Skip to main content
NCBI home page
3 Biotech logoLink to 3 Biotech
. 2019 Nov 9;9(12):442. doi: 10.1007/s13205-019-1983-2

Progress toward sourcing plants for new bioconjugation tools: a screening evaluation of a model peptide ligase using a synthetic precursor

Tunjung Mahatmanto 1,2,5,✉, Isyatul Azizah 2,3, Alex Buchberger 2,4, Nicholas Stephanopoulos 2,4
PMCID: PMC6842375  PMID: 31763120

Abstract

In the present study, leaves from 39 phylogenetically distant plant species were sampled and screened for asparaginyl endopeptidase ligase activity using mass spectrometry to test the generality of peptide ligases in plants. A modified version of the sunflower trypsin inhibitor-1 precursor was used as the substrate for reactions with leaf crude extracts and protein fractions. Masses consistent with products of asparaginyl endopeptidase activities that cleave and ligate the substrate into cyclic peptide following the reactions were detected in 8 plants: Nerium oleander and Thevetia peruviana of the family Apocynaceae; Bauhinia variegata, Dermatophyllum secundiflorum, Pithecellobium flexicaule, and Prosopis chilensis of the family Fabaceae; Morus alba of the family Moraceae; and Citrus aurantium of the family Rutaceae. This screening result represents a 20% hit rate for finding asparaginyl endopeptidase ligase activity from the arbitrary plants sampled. Analysis following a 2-h reaction of the substrate with the crude extract of D. secundiflorum leaves showed that the yield of cyclic peptide remained stable around 0.5 ± 0.1% of the substrate over the course of the reaction.

Keywords: Plant peptide ligase, Asparaginyl endopeptidase, Cyclic peptide, Protein modification, Bioconjugation tool

Introduction

Post-translational modifications of proteins give rise to diverse structural and thus functional features of proteins (Walsh et al. 2005). The diversity of protein functions that arise from these modifications has inspired scientists to develop site-specific conjugation methods for a range of applications. These include: protein labeling for live-cell imaging (Lotze et al. 2016); construction of DNA–protein conjugates for enzyme cascade reactors and biosensors (Trads et al. 2017); and the creation of new synthetic materials with advanced properties for light harvesting, molecular cargo, and water remediation (Witus and Francis 2011). The conjugation methods either target surface-exposed naturally occurring amino acid side chains on the protein target or unnatural amino acid side chains incorporated through genetic encoding (Krall et al. 2016). Both strategies come with advantages and limitations, but those mediated by enzymes allow the bioconjugation reaction to be accomplished under mild conditions: in aqueous solution, ambient temperatures, and near-neutral pH. Importantly, enzymatic processes can readily be scalable, thus providing a means to address the potential significant cost associated with large-scale protein modifications (Witus and Francis 2011).

The ligation of peptide bonds by proteases represents one type of conjugation that has long been recognized and assumed to be governed by the physicochemical conditions of the reaction, such as pH, temperature, concentration, and the nature of the substrate (Bergmann and Fraenkel-Conrat 1937). Asparaginyl endopeptidase (AEP) is a cysteine protease with a narrow substrate specificity that cleaves after asparagine (Asn) and, less preferably, aspartate (Asp) residues (Dall and Brandstetter 2016). In addition to being a proteolytic enzyme that cleaves Asx at the P1 position, AEP can also ligate peptide bonds, hence the name “peptide ligase”. For example, butelase 1 or CtAEP1, an AEP isoform found in Clitoria ternatea (Fabaceae), can perform ligation with any amino acid except proline at the P1′′ position and preferably leucine, valine, isoleucine, or cysteine at the P2′′ position as illustrated in Fig. 1a (Nguyen et al. 2014). The ability of AEPs to perform ligation has presented the opportunity to use them as a bioconjugation tool (Harris et al. 2015; Nguyen et al. 2016; Yang et al. 2017). Other proteases that display ligase activity include sortase A and V8 protease from Staphylococcus aureus (Mazmanian et al. 1999; Kumaran et al. 2000), patG from cyanobacteria (Lee et al. 2009), peptide cyclase 1 (PCY1) from Saponaria vaccaria of the family Caryophyllaceae (Barber et al. 2013), and prolyl oligopeptidase B (POPB) from Galerina marginata (Luo et al. 2014). In most of the aforementioned examples, the peptide ligases were discovered from efforts to identify the enzyme responsible for ligating cyclic peptides.

Fig. 1.

