The plant antimicrobial peptide MiAMP1 from Macadamia integrifolia and the yeast killer toxin peptide WmKT from Williopsis mrakii are structural homologues. Comparative studies of yeast mutants were performed to test their sensitivity to these two antimicrobial peptides. No differences in susceptibility to MiAMP1 were detected between wild-type and several WmKT-resistant mutant yeast strains. A yeast mutant MT1, resistant to MiAMP1 but unaffected in its susceptibility to plant defensins and hydrogen peroxide, also did not show enhanced tolerance towards WmKT. It is therefore probable that the Greek key β-barrel structure shared by MiAMP1 and WmKT provides a robust structural framework ensuring stability for the two proteins but that the specific action of the peptides depends on other motifs.
Mode of action
MiAMP1 is a novel antimicrobial peptide (AMP) found in the nut kernels of Macadamia integrifolia. It consists of 76 amino acids, including six cysteine residues engaged in three intramolecular disulphide bridges, and has an estimated isoelectric point of 10.1 . The cysteine residue spacing in MiAMP1 does not match that of any described antimicrobial peptide, and MiAMP1 is the first functionally characterized member of a new class of plant defence proteins. Other plant sequence homologues in this class include a protein found in blister rust-resistant genotypes of western white pine, Pinus monticola and deduced proteins from pathogen-induced genes in Scots pine, Pinus sylvestris. MiAMP1 has been shown to be highly inhibitory to a wide range of phytopathogens, but has no affect on the growth of plant and mammalian cells . In addition, transgenic expression of MiAMP1 in canola, Brassica napus L. provided enhanced resistance against blackleg disease caused by the fungus Leptosphaeria maculans. Therefore, MiAMP1 is potentially a useful tool for genetic engineering of disease resistance in crop plants.
The three-dimensional structure of MiAMP1 was determined to consist of eight β-strands arranged in two Greek key motifs, each containing four antiparallel β-sheets, that form a Greek key β-barrel . This structure is unique amongst plant antimicrobial peptides, but corresponds to the single domain βγ-crystallin precursor fold found in a number of diverse proteins including some with antimicrobial activity [6⇓–⇓⇓10]. The structure of MiAMP1 shows particularly strong resemblance to that of the Williopsis mrakii (formerly known as Hansenula mrakii) yeast killer toxin, WmKT (formerly known as HM-1 toxin) [5,6]. McManus et al.  noted that not only do MiAMP1 and WmKT share a common structural framework (see Fig. 9 in ), they also both have two similarly located large protein loops that are tethered to the β-sheets via disulphide bonds. These distinct structural features, common to antimicrobial peptides in different phyla, led these authors  to speculate that the three-dimensional structural conservation may indicate a direct role in the mode of action of the peptides. The three-dimensional structural homology and similar antimicrobial function is remarkable in that MiAMP1 (NCBI Accession # P80915) and WmKT (NCBI Accession # P10410) share no significant protein sequence homology and even differ both in the number of cysteine residues they contain and their respective spacing. Cysteine composition and spacing are often conserved in structurally related proteins.
The WmKT toxin is believed to interfere with cell wall biosynthesis in susceptible yeast cells by inhibiting β-(1,3)-glucan synthesis [11⇓–⇓⇓⇓16] while the mode of action of MiAMP1 is unknown. In this study we have adopted an approach using yeast mutants with respective resistance to WmKT and MiAMP1 to investigate whether these two structurally related peptides may share the same antifungal mode of action.
2 Materials and methods
2.1 Yeast Strains
The Saccharomyces cerevisiae strains used in this study are described in Table 1. Initially, twelve yeast strains commonly used in laboratory research were screened for their sensitivity to MiAMP1 and sensitivities expressed as IC50 values (peptide concentration required to inhibit 50% cell growth) ranged from 5 to >100 μg/mL (0.6 to >12.3 μM) . The strain COP161  was the most sensitive and was used to derive a mutant resistant to MiAMP1. Subsequently, WmKT-resistant yeast mutants TK201, rhk2, rhk3 and wild-type strain BJ1824 were obtained from Dr. T. Komiyama (Niigata College of Pharmacy, Niigata, Japan) while yeast deletion mutant 4007 (Invitrogen Inc.) was derived from the wild-type strain BY4741 . Coincidently, the BJ1824 and BY4741 wild-type strains had IC50 values of approximately 5 μg/mL (0.6 μM) for MiAMP1 and were therefore comparable to COP161 in their sensitivity to this peptide.
