Functional characterization of a dehydrin protein from Fagus sylvatica seeds using experimental and in silico approaches

Functional characterization of a dehydrin protein from Fagus sylvatica seeds using experimental and in silico approaches

Plant Physiology and Biochemistry 97 (2015) 246e254 Contents lists available at ScienceDirect Plant Physiology and Biochemistry journal homepage: ww...

1MB Sizes 0 Downloads 29 Views

Plant Physiology and Biochemistry 97 (2015) 246e254

Contents lists available at ScienceDirect

Plant Physiology and Biochemistry journal homepage: www.elsevier.com/locate/plaphy

Research article

Functional characterization of a dehydrin protein from Fagus sylvatica seeds using experimental and in silico approaches Ewa Marzena Kalemba*, Monika Litkowiec rnik, Poland Institute of Dendrology, Polish Academy of Sciences, Parkowa 5, 62-035 Ko

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 July 2015 Received in revised form 6 September 2015 Accepted 6 October 2015 Available online 22 October 2015

A strong increase in the level of dehydrin/response ABA transcripts expression reported from the 14th week after flowering coincident with the accumulation of 26 and 44 kDa dehydrins in the embryonic axes of developing beech (Fagus sylvatica L.) seeds. Both transcript and protein levels were strongly correlated with maturation drying. These results suggest that the 44-kDa dehydrin protein is a putative dimer of dehydrin/response ABA protein migrating as a 26-kDa protein. Dehydrins and dehydrin-like proteins form large oligomeric complexes under native conditions and are shown as several spots differing in pI through isoelectrofocusing analyses. Detailed prediction of specific sites accessible for various post-translational modifications (PTMs) in the dehydrin/response ABA protein sequence revealed sites specific to acetylation, amidation, glycosylation, methylation, myristoylation, nitrosylation, O-linked b-N-acetylglucosamination and Yin-O-Yang modification, palmitoylation, phosphorylation, sumoylation, sulfation, and ubiquitination. Thus, these results suggest that specific PTMs might play a role in switching dehydrin function or activity, water binding ability, protein-membrane interactions, transport and subcellular localization, interactions with targeted molecules, and protein stability. Despite the ability of two Cys residues to form a disulfide bond, eSH groups are likely not involved in dimer arrangement. Hisrich regions and/or polyQ-tracts are potential candidates as spatial organization modulators. Dehydrin/ response ABA protein is an intrinsically disordered protein containing low complexity regions. The lack of a fixed structure and exposition of amino acids on the surface of the protein structure enhances the accessibility to 40 predicted PTM sites, thereby facilitating dehydrin multifunctionality, which is discussed in the present study. © 2015 Elsevier Masson SAS. All rights reserved.

Keywords: Dehydrin Disorder Post-translational modification Seed development

1. Introduction Common beech (Fagus sylvatica L.) is native throughout almost all of Europe and produces seeds in large intervals. During development, beech seeds become desiccation tolerant, beginning at the 16th week after flowering (Kalemba et al., 2009), similar to all other orthodox type seeds. Beech seeds were classified into the intermediate category, reflecting the reduced longevity of these seeds during storage (Kalemba and Pukacka, 2014). Late Embryogenesis Abundant (LEA) proteins, including dehydrins (group 2 LEA proteins), are hydrophilic molecules accumulated during seed

Abbreviations: ABA, abscisic acid; IUP, intrinsically unstructured protein; LEA, late embryogenesis abundant; PTM, post-translational modification; RNS, reactive nitrogen species; ROS, reactive oxygen species; TF, transcription factors; WAF, week after flowering. * Corresponding author. E-mail address: [email protected] (E.M. Kalemba). http://dx.doi.org/10.1016/j.plaphy.2015.10.011 0981-9428/© 2015 Elsevier Masson SAS. All rights reserved.

development (reviewed in Hara, 2010; Battaglia and Covarrubias, €chter et al., 2013) as a consequence of maturation drying (Kleinwa 2014). Dehydrins are characterized by highly conserved sequences, including a lysine-rich K segment, a polyserine S-segment, a Y-segment, and a F-segment (Hara, 2010; Battaglia and Covarrubias, 2013). The K segment uniquely contains the consensus amino acid sequence EKKGIMDKIKELPG, which is present in all types of dehydrins. Using antibodies specific to the K segment (Close et al., 1993), several dehydrin-like proteins were detected in beech seeds (Kalemba et al., 2009, 2015). Among them, two heat-stable proteins were further classified as dehydrins. During development of beech seeds, 44-kDa dehydrin was predominantly accumulated in embryonic axes (Kalemba et al., 2009, 2015). The structural and functional characteristics of LEAs have been investigated to understand the roles of these proteins. LEAs prevent protein aggregation (Goyal et al., 2003), protect cellular macromolecules and membranes (Hara, 2010; Battaglia and Covarrubias,