Fig. 1

Open in a new tab

Schematics of substrate processing by AEPs. a Cleavage and ligation of a peptide substrate by CtAEP1 to produce a cyclic peptide. The substrate sequence requirements at the P1, P1′′, and P2′′ positions for the ligation reaction are indicated. b The screening approach used in this study and its major findings. The crude extract and protein fraction of 39 plants were screened for AEP ligase activity using a synthetic substrate, SFTI-NGL. The cleavage and ligation are expected to produce linear and cyclic SFTI-N from the substrate. The respective peptide sequences and monoisotopic masses are indicated

One rich source of cyclic peptides is plants. In angiosperms, cyclic peptides are found in 15 families, from Asteraceae to Schizandraceae (Mahatmanto 2015). Among the proteases found to process cyclic peptides in angiosperms, AEPs are the best studied. Interestingly, not all plants are found to contain AEP-processed cyclic peptides, presumably due to the absence of either “specialized” AEPs that have evolved to become ligases (Jackson et al. 2018) or “processable” cyclic peptide precursors. For the latter, it is assumed that the determining factor for ligation is not merely the presence of the AEP ligases, but also the processability of the peptide precursors as defined by their sequences. This assumption is supported by the naturally occurring cyclic peptide precursors and their linear counterparts in the Cucurbitaceae and Fabaceae (Mahatmanto et al. 2015; Gilding et al. 2016). In addition, in vitro, in situ, and in vivo studies show that specific precursor sequences are required for efficient ligation to occur (Mylne et al. 2011; Nguyen et al. 2014; Bernath-Levin et al. 2015; Haywood et al. 2018).

To test the generality of AEP as a model peptide ligase, here we used mass spectrometry to screen 39 phylogenetically distant plant species for ligase activity using a modified precursor of the sunflower trypsin inhibitor-1 (SFTI-1) from Helianthus annuus (Asteraceae). We observed the appearance of masses in the spectra consistent with products of AEP activities in eight plants (a summary of the work is given in Fig. 1b). Given the vast diversity and general availability of plants, our findings support the idea that plants can be sourced to expand the repertoire of peptide ligases.

Materials and methods

Plant materials

Leaves from 39 plants with species identifiers (Table 1) were collected from around the Arizona State University, Tempe campus between May and October, 2017. We focused our study on leaves because they are highly abundant, unlike flowers or fruits that are only temporally available in angiosperms, and easier to be handled compared to seeds (Mahatmanto et al. 2014).

Table 1.

List of plant species (by family order) screened for AEP ligase activity

Family and species SFTI-NGL remaining Linear SFTI-N formed Cyclic SFTI-N formed
Crude extract Protein fraction Crude extract Protein fraction Crude extract Protein fraction
Acanthaceae
 Justicia spicigera − − − − − −
Anacardiaceae
 Pistacia chinensis + + − − − −
 Rhus lancea + + − − − −
Apocynaceae
 Asclepias subulata − + − − − −
 Carissa grandiflora + + − − − −
 Nerium oleander + + + + + +
 Thevetia peruviana + − + − + −
Asparagaceae
 Nolina texana − − − − − −
Asphodelaceae
 Aloe vera − + − + − −
Bignoniaceae
 Jacaranda mimosifolia + − − − − −
Cactaceae
 Stenocereus thurberi − − − − − −
Cupressaceae
 Cupressus arizonica − − − − − −
Fabaceae
 Acacia stenophylla + + − − − −
 Bauhinia blakeana − − − − − −
 Bauhinia variegata + + + + + −
 Caesalpinia pulcherrima + + − − − −
 Dermatophyllum secundiflorum + − + − + −
 Lysiloma thornberi + + − − − −
 Olneya tesota + − + + − −
 Parkinsonia praecox − + − + − −
 Pithecellobium flexicaule + + + + − +
 Prosopis chilensis + + + + − +
 Prosopis velutina − − − − − −
Fagaceae
 Quercus suber + + − − − −
 Quercus virginiana − − − − − −
Juglandaceae
 Carya illinoinensis + + − − − −
Lamiaceae
 Vitex agnus-castus − + − + − −
Malvaceae
 Hibiscus rosa-sinensis + + − − − −
Moraceae
 Ficus benjamina + + − − − −
 Morus alba + + + + + −
Myrtaceae
 Callistemon viminalis + + − − − −
 Myrtus communis + + − − − −
Oleaceae
 Fraxinus velutina − − − − − −
 Jasminum sambac − + − − − −
 Olea europaea − + − − − −
Pinaceae
 Pinus canariensis + − − − − −
 Pinus halepensis + + − − − −
Proteaceae
 Grevillea robusta − − − − − −
Rutaceae
 Citrus aurantium + + − + − +
Open in a new tab

Positive hits were determined by discernible peak masses above 200 cps consistent with cyclic and linear SFTI-N following incubation of crude extracts or protein fractions with SFTI-NGL (n = 1)

AEP extraction

The extraction of AEP from plant materials was performed following a method developed by Nguyen et al. (2016) with modifications. Leaves were finely ground under liquid nitrogen using a mortar and pestle to preserve their protein content. Freshly prepared extraction buffer (20-mM sodium phosphate, 1-mM ethylenediaminetetraacetic acid (EDTA), 1-mM phenylmethane sulfonyl fluoride (PMSF), 5-mM 2-mercaptoethanol, pH 6.0) was added to the ground leaves to a final concentration of 100 mg/mL and mixed by rotating for 15 min to extract the proteins from the leaf granules. The mixture was then centrifuged at 3000×g for 15 min to pellet the leaf granules. The supernatant was decanted and passed through a 0.45-µm syringe filter to remove debris. We refer to this preparation as the crude extract. Ammonium sulfate was added to 1 mL of crude extract to reach 70% saturation (0.45 g) and mixed by rotating for 15 min to salt out the proteins. The mixture was then centrifuged at 16,000×g for 30 min to precipitate the proteins. The supernatant was discarded, and the protein precipitate was dissolved in 1 mL of ligation buffer (20-mM sodium phosphate, 1-mM EDTA, pH 6.0). We refer to this preparation as the protein fraction. We note that all steps following leaf grinding were performed at 4 °C to minimize protein degradation.