The WmKT peptide from W. mrakii was kindly provided by Dr. T. Komiyama (Niigata College of Pharmacy, Niigata, Japan). The DmAMP1 peptide from Dahlia merckii was kindly provided by Dr. B. Cammue (Catholic University of Leuven, Belgium). The plant defensin, CtAMP1  was obtained by microbial expression, using the vector pPIC9K (Invitrogen Inc.) in Pichia pastoris and purified by standard protein purification procedures. MiAMP1 was purified from M. integrifolia nut kernels as previously described . Peptide purity was verified using gel electrophoresis and mass spectrometry. Antimicrobial activity of protein samples against S. cerevisiae was quantitatively assayed by spectrophotometry of liquid cultures grown in microtiter plates as described previously . Cell suspensions were grown at 30 °C in SD (Synthetic Defined) growth medium [0.8 g/l CSM (Complete Supplement Mix; Bio101/26.7 g/l DOB (Drop Out Base; Bio101) with glucose] to logarithmic growth phase and then diluted to a concentration of 105 cells/mL determined by cell counts using a haemocytometer and light microscopy. Cell growth in the presence and absence of specified antimicrobial peptides or hydrogen peroxide was measured spectrophotometrically at 24 h intervals over a 96-h time period. All experiments were repeated independently at least twice and means and standard errors from 2–3 replicates are shown.
3 Results and discussion
3.1 Effect of MiAMP1 on WmKT-resistant S. cerevisiae mutants
The effects of purified MiAMP1 and WmKT on wild-type S. cerevisiae strain BJ1824 and three WmKT-resistant yeast mutant strains were tested. The yeast mutants, termed TK201, rhk2 and rhk3, were all derived from the parental strain BJ1824 [15,16]. No differences in growth rate were detected between the TK201, rhk2 and rhk3 mutants and the wild-type strain in plain SD growth medium (data not shown).
Growth of the BJ1824 (wild-type) strain was strongly inhibited by WmKT, with an IC50 value of less than 1 μg/mL (0.1 μM, Fig. 1(a)). The three mutant strains showed varying degrees of increased resistance towards WmKT, confirming previous findings . The TK201 mutant appeared to be almost unaffected by the presence of up to 5 μg/mL (0.6 μM) of WmKT (Fig. 1(a)). MiAMP1 displayed weaker inhibitory activity than WmKT against the BJ1824 in terms of the peptide concentration required to inhibit cellular growth. As the molecular masses of the mature MiAMP1 and WmKT peptides are comparable (8138 Da and 9528 Da, respectively) these results demonstrate that MiAMP1 is less potent than WmKT. Nevertheless, strong growth inhibition of the wild-type strain was detected in the presence of MiAMP1, with an IC50 value of approximately 5 μg/mL (0.6 μM, Fig. 1(b)). No significant difference in MiAMP1 susceptibility was detected between any of the mutant strains and wild-type BJ1824 cells (Fig. 1(b)). The similar growth response of wild-type and WmKT-resistant mutant cells in the presence of MiAMP1 implies that the mode of action of MiAMP1 differs from that of WmKT and involves other targets in yeast.
Growth inhibition in S. cerevisiae strains by MiAMP1 and WmKT. Yeast strains were grown in SD medium supplemented with (a) WmKT or (b) MiAMP1 for 48 h. Bars indicate ± standard error. A concentration of 1 μg/mL is equal to 105 and 123 nm of WmKT and MiAMP1, respectively.
3.2 Isolation and characterisation of a MiAMP1-resistant S. cerevisiae mutant
The S. cerevisiae strain COP161  is highly sensitive to MiAMP1 and was used for studies of its mode of action. The IC50 value of MiAMP1 for COP161 was determined by in vitro bioassays to be approximately 5 μg/mL (0.6 μM) after 48 h incubation (Fig. 2). A spontaneous yeast mutant, termed MT1, was isolated from COP161 cells grown on solid growth medium supplemented with 100 μg/mL (12.3 μM) MiAMP1. DNA fingerprinting showed that this strain was derived from COP161 (data not shown). The mutant showed a highly significant increase in MiAMP1 tolerance compared to the wild-type strain (Fig. 2). A 10-fold increase in IC50 value was detected for the MT1 mutant after 48 h incubation with MiAMP1 compared to that of the wild-type COP161 strain. No difference in growth rate was detected between the MT1 mutant and the COP161 wild-type strain in the absence of MiAMP1 (data not shown).
Growth inhibition of S. cerevisiae strains COP161 and MT1 in the presence of MiAMP1. Yeast cells were grown in SD growth medium supplemented with MiAMP1 for 48 h. Each value ± standard error is shown.
