Click on author name to view affiliation information
Huanglongbing {HLB [associated with Candidatus Liberibacter sp. (CLas)]} and asiatic citrus canker {ACC [causal organism Xanthomonas citri ssp. citri (XCC)]} are bacterial diseases that seriously threaten sustainability of the Florida citrus (Citrus sp.) industry. Sweet orange (Citrus sinensis) and grapefruit (Citrus paradisi) are highly susceptible to ACC and improvement through conventional breeding is a long-term process, making transgenic solutions attractive. No strong HLB resistance has been identified within cultivated citrus scion types: creation of transgenic citrus that would permit economic citrus production where HLB is endemic is a high priority. Little is known about the HLB pathosystem and thus broad-spectrum antimicrobial peptides (AMPs) have been the focus for current work, and identification of safe and effective transgenes is essential to our efforts. In vitro assessment of minimum inhibitory concentration (MIC) for 44 AMPs was conducted using Sinorhizobium meliloti and Agrobacterium tumefaciens as surrogates for the unculturable CLas because they are closely related alpha proteobacteria (class Alphaproteobacteria). XCC is also a gram-negative bacterium and was included in these analyses in anticipation that HLB and ACC resistance can be achieved with the same AMP transgene if expressed using non-tissue-specific promoters. Twenty AMPs from diverse sources were initially tested. AMPs with the lowest MICs included tachyplesin I from horseshoe crab (Tachypleus tridentatus), SMAP-29 from sheep (Ovis aries), D4E1 and D2A21 (which are synthetic AMPs derived through evaluation of critical amino acid residues in AMPs, overall peptide structure, and AMP effectiveness), the human (Homo sapiens) LL-37, and the honeybee (Apis mellifera) venom AMP melittin. These AMPs inhibited growth of all three test bacterial species at 1 μM or less. An additional 20 synthetic AMPs were designed based on structures of the most effective AMPs and seven of these showed effectiveness at 1 μM or less across all three test bacteria. Most AMPs were comparable in effectiveness across the three bacterial species, but some species × AMP interactions were observed. Hemolytic activity was assessed by exposure of porcine erythrocytes (from Sus scrofa) to the AMPs. Hemolysis from most AMPs was not significantly different from water, whereas melittin was highly hemolytic.
Considerable effort is being devoted to developing transgenic citrus with resistance to huanglongbing and asiatic citrus canker. These bacterial diseases seriously threaten the sustainability of the Florida citrus industry. No strong HLB resistance has been identified within cultivated citrus scion types, compelling significant effort to develop transgenic lines that permit economic citrus production where HLB is endemic. Kumquats (Fortunella sp.) are highly resistant to ACC and some mandarin (Citrus reticulata) types are tolerant (Peng et al., 2010); however, incorporation of strong ACC resistance through conventional breeding into the sweet orange and grapefruit types, which dominate the Florida citrus industry, may be a decades-long process. Cultivar characteristics are largely unchanged in transgenic plants; therefore, transgenic solutions for ACC are also an attractive option.
The CLas genomic sequence has been published (Duan et al., 2009) and is providing some targets specific to this pathogen; however, understanding of the HLB pathosystem remains quite limited. For this reason, broad-spectrum antimicrobial peptides have been the focus for much of the initial work to create HLB-resistant transgenic citrus. AMPs have considerable promise in developing transgenic plants with disease resistance and have been subjects of numerous studies (e.g., reviewed in Düring, 1996; Marcos et al., 2008; Rajasekaran et al., 2009). AMPs offer an additional advantage that many are effective against multiple microorganisms offering the possibility that a single transgene may confer some resistance to a group of gram-negative bacterial pathogens, including the causal agents for HLB and ACC as well as Xylella fastidiosa the causal agent for citrus variegated chlorosis, a disease that has not yet been reported in the United States. Some AMPs may also offer resistance to fungal pathogens [Alternaria sp., Aspergillus sp., Botrytis cinerea, Cercospora sp., Claviceps purpurea, Colletotrichum destructivum, Fusarium sp., Magnaporthe grisea, Mycosphaerella musicola, Phytophthora sp., Pythium sp., and Verticillium dahliae (Rajasekaran et al., 2009)].