Peptide synthesis

SFTI-NGL (with the sequence GRCTKSIPPICFPNGL) was synthesized using a standard 9-fluorenylmethyl carbamate (Fmoc) solid-phase peptide synthesis on pre-loaded Leu Wang resin with carboxylic acid (0.1 mmol eq. per gram) using a Liberty Blue microwave-assisted peptide synthesizer. Following synthesis, the peptide was cleaved with trifluoroacetic acid (TFA) containing 2.5% water, 2.5% triisopropylsilane (TIPS), and 5% 3,6-dioxa-1,8-octanedithiol (DODT). The cleaved peptide was purified using a reverse-phase high-performance liquid chromatography (HPLC) with a Waters 2489 instrument and a Phenomenex Aeris 5 µm PEPTIDE XB-C18 100 column (250 × 10.0 mm), using a linear gradient (0–80% over 60 min) of water to acetonitrile (ACN) containing 0.1% TFA. Isolated fractions were characterized by matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI TOF MS; Applied Biosystems 4800 MALDI TOF/TOF Analyzer), and those containing the correct peptide mass were lyophilized and stored at − 20 °C.

Peptide ligase screening

AEP activities were tested by adding 10 µL of linear SFTI-NGL (1 mg/mL in ligation buffer) to 10 µL of crude extract or protein fraction. The reaction was held at 37 °C for 1 h to allow for peptide cleavage and ligation followed by heating at 100 °C for 5 min to inactivate AEP. Samples were analyzed using MALDI TOF MS. MALDI spectra were accumulated from a minimum of two acquisitions per sample using a linear positive mode. For screening, peak heights were normalized to SFTI-NGL. A peak height of 200 counts per second (cps) was applied as the threshold for positive hits. Here, positive hits are defined as discernible peak masses consistent with cyclic and linear SFTI-N, the products of SFTI-NGL processing by AEPs, detectable above the 200-cps threshold.

Determination of ligation product

To determine the presence of cyclic SFTI-N, 50 µL of crude extract (D. secundiflorum) was added to 50 µL of SFTI-NGL (10 mg/mL in ligation buffer) and incubated at 37 °C for 0, 5, 30, 60, and 120 min. The reaction was stopped by heating at 100 °C for 5 min. Samples were then loaded into spin columns with a 3-kDa molecular weight cut-off (MWCO), diluted with 1 mL of solvent A (0.1% formic acid in water), and centrifuged at 4000 rpm for 20 min to separate the peptides from large biomolecules in the crude extract. This step was repeated five times. The collected flow-through, which contains the peptides, was desalted using a C18 solid-phase extraction (SPE) cartridge and eluted step-wise with 2 mL of 20%, 40%, 60%, and 80% solvent B (0.1% formic acid in ACN). This step-wise elution ensures that the peptides are being eluted from the C18 stationary phase. The fractions were pooled, lyophilized overnight, and resuspended with 100 µL of solvent A. This experiment was repeated three times. All samples were then analyzed using liquid chromatography–mass spectrometry (LC–MS; Agilent 1290 Infinity HPLC coupled to an Agilent 6530 Q-TOF mass spectrometer). To determine the ligation yields at the different time points above, the LC–MS extraction algorithm (Agilent MassHunter Qualitative Analysis software) was set to the following parameters: restricted m/z to 504–506 for cyclic SFTI-N, 510–512 for linear SFTI-N, and 567–569 for SFTI-NGL; used peaks with height ≥ 100 cps; applied a peptide isotope model; limited the assigned charge states to a maximum of 5; and filtered the hits with a compound quality score ≥ 70. For yield determination, peak heights for cyclic and linear SFTI-N were normalized to SFTI-NGL.

Results and discussion

Using enzymes to modify proteins at specific genetically encoded sequences affords unparalleled opportunities for selective conjugation, provided that sequence is sterically accessible. This route can often circumvent the limitations of chemical modification at mutagenetically introduced residues, which often relies on the incorporation of unnatural amino acids for driving reaction selectivity (Spicer and Davis 2014), or rare residues like cysteine that can form undesired disulfides. Enzymes capable of forming peptide bonds, or peptide ligases, have been reported since 1886 by Danilewski (Bergmann and Fraenkel-Conrat 1937), but the sortase family of enzymes is probably the most studied and extensively used for protein modification (Antos et al. 2016). Recently, plant AEPs have been receiving increased attention for this purpose, following the discovery of butelase 1 (Nguyen et al. 2014) and its application in intra- and intermolecular ligations of proteins of various sizes (Nguyen et al. 2016). The discovery of butelase 1 was subsequently followed by the successful recombinant expression and structure-based engineering of OaAEP1, an AEP isoform found in Oldenlandia affinis (Rubiaceae), into an efficient ligase (Harris et al. 2015; Yang et al. 2017).

