loops (e.g. inhibitory loop III).

Q

Q

Q

H

SmoBBI3 NSQCCPNSKFCPKGKFCTICTKSYPPSCFCTKAPPNCGVGSSCCSDGDCPKGKECSVCTKSFPPICRCG--TK

MscBBI TKSTTTACCDFCP-------CTRSIPPQCQCTDVREKC---HSAC--------KSC-LCTRSFPPQCRCYDITDFCYPSCS

A

B

GmaBBIbMELSMKVLVKVASLLFLLEFTATVVDARFDPSSFITQFLPNAEANNYYVKSTTKACCN---SC- CTKSIPPQC CSDIG

HvuBBIb

MRPQVLLVAFAVLAVVAALP--------IAHGQG------ASPWPCCD---KCG CTKSIPPQC CQDVS

SmoBBI1 MAQSIKFAPLIALLLCIAATAEVLDVPGEFFFVDGAARLPSECRPESLCCPNSKFCPK-GKYCT MPSNPPLC CTEAP

SmoBBI2 MAQSIKFAPLIALLLCIAVTAEVLDLPGEFFFVDGAASRPSGCRPNS CCPNSKFCPK-GKYCT CTKSYPPSC CTKAP

SmoBBI3 MAQSIKFAPLIALLLCIAVTAEVLDLPGEFFFVDGAASRPSACRPNS CCPNSKFCPK-GKFCT CTKSYPPSC CTKAP

SmoBBI4 MVCRPNS CCPGSKRCSKAGQFCG CTRSFPPTC CDAPN

SmoBBI5 MAQSIKFAPLIALLLCIAVTAEVLDLPGDFFFVDGAASRPSECRPNS CCPNSKFCPK-GKYCT CTRSNPPSC CTKAP

GmaBBIbET-CHSA-----------CKTC- CTRSIPPQC CSD-ITNFCYEPCNSSETEAH

HvuBBIbPTGCNSA-----------CKSC- -VRSTGGFQ CADSITNFCERRCTPAA

SmoBBI1 PN-CGVGSSCCSDGDCPKGKACS CTKSIPPIC CGPEPKIVLVSE

SmoBBI2 PN-CGVGSSCCSDGDCPKGRECS CTRSIPPIC CGAKRKLVLVSE

SmoBBI3 PN-CGVGSSCCSDGDCPKGKECS TKSFPPIC CGTKPKLVLVSE

SmoBBI4 PKRCSAASQCCSDADCLNGGSCSL TRSIPPIC CFKPGKAKLKL

SmoBBI5 PN-CGVGSSCCSDGDCPKGKECS CTKSFPPIC CGPKPKLVLVSE

I

-V

V

VC

C

V

H

-R

R

R

R

R

P

V

IC

V

I

R

V

R

R

V

F

F

F

F

* **

* * * **

* *

K67

R64

K105R90

64 90

67 105

Figure 2. S. moellendorffii BBI3 is predicted to be a BBI based sequence similarity at the inhibitory motifs and shared primary protein architecture.(A) BOXSHADE alignment of S. moellendorffii BBI-like (SmoBBI) protein sequences with soybean BBI (GmaBBIb) and barley BBI (HvuBBIb) shows conservation of the trypsin inhibitory loop (grey box). All proteins, with the exception of SmoBBI4, share a similar protein architecture with an ER signal (line above sequence) followed by Cys rich sequence with two spatially distinct conserved inhibitory motifs. Cys residues that are conserved in all sequences are indicated with asterisks.(B) Protein homology model of SmoBBI346-116 aligned with Medicago scutellata BBI (PDB code: 1MVZ) is shown along with the sequence alignment. The conserved inhibitory loops are shown as opaque while the remaining protein is translucent and the corresponding inhibitory motifs are boxed in the alignment. The P1 residues and disulfide bonds are represented as sticks.