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

2013), and might have roles as antioxidants (Hara, 2010) and membrane and protein stabilizers (Battaglia and Covarrubias, 2013). Dehydrins bind water, macromolecules (DNA, RNA, phospholipids, and cytoskeleton), ions (Ca, Cu, Ni, and Zn) and act as chaperones, molecular shields, and ion buffers, thereby preventing protein denaturation (Hara, 2010; Battaglia and Covarrubias, 2013; Graether and Boddington, 2014) and playing an essential role in embryogenesis and the regulation of cellular osmotic potential in seeds (Kleinw€ achter et al., 2014). Dehydrins are peripheral proteins (Danyluk et al., 1998) that might function as dimers (Lin et al., 2012; nchez et al., 2014). The functional versatility in these Hern andez-Sa proteins might be associated with diverse post-translational modifications (PTMs). For example, phosphorylation was required for dehydrin function (Brini et al., 2007) and membrane binding (Eriksson et al., 2011). Protein activity and localization are determined through reversible PTMs (Traverso et al., 2008; Park et al., 2011; Running, 2014). According to statistics data, phosphorylation is the most frequently occurring PTM type experimentally reported, whereas non-experimental data suggest that N-linked glycosylation and phosphorylation are the main PTMs (Khoury et al., 2011). The role of protein phosphorylation has been well documented in seed physiology; for example, the phosphorylation of cruciferin, a storage protein, exclusively occurs during Arabidopsis seed maturation and has not been observed during germination (Arc et al., 2011). Phosphorylation in dehydrins has been associated with the presence of the S-segment (Brini et al., 2007). Dehydrins and dehydrin-like proteins were present in phosphorylated forms in developing beech seeds (Kalemba et al., 2015). Public databases contain one unique dehydrin gene and protein sequence originating from beech seeds, called dehydrin/response ABA pronez et al., 2008), and tein with accession number CAE54590.1 (Jime partial sequences of this protein. Dehydrin/response ABA protein nez et al., 2008), and has been classified in the YnSK2 group (Jime the molecular mass predicted from the amino acid sequence (21.5 kDa) is similar to the molecular weight of the beech 26-kDa dehydrin after enzymatic dephosphorylation (Kalemba et al., 2015). During beech seed development dehydrin-like proteins are localized to the cytosol and amyloplasts, inside nuclei, along the cell plasma membrane, and in the electron-dense vacuoles and vesicles. Dehydrins have also been associated with small vesicles and small membrane structures throughout the cytoplasm (Kalemba et al., 2015). Among all PTMs, acetylation, phosphorylation (Jensen et al., 1998), palmitoylation (Running, 2014) and sumoylation (Park et al., 2011) are involved in the determination of protein subcellular localization. Several putative sites that interact with DNA, proteins, and lipid membranes have been predicted in dehydrin/response ABA proteins derived from beech seeds (Kalemba et al., 2015). Dehydrin-membrane interactions could be alleviated through protein lipidation, including myristoylation, (Traverso et al., 2008), palmitoylation and prenylation (Running, 2014). Conversely, lipidation has not been studied in dehydrin proteins. Experimental data and in silico predictions have suggested that dehydrins and dehydrin-like proteins are involved in the reduction of desiccation-induced cellular damage of the plasma membrane, organelle membranes, nuclei, and storage material reservoirs (Kalemba et al., 2015); however, the mechanism of dehydrin action is not fully understood. Through a search of the Protein Data Bank (PDB) database, two unique three-dimensional structures (PDB IDs: 1YYC and 1XO8) related to the LEA family were identified. The structure 1YYC is a member of the LEA-2 family, comprising one a-helix and seven bsheets. Helical structures might be introduced in LEA proteins after using sodium dodecyl sulfate (Koag et al., 2009) or after slow and fast drying (Goyal et al., 2003; Shih et al., 2012). Despite the conserved segments, dehydrins are intrinsically unstructured

247

proteins (IUPs) (Kovacs et al., 2008; Hara, 2010; Shih et al., 2012; Battaglia and Covarrubias, 2013; Graether and Boddington, 2014). The transition from a disordered to an ordered state has been associated with dehydrin function (Hara, 2010) or interactions with specific partner molecules (Tompa et al., 2006), such as membrane targets (Koag et al., 2009). Conformational changes suggest potential interactions with a broad range of molecules and might specify more than one function (Battaglia and Covarrubias, 2013). The sequencing and functional characterization of dehydrins and dehydrin-like proteins derived from beech seeds have only been partially achieved. Therefore, in the present study, we performed several biochemical and bioinformatics studies to obtain additional information about dehydrin proteins. 2. Material and methods 2.1. Material collection Common beech (F. sylvatica L.) seeds were collected every 7 or 10 days from the 12th to the 19th week after flowering (WAF) from rnik Arboretum (Western Poland) during the a single tree in Ko cropping season in 2009. At the 19th WAF, the seeds were completely matured and shed from the tree. The seeds intended for protein isolation were prepared for storage at 80  C. The seed coats were removed, and the embryonic axes were separated from the cotyledons. The embryonic axes were used in all experiments. The water content (WC, dry weight basis) was determined during seed collection (Table 1, Kalemba et al., 2015). 2.2. Protein extraction The embryonic axes were ground to a powder in liquid nitrogen. To obtain soluble proteins, the dried powder was homogenized at 4  C in a 1:2 (w:v) extraction buffer containing 50 mM phosphate buffer (pH 7.0), 1% protease inhibitor cocktail (SigmaeAldrich, St. Louis, USA) and 1.5% polyvinylpolypyrrolidone (SigmaeAldrich, St. Louis, USA). The samples were further centrifuged at 20,000g for 20 min at 4  C. The protein concentration was measured according to the Bradford (1976) method using bovine albumin as a standard. Proteins intended for 2D-electrophoresis were prepared in buffer containing 7 M urea, 2 M thiourea, 40 mM dithiothreitol (DTT), 0.5% carrier ampholytes, and 4% CHAPS. The protein concentration was measured using a 2-D Quant Kit (GE Healthcare, Piscataway, USA). 2.3. Electrophoresis The proteins were separated using 4e16% precast gels (Bio-Rad Laboratories, Hercules California, USA) and the Mini-PROTEAN® Tetra Cell (Bio-Rad Laboratories, Hercules California, USA) system. Tris-glycine and Tris-glycine-sodium dodecyl sulfate (SDS) buffers were used to run Native-PAGE and SDS-PAGE, respectively. The gels were stained with Coomassie Brilliant Blue or silver nitrate according to the method of Sinha et al. (2001). For the diagonal 2D redox SDS-PAGE, the samples were incubated with 100 mM DTT or 10 mM hydrogen peroxide (H2O2) for 30 min at room temperature, with subsequent iodoacetamide treatment for 30 min in the dark. Thirty micrograms of protein was subjected to diagonal 2D redox €her and Dietz SDS-PAGE analysis according to the protocol of Stro (2008). For 2D-electrophoresis, isoelectrofocusing was conducted using 60 mg of protein loaded onto 7 cm, 4e7 pH ready gel strips and the Multiphor II Electrophoresis System (GE Healthcare, Piscataway, USA). Spectra™ Multicolor Broad Range Protein Ladder and NativeMark™ standards (Thermo Scientific, Rockford, USA) were used to estimate the protein molecular mass.