The limited number of currently available peptide ligases for bioconjugation prompted us to test the generality of ligase activity in plants using AEP as a model. To identify AEP with ligase activity, we used the SFTI-1 peptide precursor because it has a short sequence (Mylne et al. 2011) and represents a substrate from a distal plant family branch (in the order Asterales) known to contain cyclic peptides (Mahatmanto 2015). In addition, it is unlikely that the peptide will be present in other plant species as suggested by its confinement in H. annuus (Elliott et al. 2014). To accommodate the preference of AEP for Asn at the P1 position (Dall and Brandstetter 2016), we changed the original sequence of the proto-C-terminal Asp to an Asn but preserved the neighboring Gly and Leu at the P1′ and P2′ positions, respectively (Mylne et al. 2011), as both define the minimum requirements for cyclization (Harris et al. 2015; Poon et al. 2018) and are conserved in cyclotides (Wang et al. 2008), a unique class of plant cyclic peptides (Craik et al. 1999). We refer to the substrate as SFTI-NGL. This peptide has a monoisotopic mass of 1699.86 Da (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

Representative MALDI spectra acquired from plant ligase screening. Shown here are the positive hit spectra (n = 1) from three plants. For clarity, the relative peak heights of the crude extracts (left panel, a) and protein fractions (right panel, b) were normalized to the highest peak of the respective plant, not to the initial SFTI-NGL (c). Masses consistent with cyclic SFTI-N, linear SFTI-N, and SFTI-NGL are labeled C, L, and S, respectively. A sodium adduct of the SFTI-NGL is indicated with an asterisk

We chose plants for our search because they are abundant, with over 350,000 species to screen from (The Plant List 2010), they are readily accessible, and they represent a rich source of AEPs, with a single plant containing multiple isoforms (Dall and Brandstetter 2016). We collected leaves from 39 species across 19 plant families (Table 1) to capture the diversity of plants being sampled. Then, we checked the background signals of the crude extracts and protein fractions from the 39 plants to set the peak height threshold for linear SFTI-N (monoisotopic mass 1529.75 Da) and cyclic SFTI-N (monoisotopic mass 1511.75 Da), the two products of SFTI-NGL processing by AEPs. Based on the background signal, we set a peak height threshold of ≥ 200 cps for positive hits. We then performed ligation reaction by adding SFTI-NGL to the crude extracts and protein fractions. Following the reaction, we analyzed the presence of masses for cyclic and linear SFTI-N as well as the parent SFTI-NGL substrate. From this analysis, we found eight positive hits suggestive of ligase activity that can be attributed to AEP. A summary of the cyclic and linear SFTI-N hits as well as the remaining SFTI-NGL present in the crude extracts and protein fractions is given in Table 1.

The screening result shows the appearance of masses consistent with linear and cyclic SFTI-N produced from AEPs cleaving and ligating the SFTI-NGL substrate (Fig. 2). Interestingly, in some cases, the product masses were detectable only in certain crude extracts, while others only in certain protein fractions (Table 1). It is possible that naturally occurring factors present in the crude extracts are required for (in the former case) or inhibited (in the latter case) the AEP activities, and that these factors were washed away during protein precipitation and thus were absent in the protein fractions. The ligase activity of AEPs is known to be mediated either by an endogenous metastable succinimide (Dall et al. 2015) or through cleavage-coupling transpeptidation (Bernath-Levin et al. 2015). It would be interesting to determine what the factors are in future studies. Another factor that could contribute to the absence of AEP activities in the crude extracts but was present in the protein fractions is the ammonium sulfate treatment. Protein fractionation using ammonium sulfate exposes proteins to high concentration of salts (in our case ~ 3.4 M ammonium sulfate). This high ionic strength exposure could help to induce the activation of AEP from its inactive zymogen form through electrostatic removal of the legumain stabilization and activity modulation (LSAM) domain that would generally be attained at low pH (Dall and Brandstetter 2016). If this was the case, it would be restricted to AEPs that do not get activated at pH 6 of the reaction buffer used in this study. On the other hand, LSAM is instrumental to the stabilization of AEP at near-neutral pH and its release has been shown to contribute to the conformational destabilization of AEP, which critically affects its activities (Dall and Brandstetter 2012, 2013). This might explain the absence of AEP activities in the protein fractions but were present in the crude extracts. In addition, AEPs are known to exist as dimers (Zhao et al. 2014; Zauner et al. 2018a) and recently it has been shown that the dimeric state of AtLEGγ, an AEP isoform from Arabidopsis thaliana, is pH dependent and contributes to keeping the enzyme silent (Zauner et al. 2018a). This oligomeric nature might be responsible for the absence of AEP activities in the crude extracts where the dimeric state may naturally occur or in the protein fractions where high salt concentration has driven proteins to self-associate (Wingfield 2016) and renders the AEPs in their dimeric state.

