Genes encoding 4-Cys antimicrobial peptides in wheat Triticum kiharae Dorof. et Migush.: multimodular structural organization, instraspecific variability, distribution and role in defence



A novel family of antifungal peptides was discovered in the wheat Triticum kiharae Dorof. et Migusch. Two members of the family, designated Tk-AMP-X1 and Tk-AMP-X2, were completely sequenced and shown to belong to the α-hairpinin structural family of plant peptides with a characteristic C1XXXC2-X(n)-C3XXXC4 motif. The peptides inhibit the spore germination of several fungal pathogens in vitro. cDNA and gene cloning disclosed unique structure of genes encoding Tk-AMP-X peptides. They code for precursor proteins of unusual multimodular structure, consisting of a signal peptide, several α-hairpinin (4-Cys) peptide domains with a characteristic cysteine pattern separated by linkers and a C-terminal prodomain. Three types of precursor proteins, with five, six or seven 4-Cys peptide modules, were found in wheat. Among the predicted family members, several peptides previously isolated from T. kiharae seeds were identified. Genes encoding Tk-AMP-X precursors have no introns in the protein-coding regions and are upregulated by fungal pathogens and abiotic stress, providing conclusive evidence for their role in stress response. A combined PCR-based and bioinformatics approach was used to search for related genes in the plant kingdom. Homologous genes differing in the number of peptide modules were discovered in phylogenetically-related Triticum and Aegilops species, including polyploid wheat genome donors. Association of the Tk-AMP-X genes with A, B/G or D genomes of hexaploid wheat was demonstrated. Furthermore, Tk-AMP-X-related sequences were shown to be widespread in the Poaceae family among economically important crops, such as barley, rice and maize.


Nucleotide sequence data have been deposited in the EMBL database under accession numbers: HF562347, HF562348, HF562349, HF562350, HF562351, HF562352, HF562353, HF562354, HF562355, HF562356, HF562357, HF562358, HF562359, HF562360, HF562361, HF562362, HF562363 and HF562364.


antimicrobial peptide


trifluoroacetic acid


Antimicrobial peptides (AMPs) play a crucial role in the innate immune system of all living organisms, including plants [1-4]. In recent decades, they have been the focus of intensive research as a result of the development of antibiotic resistance and a growing need for highly efficient genes for disease control in crops [2, 5-8]. Unexpected abundance of AMP genes (approximately 2–3% of the protein-coding genes) in plant genomes was disclosed by bioinformatics tools [9, 10], which reflects long-term plant-pathogen coevolution and highlights the role of AMPs in defence and possibly other vital functions. Exploiting diverse mechanisms, AMPs impede the growth of pathogenic microorganisms by disrupting microbial plasma membranes and/or acting on intracellular targets [11-13]. Despite considerable variation in amino acid sequences and three-dimensional structures, AMPs share similar properties, such as an amphiphilic nature and a positive charge of the molecule, facilitating their interaction with microbial plasma membranes. The vast majority of plant AMPs belong to cysteine-rich peptides with three or four disulfide bridges. On the basis of structural similarity and the so-called cysteine motifs (i.e. the arrangement of cysteine residues in the polypeptide chain), AMPs are classified into several families: thionins, defensins, hevein- and knottin-like peptides, and nonspecific lipid-transfer proteins. Further research into AMPs from plants revealed novel molecular types that do not fall into the above-mentioned families. Although dozens of peptides have been isolated in recent years, gene structure analysis lags behind biochemical studies and microbiological assays.

During comprehensive analysis of AMPs expressed in seeds of the wheat Triticum kiharae Dorof. et Migush., we previously isolated and characterized 24 novel AMPs, which were assigned to known as well as new plant AMP families [14]. Among them, two short (approximately 30 amino acid residues) peptides, named Tk-AMP-X1 and Tk-AMP-X2, with a peculiar 4-Cys motif were partially sequenced. In the present study, we report the complete amino acid sequences of these peptides, as well as their secondary structure and biological activity. We show that they belong to a newly acknowledged α-hairpinin family of plant AMPs. Details of the peculiar structure of cDNA encoding Tk-AMP-X peptides are presented. We demonstrate that these peptides are synthesized as precursor proteins of unique modular structure comprising several peptide domains with the same 4-Cys motif. A whole family of 4-Cys peptides was predicted from T. kiharae cDNA and genomic DNA sequences. Some family members were previously isolated from seeds. In the present study, we report their complete amino acid sequences. Furthermore, we demonstrate the widespread distribution of homologous genes among cereals and provide evidence for their role in response to biotic and abiotic stress.


Isolation and characterization of Tk-AMP-X1 and Tk-AMP-X2 peptides: amino acid sequence, secondary structure and biological activity

Peptides Tk-AMP-X1 and Tk-AMP-X2 were isolated from T. kiharae seeds using a previously developed multidimensional chromatographic procedure combining different types of HPLC (affinity, size-exclusion and RP) [14] (Fig. 1). Nine peak fractions were collected. Peak fractions designated 2 and 3 were re-chromatographed on the same RP column using a shallower acetonitrile gradient (not shown). Two peptides were purified from each peak. The peptides from peak fraction 2, designated as Tk-AMP-X1 and Tk-AMP-X2, had monoisotopic molecular masses of 3857.6 and 3518.3 Da, respectively. The yield of peptides was 5 and 7 μg·g−1 flour for Tk-AMP-X1 and Tk-AMP-X2, respectively. The peptides of peak fraction 3, named Tk-AMP-G6 and Tk-AMP-G7, had monoisotopic molecular masses of 4360.2 and 4603.1 Da, respectively, and similar Gly- and His-rich N-terminal amino acid sequences, as reported previously [14]. Peptides Tk-AMP-X1 and Tk-AMP-X2 were studied in more detail. Comparison of molecular masses of Tk-AMP-X peptides before and after reduction followed by alkylation indicated that both peptides possess four cysteine residues (not shown). Alkylation without prior reduction of both Tk-AMP-X peptides did not lead to molecular mass changes, indicating the absence of free thiol groups. Thus, Tk-AMP-X peptides possess four cysteine residues each, and all of them are involved in the formation of disulfide bridges.

Figure 1.