To test whether the MT1 mutant had generally increased tolerance to abiotic stresses we tested its response to oxidative stress using hydrogen peroxide. No difference in the sensitivity to hydrogen peroxide was observed between MT1 and the wild-type COP161 strain (Fig. 3(a)). To examine whether MT1 showed enhanced resistance to cytocidal peptides structurally unrelated to MiAMP1, the MT1 mutant was also assayed against the plant defensins, DmAMP1 and CtAMP1 . In contrast to its resistance to MiAMP1 (Fig. 2), the MT1 mutant did not show any enhanced resistance to the plant defensins (Fig. 3(b) and (c)). Interestingly, MT1 was found to be more susceptible to DmAMP1 than the wild-type strain whereas there was no difference in susceptibility to CtAMP1 between the wild-type and the mutant strain (Fig. 3(b) and (c)). To further test whether there was any relationship between the mode of action of plant defensins that inhibit S. cerevisiae and the action of MiAMP1 we also examined the effect of MiAMP1 on a yeast deletion mutant (Table 1) that has enhanced resistance to DmAMP1 [21,22]. Sensitivity to DmAMP1 is determined by the IPT1 gene that encodes an inositol transferase involved in the final step of synthesis of the sphingolipid mannosyldiinositolphosphorylceramide [21,22]. A yeast deletion mutant for the IPT1 gene and the wild-type strain had IC50 values of 4.2 ± 0.1 μg/mL (0.5 μM) and 3.8 ± 0.3 μg/mL (0.5 μM), respectively, for MiAMP1 and therefore the IPT1 gene and this sphingolipid class are not involved in MiAMP1 action.
Growth inhibition in S. cerevisiae wild-type and mutant strains by hydrogen peroxide, DmAMP1, CtAMP1 and WmKT. Wild-type (COP161) and mutant (MT1) strains were grown in SD growth medium supplemented with (a) hydrogen peroxide, (b) DmAMP1, (c) CtAMP1 or (d) WmKT for 48 h. Each value ± standard error is shown. 1 μg/mL of DmAMP1 and CtAMP1 is equivalent to 200 nM and 1 μg/mL of WmKT equals 105 nm.
These results suggest that MT1 does not have enhanced resistance towards oxidative stress and other cytocidal peptides in general, but that the mutation in MT1 probably affects MiAMP1 tolerance specifically. Therefore, an experiment was conducted to examine whether the enhanced resistance of MT1 for MiAMP1 also corresponded to resistance to WmKT. Importantly, the MT1 mutant showed no significant difference in susceptibility to WmKT compared to the wild-type strain, COP161 (Fig. 3(d)), further substantiating the notion that MiAMP1 and WmKT differ in their mode of action.
Extensive studies on the cytocidal effects of WmKT have demonstrated that WmKT perturbs cell wall biosynthesis in sensitive yeast, by inhibiting β-glucan synthesis at budding sites or conjugation tubes, which leads to cell lysis [13,14,16]. Based on the resemblance in globular folding between MiAMP1 and WmKT  it was hypothesized that MiAMP1 and WmKT may have functional similarities. However, no increase in resistance towards MiAMP1 was detected in any of three WmKT-resistant yeast mutant strains. Two of the mutant yeast strains, TK201 and rhk2, are defective in different parts of protein N-glycosylation, resulting in underglycosylation. This is presumed to cause alterations in cell wall or mannoprotein properties that may act as receptors. Modification of such putative receptors is thought to prevent or impede WmKT from binding to the cell and interfering with β-(1,3)-glucan synthesis . The mutation in the third yeast mutant, rhk3, is currently unknown. The MT1 mutant with increased resistance to MiAMP1, isolated in the present study, also did not show increased resistance to WmKT. The observation that mutations for resistance to WmKT have no effect on sensitivity to MiAMP1 suggests that there is little functional similarity in the mode of action of WmKT and MiAMP1. It is most probable that the structural resemblance between MiAMP1 and WmKT is unrelated to protein function and probably reflects a robust structural framework that provides stability for the two proteins with the specific mode of action determined by specific protein motifs. It is possible that the different responses of the yeast mutants to MiAMP1 and WmKT may reflect altered receptor specificities for the two peptides. Fungal receptors for AMPs are as yet unknown, while receptors for some bacteriocin AMPs have been identified in bacteria via mutational studies [23,24]. The screening of yeast deletion mutant libraries, now established for studying the mode of action of chemical inhibitors of yeast [25,26] represents a new and powerful approach to identify key genes determining fungal susceptibility to antimicrobial peptides such as MiAMP1 and WmKT.
We thank Dr. T. Komiyama for provision of WmKT and WmKT-resistant yeast strains and Dr. B. Cammue for providing DmAMP1. This study was supported by an Australian Postgraduate Award scholarship.
(2003) Isolation of a novel antimicrobial peptide gene (Sp-AMP) homologue from Pinus sylvestris (Scots pine) following infection with the root rot fungus Heterobasidion annosum. FEMS Microbiol. Lett. 22827–31.
(1998) NMR structure of the Streptomyces metalloproteinase inhibitor, SMPI, isolated fromStreptomyces nigrescens TK-23: another example of an ancestral βγ-crystallin precursor structure. J. Mol. Biol. 282421–433.
(2000) A gene encoding a sphingolipid biosynthesis enzyme determines the sensitivity of Saccharomyces cerevisiae to an antifungal plant defensin from dahlia (Dahlia merckii). Proc. Natl. Acad. Sci. USA 979531–9536.