The absence of masses for cyclic SFTI-N alone or both cyclic and linear SFTI-N but with the presence of SFTI-NGL mass in the samples (Table 1) suggests that AEP activities were limited. This limitation could arise from a low expression level of the enzyme, low yield of the extracted enzyme, the presence of natural inhibitors, or incompatibility of the enzyme with the reaction buffer. Another possible reason could be the occurrence of protease cannibalism and degradation. On the other hand, the absence of masses for cyclic and linear SFTI-N as well as SFTI-NGL suggests that the substrate could be degraded by proteases. This result is consistent with the rapid degradation of synthetic SFTI-GLDN substrate (with the sequence GRCTKSIPPICFPDGLDN) in the crude extract of sunflower seeds (Bernath-Levin et al. 2015). Thus, some crude extracts appear to contain higher levels of proteolytic activities than others, and this factor should be taken into consideration when working with crude extracts. For example, should one attempt to label proteins using crude extracts (or protein fractions), potential side reactions (see the multiple peaks in Fig. 2) from other proteases must be anticipated.

Taken together, the positive hits we observed highlight that the products of SFTI-NGL processing can still be detected in 20–30% of the cases even with interfering proteases, albeit at a low yield for cyclic SFTI-N (discussed below). On the contrary, the negative hits highlight that screening for ligase activity can be abruptly affected by the presence of proteases and will likely require iterative attempts.

Following the screen, we sought to determine the presence of cyclic and linear SFTI-N from SFTI-NGL processing by AEPs using LC–MS. We selected the Fabaceous D. secundiflorum, a flowering shrub native to the American Southwest, and performed a time course reaction between its crude extract and SFTI-NGL. Of note, we chose D. secundiflorum because it was the first plant that we tested and showed positive hits for cyclic SFTI-N and we consequently used it to establish our screening protocol. From the LC–MS analysis, we found hit compounds with monoisotopic masses 1511.74 and 1529.75 Da consistent with cyclic and linear SFTI-N, respectively. We also sought to determine the yields of the cyclic and linear SFTI-N from the reaction. Cyclic SFTI-N appeared after 5 min of reaction, but its yield remained low, ~ 0.5 ± 0.1% (n = 3) of the substrate, over the 2-h reaction (Fig. 3). On the other hand, linear SFTI-N appeared at 0 min (the time between mixing the reaction and reaching the inactivation temperature), and its yield increased linearly from 2.1 ± 0.1% to 14.5 ± 3.5% (n = 3) of the substrate over the course of the reaction, suggesting that the SFTI-NGL was readily cleaved into linear SFTI-N and that this cleavage product was stable in the crude extract. Of note, after the 2-h reaction, the remaining SFTI-NGL was still high, as the initial concentration used was 10 times that of the screening.

Fig. 3.

Fig. 3

Open in a new tab

Cyclic and linear SFTI-N yields from reactions with D. secundiflorum crude extract. The relative intensities of linear (square) and cyclic (triangle) SFTI-N were normalized to SFTI-NGL. The yield of linear SFTI-N increased following a linear function, whereas the yield of cyclic SFTI-N remained stably low over the course of the reaction. The error bars represent the standard deviation from n = 3 repeats

The increasing yield of linear SFTI-N might have arisen from the activity of AEP isoforms that preferentially cleave SFTI-NGL at Asn and out-compete the AEP isoform that produces cyclic SFTI-N, which interestingly remained constant during the course of the reaction. The stability of the cyclic SFTI-N may be due to the cis–trans isomerization of Pro at the P2 position upon cyclization that prevented the cyclic peptide product from re-binding to and re-cleavage by AEP as suggested by a recent structural study using SFTI-1 and AtLEGγ (Zauner et al. 2018b). However, we also note that these results are in contrast to those observed in situ (in the crude extract of sunflower seeds) and in planta (in Nicotiana benthamiana leaves), where disulfide acyclic SFTI and cyclic SFTI-N or its precursor, respectively, are prone to rapid degradation (Bernath-Levin et al. 2015; Jackson et al. 2019). These contrasting results highlight that different biochemical compositions of plants affect the protease–ligase activities observed and that this nature should be considered when interpreting results from different samples. Another factor that may have affected the low yield of cyclic SFTI-N is pH, as AEP activities are pH dependent (Dall and Brandstetter 2016)—each isoform may have different pH requirements for performing ligation. Indeed, it was recently shown that AtAEP2 from A. thaliana and RcAEP1 from Ricinus communis can perform ligation of SFTI(D14N)-GLDN (with the sequence GRCTKSIPPICFPNGLDN) at pH 6.5 (Haywood et al. 2018), an activity that could not be observed previously at a pH of 5 (Bernath-Levin et al. 2015). In addition, the SFTI-NGL substrate might not have met the sequence requirements for efficient ligation by the AEP in D. secundiflorum. Nevertheless, these data suggest the possibility of using substrates with foreign sequences to screen for peptide ligases in crude extracts or protein fractions.