Because development of transgenic citrus and screening for disease resistance is extremely laborious, a series of AMPs were screened in vitro to assess their potential for providing resistance to HLB and ACC. Some AMPs also disrupt mammalian cells; therefore, screening for hemolysis of porcine corpuscles was also conducted.
The AMPs listed in Table 1 include many reported in the literature as occurring naturally in diverse organisms and were custom-synthesized based on reported amino acid sequences: apidaecin, drosocin, histatin-5, indolicidin, LL-37, pyrrhocoricin, and SMAP-29 were synthesized by AnaSpec (Fremont, CA); spinach defensin 2 (SoD2 from Spinacia oleracea), Cn-AMP-1, Ib-AMP-4, and P4c were synthesized by GenScript USA (Piscataway, NJ); and tachyplesin I was synthesized by Bachem Americas (Torrance, CA). The well-studied standard AMPs cecropin A, cecropin B, magainin 1, magainin 2, and melittin were purchased from Sigma Chemical Co. (St. Louis, MO). The synthetic AMPs D4E1 and D2A21 were provided by AgroMed (Annapolis, MD).
Additional synthetic AMPs were designed by one of the authors (J. Jaynes) based on characteristics of the most promising animal AMPs. A series of 24 additional AMPs were synthesized (PolyPeptide Laboratories, San Diego, CA) through agreement with AgroMed and Southern Gardens Citrus (Clewiston, FL). The identities of these AMPs are indicated in Table 2 and are covered by either issued or pending patents.
Test organisms were XCC (isolate Xsc 67D), A. tumefaciens [AT (isolate EHA 105)], and S. meliloti [SM (isolate Sm 1021)]. SM and AT were used as surrogates for the unculturable CLas because they are closely related alpha proteobacteria (Bastianel et al., 2005). Xsc 67D is a field isolate from symptomatic grapefruit leaves made on 28 Apr. 2006 from the U.S. Department of Agriculture, Ft. Pierce, FL, farm, with its identity confirmed by sequencing the product of the VM3/VM4 primer pair (Mavrodieva et al., 2004) and the pathogenicity of the isolate confirmed by presence of hyperplastic lesions on inoculated grapefruit leaves. All studies used bacteria no more than two transfers from the stock culture.
AMPs were tested at concentrations of 0, 0.03, 0.1, 0.3, 1, 3, 10, and 30 μM. Assays were performed in 96-well plates with lids (Nunc Nunclon D; Thermo Scientific, Rochester, NY). Each well contained 160 μL of Luria Broth (Becton, Dickinson and Co., Franklin Lakes, NJ), 20 μL bacterial suspension (targeting initial 107 colony-forming units/mL in each well), and 20 μL AMP solution.
Plates were incubated at 20 °C and shaken for 20 s and absorbance was read at A630 at 0, 16, 20, 24, 40, 44, 48, and 64 h of incubation using an Ultramark EX microplate imaging system (Bio-Rad Laboratories, Hercules, CA).
There were three replicates of each AMP concentration. Each plate contained three control wells with no bacteria. Assays of all three bacteria were made on a single 96-well plate. Each experiment was performed twice and results were identical or varied by a single dilution step in a few cases. MIC values were determined by examining plots of growth data and identifying the lowest concentration where there was no increase in A630 between 0 and 48 h.
Hemolytic activities of AMPs were measured using a slight modification of the method published by Park et al. (2007). Two milliliters of porcine blood (Lampire Biological Laboratories, Pipersville, PA) were washed three times by rinsing cells with 4 mL phosphate-buffered saline [PBS (pH 7.4)], mixing by gentle inversion, and centrifugation at 500 gn for 10 min at room temperature. Rinsed cells were mixed in 12 mL PBS and placed at 4 °C for 0.5 h to cool. Stocks of porcine blood were used within 1 week of receipt.