Purification of Tk-AMP-X peptides by RP-HPLC on a Vydac C18 column in a 60-min linear acetonitrile gradient (10–40% B, B: 80% acetonitrile in 0.1% TFA). Fraction 2 containing Tk-AMP-X1 and Tk-AMP-X2 peptides, as well as fraction 3 containing Tk-AMP-G6 and Tk-AMP-G7 peptides, are marked with arrows. For details, see 'Results'.

Amino acid sequences of the reduced and alkylated Tk-AMP-X1 and Tk-AMP-X2 peptides were determined by automated Edman degradation (Fig. 2). Tk-AMP-X1 peptide consists of 31 amino acid residues, whereas Tk-AMP-X2 is by three residues shorter (28 amino acid residues). Both peptides are basic and highly similar, differing by only two amino acid substitutions. The predicted pI of the peptides is approximately 9 for Tk-AMP-X1 and 10.5 for Tk-AMP-X2, respectively, and the corresponding charges at pH 7.0 are +1 and +4. The calculated molecular masses of the peptides (3861.3 Da for Tk-AMP-X1 and 3519.0 Da for Tk-AMP-X2) agree well with the measured values, providing evidence for the absence of other post-translational modifications except disulfide bonds. A search in the public protein databases yielded no characterized sequences with significant similarity to Tk-AMP-X peptides, indicating that they are new plant peptides.

Figure 2.

Amino acid sequence alignment of Tk-AMP-X peptides with selected α-hairpinins. Accession numbers of the presented sequences in the Uniprot database are: MBP-1 (AAB23306), MiAMP2c and MiAMP2d (Q9SPL4), VhTI (P85981), EcAMP1 (P86698), BWI-2c (P86794). Conserved cysteine residues are indicated by inverted text. The cysteine motif is included at the bottom of the alignment.

Disulfide connectivity was studied in Tk-AMP-X2. The peptide was cleaved at Met residues by cyanogen bromide, and the resulting mixture was analyzed by MS. Two fragments were identified with molecular masses of 1628 and 1831 Da. The latter corresponds to the middle part of the peptide (residues 9–22) containing the ‘inner’ Cys9-Cys21 disulfide bond. The former corresponds to the two flanking parts (residues 1–8 and 23–28) held together by the ‘outer’ Cys5-Cys25 disulfide bond. Other variants of S-S connectivities were ruled out because they would not allow separation of CNBr fragments. Note that methionine residues were converted to homoserine lactone residues.

The secondary structure of Tk-AMP-X1 and Tk-AMP-X2 peptides was studied by CD (Fig. S1). Calculation of the secondary structure content showed: α-helix = 59.9% and 46.6%; β-structure = 9.7% and 17.7%; and random coil = 30% and 35% for Tk-AMP-X1 and Tk-AMP-X2, respectively. Thus, these wheat peptides belong to α-helical peptides.

The peptides Tk-AMP-X1 and Tk-AMP-X2 possess four cysteine residues arranged in an unusual manner: the first two (Cys5 and Cys9), as well as the last two (Cys21 and Cys25), are separated by three amino acid residues that are not cysteines. The cysteine motif of these peptides may therefore be presented as: C1XXXC2-X(n)-C3XXXC4. The same cysteine pattern is shared by AMPs from two other cereals, maize (MBP-1) [15] and barnyard grass Echinochloa crus-galli (EcAMP1) [16], a dicotyledonous plant the Queensland nut Macadamia integrifolia (MiAMP2) [17], trypsin inhibitors from the ivy-leaved speedwell Veronica hederifolia (VhTI) [18] and buckwheat Fagopyrum esculentum (BWI-2b and BWI-2c) [19, 20]. Moreover, we established the disulfide connectivity in Tk-AMP-X2, which was found to be identical to that in EcAMP1, VhTI, BWI-2b and BWI-2c. Despite the identical cysteine motif, sequence similarity between the members of this peptide family named α-hairpinins is very low (Fig. 2).

The biological activity of the Tk-AMP-X1 and Tk-AMP-X2 peptides against several fungal pathogens was assayed by measuring inhibition of spore germination in liquid nutrient medium. The results are shown in Table 1. The peptides were active against both Fusarium species tested and Diplodia maydis. The fungus Colletotrichum graminicola was not inhibited by the peptides at concentrations below 30 μg·mL−1. As shown in Table 1, the antifungal activity of both peptides is similar, with the Tk-AMP-X2 peptide being more active against two of the fungi tested. This indicates that even minor sequence variations have a detectable impact on the inhibitory activity of the Tk-AMP-X peptides.

Table 1. Antifungal activity of Tk-AMP-X1 and Tk-AMP-X2 in comparison with EcAMP1. Inhibition of spore germination is expressed in IC50 (μg·mL−1) values
FungiTk-AMP-X1Tk-AMP-X2EcAMP1 [16]
Fusarium graminearum 7.57.519.3
Fusarium verticillioides
Diplodia maydis 30.017.0> 43.0
Colletotrichum graminicola > 30.0> 30.0> 43.0

Tk-AMP-X1 and Tk-AMP-X2 peptides are derived from multimodular precursor proteins

The cDNA sequence encoding Tk-AMP-X peptides was determined by a combination of 3′- and 5′-RACE. Schematic representation of the cloning procedure is shown in Fig. S2 and primers are listed in Table S1. As a result, two fragments of approximately 850 and 1100 bp were amplified. They represented the full-length cDNA sequences encoding Tk-AMP-X precursors. Precursors of Tk-AMP-X1 and Tk-AMP-X2 peptides are multimodular and consist of a signal peptide of 25 amino acid residues, several peptide domains separated by linkers, and a C-terminal prodomain (Fig. 3A). Despite overall structural similarity, the Tk-AMP-X precursors could be divided into two classes according to the number of peptide domains. Precursors of the first class (247–266 amino acid residues in length), named ‘short’ (S) precursors, contain five α-hairpinin (4-Cys) modules (Fig. 4A). Precursors of the second class (346–362 residues in length), designated ‘long’ (L) precursors, consist of seven 4-Cys peptide modules (Fig. 4C).

Figure 3.