0

0.2

0.4

0.6

0.8

1.0

1.2

WT K67A K77A K105A K77A

K105A

K67A

K105A

K67A

K77A

K105A

NIC

Try

psin

activity (

OD

410)

SmoBBI3 WT CTICTKSYPPSCFCTKAPPNC CTKSFPPICRCK67A CTICTASYPPSCFCTKAPPNC CTKSFPPICRCK77A CTICTKSYPPSCFCTAAPPNC CTKSFPPICRCK105A CTICTKSYPPSCFCTKAPPNC CTASFPPICRCK77A K105A CTICTKSYPPSCFCTAAPPNC CTASFPPICRCK67A K105A CTICTASYPPSCFCTKAPPNC CTASFPPICRCK67A K77A K105A CTICTASYPPSCFCTAAPPNC CTASFPPICRC

... ...

...

...

.........

.........

... ...

...

......

...

...

...

...

...

...

67 77 105

A

B

C

10 20 30 40

50 60 70 80 90

100 110 120

| | | |

|||||

| | |

MAQSIKFAPLIALLLCIAVTAEVLDLPGEFFFVDGAASRPSACRPNS

QCCPNSKFCPKGKFCTICTKSYPPSCFCTKAPPNCGVGSSCCSDGDC

PKGKECSVCTKSFPPICRCGTKPKLVLVSE

Figure 3. Trypsin inhibition by S. moellendorffii BBI3 and mutants

(A) Full length wildtype BBI3 showing predicted inhibitory residues (bold).

(B) Partial BBI sequences showing mutations (black highlight).

(C) Inhibitors were incubated with trypsin and a colorogenic substrate of trypsin whose activity is

measurable as OD410

. Error bars represent standard deviation for three technical replicates in a

single microtiter plate. Trypsin incubated with BAPNA substrate but no inhibitor was used as a

control (NIC).

5

5 basal angiospermgymnosperm

mon

ocot

eudi

cot

rosi

d II

aste

ridro

sid

IM

alpi

ghia

les

lycopod

bryphyte

species # Physcomitrella patens 0 Sphagnum fallax 0 Selaginella moellendorffii

Pinus taeda 0 Amborella trichopoda

Spirodela polyrhiza 0 Phalaenopsis equestris 1 Elaeis guineensis 8 Musa acuminata 10 Brachypodium distachyon 5 Brachypodium stacei 4 Oryza sativa 10 Panicum hallii 2 Setaria viridis 8 Zea mays 10 Sorghum bicolor 10 Ananas comosus 2 Aquilegia coerulea 3 Kalanchoe laxiflora 0 Vitis vinifera 0 Linum usitatissimum 0 Ricinus communis 0 Manihot esculenta 0 Populus trichocarpa 0 Salix purpurea 0 Glycine max 13 Phaseolus vulgaris 3 Medicago truncatula 9 Ziziphus jujuba 11 Fragaria vesca 0 Prunus persica 0 Cucumis sativus 0 Eucalyptus grandis 0 Citrus sinensis 0 Citrus clementina 0 Gossypium raimondii 0 Carica papaya 0 Arabidopsis lyrata 0 Arabidopsis thaliana 0 Boechera stricta 0

0 Capsella grandiflora 0 Capsella rubella

Brassica rapa

0

Eutrema salsugineum

0 Amaranthus hypochondriacus 0 Beta vulgaris 0 Primula vulgaris 0 Solanum lycopersicum 0 Solanum tuberosum 0 Mimulus guttatus 0 Carthamus tinctorius 0

X

X

XX

X

X

X

X

X

Figure 4. Distribution of BBIs in land plants. The number of BBI-like sequences identified in each plant genome searched. Accession numbers for all sequences are included in Supplemental Dataset 1. Predicted gene loss events are indicated with an X. The phylogeny of the plant genomes investigated is derived from species trees available from Cogepedia and APG III.