248

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

Table 1 The water content (WC, dry weight basis) determined in embryonic axes of common beech (Fagus sylvatica L.) seeds collected every 7 or 10 days from the 12th to the 19th week  rnik Arboretum, Western Poland (Kalemba et al., 2015). after flowering (WAF) from a single tree in Ko WC

12 WAF

13 WAF

14 WAF

15 WAF

16 WAF

17.5 WAF

19 WAF

Embryonic axes

79.3 ± 1.4

75.2 ± 2.3

72.2 ± 3.1

63.1 ± 1.7

60.4 ± 2.8

51.3 ± 1.9

17.7 ± 0.8

2.4. Western blot analysis The protein extract (12.5 mg) was loaded onto a gel. The protein concentration was calculated based on the calibration curve data, in which the relationship between the Western blot band intensities and the amount of loaded protein was linear (correlation coefficient, r ¼ 0.98). The fractionated proteins were transferred onto a polyvinylidene fluoride membrane Immobilon™-P (Millipore, Billerica, USA) at 350 mA for 1 h and subsequently blocked and incubated with a primary antibody (dilution 1:1000) against the dehydrin consensus K segment (Close et al., 1993). The primary antibody was raised in rabbit. The secondary antibody was conjugated with alkaline phosphatase (SigmaeAldrich, St. Louis, USA) and used at a 1:10,000 dilution. The protein bands were visualized on the membrane using an alkaline phosphate substrate (5-bromo4-chloro-3-indolyl phosphate/nitro blue tetrazolium) (SigmaeAldrich, St. Louis, USA). Control immunoassays were also performed using a dehydrin K-segment-specific antibody blocked with the synthetic peptide TGEKKGIMDKIKEKLPGQH (Close et al., 1993). 2.5. Transcript analyses Total RNA was isolated using the SV Total RNA Isolation System (Promega, Madison, USA). The quality and quantity of the purified RNA was determined using Biofotometer Plus (Eppendorf, Hamburg, Germany) based on A260/280 and A260/230 parameters. The cDNA was synthesized using the Verte Kit (Novazym, Poznan, Poland) containing oligo(dT)15 primers and MMLV reverse transcriptase supplemented with RNase inhibitor (Promega, Madison, USA). Primers specific to actin and dehydrin-conserved segments were examined in parallel, and the QIAGEN Multiplex PCR Kit was used to obtain multiple copies of specific genes. Semi-qRT-PCR was performed using the Veriti® 96-Well Thermal Cycler (Applied Biosystems, Foster City, USA).

and the number of pixels inside the area of the spot.

2.8. Statistical analysis The data are presented as the means ± standard deviation of three biological replicates. The significant differences between particular parameters were examined using Pearson's correlation coefficient analysis. The significance among the means of components was verified using analysis of variance (ANOVA), followed by Tukey's test at P  0.05, and significantly different values were marked with different letters in the graph.

2.9. Bioinformatics tools 2.9.1. Detection of accessible sites of PTM All in silico analyses were performed using the F. sylvatica L. dehydrin/response ABA protein sequence, accession number nez et al., 2008), as query. Phosphorylation, CAE54590.1 (Jime myristoylation, amidation, sumoylation, sulfation, acetylation, Nglycosylation, O-glycosylation, O-linked b-N-acetylglucosamination (O-b-GlcNac), ubiquitination, S-nitrosylation, S-palmitoylation, and methylation sites were predicted. For more details see the Supplemental Data file. 2.9.2. Structure analyses Protein disorder was assessed through the MetaDisorder Web Service (Kozlowski and Bujnicki, 2012). Moreover low complexity regions, non-ordinary and helical secondary structures, hydrophobic clusters disulfide-bonding states of Cys residues were predicted. For more details see the Supplemental Data file. 3. Results

2.6. Transcript sequencing

3.1. Detection of dehydrin and dehydrin-like proteins

The semi-qRT-PCR products were purified using MultiScreen plates with specialty membranes (Millipore, Billerica, USA) and a vacuum pump (20 min). The products were sequenced using an ABI PRISM 3130 automatic sequencer (Applied Biosystems, Foster City, USA). The sequencing PCR reaction was performed using the standard protocol for BigDye® Terminator v3.1 (Life Technologies, Carlsbad, USA) and selected primers. Multiple sequence alignment (MSA) of the sequenced transcripts was performed to identify the consensus sequence used for the search for highly identical sequences.

Analyses performed under reducing conditions and SDS-PAGE electrophoresis, followed by Western blotting, indicated the presence of 26, 35, 40, and 44 kDa dehydrin-like proteins, among which 26 and 44 kDa proteins accumulated during seed maturation (Fig. 1A) and were classified as dehydrins because these proteins were heat-stable. Protein extracts isolated from embryonic axes of maturing beech seeds were separated under native conditions, followed by Western blotting using an antibody specific to the dehydrin K-segment (Fig. 1B). It has been reported that in the native state, dehydrin proteins might perform high molecular aggregates, approximated using NativeMark™ protein standards. 2D electrophoresis showed that all dehydrins and dehydrin-like proteins are acidic proteins (Fig. 1C). The isoelectric point (pI) ranged between 5 and 6, except for the 35-kDa protein. Interestingly, 40 and 44 kDa proteins exhibited several 0.08e0.3 pI shifts, indicating the presence of PTMs, including phosphorylation, as previous studies have shown that dehydrin and dehydrin-like proteins in beech seeds exist in a phosphorylated state (Kalemba et al., 2015). Moreover, a strong correlation was identified between the 26- and 44-kDa dehydrins and the desiccation levels in seeds (Fig. 1D).

2.7. Densitometry analysis Images were analyzed in triplicate using the UviBand (UviTec, Cambridge, United Kingdom) program in the Fire Reader Gel Documentation System. Densitometry image analysis was based on the digitalization of the image in pixels, and the intensity was coded on a scale of 256 gray levels. The density, based on the volume (V) of the spot, was calculated as the sum of all 3D intensities (I). The data are presented in relative units obtained from V ¼ SniI

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

249

Fig. 1. Detection of dehydrin-like proteins (indicated as arrows) in the embryonic axes of beech seeds (Fagus sylvatica L.) using antibodies specific to the K-segment and Western blot (Wb) method: A) during development between the 12th and 19th week after flowering (WAF) after protein separation through SDS-PAGE electrophoresis, B) at the 19th WAF after protein separation under native conditions (N), C) after protein separation through 2D electrophoresis. Each gel shows a representative of three independent experiments with similar results. Spectra™ multicolor broad range protein ladder (A) and NativeMark™ (B) standards were used to assess the molecular weight of the detected proteins. The accumulation of 26 kDa and 44 kDa dehydrins was correlated with the desiccation levels (water content, dry weight basis) of embryonic axes during seed development using Pearson's correlation coefficient analysis (D).