Structure and processing of Tk-AMP-X precursors (A) Scheme of T. kiharae Tk-AMP-X precursor structure. Peptide domains with a characteristic 4-Cys motif are numbered from 1 to 7. (B) Putative processing of the L-2 precursor. The signal peptide is boxed. Predicted and isolated 4-Cys peptides are shown in bold. Peptides Tk-AMP-X2, Tk-AMP-G7 and Tk-AMP-G6 isolated from seeds are underlined (partial sequences for Tk-AMP-G7 and Tk-AMP-G6). Cysteines are shaded in black. Asparagine residues are shaded in grey. Negatively-charged amino acid residues, which are important for precursor processing, are shown in italics. A line with arrows at both ends shows the EtoR motif. ↓, cleavage site; ←, cleavage site for carboxypeptidases; ▼, alternative cleavage site; ◄, cleavage of a glycine residue from the COOH-terminus.

Figure 4.

Sequence alignment of deduced precursors: (A) short, (B) medium, (C) long. Signal peptides are boxed. Conserved cysteine residues are indicated by inverted text to show the characteristic motif. Predicted peptides are shown in bold. In short precursors (A), peptides isolated from T. kiharae seeds in the order they appear in the precursor, Tk-AMP-X3 [14] and Tk-AMP-X1 are underlined. In medium precursors (B), underlined sequences correspond to Tk-AMP-X3-like peptide, Tk-AMP-X2 and Tk-AMP-G7-like peptide. In long precursors (C), underlined sequences correspond to Tk-AMP-X3-like peptide and Tk-AMP-X2, Tk-AMP-G7 and Tk-AMP-G6. Only partial sequences are underlined for Tk-AMP-X3, Tk-AMP-G7 and Tk-AMP-G6.

Among short precursors, four homologous sequences designated from S-1 to S-4 were distinguished. Sequence alignment of short precursors is given in Fig. 4A. All of them contain five peptide modules with a typical 4-Cys motif. Of particular interest is the fourth peptide with an additional fifth cysteine residue. We assume that, in plants, such 5-Cys peptides may form dimers. In general, S-type precursors differ by single substititutions and insertions/deletions. The second peptide domain exists in two variants with a more pronounced sequence variation: variant 1 is found in S-1 and S-2 precursors, and variant 2 is found in S-3 and S-4 precursors. The Tk-AMP-X1 peptide is produced by processing of both S-1 and S-2 precursors. S-3 and S-4 precursors carry two highly similar peptides with Gly instead of Thr at the NH2-terminus and a single additional amino acid substitution in S-4. Among the constituent peptides of the short precursors, we identified the peptide Tk-AMP-X3 that was isolated earlier from T. kiharae seeds [14].

Sequences of long precursors were less variable (Fig. 4C). Two main prepropeptides named L-1 and L-2 were identified. Similarly to short precursors, the fourth peptide in long precursors has an additional cysteine residue. Sequence identity between the two long precursors amounts to 83%. It should be noted that the fifth and the sixth peptides in L-1 and L-2 precursor proteins are highly similar and correspond to the earlier isolated peptides: Tk-AMP-G7 and Tk-AMP-G6, respectively [14]. The Tk-AMP-X2 peptide is derived from the L-2 precursor. The sequences of peptide domains 4, 5, 6 and 7 coincide in both L precursors. Peptide domains 1 and 3 differ by single amino acid substitutions and peptide domain 2 is the most variable part of the precursor as in short precursors.

Prediction of proteolytic processing sites in Tk-AMP-X precursor proteins

The issue of putative processing sites in Tk-AMP-X precursors deserves special attention. It should be noted that the processing machinery of plant precursor proteins has received little attention. Recent studies have implicated the role of subtilases in the processing of signalling peptides. They recognize dibasic residues upstream of the mature peptide sequences (positions P2–P4). Thus, subtilase-mediated processing of the phytosulfokine PSK4 propeptide was reported [21]. Dibasic subtilase recognition sites were also discovered in precursors of rapid alkalinization factors involved in plant development [22]. Other processing enzymes were shown to recognize sites composed of two or more negatively-charged residues, as described for Impatiens balsamina Ib-AMPs precursors [23]. A third group of processing enzymes cleave at an Asn residue (position P1) [24, 25]. An additional role of carboxypeptidases in the processing of some plant hormones has been suggested [26].

For those peptides that were directly isolated from T. kiharae seeds (Tk-AMP-X1, Tk-AMP-X2, Tk-AMP-X3, Tk-AMP-G6 and Tk-AMP-G7), the cleavage sites could be predicted reliably. Other sites are putative, and they were suggested as a result of the cleavage specificities of known enzymes, with most data originating from animals. We used the comprehensive MEROPS database [27] to identify target sequences of known proteases. Besides these general considerations, we acquired the total mass spectrum of T. kiharae seed extract, and performed a search for masses that matched the masses of the predicted mature peptides with high accuracy. The results of such analysis for long precursors are given in Table S2. Interestingly, length variants are detected in the seed extract for most peptides, which may result from carboxypeptidase activity. The structure and putative processing of Tk-AMP-X2 L-2 precursor are shown schematically in Fig. 3B, and the summarized data for all precursors are presented in Table S3.

Structure of T. kiharae Tk-AMP-X genes

In addition to cDNA cloning, genomic DNA encoding Tk-AMP-X precursors was isolated. For Tk-AMP-X gene cloning, T. kiharae genomic DNA was PCR amplified, with primers 5e3 and 5r used for full-length cDNA cloning (Table S1). PCR fragments of approximately 850 and 1100 bp were generated and sequenced. The obtained sequences matched precisely the cDNA sequences. Accordingly, the protein-coding regions of Tk-AMP-X peptide genes encoding long and short precursors have no introns. Together with S- and L- type precursor genes, a gene for a precursor with six peptide domains named M-type precursor (M for ‘medium-sized’) was found (Fig. 4B). The putative processing sites in the M-type precursor were predicted on the basis of the same considerations as those used for S- and L-type precursors. It should be noted that the M-type precursor was not found during PCR amplification from cDNA. This may be a result of this gene not being expressed in immature seeds.