In closing our discussion, we emphasize the following points. First, the definitive proof for AEP ligation (in crude extracts or protein fractions) would be: (1) the cyclic SFTI-N being purified from the reaction and tested biochemically and biophysically; (2) the activation status of the AEP being tested using Western blot; and (3) the correlation between protease–ligase activities of the AEP being tested using a fluorescent probe. For the purpose of screening, we did not perform these assays. Second, false positives might be unintentionally assigned by the screening approach. For example, the reduction of 18 Da, which we used as an indicator for peptide bond formation, could also arise from other reactions such as the formation of dehydroalanine from serine, although uncommon in plants (Siodłak 2015). Thus, confirmation of selected hits should be performed before proceeding further. Comparison with known peptide ligase-producing plants such as species of the family Violaceae (Burman et al. 2015), if available, could also be performed. Third, in our discussion above, we have assumed that AEPs were responsible for the reactions observed. This assumption does not rule out the possibility of other proteases carrying out the chemistry. To narrow down the identification of the peptide ligase, one could perform the assays with the addition of inhibitors for specific classes of proteases or the removal of potential required factors. As our knowledge in this field is still in its infancy, we wish to conclude with a thought on future studies to advance the field.

Conclusion

Based on this study and the diversity of cyclic peptides distributed across phylogenetically distant plant species (Arnison et al. 2013), we speculate that proteases capable of ligating peptide bonds might be more common than previously thought. By identifying the suitable substrate sequences, in addition to the physicochemical conditions of the reaction to allow for ligation (Bergmann and Fraenkel-Conrat 1937), it should be possible to harness plants as a source for bioconjugation tools. For example, using a similar approach, plant ligase activity could be screened with libraries of peptide substrates. To avoid the prohibitive cost of pure synthetic peptide libraries, synthetic combinatorial libraries could be used instead, although this approach imposes the additional challenge of identifying the sequence of peptide hits (Falciani et al. 2005). Alternatively, biological display libraries could be used as well, as this approach offers direct access for acquiring sequence information from positive hits.

For example, future experiments could use the following workflow to identify new peptide ligases:

  1. Prepare the main components: (A) cells expressing combinatorial peptide libraries on the extracellular part of the N- or C-terminus of a membrane protein; (B) fluorescently labeled peptides with defined sequences; and (C) fresh plant crude extracts or protein fractions for ligase activity screening.

  2. Perform the ligation reaction by mixing and incubating components A–C.

  3. Separate cells by centrifugation, sort labeled cells by fluorescence-activated cell sorting (FACS), culture sorted cells, and sequence the colonies.

  4. Confirm the processability of the peptide sequence hits by intramolecular ligation and pursue positive hits for new ligase identification.

Although the above workflow might seem straightforward, it carries several caveats. These include: (1) the requirement for fluorescently labeled peptides, which can be expensive; (2) the use of complex mixtures containing proteases that can degrade the fluorescently labeled peptides and ligated products; (3) the use of cells that might express extracellular and membrane-bound proteases with unknown ligase activity, which might introduce false positives; (4) the dependence on exposure of the peptide libraries to the extracellular milieu but with no control over the steric hindrance presented by other molecules on the membrane; and (5) the dependence on detection, and thus number, of ligated fluorescently labeled peptides, which must attain a discernible signal–noise ratio and a minimum intensity above the FACS detection limit.

Finally, we wish to reiterate that conjugation tools for orthogonal protein modification are highly sought across a wide range of fields, from chemical biology to materials science (Sletten and Bertozzi 2009; Witus and Francis 2011). Some proteases such as sortases and AEPs have already provided attractive options for site-specific protein modification purposes (Antos et al. 2016; Nguyen et al. 2016). Efforts to further expand the diversity and availability of bioconjugation tools will likely open opportunities to new applications and benefit many fields.

Acknowledgements

We thank Rizal F. Hariadi (Arizona State University), Edward K. Gilding (The University of Queensland) and Mark A. Jackson (The University of Queensland) for the helpful discussion, Gabrielle R. Hirneise (Arizona State University) for help with the manuscript, and the Biodesign Institute, Arizona State University, for access to their mass spectrometry facility.

Author contributions

Conceived and designed the experiments: Tunjung Mahatmanto, Isyatul Azizah. Performed the experiments: Tunjung Mahatmanto, Isyatul Azizah, Alex Buchberger. Analyzed the data: Tunjung Mahatmanto, Isyatul Azizah. Contributed reagents/materials/analysis tools: Nicholas Stephanopoulos. Wrote the paper: Tunjung Mahatmanto, Isyatul Azizah, Nicholas Stephanopoulos.

Funding

This study was supported by the Arizona State University Start-Up Funds to Rizal F. Hariadi and Nicholas Stephanopoulos.