E

T

A

A

*

* * * * * * * *

* * * * *GmaBBIb MELSMKVLVKVASLLFLLEFTATVV--DARFDPSSFITQFLPN-------AEANNYYVKSTTKACCNSCP CTKSIPPQC CSD-IGE

IdrBBI MATKMGLLILMAVILLCYKSAAGLVCR-----------------PDSECCPKSKFCEAGKFCT CDRSAHPYC CSG-PSS

MacBBI1 MRSAGLVVAFLALV---FLVSLSAAREDPD----IFLPS-------QGIGEEVGGEKPWACCDSCS CTKSIPPQC CTDQLIG

MacBBI2 MRSGGLVVTFLALV---FLVSLSTARVDPHL--LLLLPS-------QGNGEGLAGEKPWACCDMCL CTRSFPPQC CTDELIG

MacBBI3 MRYNMVVFSLVLMVAAAFFASATTTASSSHPELRSALSTKGHEEDGEGVGERSRQRRTWPCCDRCG CTKS PPQC CQD-MVR

GmaBBIb TCHSACKT-----------C-- CTRSIPPQ CSDITN-FCYEPCNSSETE H

IdrBBI DCRAESQCCSDQDCPHGSSCS- CTKSIPPIC CTGVYLDADN

MacBBI1 GCDPNCKT-----------C-- CTRSYPPKC CYDIINDYCGERCNPEQ

MacBBI2 GCHPNCKN-----------C-- CTRSFPPK CRDIIYEDCGDRCHP

MacBBI3 SCHPSCRH-----------CVR PLSVSPPL CMDRIPNYCRRRCTP PLL Q

R

R

R

R

Q

-

A

-

-

G

I

I

I

H

S

CH

CY

YQ

R

R

Figure 5. BBI-like sequences identified from quillwort and banana.Alignment of translated sequences of genes cloned from genomic DNA of banana (Musa acuminata, Mac) and cDNA of quillwort (Isoetes drummondii, Idr). Sequences are aligned with a BBI encoded by a gene from soybean (GmaBBIb) to illustrate conserved regions including the inhibitory loops (grey boxes). Conserved Cys residues are indicated by asterisks.