Tk-AMP-X family comprises seven groups of structurally related peptides

Amino acid sequence alignment of S-, L- and M-type precursors is shown in Fig. S3. Some regions of different precursors are seen to exhibit high sequence similarity. Thus, in the L-1 precursor, the second peptide domain exhibits high sequence similarity to the second peptide domain in S-3 and S-4 precursors, whereas the seventh peptide domain is homologous to the fifth peptide in S-2 and S-3. In the M-1 precursor, the first peptide domain exhibits sequence similarity to the first peptide domain of the S-4 precursor, whereas the second peptide domain is structurally homologous to the second peptide domain both in S-1 and S-2 precursors. In general, L-type precursors differ from S-type precursors by the presence of two additional peptide domains (the fifth and sixth), and from M-type precursors by the presence of the sixth peptide domain. If we number peptide domains in L-type precursors from 1 to 7, its peptide composition may be presented as 1-2-3-4-5-6-7; in M-type precursors: 1-2-3-4-5-7, in S-type precursors: 1-2-3-4-7 (Fig. 3A).

The deduced sequences of α-hairpinins derived from all three types of precursor proteins (S-, L- and M-) allocate into seven groups of closely-related peptides. Their main characteristics are shown in Table 2. Most peptides are positively charged, except for group 2 peptides and selected peptides from group 1. Sequence similarity within each group is very high (76–100%) but, between the groups, it is rather low. In general, each group comprises members of a particular peptide domain. The exceptions are the second peptide domain, which forms two groups of homologous peptides, and domains 5 and 6 in long precursors, which are extremely similar and form single group of closely-related peptides. The last peptide domain in all precursors (domain 5 in short, domain 6 in medium and domain 7 in long precursors) also consists of closely-related peptides, which fall into one group.

Table 2. Main characteristics of peptides predicted from Tk-AMP-X precursor sequences. Peptides are designated as: S, M or L, short, medium or long precursor; the first number indicates the sequence number; and the second number is the number of the peptide domain in the precursor from the NH2- to the COOH-terminus. In the consensus sequence, ‘X’ is any amino acid; ‘−’ represents negatively–charged residues; ‘+’ represents positively-charged residues; cysteines forming a typical 4-Cys motif are indicated by inverted text
GroupPeptidesConsensus sequenceCharge rangeaSimilarity percentb
  1. a

     The minimal and maximal charge of group members are indicated.

  2. b

     Calculated with the alibee-multiple alignment, version 2.0 (

1S-1-1, S-2-1, S-3-1, S-4-1, L-1-1, L-2-1 (M-1-1) IRXCXXXC–WKAGXDTGKAREC+EXCERXXXXXXXXX −2 … +276.6
3S-1-2 (Tk-AMP-X3), S-2-2, L-2-2 (M-1-2) GDSFDSCVSXCRGHGGWWGKERW-RCRXICRQSQE +2 … +3100.0
4S-1-3 (S-2-3, Tk-AMP-X1), L-2-3 (M-1-3, Tk-AMP-X2), S-3-3, S-4-3, L-1-3 XDDRCERMCXXYHDRREKKXCMKGCRYXXXX +1 … +494.2
5S-1-4 (S-2-4), S-3-4, S-4-4, L-1-4 (L-2-4, M-1-4) GHEHGDRCQXQCKRXRPGSYDRQQCIE+CQCQQ +2100.0
6L-1-5 (L-2-5, Tk-AMP-G7), L-1-6 (L-2-6, T-AMP-G6), M-1-5 HHGGGXXXHGDRCQXQCKRXPRGSYDRWQCTERCQSHQQD +2 … +3100.0
7S-1-5 (S-4-5), S-2-5 (S-3-5), L-1-7 (L-2-7), M-1-6 XXXXXXXXHXXSSCEQKCQQRXRHEY-+XQCVRDCKSGGXGGAXXXXXX +2 … +383.5

Tk-AMP-X genes are upregulated by pathogens and abiotic stress in seedlings

The expression of genes coding for T. kiharae 4-Cys peptides under stressful conditions was studied. Biotic stress was induced by four phytopathogenic fungi: Aspergillus niger, Bipolaris sorokiniana, Fusarium graminearum and Fusarium oxysporum. Expression of β-actin was used as a control (Fig. 5A). The results obtained showed that, in most cases, fungal infection upregulated α-hairpinin genes (Fig. 5B,C). However, the response to different fungi varied. Thus, infection with F. oxysporum caused a 5.8-fold increase in 4-Cys gene expression. The fungi A. niger and B. sorokiniana enhanced α-hairpinin gene expression by a factor of 2.7 and 2.2, respectively. Infection with F. graminearum did not cause significant changes in the transcript levels of the target genes.

Figure 5.

Expression of Tk-AMP-X genes in response to pathogens, elevated salt concentrations and different temperatures. (A) Expression of β-actin gene, which was used as a control. (B, C) Expression of Tk-AMP-X genes in response to pathogens. A.n., A. niger; F.g., F. graminearum; B.s., B. sorokiniana; F.o., F. oxysporum. (D) Expression of Tk-AMP-X genes in response to elevated NaCl concentrations. (E) Expression of Tk-AMP-X genes in response to extreme temperatures. The data are the mean values of four repetitions. Error bars indicate the SDs. Significance of differences was determined by Student's t-test. Significant differences (P < 0.01) in (C) to (E) are marked with ‘a’.

Abiotic stressful stimuli (i.e. elevated temperatures and high salt concentrations) activated 4-Cys gene expression (Fig. 5D,E). Thus, at 37 °C, the expression of 4-Cys genes increased by a factor of 2.8, whereas low temperatures (4 °C) downregulated these genes (Fig. 5E). High salinity (200 mm NaCl) elevated the target gene expression level by 2.1-fold (Fig. 5D). No significant changes in gene expression were observed at lower salt concentrations (100 mm NaCl).

Tk-AMP-X gene homologues are widely distributed among Triticum and Aegilops species

The distribution of Tk-AMP-X gene homologues among Triticum and Aegilops species was studied. Several diploid, tetraploid and hexaploid species were examined. Among them there are the putative donors of polyploid wheat genomes: Triticum boeoticum Boiss., Triticum monococcum L., Triticum urartu Thum. ex Gandil. (genome A), Aegilops tauschii Coss. (genome D) and Aegilops speltoides Tausch. (genome B). The tetraploid wheat Triticum timopheevii Zhuk. (AbAbGG) and the hexaploid wheat Triticum aestivum L. (cv. Khakasskaya) (AuAuBBDD) were also included in the analysis. It is generally accepted that bread wheat T. aestivum originated by hybridization of cultivated allotetraploid emmer wheat Triticum turgidum ssp. dicoccum (genome AuAuBB) with diploid Ae. tauschii [28, 29]. Wild emmer itself was derived through hybridization between two wild diploid species: T. urartu contributing the Au genome and Ae. speltoides donating the B genome. T. kiharae is a synthetic allopolyploid produced by crossing T. timopheevii with Ae. tauschii. T. timopheevii is assumed to originate from hybridization between the ancestral diploid Ab and G genome donors. Analysis of 4-Cys peptide genes in genomes of Triticum and Aegilops species may shed light on wheat evolution.