3.2. Analyses of transcripts related to dehydrin protein The search for genes encoding dehydrins and dehydrin-like proteins in the beech genome based on degenerated primers and several primer pairs specific to conserved segments revealed one clear DNA product (data not shown). The PCR product was subjected to sequencing. Sequencing results (Suppl. Fig. 1A) were subsequently used as query in a Blast search for highly identical sequences deposited in the databases. The results showed 99% identity with the F. sylvatica dhn1 gene encoding mRNA for dehydrin/response ABA protein and the F. sylvatica partial dhn gene encoding exons 1e2 of dehydrin (Suppl. Fig. 1B). Based on this result, we analyzed the expression of this gene at the mRNA level. Samples of embryonic axes were examined separately during beech seed maturation from 12 to 17.5 WAF. The transcript level increased during seed maturation (Fig. 2A). The particular changes, based on independent biological replicates were graphed (Fig. 2B). The level of semi-qRT-PCR product was 3 times higher at the end of maturation than at the beginning of maturation (12e13 WAF). A strong correlation (r ¼ 0.90716, p ¼ 0.01253) was observed between the transcript level and the desiccation levels (Fig. 2C). To determine the dehydrin protein corresponding to this transcript, we compared the transcript levels to the level of 26 and 44 kDa dehydrins. A strong correlation was assigned to both dehydrins (Fig. 2D). 3.3. Detection of sites specific to post-translational modifications in dehydrin protein A detailed prediction of specific sites in the dehydrin/response ABA protein sequence accessible for PTM was performed (Fig. 3).

Analyses showed 22 potential sites phosphorylated through specific kinases (Suppl. Fig. 2A), and 12 non-specific phosphorylation sites (Suppl. Fig. 2B). The phosphorylation is highly likely, as the phosphorylation potential was nearly 1 using 0.5 as the threshold (Suppl. Fig. 2C). Considering the fact that the lack of tertiary and quaternary structure promotes post-translational modifications, analyses using a disorder-enhanced phosphorylation sites predictor were executed. Among 28 potential phosphorylation sites, 11 sites might be phosphorylated with high potential (Suppl. Fig. 2D). Accessible servers and programs listed in the ExPASy bioinformatics resource portal, recognizing PTMs sites, were run. The distribution of all predicted PTMs in the dehydrin/response ABA protein sequence is shown in Fig. 3. In silico predictions showed that in dehydrin/response ABA protein Ser and Tyr residues undergoes phosphorylation (Fig. 3, Suppl. Fig. 2). Further predictions revealed that the dehydrin/response ABA protein contains three specific myristoylation sites, one specific amidation site (Suppl. Fig. 3A), five sumoylation sites (Suppl. Fig. 3B), two sulfation sites (Suppl. Fig. 3C), one N-terminal acetylation site (Suppl. Fig. 3D), one N-glycosylation site (Suppl. Fig. 3E), one Oglycosylation site (Suppl. Fig. 3F), eight O-b-GlcNac modification sites, among which three sites were also predicted as phosphorylation sites and therefore considered as three sites of Yin-O-Yang type PTM (Suppl. Fig. 3G), five ubiquitination sites (Suppl. Fig. 3H), one S-nitrosylation site (Suppl. Fig. 3I), one S-palmitoylation site (Suppl. Fig. 3J), and two methylation sites (Suppl. Fig. 3K). All the PTMs sites listed above were also detected in other dehydrin proteins grouped into YnSKn, YnKn, SKn, Kn classes (Suppl. Table 1).

250

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

Fig. 2. Detection of dehydrin-related transcripts in the embryonic axes of beech seeds (Fagus sylvatica L.) (A) using semi-qRT-PCR and primers specific to conserved dehydrin segments and actin as a reference probe during development between the 12th and 17.5th week after flowering (WAF). Each gel shows a representative result obtained from three independent experiments with similar results. The transcript levels shown in the graph were derived from three independent experiments after digitalization of the results and densitometry analyses of the images (B). Statistically significant differences are indicated with different letters (one-way ANOVA, followed by Tukey's test at p  0.05). Changes in the transcript levels were correlated with the desiccation levels (water content, dry weight basis) of embryonic axes during seed development time using Pearson's correlation coefficient analysis (C) to determine whether mRNA synthesis is related to maturation drying. The level of 26 kDa and 44 kDa dehydrins was correlated with the transcript levels during development to identify the protein related to this transcript (D).

3.4. Structure analyses of dehydrin protein The results obtained from the MetaDisorder web service determined that the dehydrin/response ABA protein was disordered because the value of disorder tendency for meta-analyses ranged from 0.5 to 1 (Fig. 4). The region at positions 1e84 was partially unstructured, whereas the area of amino acids at positions 85e183 was completely unstructured, characterized by a disorder tendency higher than 0.7. The detailed analysis of the structural elements showed that the amino acids at positions 97e146 and 167e179, which overlaps with the S- and K-segments, were considered low complexity regions (Suppl. Fig. 4A). Protein

structure predictions revealed that non-ordinary secondary structures comprise a majority of the dehydrin protein, whereas the 8 amino acids overlapping the K-segment (at positions 171e179) might form helical structures (Suppl. Fig. 4B). Solvent accessibility analyses showed that all amino acids were exposed on the surface of the protein structure. Hydrophobic amino acids were clustered into small groups, indicating a lack of regular secondary structures (Suppl. Fig. 4C). Dehydrin/response ABA protein contains two neighboring Cys residues (at positions 58e59), predicted to form disulfide bonds with high scores using independent algorithms (Suppl. Fig. 4D). Based on the presence of two Cys residues forming disulfide

Fig. 3. The contribution of potential posttranslational modifications (PTMs) predicted in dehydrin/response ABA proteins using the accessible and high-accuracy servers and programs listed at ExPASy bioinformatics resource portal. Conserved sequences, including two K-segments, two Y-segments, and poly-serine track (bolded type), are indicated in the sequence. The following PTMs were predicted: amidation (Amid), methylation (Met), myristoylation (Myr), N-glycosylation (N-Glyc), N-terminal acetylation (N-Acet), Oglycosylation (N-Glyc), O-linked b-N-acetylglucosamination (O-b-GlcNac), O-linked b-N-acetylglucosamination and phosphorylation (Yin-O-Yang), phosphorylation (P), S-nitrosylation (SNO), S-palmitoylation (Palm), sulfation (Sulfat), sumoylation (SUMO), and ubiquitination (Ub). Additional data are available in Suppl. Figs. 2 and 3 and Suppl. Table 1.