0.4

Zea mays 2

Amborella trichopoda 1

Selaginella moellendorffii 2

Medicago truncatula 7


Brachypodium stacei 3


Brachypodium distachyon 3



Elaeis guineensis 6Elaeis guineensis 2

Musa acuminata 1

Ziziphus jujuba 10

Musa acuminata 5

Phaseolus vulgaris 1


Zea mays 6

Aquilegia coerulea 2

Glycine max10

Oryza sativa 2

Ziziphus jujuba 5

Sorghum bicolor 7

Musa acuminata 3



Oryza sativa 5

Glycine max12

Sorghum bicolor 12

Setaria viridis 1

Panicum hallii 1

Isoetes drummondii 1

Glycine max 5

Oryza sativa 8

Setaria viridis 3

Zea mays 10

Glycine max 8


Glycine max 7

Sorghum bicolor 4

Zea mays 1

Setaria viridis 9

Oryza sativa 3



Sorghum bicolor 2Sorghum bicolor 1

Elaeis guineensis 3

Musa acuminata 4

Ziziphus jujuba 1

Zea mays 3


Panicum hallii 2





Ananas comosus 1


Oryza sativa 7



Elaeis guineensis 8

Aquilegia coerulea 1

Ananas comosus 2

Musa acuminata 2


Glycine max 1


Zea mays 5

Glycine max 6

Elaeis guineensis 7

Setaria viridis 8

Ziziphus jujuba 11

Phalaenopsis equestris 1

Ziziphus jujuba 8

Setaria viridis 2

Oryza sativa 1

Sorghum bicolor 10

Sorghum bicolor 3

Oryza sativa 4

Setaria viridis 7


Ziziphus jujuba 6

Elaeis guineensis 4

Sorghum bicolor 8


Ziziphus jujuba 7

Zea mays 9

Elaeis guineensis 5

Oryza sativa 10

Zea mays 4


Setaria viridis 6

Glycine max13

Oryza sativa 6

Elaeis guineensis 1

Setaria viridis 4

Zea mays 8

Ziziphus jujuba 3


Ziziphus jujuba 2

Ziziphus jujuba 9


10051.4

98.0

58.5

97.0

99.599.0

77.7

50.0

10066.1

56.7

99.084.1

61.1

99.873.5

70.7

66.269.7

10096.5

10062.8

67.4

85.0

70.4

68.560.8

56.4

94.3

94.8

85.498.7

99.183.9

99.9

100

99.896.5

66.883.3

63.3

58.794.7

63.2

62.4

79.378.8

88.5

88.6

100 92.897.0

53.081.973.2

10064.8

92.1

60.6

97.7

72.888.4

99.7

88.994.3

56.2

67.3

98.562.3

monocot clade

dicot clade

Poaceae clade II

Poaceae clade I

Amborella trichopoda clade

lycopod clade

Figure 6. Phylogeny of BBI-like sequences in plants. Neighbour-joining tree from a protein distance matrix of 104 BBI-like sequences. Sequences were identified by searches within plant genome assemblies. Major clades are indicated and defined by branching patterns. The five S. moellendorffii sequences and the I. drummondii sequence were used to root the tree with the assumption they were evolutionarily most distant from all other sequences. Percent bootstrap values from 1000 replicates of a distance neighbour joining analysis are indicated at each node (only nodes with >50% bootstrap support are labelled). Identifiers and accession numbers are listed in Supplemental Data Set 1. The scale bar represents number of substitutions per site.

Parsed Citations

Albalat, R., and Canestro, C. (2016). Evolution by gene loss. Nat. Rev. Genet. 17, 379-391.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Angiosperm Phylogeny Group. (2016). An update of the Angiosperm Phylogeny Group classification for the orders and families offlowering plants: APG IV. Bot. J. Linn. Soc. 181, 1-20.

Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Arrigo, N., Therrien, J., Anderson, C.L., Windham, M.D., Haufler, C.H., and Barker, M.S. (2013). A total evidence approach tounderstanding phylogenetic relationships and ecological diversity in Selaginella subg. Tetragonostachys. Am. J. Bot. 100, 1672-1682.


Banks, J.A. (2009). Selaginella and 400 million years of separation. Annu. Rev. Plant Biol. 60, 223-238.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Banks, J.A., Nishiyama, T., Hasebe, M., Bowman, J.L., Gribskov, M., dePamphilis, C., Albert, V.A., Aono, N., Aoyama, T., Ambrose,B.A., Ashton, N.W., Axtell, M.J., Barker, E., Barker, M.S., Bennetzen, J.L., Bonawitz, N.D., Chapple, C., Cheng, C., Correa, L.G.G.,Dacre, M., DeBarry, J., Dreyer, I., Elias, M., Engstrom, E.M., Estelle, M., Feng, L., Finet, C., Floyd, S.K., Frommer, W.B., Fujita, T.,Gramzow, L., Gutensohn, M., Harholt, J., Hattori, M., Heyl, A., Hirai, T., Hiwatashi, Y., Ishikawa, M., Iwata, M., Karol, K.G., Koehler,B., Kolukisaoglu, U., Kubo, M., Kurata, T., Lalonde, S., Li, K., Li, Y., Litt, A., Lyons, E., Manning, G., Maruyama, T., Michael, T.P.,Mikami, K., Miyazaki, S., Morinaga, S.i., Murata, T., Mueller-Roeber, B., Nelson, D.R., Obara, M., Oguri, Y., Olmstead, R.G., Onodera,N., Petersen, B.L., Pils, B., Prigge, M., Rensing, S.A., Riaño-Pachón, D.M., Roberts, A.W., Sato, Y., Scheller, H.V., Schulz, B., Schulz,C., Shakirov, E.V., Shibagaki, N., Shinohara, N., Shippen, D.E., Sørensen, I., Sotooka, R., Sugimoto, N., Sugita, M., Sumikawa, N.,Tanurdzic, M., Theißen, G., Ulvskov, P., Wakazuki, S., Weng, J.K., Willats, W.W.G.T., Wipf, D., Wolf, P.G., Yang, L., Zimmer, A.D., Zhu,Q., Mitros, T., Hellsten, U., Loqué, D., Otillar, R., Salamov, A., Schmutz, J., Shapiro, H., Lindquist, E., Lucas, S., Rokhsar, D., andGrigoriev, I.V. (2011). The Selaginella genome identifies genetic changes associated with the evolution of vascular plants. Science332, 960-963.


Bell, C.D., Soltis, D.E., and Soltis, P.S. (2010). The age and diversification of the angiosperms re-revisited. Am. J. Bot. 97, 1296-1303.