Tk-AMP-X gene homologues were revealed by PCR amplification from genomic DNA with primers 5e3 and 5r. The results showed that α-hairpinin genes were present in all species studied except for Ae. speltoides. A more detailed analysis revealed that the discovered genes differed in the number of peptide domains in precursor proteins. Thus, in T. timopheevii and T. aestivum, the 4-Cys peptide genes encoded L-type precursors, and, in T. aestivum and Ae. tauschii, S-type precursors. Furthermore, in all species studied, except Ae. tauschii, genes for M-type precursors were found. Deduced sequences of α-hairpinin precursors found in Triticum and Aegilops genomes are presented in Fig. S3. They form a family of closely-related precursor polypeptides, which differ in substitutions and deletions/insertions of single amino acid residues or short sequences. A phylogenetic tree based on the analysis of Tk-AMP-X precursor sequences is shown in Fig. 6. All precursor polypeptides form three large clusters according to the number of peptide modules. Within clusters, a high level of sequence similarity is observed among M-type precursors, and a lower level is observed among L- or S-type precursors.

Figure 6.

Phylogenetic analysis of species belonging to the Poaceae family based on sequence data from α-hairpinin precursors. Letters on branches designate genomes and bootstrap values are shown at the nodes.

In addition to isolation of 4-Cys peptide genes from genomic DNA of Triticum and Aegilops species, we also performed a bioinformatic search for homologous sequences in NCBI databases ( First, we used the blastx 2.2.25 algorithm and s-1 cDNA sequence as a query. Sequences homologous to wheat 4-Cys peptide genes were found in genomes of other cereals: Hordeum vulgare, Zea mays, Oryza sativa and Sorghum bicolor. The highest sequence similarity (91.5%) was with the barley gene BAK06493. The precursor protein in barley contains five peptide modules, similar to the wheat S-1 precursor. In maize and rice genomes, the precursor polypeptides with eight peptide domains were found (NP_001142639, EEE52377, EAY81437, ABC74439) and in sorghum, with nine peptide domains (XP_002449751). Thus, the modular structure of 4-Cys peptide genes is characteristic of cereals. A bioinformatics search also revealed storage protein (vicilin) genes containing regions for several peptide domains with a characteristic 4-Cys peptide motif; among them, vicilin genes of Carya illinoinensis (ABV49590) and Gossypium hirsutum (AEO27685). However, it remains to be determined whether the predicted peptides possess antimicrobial activity.

Analysis of expressed sequence tag databases including TIGR [30] revealed target gene homologues not only in T. aestivum and H. vulgare, but also in a dicotyledonous species Solanum tuberosum (CK258939). Potato cDNA codes for a precursor polypeptide with six 4-Cys peptide modules, which is homologous to T. kiharae M-1 precursor, and the highest variation was observed in the C-terminal region of the prepropeptide. These findings expand the distribution of 4-Cys peptide genes from cereals to other plant families; however, their role in plant immunity remains to be investigated (Fig. S4).


In a previous study, T. kiharae seeds were analyzed by a peptidomic approach. We discovered 24 novel AMPs, which were partially characterized, including their N-terminal sequencing [14]. Among them, two peptides, Tk-AMP-X1 and Tk-AMP-X2, contained four cysteine residues. In the present study, we completely sequenced these peptides and showed that they are small (28 and 31 residues long), closely-related, basic molecules, which inhibit the growth of several fungal pathogens in vitro. Their potent inhibitory activity makes Tk-AMP-X highly attractive for genetic manipulations with crops. These peptides are considered new because they show no significant sequence similarity to proteins in the available databases. However they share the same cysteine motif C1XXXC2-X(n)-C3XXXC4 with several plant peptides described earlier: AMPs MBP-1 from maize [15], EcAMP1 from E. crus-galli [16] and MiAMP2 from M. integrifolia [17], protease inhibitors VhTI from V. hederifolia [18], and BWI-2b and BWI-2c from F. esculentum [19, 20], suggesting that they belong to the same plant peptide family named α-hairpinins [20]. Determination of the secondary structure of Tk-AMP-X peptides by CD confirmed that they are helical, and therefore are likely to adopt a helix-loop-helix conformation stabilized by two disulfide bridges similar to that described for α-hairpinins [16, 20]. The disulfide linkage analyzed for Tk-AMP-X2 is also shared by other α-hairpinins. Wheat Tk-AMP-X peptides are antifungal, comparable to similar peptides from maize, barnyard grass and Macadamia, and they do not exhibit protease inhibitory activity (data not shown). However, their antifungal activity against plant pathogens varies considerably from the maize and barnyard grass AMPs. For example, similar to EcAMP1 from barnyard grass, Tk-AMP-X1 and Tk-AMP-X2 are active against Fusarium species; however the wheat peptides appear to be more active than EcAMP1 [16]. Furthermore, they are likely to show a broader activity spectrum than the barnyard grass peptide, as indicated by their activity towards D. maydis. Thus, sequence variations between wheat and barnyard grass peptides are responsible for the antifungal activity strength and spectrum (Table 1).

The mode of action of wheat Tk-AMP-X1 and Tk-AMP-X2 peptides, similar to other antifungal α-hairpinins from maize and Macadamia, remains unknown. The first insight comes from the study of the barnyard grass peptide EcAMP1, which was shown to accumulate in the fungal cell cytoplasm without interfering with the plasma membrane integrity [16]. Given the overall structural similarity to EcAMP1, we may speculate that the wheat peptides also penetrate into fungal cell cytoplasm and interact with an intracellular target.