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

251

Fig. 4. Disorder plot of dehydrin/response ABA protein. Metaprediction (MetaDisorderMD2) was based on the results of 13 disorder predictors using GeneSilico MetaDisorder web service. The residues with disorder probabilities higher than 0.5 were considered as disordered.

bonds, we performed diagonal 2D redox SDS-PAGE experiments to detect proteins forming intermolecular and/or intramolecular disulfide bonds. Interestingly, after reduction of protein samples, proteins resolved as a 26 kDa dehydrin after first dimension electrophoresis were subsequently separated into a 44 and 26 kDa forms after second dimension electrophoresis (Fig. 5), suggesting that 44 kDa dehydrin might be a dimer of 26 kDa dehydrin molecules. Thus, it is reasonable to hypothesize that under reducing conditions, monomers might spontaneously form dimers; however, reducing conditions are not sufficient to disrupt dimers into monomers, and oxidizing conditions cannot completely transform monomers into dimers; thus, dimer formation does not depend on Cys.

strongly influenced through ABA (Graether and Boddington, 2014), i.e., the level of 44 kDa protein synthesis depends on drying and ABA (Kalemba et al., 2009). These results support the findings that dehydrin mRNA is synthesized during seed maturation (Lin et al., 2012) in response to maturation drying (Kleinw€ achter et al., 2014) and high ABA concentrations (Kalemba et al., 2009). An increased correlation between the transcript (Fig. 2C) and protein levels (Fig. 1D) and a tendency for dimer formation (Fig. 5) suggested that the 44 kDa dehydrin protein is a putative dimer of dehydrin/response ABA proteins migrating as 26-kDa proteins. Recently, natural antisense dehydrins transcripts were reported (Vaseva and Feller, 2013), indicating that the regulation of dehydrin expression is more complex.

4. Discussion

4.2. Possible roles of PTMs in determining dehydrin function

4.1. Dehydrin expression in relation to maturation drying

Five spots of 44-kDa dehydrin and three forms of 40-kDa dehydrin-like protein differing in pI have been reported (Fig. 1C). The shift in pI was presumably influenced through PTMs, including phosphorylation (Kalemba et al., 2015). The phosphorylation status of dehydrins differs under various environmental conditions (Jensen et al., 1998). Several dephosphorylation states have been reported in dehydrins and dehydrin-like proteins expressed in the embryonic axes of developing beech seeds (Kalemba et al., 2015), indicating that phosphorylation strongly affects the gel mobility of dehydrins and dehydrin-like proteins, and several phosphorylation sites exist in these proteins. Phosphorylation is required for dehydrin function, nuclear targeting and ion and membrane binding (Brini et al., 2007; Jensen et al., 1998; Eriksson et al., 2011). Bioinformatics predictions showed that 11 sites might be phosphorylated with high potential in the dehydrin/response ABA protein (Fig. 3, Suppl. Fig. 2D) and therefore might contribute to protein function or cooperate with other PTMs. Beyond phosphorylation, the PTMs of dehydrin-like proteins are poorly understood, primarily reflecting the difficulties in detection. Myristate moieties modulate reversible protein-membrane interactions (Traverso et al., 2008). Three specific myristoylation sites predicted in the dehydrin/response ABA protein (Fig. 3) would facilitate the association of these proteins with small vesicles located near the cell membrane, small membrane structures throughout the cytoplasm, and the mitochondrial envelope (Kalemba et al., 2015). Extensive myristoylation might be significant in protection of membranes by cold-induced dehydrins

The accumulation of 26 and 44 kDa dehydrin proteins in embryonic axes of maturing beech seeds has previously been documented (Kalemba et al., 2009, 2015), and the accumulation of 44kDa dehydrin has been associated with the acquisition of desiccation tolerance (Kalemba et al., 2009). Here, we showed that the level of both dehydrins is strongly associated with the desiccation levels in seeds (Fig. 1D), suggesting that dehydrin accumulation is a consequence of maturation drying (Kleinw€ achter et al., 2014). The gel migration of dehydrins and dehydrin-like proteins in SDS-PAGE is abnormal, showing up to 1.8 times higher experimental molecular masses than those calculated from the sequence (Receveurchot et al., 2006). Under native conditions, gel migration Bre might yield similar symptoms (Fig. 1B), further suggesting the formation of oligomeric structures that migrate slowly. Monomer, ndez-Sa nchez dimer, and trimer forms (Goyal et al., 2003; Herna ndez-Sa nchez et al., 2014) and large multimeric complexes (Herna et al., 2014) of LEA proteins have been experimentally detected. According to GO terms, responses to water and stress are the main biological processes involving dehydrin genes. Experimental data suggest that dehydrin expression is associated with water stress (Shih et al., 2012; Graether and Boddington, 2014; €chter et al., 2014). The increased synthesis of the dehyKleinwa drin/response ABA mRNA was detected beginning from the 14th WAF (Fig. 2A and B), coincident with the intensification of protein synthesis (Fig. 1A). In dehydrins, the level of mRNA and protein is

252

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

Fig. 5. The diagonal 2D redox SDS-PAGE method illustrating the separation of thiol-disulfide transition proteins with intra- or intermolecular disulfide bonds was used to determine whether disulfide bonds might be formed in the 26-kDa dehydrin. Protein extracts from embryonic axes of mature beech (Fagus sylvatica L.) collected at the 19th WAF were used. Prior to first separation proteins were treated with 100 mM DTT to introduce reducing conditions or 10 mM H2O2 to introduce oxidative conditions. Then thiol reshuffling was prevented by alkylation with iodoacetamide. Migration differences dependent on thiol-disulfide switches are observed only after protein oxidation before first separation and are €her and Dietz (2008). Protein spots localized below and above diagonal line correspond to proteins with intermolecular and visualized on the scheme (1) prepared according to Stro intramolecular disulfide bridges, respectively. The representative gels stained with Coomassie Brilliant Blue obtained after reduction and oxidation of protein samples before diagonal 2D redox SDS-PAGE separation (2). First (non-reducing) and second (reducing) separations are indicated with arrows and the 1st D and 2nd D, respectively. Western blot analyses were performed to visualize dehydrins and dehydrin-like proteins in the pool of all proteins separated by diagonal 2D redox SDS-PAGE (3). “A” corresponds to protein that migrated as 44 kDa protein in the first and the second separation, “B” corresponds to protein that migrated as 26 kDa protein in the first and the second separation, “C” corresponds to protein that migrated as 26 kDa protein in the first separation and as a 44 kDa protein in the second separation.