Compliance with ethical standards

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

References

  1. Antos JM, Truttmann MC, Ploegh HL. Recent advances in sortase-catalyzed ligation methodology. Curr Opin Struct Biol. 2016;38:111–118. doi: 10.1016/j.sbi.2016.05.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arnison PG, Bibb MJ, Bierbaum G, et al. Ribosomally synthesized and post-translationally modified peptide natural products: overview and recommendations for a universal nomenclature. Nat Prod Rep. 2013;30:108–160. doi: 10.1039/C2NP20085F. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Barber CJS, Pujara PT, Reed DW, et al. The two-step biosynthesis of cyclic peptides from linear precursors in a member of the plant family Caryophyllaceae involves cyclization by a serine protease-like enzyme. J Biol Chem. 2013;288:12500–12510. doi: 10.1074/jbc.M112.437947. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bergmann M, Fraenkel-Conrat H. The role of specificity in the enzymatic synthesis of proteins: syntheses with intracellular enzymes. J Biol Chem. 1937;119:707–720. [Google Scholar]
  5. Bernath-Levin K, Nelson C, Elliott AG, et al. Peptide macrocyclization by a bifunctional endoprotease. Chem Biol. 2015;22:571–582. doi: 10.1016/j.chembiol.2015.04.010. [DOI] [PubMed] [Google Scholar]
  6. Burman R, Yeshak MY, Larsson S, et al. Distribution of circular proteins in plants: large-scale mapping of cyclotides in the Violaceae. Front Plant Sci. 2015 doi: 10.3389/fpls.2015.00855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Craik DJ, Daly NL, Bond T, Waine C. Plant cyclotides: a unique family of cyclic and knotted proteins that defines the cyclic cystine knot structural motif. J Mol Biol. 1999;294:1327–1336. doi: 10.1006/jmbi.1999.3383. [DOI] [PubMed] [Google Scholar]
  8. Dall E, Brandstetter H. Activation of legumain involves proteolytic and conformational events, resulting in a context-and substrate-dependent activity profile. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2012;68:24–31. doi: 10.1107/S1744309111048020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dall E, Brandstetter H. Mechanistic and structural studies on legumain explain its zymogenicity, distinct activation pathways, and regulation. Proc Natl Acad Sci. 2013;110:10940–10945. doi: 10.1073/pnas.1300686110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dall E, Brandstetter H. Structure and function of legumain in health and disease. Biochimie. 2016;122:126–150. doi: 10.1016/j.biochi.2015.09.022. [DOI] [PubMed] [Google Scholar]
  11. Dall E, Fegg JC, Briza P, Brandstetter H. Structure and mechanism of an aspartimide-dependent peptide ligase in human legumain. Angew Chemie Int Ed. 2015;54:2917–2921. doi: 10.1002/anie.201409135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Elliott AG, Delay C, Liu H, et al. Evolutionary origins of a bioactive peptide buried within preproalbumin. Plant Cell. 2014;26:981–995. doi: 10.1105/tpc.114.123620. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Falciani C, Lozzi L, Pini A, Bracci L. Bioactive peptides from libraries. Chem Biol. 2005;12:417–426. doi: 10.1016/j.chembiol.2005.02.009. [DOI] [PubMed] [Google Scholar]
  14. Gilding EK, Jackson MA, Poth AG, et al. Gene coevolution and regulation lock cyclic plant defence peptides to their targets. New Phytol. 2016;210:717–730. doi: 10.1111/nph.13789. [DOI] [PubMed] [Google Scholar]
  15. Harris KS, Durek T, Kaas Q, et al. Efficient backbone cyclization of linear peptides by a recombinant asparaginyl endopeptidase. Nat Commun. 2015;6:10199. doi: 10.1038/ncomms10199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Haywood J, Schmidberger JW, James AM, et al. Structural basis of ribosomal peptide macrocyclization in plants. Elife. 2018;7:e32955. doi: 10.7554/eLife.32955. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jackson MA, Gilding EK, Shafee T, et al. Molecular basis for the production of cyclic peptides by plant asparaginyl endopeptidases. Nat Commun. 2018;9:2411. doi: 10.1038/s41467-018-04669-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Jackson MA, Yap K, Poth AG, et al. Rapid and scalable plant-based production of a potent plasmin inhibitor peptide. Front Plant Sci. 2019 doi: 10.3389/fpls.2019.00602. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Krall N, Da Cruz FP, Boutureira O, Bernardes GJL. Site-selective protein-modification chemistry for basic biology and drug development. Nat Chem. 2016;8:103–113. doi: 10.1038/nchem.2393. [DOI] [PubMed] [Google Scholar]
  20. Kumaran S, Datta D, Roy RP. An enigmatic peptide ligation reaction: protease-catalyzed oligomerization of a native protein segment in neat aqueous solution. Protein Sci. 2000;9:734–741. doi: 10.1110/ps.9.4.734. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lee J, Mcintosh J, Hathaway BJ, Schmidt EW. Using marine natural products to discover a protease that catalyzes peptide macrocyclization of diverse substrates. J Am Chem Soc. 2009;131:2122–2124. doi: 10.1021/ja8092168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lotze J, Reinhardt U, Seitz O, Beck-Sickinger AG. Peptide-tags for site-specific protein labelling in vitro and in vivo. Mol BioSyst. 2016;12:1731–1745. doi: 10.1039/C6MB00023A. [DOI] [PubMed] [Google Scholar]