Seven genes encoding Tk-AMP-X peptides that represent a novel gene family were isolated from T. kiharae by genomic and cDNA cloning. Six of them are expressed in immature seeds. The discovered genes are remarkable for their unique structural organization. First, no introns were spotted in the protein-coding regions of these genes. Second, all of them code for prepropeptides of similar multidomain structure: a signal peptide region, a region consisting of several (5–7) peptide modules with a characteristic 4-Cys motif separated by linkers of variable length and a C-terminal prodomain. The presence of signal peptides most likely indicates that precursors enter the usual secretory pathway of the cell. Of particular interest is the fourth peptide with an additional fifth cysteine residue. Third, the Tk-AMP-X genes differ in the number of peptide modules in precursor proteins. The gene structure of AMPs with a C1XXXC2-X(n)-C3XXXC4 motif from a native Australian plant M. integrifolia has been investigated. These α-hairpinins were shown to be derived from the N-terminal hydrophilic region of a storage protein (vicilin) precursor by post-translational proteolytic processing, resulting in simultaneous production of four homologous antifungal peptides [17]. Quite similarly, two 4-Cys peptides are processed from a vicilin precursor in the squash Cucurbita maxima [31]. α-Hairpinin modules in vicilin genes were also discovered bioinformatically in a number of plant species (see above). By contrast, wheat Tk-AMP-X precursors show no sequence homology to storage proteins, and thus represent products of genes unrelated to storage protein genes. We discovered sequences homologous to Tk-AMP-X precursor genes by genomic and expressed sequence tag database searches in other cereals and even in the phylogenetically distant Solanaceae family. Accordingly, we may postulate at least two types of gene organization for 4-Cys plant AMPs with a C1XXXC2-X(n)-C3XXXC4 motif: as complex multimodular precursors encoding only α-hairpinins (wheat, other cereals) or as part of storage vicilin precursors (Macadamia and some other dicotyledonous plants). The AMPs from seeds of I. balsamina were shown to be derived from a single precursor protein [23]. However, in contrast to wheat, basic mature peptide domains in I. balsamina have a different cysteine motif and are separated by acidic propeptide domains with at least five acidic residues arranged in doublets in the vicinity of the cleavage sites. In T. kiharae, the linker regions are more variable and, in addition to clusters of acidic amino acid residues, some of them possess stretches of positively-charged residues or specific amino acid residues/sequences recognized by other processing enzymes. These observations indicate that different types of proteases are involved in the processing of precursor polypeptides in T. kiharae. Protease inhibitors from stigmas of Nicotiana alata, assumed to be involved in defence against herbivorous insects, comprise another known example of plant peptides being encoded within a single transcript; however, cleavage occurs in the identical linker region EEKKND of the precursor repeated six times, resulting in the release of five protease inhibitors [32]. Accordingly, no structures resembling wheat Tk-AMP-X cDNA have been reported in plants.

Sequence similarity between all discovered Tk-AMP-X cDNA and genomic sequences point to their common evolutionary origin. However, the encoded Tk-AMP-X precursors differ in the number of peptide modules (5, 6 and 7 in short, medium and long precursors, respectively), suggesting deletions/insertions of sequences coding for 4-Cys peptide modules in the evolution of the ancestral gene. Extreme sequence similarity between peptide domains in a particular precursor protein (e.g. the fifth and sixth peptide domains in long precursors) may indicate the duplication of nucleotide sequences as another plausible evolutionary scenario.

A whole family of 27 peptides with a characteristic 4-Cys pattern is predicted from the Tk-AMP-X precursor sequences. The members fall into seven groups made up of 2–6 peptides. Most of the peptides are basic. Within each, group sequence identity is rather high, whereas it is very low between the groups. The length of the loop between the second and third cysteines varies from 11 to 14 residues in peptides belonging to different groups. Other variations include amino acid substitutions and the deletion/insertion of short amino acid sequences. It is of interest that, among the deduced peptides, we found several peptides previously isolated from T. kiharae seeds [14]; most other peptides were identified in seed extracts by MS. The variations in amino acid sequences of α-hairpinins in T. kiharae are likely to affect their biological activity. Our preliminary results are consistent with this suggestion. Thus, multimodular structure of Tk-AMP-X propeptides ensures the simultaneous release of a plethora of defence molecules with different activity spectra.

The possible function of the discovered Tk-AMP-X genes was explored by examining their expression patterns in response to biotic and abiotic stress. We have shown that the genes are upregulated by pathogens, elevated temperatures and salt stress. Together with in vitro antimicrobial activity of Tk-AMP-X peptides, these data point to the involvement of the new gene family in the stress response. It still needs to be determined whether complete processing of the precursor polypeptides occurs under stressful conditions or whether some peptide modules remain associated with each other.

Analysis of distribution of Tk-AMP-X homologues in Triticum and Aegilops species adds to the evolutionary picture of polyploid wheats. We have shown that the heterogeneity of genes encoding Tk-AMP-X precursors in T. kiharae is associated with its hexaploid nature. Allopolyploid wheat species acquired a complete diploid set of chromosomes from each parental species involved in its origin (A, B/G and D-genome donors). At least five diploid species are assumed to have contributed their genomes to polyploid wheats. T. kiharae, a synthetic allopolyploid, obtained full sets of chromosomes from T. timopheevii and Ae. tauschii. Analysis of Triticum and Aegilops species by PCR from genomic DNA allowed us to identify Tk-AMP-X gene homologues in all species studied except Ae. speltoides. A single PCR product was obtained with gene-specific primers from the genomic DNA of Ae. tauschii (D genome) and A-genome donors (T. urartu, T. monococcum and T. boeoticum). However, the discovered genes differed in the length of the encoded precursors. This allowed us to associate five-modular precursors with genome D originated from Ae. tauschii, six-modular precursors with genome A (T. monococcum, T. boeoticum and T. urartu) and seven-modular precursors with genome B or the closely-related genome G. The absence of 4-Cys peptide genes in Ae. speltoides, a presumable B genome donor to polyploid wheats, may indicate that present-day Ae. speltoides is different from the ancestor involved in the evolution of polyploid forms. Numerous studies clearly demonstrate that genome B of polyploid wheats is more variable than any of the other genomes [33] and its origin remains unclear. Thus, in T. kiharae and other allopolyploids precursors with different numbers of peptide modules are likely to be encoded by homologous genes. Our results show that the number of 4-Cys peptide genes discovered in T. kiharae exceeds that found in diploids. This may be associated either with the polyphyletic origin of wheat or with genomic rearrangements that occurred during polyploid formation. Rapid genomic changes in chromosome- and genome-specific sequences were shown in newly-formed allopolyploids of the wheat group [34].