Birk, Y. (1961). Purification and some properties of a highly active inhibitor of trypsin and a-chymotrypsin from soybeans. Biochim.Biophys. Acta 54, 378-381.


Birk, Y., Gertler, A., and Khalef, S. (1963). A pure trypsin inhibitor from soya beans. Biochem. J. 87, 281.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Bowman, D.E. (1946). Differentiation of soy bean antitryptic factors. Proc. Soc. Exp. Biol. Med. 63, 547-550.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Brauer, A.B.E., and Leatherbarrow, R.J. (2003). The conserved P1' Ser of Bowman-Birk-type proteinase inhibitors is not essentialfor the integrity of the reactive site loop. Biochem. Biophys. Res. Commun. 308, 300-305.


Brauer, A.B.E., Domingo, G.J., Cooke, R.M., Matthews, S.J., and Leatherbarrow, R.J. (2002). A conserved cis peptide bond isnecessary for the activity of Bowman-Birk inhibitor protein. Biochemistry 41, 10608-10615.


Clemente, A., Arques, M.C., Dalmais, M., Le Signor, C., Chinoy, C., Olias, R., Rayner, T., Isaac, P.G., Lawson, D.M., Bendahmane, A.,and Domoney, C. (2015). Eliminating anti-nutritional plant food proteins: the case of seed protease inhibitors in pea. PLoS One 10,e0134634.

Pubmed: Author and Title

CrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

D'Hont, A., Denoeud, F., Aury, J.M., Baurens, F.C., Carreel, F., Garsmeur, O., Noel, B., Bocs, S., Droc, G., Rouard, M., Da Silva, C.,Jabbari, K., Cardi, C., Poulain, J., Souquet, M., Labadie, K., Jourda, C., Lengelle, J., Rodier-Goud, M., Alberti, A., Bernard, M.,Correa, M., Ayyampalayam, S., McKain, M.R., Leebens-Mack, J., Burgess, D., Freeling, M., Mbeguie, A.M.D., Chabannes, M., Wicker,T., Panaud, O., Barbosa, J., Hribova, E., Heslop-Harrison, P., Habas, R., Rivallan, R., Francois, P., Poiron, C., Kilian, A., Burthia, D.,Jenny, C., Bakry, F., Brown, S., Guignon, V., Kema, G., Dita, M., Waalwijk, C., Joseph, S., Dievart, A., Jaillon, O., Leclercq, J., Argout,X., Lyons, E., Almeida, A., Jeridi, M., Dolezel, J., Roux, N., Risterucci, A.M., Weissenbach, J., Ruiz, M., Glaszmann, J.C., Quetier, F.,Yahiaoui, N., and Wincker, P. (2012). The banana (Musa acuminata) genome and the evolution of monocotyledonous plants. Nature488, 213-217.


Domingo, G.J., Leatherbarrow, R.J., Freeman, N., Patel, S., and Weir, M. (1995). Synthesis of a mixture of cyclic peptides based onthe Bowman-Birk reactive site loop to screen for serine protease inhibitors. Int. J. Pept. Protein Res. 46, 79-87.


Elliott, A.G., Delay, C., Liu, H., Phua, Z., Rosengren, K.J., Benfield, A.H., Panero, J.L., Colgrave, M.L., Jayasena, A.S., Dunse, K.M.,Anderson, M.A., Schilling, E.E., Ortiz-Barrientos, D., Craik, D.J., and Mylne, J.S. (2014). Evolutionary origins of a bioactive peptideburied within preproalbumin. Plant Cell 26, 981-995.


Engelken, J., Brinkmann, H., and Adamska, I. (2010). Taxonomic distribution and origins of the extended LHC (light-harvestingcomplex) antenna protein superfamily. BMC Evol. Biol. 10, 233.


Gariani, T., McBride, J.D., and Leatherbarrow, R.J. (1999). The role of the P2' position of Bowman-Birk proteinase inhibitor in theinhibition of trypsin. Studies on P2' variation in cyclic peptides encompassing the reactive site loop. Biochim. Biophys. Acta 1431,232-237.