(Suppl. Tab. 1). Among all protein lipidation modifications assumed as important in plant developmental processes, palmitoylation is a reversible modification. Palmitoylation alters the subcellular locations, proteineprotein interactions, binding capacities and catalytic activities of proteins (Running, 2014). Palmitoylated Cys residues are commonly identified in close proximity to myristoylated Gly residues. In the model of PTM contribution to dehydrin/response ABA protein proposed in the preset study, the space between these two PTMs comprises 21 amino acids (Fig. 3), making both modifications more likely. The K-segment is involved in the binding or

interactions with membranes (Koag et al., 2009; Kalemba et al., 2015), and the myristoylation of dehydrin/response ABA protein in the region of the two K-segments (Fig. 3) might as well assist in these interactions. Amides enhance the water binding and storage ability of dehydrins (Tompa et al., 2006). Sumoylation alters protein function or activity, subcellular location, and protein interaction (Park et al., 2011). The sumoylation of dehydrin/response ABA protein at five potential sites (Fig. 3) might facilitate the translocation of this protein to the nucleus or might be associated with the localization

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

(Kalemba et al., 2015). However, the sumoylation seems to be more important in Kn class dehydrins (Suppl. Tab. 1). Secreted proteins predominantly undergo sulfation, a PTM essential for strengthening proteineprotein interactions and regulating metabolic enzymes and plant stress responses (Yang et al., 2015). The sulfation of dehydrins might determine their final destination in vacuoles, as a possible endpoint of the secretory pathway. The O-linked glycosylation of dehydrin was reported in blueberry (Levi et al., 1999) and pistachio (Golan-Goldhirsh, 1998). Nglycosylation and O-glycosylation sites were predicted in the dehydrin/response ABA protein (Fig. 3). In a recent study, Saibi et al. (2015) used several experimental techniques but did not detect the glycosylation of the dehydrin protein. However, glycosylation might modify protein folding, interaction, stability, and mobility, thereby facilitating dehydrin utility. Particularly, YnKn class dehydrin proteins that contain notably more O-glycosylation sites (Suppl. Tab. 1). Beyond phosphorylation nitrosylation, sumoylation (Park et al., 2011), myristoylation (Traverso et al., 2008), and acetylation have all been implicated in cell signaling, suggesting that additional studies of dehydrins are needed. 4.3. Dehydrin protein structures The formation of dimeric (Fig. 5) and oligomeric (Fig. 1B) dehydrin/response ABA protein structures was experimentally demonstrated, however disulfide bonds are likely not involved in dimer arrangement (Fig. 5). K-segments and His-rich regions are crucial to the formation of dehydrin dimers in Opuntia streptacannchez et al., 2014). The His content tha (OpsDHN1) (Hern andez-Sa in dehydrin/response to ABA protein is 6.55%, similar to that in OpsDHN1 (7.25%). Interestingly, the dehydrin/response ABA protein contains a poly-glutamine (Q) tract (Fig. 3). PolyQ-tracts are characteristics of regulatory proteins, particularly TFs. Recent studies have demonstrated that polyQ-containing TFs recruit other TFs with polyQ-tracts (Atanesyan et al., 2012), suggesting that polyQtract might be involved in dimer and oligomer structure formation in the dehydrin/response ABA protein. Moreover, the polyQtract enhances TF activity (Atanesyan et al., 2012). The role of polyQ-tract in dehydrin proteins has received much less attention compared with K-, S-,Y-, and F-segments and His-rich regions. Based on the results of a recent study (Atanesyan et al., 2012), the polyQ-tract might be a good candidate as a modulator of the spatial organization and function of dehydrin-like proteins. Disulfide bond formation was predicted in dehydrin/response ABA proteins (Suppl. Fig. 4D). Disulfide bonds are likely intramolecular bonds, suggesting interactions with eSH-containing proteins rather than homodimer formation. Disulfide bonds are considered as a protein-stabilization method (Beeby et al., 2005). Therefore, the incorporation of disulfide bonds might be helpful for the reinforcement of larger oligomeric structures of dehydrins and dehydrin-like proteins. Dehydrin/response ABA protein is disordered (Fig. 4). Chaperone activity has been suggested as a general function of IUPs. The protective effects of dehydrins, enabling plant survival under stressful conditions (Hara, 2010; Battaglia and Covarrubias, 2013; Graether and Boddington, 2014), suggested that these proteins might also act as chaperones. Nevertheless, the specific functions and mechanisms remain unknown for many proteins of the LEA chot et al., family. IUPs are easy targets of proteolysis (Receveur-Bre 2006). Dehydrin/response ABA protein might be targeted to proteasomal degradation in a sumoylation- or ubiquitinationdependent manner because these two PTMs engage at identical or closely located K residues in the protein sequence (Fig. 3). Low complexity regions were detected in the central region and C-