In conclusion, a whole variety of novel AMPs belonging to the α-hairpinin family of plant peptides was discovered in T. kiharae. Two of them, named Tk-AMP-X1 and Tk-AMP-X2, were completely sequenced and shown to inhibit the growth of phytopathogenic fungi in the micromolar range. cDNA and genomic sequences encoding Tk-AMP-X peptides were determined. The peptides are synthesized as precursor proteins of unique modular structure containing several homologous peptide domains with a characteristic 4-Cys motif flanked by a signal peptide and a C-terminal prodomain. Thus, a novel principle of structural organization of plant AMP genes with a C1XXXC2-X(n)-C3XXXC4 motif was discovered. Despite overall structural similarity, three types of precursor polypeptides that differ in the number of peptide modules (5, 6 or 7) were identified. The discovered genes were shown to be upregulated by biotic and abiotic stress, thus providing evidence for their role in stress response. To the best of our knowledge, this is the first report of the role of α-hairpinins in the abiotic stress response. Analysis of the distribution of homologous genes in the plant kingdom confirmed that they are not specific to wheat but, instead, are widespread among cereals and even some phylogenetically unrelated dicotyledonous plants.

Experimental procedures

Biological material

Seeds of T. kiharae [genome composition (AbAbGGDD)], T. timopheevii (AbAbGG), T. urartu (AuAu), T. boeoticum (AbAb), T. monococcum (AbAb), Ae. tauschii (DD), Ae. speltoides (BB) and T. aestivum (cv. Khakasskaya) (AuAuBBDD) were obtained from the collection of the Vavilov Institute of General Genetics of the Russian Academy of Sciences (Moscow, Russia). F. oxysporum strain 16/10 was obtained from the collection of the Russian State Agrarian University (Moscow, Russia); F. graminearum strain VKM F-1668, B. sorokiniana strain VKM F-1446 and A. niger strain VKM F-33 were obtained from the All-Russian Collection of Microorganisms (Moscow, Russia). The fungi Fusarium verticilloides, C. graminicola and D. maydis were obtained from the Collection of Pioneer Hi-Bred (Johnston, IA, USA).

Isolation of Tk-AMP-X1 and Tk-AMP-X2 peptides

Tk-AMP-X1 and Tk-AMP-X2 peptides were isolated from T. kiharae seeds by a combination of affinity, size-exclusion and RP-HPLC as described previously [14]. Briefly, wheat flour was extracted with a mixture of 1 м HCl, 5% (v/v) formic acid, 1% (v/v) trifluoroacetic acid (TFA) and 1% (w/v) NaCl at a seed/solvent ratio of 1 : 4 in the presence of the proteinase inhibitor cocktail for plant extracts (Sigma-Aldrich, St Louis, MO, USA). The extract was desalted by RP-HPLC on an Aquapore C8 RP-300 cartridge (10 × 30 mm; Applied Biosystems, Foster City, CA, USA) equilibrated with 0.1% TFA (solvent A). Proteins and peptides were eluted with 70% of solvent B (80% acetonitrile in solvent A), freeze-dried and loaded onto a HiTrapHeparin HP column (1 × 5 mL) (GE Healthcare, Milwaukee, WI, USA). The column was washed with 10 mm Tris-HCl buffer (pH 7.5) (solvent C) to remove unbound proteins. The adsorbed proteins were eluted with a stepwise NaCl gradient (100 and 500 mm) in solvent C at a flow rate of 0.7 mL·min−1. The unbound, 100- and 500-mm fractions were desalted by RP-HPLC on the C8 cartridge. The fraction that eluted at 500 mm NaCl was further separated by size-exclusion chromatography on a Superdex Peptide HR 10/30 column (1 × 30 cm) (GE Healthcare). Proteins and peptides were eluted with 5% acetonitrile and 0.05% TFA (solvent D) at a flow rate of 250 μL·min−1 and monitored by absorbance at 214 nm. The peptide-containing fraction was subsequently separated by RP-HPLC on a Vydac C18 column (4.6 × 250 mm) (Separations Group, Hesperia, CA, USA) or a Reprosil C18 column (4 × 250 mm) (Dr Maisch GmbH, Ammerbuch-Entringen, Germany) in a 60-min linear acetonitrile gradient (10–40% B) at a flow rate of 1 or 0.75 mL·min−1 at 40 °C and detection at 214 nm.

Analytical methods

The peptide concentration was determined spectrophotometrically. Absorption spectra were recorded on a U-3210 spectrophotometer (Hitachi, Tokyo, Japan). The molar extinction coefficients were determined using gpmaw software (Lighthouse data, Denmark; website: Reduction of peptides with dithiothreitol and alkylation with 4-vinylpyridine was conducted as described previously [14]. CNBr cleavage of Tk-AMP-X2 was performed in accordance with the published guidelines [35]. Amino acid sequencing of reduced and alkylated peptides was performed by automated Edman degradation on a model 492 Procise sequencer (Applied Biosystems) in accordance with the manufacturer's instructions. A homology search was carried out using Swiss-Prot and TrEMBL databases with a blast algorithm. Mass spectra were acquired on an Ultraflex MALDI-TOF mass spectrometer (Bruker Daltonics, Ettlingen, Germany) in a positive ion mode. 2,5-Dihydroxybenzoic acid was used as a matrix. Mass spectra were analyzed with dataanalysis for tof software (Bruker Daltonics).


CD spectra were acquired on a JASCO model J-500C spectropolarimeter (Jasco Corp., Tokyo, Japan) with an optical path length of 0.01 cm. Peptides were dissolved at a concentration of 1 mg·mL−1 in 110 mm NaCl, 50 mm Na2HPO4/NaH2PO4 (pH 7.5). Five scans for each sample were obtained and averaged over the wavelength range of 190–240 nm. The content of α-helix, β-sheet and random coil conformation was estimated using contin software [36].

Antifungal assays

Antifungal activity was determined as described previously [37]. Inhibition of spore germination was estimated spectrophotometrically at 620 nm in microtitre plates after incubation of spore suspension (2 × 104 conidia·mL−1) with the peptide solution for 48 h at 22 °C. IC50 values showing the protein concentration required for 50% growth inhibition were calculated.