  23. Luo H, Hong SY, Sgambelluri RM, et al. Peptide macrocyclization catalyzed by a prolyl oligopeptidase involved in α-amanitin biosynthesis. Chem Biol. 2014;21:1610–1617. doi: 10.1016/j.chembiol.2014.10.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mahatmanto T. Seed biopharmaceutical cyclic peptides: from discovery to applications. Biopolymers. 2015;104:804–814. doi: 10.1002/bip.22741. [DOI] [PubMed] [Google Scholar]
  25. Mahatmanto T, Poth AG, Mylne JS, Craik DJ. A comparative study of extraction methods reveals preferred solvents for cystine knot peptide isolation from Momordica cochinchinensis seeds. Fitoterapia. 2014;95:22–33. doi: 10.1016/j.fitote.2014.02.016. [DOI] [PubMed] [Google Scholar]
  26. Mahatmanto T, Mylne JS, Poth AG, et al. The evolution of Momordica cyclic peptides. Mol Biol Evol. 2015;32:392–405. doi: 10.1093/molbev/msu307. [DOI] [PubMed] [Google Scholar]
  27. Mazmanian SK, Liu G, Ton-That H, Schneewind O. Staphylococcus aureus sortase, an enzyme that anchors surface proteins to the cell wall. Science. 1999;285:760–763. doi: 10.1126/science.285.5428.760. [DOI] [PubMed] [Google Scholar]
  28. Mylne JS, Colgrave ML, Daly NL, et al. Albumins and their processing machinery are hijacked for cyclic peptides in sunflower. Nat Chem Biol. 2011;7:257–259. doi: 10.1038/nchembio.542. [DOI] [PubMed] [Google Scholar]
  29. Nguyen GKT, Wang S, Qiu Y, et al. Butelase 1 is an Asx-specific ligase enabling peptide macrocyclization and synthesis. Nat Chem Biol. 2014;10:732–738. doi: 10.1038/nchembio.1586. [DOI] [PubMed] [Google Scholar]
  30. Nguyen GKT, Qiu Y, Cao Y, et al. Butelase-mediated cyclization and ligation of peptides and proteins. Nat Protoc. 2016;11:1977–1988. doi: 10.1038/nprot.2016.118. [DOI] [PubMed] [Google Scholar]
  31. Poon S, Harris KS, Jackson MA, et al. Co-expression of a cyclizing asparaginyl endopeptidase enables efficient production of cyclic peptides in planta. J Exp Bot. 2018;69:633–641. doi: 10.1093/jxb/erx422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Siodłak D. α, β-Dehydroamino acids in naturally occurring peptides. Amino Acids. 2015;47:1–17. doi: 10.1007/s00726-014-1846-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Sletten EM, Bertozzi CR. Bioorthogonal chemistry: fishing for selectivity in a sea of functionality. Angew Chemie Int Ed. 2009;48:6974–6998. doi: 10.1002/anie.200900942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Spicer CD, Davis BG. Selective chemical protein modification. Nat Commun. 2014;5:4740. doi: 10.1038/ncomms5740. [DOI] [PubMed] [Google Scholar]
  35. The Plant List (2010). In: Version 1. http://www.theplantlist.org/. Accessed 8 Jun 2018
  36. Trads JB, Tørring T, Gothelf KV. Site-selective conjugation of native proteins with DNA. Acc Chem Res. 2017;50:1367–1374. doi: 10.1021/acs.accounts.6b00618. [DOI] [PubMed] [Google Scholar]
  37. Walsh CT, Garneau-Tsodikova S, Gatto GJ. Protein posttranslational modifications: the chemistry of proteome diversifications. Angew Chemie Int Ed. 2005;44:7342–7372. doi: 10.1002/anie.200501023. [DOI] [PubMed] [Google Scholar]
  38. Wang CKL, Kaas Q, Chiche L, Craik DJ. CyBase: a database of cyclic protein sequences and structures, with applications in protein discovery and engineering. Nucleic Acids Res. 2008;36:D206–D210. doi: 10.1093/nar/gkm953. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wingfield PT (2016) Protein precipitation using ammonium sulfate. Curr Protoc Protein Sci 84:A.3F.1–A.3F.9. 10.1002/0471140864.psa03fs84 [DOI] [PubMed]
  40. Witus LS, Francis MB. Using synthetically modified proteins to make new materials. Acc Chem Res. 2011;44:774–783. doi: 10.1021/ar2001292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Yang R, Wong YH, Nguyen GKT, et al. Engineering a catalytically efficient recombinant protein ligase. J Am Chem Soc. 2017;139:5351–5358. doi: 10.1021/jacs.6b12637. [DOI] [PubMed] [Google Scholar]
  42. Zauner FB, Dall E, Regl C, et al. Crystal structure of plant legumain reveals a unique two-chain state with pH-dependent activity regulation. Plant Cell. 2018;30:686–699. doi: 10.1105/tpc.17.00963. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zauner FB, Elsässer B, Dall E, et al. Structural analyses of Arabidopsis thaliana legumain γ reveal differential recognition and processing of proteolysis and ligation substrates. J Biol Chem. 2018;293:8934–8946. doi: 10.1074/jbc.M117.817031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhao L, Hua T, Crowley C, et al. Structural analysis of asparaginyl endopeptidase reveals the activation mechanism and a reversible intermediate maturation stage. Cell Res. 2014;24:344–358. doi: 10.1038/cr.2014.4. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from 3 Biotech are provided here courtesy of Springer

RESOURCES