253

terminus of the dehydrin/response ABA protein sequence (Suppl. Fig. 4A). Coletta et al. (2010) reported that proteins containing terminal low-complexity regions are important for stress responses and translation and transport processes and are enriched in protein complexes, whereas proteins containing central lowcomplexity regions are important in transcription and transcription regulation. Thus, the multifunctionality of dehydrin proteins is more persuasive. 5. Conclusion Dehydrin/response ABA protein lacks a fixed structure. Disordered arrangement enhances accessibility to 40 potential sites of PTMs, which might modulate dehydrin function. Dehydrin/ response ABA protein has oligomer-forming properties; thus, the 44-kDa protein is likely a homodimer comprising two monomeric forms of dehydrin/response ABA protein (26 kDa protein). Dimer formation might be favored through His residues and polyQ-tracts, whereas disulfide bonds might contribute to the stability of larger dehydrin oligomers. The formation of disulfide bonds requires certain conditions interfered through potential PTMs on Cys residues, including S-nitrosylation and/or palmitoylation. The detection of five protein spots of 44 kDa dehydrin differing in pI suggests that PTMs indicate potential switches in dehydrin function, assisting in the modulation of subcellular localization, membrane targeting, cell signaling, and particularly the induction/support/ enhancement of interactions with target molecules. Conflicts of interest The authors declare no conflict of interest. Author contributions The conception and design of the study EMK, acquisition of data EMK and ML, analysis and interpretation of data EMK and ML, writing the article EMK. All authors approved the final version of submitted manuscript. Acknowledgments The antiserum raised against the K segment was a kind gift from T.J. Close (Department of Botany and Plant Sciences, University of California, Riverside, CA, USA). This work was financially supported through a grant from The Ministry of Science and Higher Education (Poland)/National Centre of Sciences (Poland) (grant no. N N309 136535). Additional funding was provided by the Institute of Dendrology of the Polish Academy of sciences. 2D-electrophoresis was performed in the Proteomics Laboratory using devices purchased with funds provided through a grant from The Ministry of Science and Higher Education (Poland) (grant no. PO4G 075 25). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.plaphy.2015.10.011. References Arc, E., Galland, M., Cueff, G., Godin, B., Lounifi, I., Job, D., Rajjou, L., 2011. Reboot the system thanks to protein post-translational modifications and proteome diversity: how quiescent seeds restart their metabolism to prepare seedling establishment. Proteomics 11, 1606e1618. http://dx.doi.org/10.1002/ pmic.201000641. Atanesyan, L., Günther, V., Dichtl, B., Georgiev, O., Schaffner, W., 2012. Polyglutamine tracts as modulators of transcriptional activation from yeast to

254

E.M. Kalemba, M. Litkowiec / Plant Physiology and Biochemistry 97 (2015) 246e254

mammals. Biol. Chem. 393 (1e2), 63e70. http://dx.doi.org/10.1515/BC-2011252. Battaglia, M., Covarrubias, A.A., 2013. Late Embryogenesis Abundant (LEA) proteins in legumes. Front. Plant Sci. 4, 190. http://dx.doi.org/10.3389/fpls.2013.00190. eCollection 2013. Beeby, M., O'Connor, B.D., Ryttersgaard, C., Boutz, D.R., Perry, L.J., Yeates, T.O., 2005. The genomics of disulfide bonding and protein stabilization in thermophiles. PLoS Biol. 3 (9), e309. http://dx.doi.org/10.1371/journal.pbio.0030309. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities utilizing the principle of protein dye binding. Anal. Biochem. 72, 248e254. Brini, F., Hanin, M., Lumbreras, V., Irar, S., Pages, M., Masmoudi, K., 2007. Functional characterization of DHN-5, a dehydrin showing a differential phosphorylation pattern in two Tunisian durum wheat (Triticum durum Desf.) varieties with marked differences in salt and drought tolerance. Plant Sci. 172 (1), 20e28. http://dx.doi.org/10.1016/j.plantsci.2006.07.011. Close, T.J., Fenton, R.D., Moonan, F., 1993. A view of plant dehydrins using antibodies specific to the carboxy terminal peptide. Plant Mol. Biol. 23 (2), 279e286. http://dx.doi.org/10.1007/BF00029004. Coletta, A., Pinney, J.W., Solís, D.Y., Marsh, J., Pettifer, S.R., Attwood, T.K., 2010. Lowcomplexity regions within protein sequences have position-dependent roles. BMC Syst. Biol. 4, 43. http://dx.doi.org/10.1186/1752-0509-4-43. Danyluk, J., Perron, A., Houde, M., Limin, A., Fowler, B., Benhamou, N., Sarhan, F., 1998. Accumulation of an acidic dehydrin in the vicinity of the plasma membrane during cold acclimation of wheat. Plant Cell 10, 623e638. http://dx.doi. org/10.1105/tpc.10.4.623. € bner, G., Harryson, P., 2011. Tunable memEriksson, S.K., Kutzer, M., Procek, J., Gro brane binding of the intrinsically disordered dehydrin Lti30, a cold-induced plant stress protein. Plant Cell 23 (6), 2391e2404. http://dx.doi.org/10.1105/ tpc.111.085183. Golan-Goldhirsh, A., 1998. Developmental proteins of Pistacia vera L. bark and bud and their biotechnological properties: a review. J. Food Biochem. 22 (5), 375e382. http://dx.doi.org/10.1111/j.1745-4514.1998.tb00251.x. Goyal, K., Tisi, L., Basran, A., Browne, J., Burnell, A., Zurdo, J., Tunnacliffe, A., 2003. Transition from natively unfolded to folded state induced by desiccation in an anhydrobiotic nematode protein. J. Biol. Chem. 278 (15), 12977e12984. http:// dx.doi.org/10.1074/jbc.M212007200. Graether, S.P., Boddington, K.F., 2014. Disorder and function: a review of the dehydrin protein family. Front. Plant Sci. 5, 576. http://dx.doi.org/10.3389/ fpls.2014.00576. Hara, M., 2010. The multifunctionality of dehydrins, an overview. Plant Signal. Behav. 5, 503e508. http://dx.doi.org/10.4161/psb.11085. ndez-Sa nchez, I.E., Martynowicz, D.M., Rodríguez-Hern rezHerna andez, A.A., Pe nez-Bremont, J.F., 2014. A dehydrin-dehydrin Morales, M.B., Graether, S.P., Jime interaction: the case of SK3 from Opuntia streptacantha. Front. Plant Sci. 5, 520. http://dx.doi.org/10.3389/fpls.2014.00520. Jensen, A.B., Goday, A., Figueras, M., Jessop, A.C., Pages, M., 1998. Phosphorylation mediates the nuclear targeting of the maize Rab17 protein. Plant J. 13, 691e697. http://dx.doi.org/10.1046/j.1365-313X.1998.00069.x. nez, J.A., Alonso-Ramírez, A., Nicol Jime as, C., 2008. Two cDNA clones (FsDhn1 and FsClo1) up-regulated by ABA are involved in drought responses in Fagus sylvatica L. seeds. J. Plant Physiol. 165 (17), 1798e1807. http://dx.doi.org/10.1016/ j.jplph.2007.11.013. Kalemba, E.M., Janowiak, F., Pukacka, S., 2009. Desiccation tolerance acquisition in developing beech (Fagus sylvatica L.) seeds: the contribution of dehydrin-like protein. Trees 23, 305e315. http://dx.doi.org/10.1007/s00468-008-0278-8. Kalemba, E.M., Pukacka, S., 2014. Carbonylated proteins accumulated as vitality decreases during long-term storage of beech (Fagus sylvatica L.) seeds. Trees 28 (2), 503e515. http://dx.doi.org/10.1007/s00468-013-0967-9. Kalemba, E.M., Bagniewska-Zadworna, A., Ratajczak, E., 2015. Multiple subcellular localizations of dehydrin-like proteins in the embryonic axes of common beech (Fagus sylvatica L.) seeds during maturation and dry storage. J. Plant Growth