Cloning of Tk-AMP-X cDNA from T. kiharae

Total RNA was isolated from 100 μg of T. kiharae immature seeds (from 10 to 15 days post-anthesis) using the Trizol RNA Prep 100 kit (IsoGen, Moscow, Russia) in accordance with the manufacturer's instructions. The first-strand cDNA was synthesized from 2 μg of total RNA using the Mint kit (Evrogen, Moscow, Russia) at 42 °C for 1.5 h in accordance with the manufacturer's instructions.

cDNA clones encoding Tk-AMP-X peptides were produced by a combination of 3′- and 5′-RACE [38]. Initially, several direct degenerate primers were designed on the basis of amino acid sequences of Tk-AMP-X1 and Tk-AMP-X2 peptides: Dir1, Dir1-2, X1, X2 (Table S1). For 3′-RACE, PCR amplification with the universal primer T7cap and gene-specific degenerate primers Dir1 and Dir1-2 was carried out. The products of the reaction were diluted 100 times and used as a template in PCR with primers X1, X2 and T7cap. All PCR products were cloned into the pAL-TA vector (Evrogen) and sequenced on an ABI PRISM 3730 (Applied Biosystems). At least 10 independent clones were analyzed. For 5′-RACE, on the basis of the nucleotide sequences obtained during 3′-RACE several reverse primers were constructed (38rs, 67rs1, 67rs2) (Table S1), which were used in combination with the universal primer T7cap in PCR with the cDNA template. cDNA fragments were cloned into the pAL-TA vector and sequenced. To obtain the full-length cDNA encoding Tk-AMP-X peptides, several direct primers to the 5′-ends of the target genes and additional reverse primers to their 3′-ends were synthesized. With a combination of primers 5e3 and 5r (Table S1), the complete nucleotide sequences of the protein-coding regions of the target genes were determined. All PCR amplifications were run on a PTC-200 automated thermal cycler (MJ Research, St Bruno, Quebec, Canada) using 1–10 ng of DNA or the first-strand cDNA as a template, 0.25 mm dNTPs, oligonucleotide primers (10 pmol each) and Taq DNA polymerase (Evrogen) in accordance with the manufacturer's instructions.

Cloning of genomic DNA encoding Tk-AMP-X peptides from T. kiharae

To amplify genomic DNA encoding Tk-AMP-X peptides, two primers, 5e3 and 5r (Table S1), were used. The PCR amplification was performed using Taq-polymerase (Evrogen) in accordance with the manufacturer's instructions. The PCR conditions were: 2 min of denaturation at 94 °C, followed by 34 cycles of denaturation at 94 °C for 20 s and primer annealing at 64 °C for 20 s, and primer extension at 72 °C for 90 s, with a final 5-min extension at 72 °C. PCR amplification from the cDNA was carried out in parallel. The amplified genomic fragment was cloned into the pAL-TA vector. No <3 different clones were sequenced to obtain full-length genomic DNA encoding each class of precursor of the T. kiharae 4-Cys peptides.

Identification of Tk-AMP-X gene homologues in genomes of Triticum and Aegilops species

Genomic DNA was isolated from 5-day-old seedlings of T. monococcum, T. boeoticum, T. urartu, Aespeltoides, Ae. tauschii, T. timopheevii and T. aestivum as described for T. kiharae. For amplification of the coding region of the genes, a pair of specific primers 5e3 and 5r was used with genomic DNA as a template. PCR products were cloned into the pAL-TA vector. At least three clones for each fragment were sequenced.

Biotic stress induction

Biotic stress was induced by infecting T. kiharae seedlings with spore suspensions of phytopathogenic fungi essentially as described previously [8]. Several fungal species were used: F. oxysporum, F. graminearum, B. sorokiniana and A. niger. Fungi were grown on solid nutrient agar for 8–10 days at 23–25 °C. Conidia were collected by washing colonies with 10% potato-carrot broth. The obtained suspensions were diluted to spore concentrations of: F. oxysporum, 5 × 104; F. graminearum, 5 × 104; B. sorokiniana, 2 × 103; and A. niger 3 × 104. T. kiharae seeds were surface-sterilized with 70% ethanol for 2 min, soaked in spore suspensions for 1 min, and placed on the filter paper in Petri dishes for germination. Control seeds were incubated in sterile water and treated similarly. Four-day-old infected and control seedlings were washed with water and used for RNA isolation. The experiment was performed with four replicates.

Abiotic stress induction

The effect of temperature and salt stress on Tk-AMP-X gene expression was studied as described previously [8]. To explore the influence of temperature, T. kiharae 3-day-old seedlings grown at room temperature (+22 °C) were divided into three groups, which were kept for the next 48 h at different temperature regimes. Plants of the first group were kept at +4 °C. Second group plants were grown at +22 °C (control), and those of the third group were grown at +37 °C. In salt stress experiments, seedlings grown in salt solution (100 mm NaCl or 200 mm NaCl) for 3 days under a 16 : 8 h light/dark cycle were used. Control seedlings were grown in water. All experiments were performed with four replicates.

RT-PCR analysis

To determine the Tk-AMP-X mRNA expression level, a semiquantitative RT-PCR was used. RNA was isolated from stressed seedlings using the Trizol RNA Prep 100 kit, as described above. Three or four plants were used for each RNA preparation. Reverse transcription was performed with the standard kit for the first-strand cDNA synthesis (Fermentas, Vilnius, Lithuania) using the oligo dT18 primer. cDNA was then used as a template in a PCR with gene-specific primers 4d and 38rs to the identical regions of the target genes (Table S1). A PCR product of approximately 370 bp was generated. The β-actin gene was used as an internal control for sample normalization. To exclude amplification from genomic DNA, PCR for 40–44 cycles from all RNA preparations was performed. Accumulation of PCR products was quantified using biocapt mv, version 1.1 (Vilber Lourmat, France) in the image quantification regime.


The work carried out in the present study was supported in part by the Russian Foundation for Basic Research (grant numbers 11-04-00190-a and 12-04-00117-a); Russian Federal Programs of the Ministry of Education and Science (contract numbers 16.512.11.2156 and 16.740.11.0424); the Biodiversity Program of the Russian Academy of Sciences; AAS and AAV are recipients of the stipend of the President of Russian Federation.