Habib, H., and Fazili, K.M. (2007). Plant protease inhibitors: a defense strategy in plants. Biotechnol. Mol. Biol. Rev. 2, 68-85.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Hall, T.A. (1999). BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. NucleicAcids Symp. S. 41, 95-98.


Hazkani-Covo, E., Altman, N., Horowitz, M., and Graur, D. (2002). The Evolutionary History of Prosaposin: Two Successive Tandem-Duplication Events Gave Rise to the Four Saposin Domains in Vertebrates. J. Mol. Evol. 54, 30-34.


Hirsh, A.E., and Fraser, H.B. (2001). Protein dispensability and rate of evolution. Nature 411, 1046-1049.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Jayasena, A.S., Secco, D., Bernath-Levin, K., Berkowitz, O., Whelan, J., and Mylne, J.S. (2014). Next generation sequencing and denovo transcriptomics to study gene evolution. Plant Methods 10, 34.


Krylov, D.M., Wolf, Y.I., Rogozin, I.B., and Koonin, E.V. (2003). Gene loss, protein sequence divergence, gene dispensability,expression level, and interactivity are correlated in eukaryotic evolution. Genome Res. 13, 2229-2235.


Li, J., Zhang, C., Xu, X., Wang, J., Yu, H., Lai, R., and Gong, W. (2007). Trypsin inhibitory loop is an excellent lead structure to designserine protease inhibitors and antimicrobial peptides. FASEB J. 21, 2466-2473.


Luckett, S., Garcia, R.S., Barker, J.J., Konarev, A.V., Shewry, P.R., Clarke, A.R., and Brady, R.L. (1999). High-resolution structure ofa potent, cyclic proteinase inhibitor from sunflower seeds. J. Mol. Biol. 290, 525-533.


McBride, J.D., Brauer, A.B.E., Nievo, M., and Leatherbarrow, R.J. (1998). The role of threonine in the P2 position of Bowman-Birkproteinase inhibitors: studies on P2 variation in cyclic peptides encompassing the reactive site loop. J. Mol. Biol. 282, 447-457.


McBride, J.D., Watson, E.M., Brauer, A.B., Jaulent, A.M., and Leatherbarrow, R.J. (2002). Peptide mimics of the Bowman-Birkinhibitor reactive site loop. Biopolymers 66, 79-92.


Mello, M.O., Tanaka, A.S., and Silva-Filho, M.C. (2003). Molecular evolution of Bowman-Birk type proteinase inhibitors in floweringplants. Mol. Phylogenet. Evol. 27, 103-112.


Mylne, J.S., Colgrave, M.L., Daly, N.L., Chanson, A.H., Elliott, A.G., McCallum, E.J., Jones, A., and Craik, D.J. (2011). Albumins andtheir processing machinery are hijacked for cyclic peptides in sunflower. Nat. Chem. Biol. 7, 257-259.


Mylne, J.S., Chan, L.Y., Chanson, A.H., Daly, N.L., Schaefer, H., Bailey, T.L., Nguyencong, P., Cascales, L., and Craik, D.J. (2012).Cyclic peptides arising by evolutionary parallelism via asparaginyl-endopeptidase-mediated biosynthesis. Plant Cell 24, 2765-2778.


Nishino, N., Aoyagi, H., Kato, T., and Izumiya, N. (1977). Studies on the synthesis of proteinase inhibitors. I. Synthesis and activity ofnonapeptide fragments of soybean Bowman-Birk inhibitor. J. Biochem. 82, 901-909.


Odani, S., and Ikenaka, T. (1973). Studies on soybean trypsin inhibitors VIII. Disulfide bridges in soybean Bowman-Birk proteinaseinhibitor. J. Biochem. 74, 697-715.


Park, E.Y., Kim, J.A., Kim, H.W., Kim, Y.S., and Song, H.K. (2004). Crystal structure of the Bowman-Birk inhibitor from barley seedsin ternary complex with porcine trypsin. J. Mol. Biol. 343, 173-186.


Petersen, T.N., Brunak, S., von Heijne, G., and Nielsen, H. (2011). SignalP 4.0: discriminating signal peptides from transmembraneregions. Nat. Meth. 8, 785-786.