Regul. 34 (1), 137e149. http://dx.doi.org/10.1007/s00344-014-9451-z. Khoury, G.A., Baliban, R.C., Floudas, C.A., 2011. Proteome-wide post-translational modification statistics: frequency analysis and curation of the swiss-prot database. Sci. Rep. 1, 90. http://dx.doi.org/10.1038/srep00090. €chter, M., Radwan, A., Hara, M., Selmar, D., 2014. Dehydrin expression in Kleinwa seeds: an issue of maturation drying. Front. Plant Sci. 5, 402. http://dx.doi.org/ 10.3389/fpls.2014.00402. Koag, M.C., Wilkens, S., Fenton, R.D., Resnik, J., Vo, E., Close, T.J., 2009. The Ksegment of maize DHN1 mediates binding to anionic phospholipid vesicles and concomitant structural changes. Plant Physiol. 150, 1503e1514. http:// dx.doi.org/10.1104/pp.109.136697. Kovacs, D., Kalmar, E., Torok, Z., Tompa, P., 2008. Chaperone activity of ERD10 and ERD14, two disordered stress related plant proteins. Plant Physiol. 147, 381e390. http://dx.doi.org/10.1104/pp.108.118208. Kozlowski, L.P., Bujnicki, J.M., 2012. MetaDisorder: a meta-server for the prediction of intrinsic disorder in proteins. BMC Bioinform. 13, 111. http://dx.doi.org/ 10.1186/1471-2105-13-111. Levi, A., Panta, G.R., Parmentier, C.M., Muthalif, M.M., Arora, R., Shanker, S., Rowland, L.J., 1999. Complementary DNA cloning, sequencing, and expression of a unique dehydrin from blueberry floral buds. Physiol. Plant 107, 98e109. http://dx.doi.org/10.1034/j.1399-3054.1999.100114.x. Lin, C.H., Peng, P.H., Ko, C.Y., Markhart, A.H., Lin, T.Y., 2012. Characterization of a novel Y2K-type dehydrin VrDhn1 from Vigna radiata. Plant Cell Physiol. 53 (5), 930e942. http://dx.doi.org/10.1093/pcp/pcs040. Park, H.J., Kim, W.Y., Park, H.C., Lee, S.Y., Bohnert, H.J., Yun, D.J., 2011. SUMO and SUMOylation in plants. Mol. Cells 32 (4), 305e316. http://dx.doi.org/10.1007/ s10059-011-0122-7. chot, V., Bourhis, J.M., Uversky, V.N., Canard, B., Longhi, S., 2006. Receveur-Bre Assessing protein disorder and induced folding. Proteins 62 (1), 24e45. http:// dx.doi.org/10.1002/prot.20750. Running, M.P., 2014. The role of lipid post-translational modification in plant developmental processes. Front. Plant Sci. 5, 50. http://dx.doi.org/10.3389/ fpls.2014.00050. Saibi, W., Drira, M., Yacoubi, I., Feki, K., Brini, F., 2015. Empiric, structural and in silico findings give birth to plausible explanations for the multifunctionality of the wheat dehydrin (DHN-5). Acta Phys. Plant 37 (3), 1e8. http://dx.doi.org/ 10.1007/s11738-015-1798-7. Shih, M.D., Hsieh, T.Y., Jian, W.T., Wu, M.T., Yang, S.J., Hoekstra, F.A., Hsing, Y.I., 2012. Functional studies of soybean (Glycine max L.) seed LEA proteins GmPM6, GmPM11, and GmPM30 by CD and FTIR spectroscopy. Plant Sci. 196, 152e159. http://dx.doi.org/10.1016/j.plantsci.2012.07.012. € lzer, M., Rabilloud, T., 2001. A new silver staining appaSinha, P., Poland, J., Schno ratus and procedure for matrix-assisted laser desorption/ionization-time of flight analysis of proteins after two-dimensional electrophoresis. Proteomics 1 (7), 835e840. http://dx.doi.org/10.1002/1615-9861(200107)1:7<835::AIDPROT835>3.0.CO;2e2. €her, E., Dietz, K.J., 2008. The dynamic thiol-disulfide redox proteome of the Stro Arabidopsis thaliana chloroplast as revealed by differential electrophoretic mobility. Physiol. Plant 133, 566e583. http://dx.doi.org/10.1111/j.13993054.2008.01103.x. nki, P., Bokor, M., Kamasa, P., Kov Tompa, P., Ba acs, D., Lasanda, G., Tompa, K., 2006. Protein-water and protein-buffer interactions in the aqueous solution of an intrinsically unstructured plant dehydrin: NMR intensity and DSC aspects. Biophys. J. 91 (6), 2243e2249. http://dx.doi.org/10.1529/biophysj.106.084723. Traverso, J.A., Meinnel, T., Giglione, C., 2008. Expanded impact of protein N-myristoylation in plants. Plant Signal. Behav. 3 (7), 501e502. http://dx.doi.org/ 10.4161/psb.3.7.6039. Vaseva, I.I., Feller, U., 2013. Natural antisense transcripts of Trifolium repens dehydrins. Plant Signal. Behav. 8 (12), e27674. http://dx.doi.org/10.4161/psb.27674. Yang, Y.S., Wang, C.C., Chen, B.H., Hou, Y.H., Hung, K.S., Mao, Y.C., 2015. Tyrosine sulfation as a protein post-translational modification. Molecules 20 (2), 2138e2164. http://dx.doi.org/10.3390/molecules20022138.