Prakash, B., Selvaraj, S., Murthy, M.R., Sreerama, Y.N., Rao, D.R., and Gowda, L.R. (1996). Analysis of the amino acid sequences ofplant Bowman-Birk inhibitors. J. Mol. Evol. 42, 560-569.


Prasad, E.R., Merzendorfer, H., Madhurarekha, C., Dutta-Gupta, A., and Padmasree, K. (2010). Bowman-Birk Proteinase inhibitorfrom Cajanus cajan seeds: purification, characterization, and insecticidal properties. J. Agric. Food Chem. 58, 2838-2847.


Qi, R.F., Song, Z.W., and Chi, C.W. (2005). Structural features and molecular evolution of Bowman-Birk protease inhibitors and theirpotential application. Acta Biochim. Biophys. Sin. 37, 283-292.


Rawlings, N.D., Barrett, A.J., and Finn, R. (2016). Twenty years of the MEROPS database of proteolytic enzymes, their substratesand inhibitors. Nucleic Acids Res. 44, D343-D350.


Roberts, T.H., and Hejgaard, J. (2008). Serpins in plants and green algae. Funct. Integr. Genomics 8, 1-27.Pubmed: Author and TitleCrossRef: Author and TitleGoogle Scholar: Author Only Title Only Author and Title

Schechter, I., and Berger, A. (1967). On the size of the active site in proteases. I. Papain. Biochem. Biophys. Res. Commun. 27, 157-162.


Soltis, D.E., Albert, V.A., Leebens-Mack, J., Palmer, J.D., Wing, R.A., dePamphilis, C.W., Ma, H., Carlson, J.E., Altman, N., Kim, S.,Wall, P.K., Zuccolo, A., and Soltis, P.S. (2008). The Amborella genome: an evolutionary reference for plant biology. Genome Biol 9,402.


Song, H.K., Kim, Y.S., Yang, J.K., Moon, J., Lee, J.Y., and Suh, S.W. (1999). Crystal structure of a 16 kDa double-headed Bowman-Birk trypsin inhibitor from barley seeds at 1.9 Å resolution. J. Mol. Biol. 293, 1133-1144.


Tan, G., Muffato, M., Ledergerber, C., Herrero, J., Goldman, N., Gil, M., and Dessimoz, C. (2015). Current Methods for AutomatedFiltering of Multiple Sequence Alignments Frequently Worsen Single-Gene Phylogenetic Inference. Syst. Biol. 64, 778-791.


Terada, S., Sato, K., Kato, T., and Izumiya, N. (1978). Inhibitory properties of nonapeptide loop structures related to reactive sitesof soybean Bowman-Birk inhibitor. FEBS Lett. 90, 89-92.


Tran, L.T., Taylor, J.S., and Constabel, C.P. (2012). The polyphenol oxidase gene family in land plants: Lineage-specific duplicationand expansion. BMC Genomics 13, 395.


Zheng, L., Baumann, U., and Reymond, J.L. (2004). An efficient one-step site-directed and site-saturation mutagenesis protocol.Nucl. Acids Res. 32, e115.


DOI 10.1105/tpc.16.00831; originally published online March 14, 2017;Plant Cell

James Whelan, Charles Bond and Joshua S MylneAmy Midori James, Achala S Jayasena, Jingjing Zhang, Oliver Berkowitz, David Secco, Gavin J Knott,

Evidence for Ancient Origins of Bowman-Birk Inhibitors from Selaginella moellendorffii

This information is current as of April 14, 2021

Supplemental Data /content/suppl/2017/04/17/tpc.16.00831.DC2.html /content/suppl/2017/03/14/tpc.16.00831.DC1.html

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists

https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

http://www.plantcell.org/cgi/alerts/ctmain

http://www.plantcell.org/cgi/alerts/ctmain

http://www.aspb.org/publications/subscriptions.cfm

New Evidence for Ancient Origins of Bowman-Birk Inhibitors from … · 2017. 3. 14. · 1 RESEARCH...

Documents

Transcript of New Evidence for Ancient Origins of Bowman-Birk Inhibitors from … · 2017. 3. 14. · 1 RESEARCH...