It was determined (data not shown) that 10 μM was the minimum concentration required to achieve 100% hemolysis with melittin, which has been extensively studied for its high hemolytic activity (Blondelle and Houghten, 1991), and so all AMPs were tested at this concentration. Stocks were made of each AMP at 100 μM in PBS to produce test solutions. The positive control was 0.1% Triton X-100 (Dow Chemical Co., Midland, MI) and the negative control was PBS. Test solutions were 20 μL of AMP solution added to 0.5-mL microcentrifuge tubes and briefly cooled on ice before adding 180 μL cooled suspended porcine cells to each tube. Solutions were mixed by drawing suspension gently back into the pipette and gently expelling one time. Tubes were incubated at 38.5 °C for 30 min, centrifuged 5 min at 4000 gn, then 100 μL of supernatant was added to 900 μL PBS and absorbance was read at A580. Percent hemolysis was calculated as: [(sample – blank)/control] × 100.
All experiments were highly reproducible for assessment of MIC and percent hemolysis. In all cases, data are presented from two identical experiments with three replicates of each treatment in each experiment.
In comparison of widely diverse AMPs (Table 1), the lowest MICs (greatest ability to suppress microbial growth) were observed from tachyplesin I, SMAP-29 from sheep, D4E1 and D2A21 (which are synthetic AMPs derived through evaluation of critical amino acid residues in AMPs, overall peptide structure, and AMP effectiveness), the human LL-37, and the AMP melittin found in bee venom. All suppressed growth of the three bacterial species at 1 μM or less and had hemolytic activity not significantly different from the buffer control, except melittin, which was as disruptive to porcine erythrocytes at 10 μM as was 0.1% Triton X-100. The MICs observed in our study were comparable to those reported in earlier reports on in vitro activity of AMPs against plant pathogens (Badosa et al., 2007; Hammerschlag, 2004; Kuzina et al., 2006). The AMPs in this most active group displayed bacterial suppression at concentrations 0.1 to 0.33 times the MIC observed for cecropins A and B, which are among the most highly studied AMPs for inducing transgenic plant resistance to bacterial diseases (Düring, 1996; Rajasekaran et al., 2009).
Because use of animal or human transgenes in food products might generate considerable consumer concern, we are committed to using transgenes of plant or synthetic origin. Unfortunately, none of the plant-derived AMPs selected for evaluation displayed strong activity in the in vitro assay. Therefore, efforts were made to identify synthetics AMPs with enhanced antimicrobial activity.
Twenty-four potentially enhanced synthetic AMPs were designed (Table 2), many as modifications of the D4E1 and D2A21 synthetic peptides, which looked promising in the initial comparison. Many of these variants resulted in negligible changes in MIC or hemolysis, and none resulted in significantly enhanced suppression of the test species. Influences of structural variants on MIC were interesting with some having no effect, some increasing MIC for some but not all test species, and some variants increasing MIC for all three test species (Table 3). Changes in hemolysis showed little relationship to effects on MIC.
By assessing the hydrophobicity/hydrophilicity, charge, and volume of each amino acid in tachyplesin I, new amino acid sequences with similar physiochemical traits were developed.
Figure 1 depicts a proprietary system developed by J. Jaynes of AgroMed to compare and substitute amino acids to create peptides with similar physiochemical characteristics (described in Rajasekaran et al., 2009). Figure 2 shows the structures of the AGM 181 through 184 modeled on tachyplesin I. The design goal was to maximize activity while at the same time achieving negligible toxicity. AMPs where both MICs and hemolysis in vitro are very low would then be those peptides selected for further testing.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 2; 10.21273/JASHS.138.2.142
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 2; 10.21273/JASHS.138.2.142
All newly synthesized AMPs had very similar MICs to tachyplesin I, but AMG181 had markedly higher hemolysis than the others. Similarity of MIC values within this group validates the physiochemical substitution system used in creation of these synthetic AMPs.
Efforts to create improved synthetic AMPs resulted in no increase in activity against target bacteria, as indicated by MIC from the in vitro assays. This may indicate a ceiling for activity inherent in AMPs. However, it is possible that some of these synthetics will have other desirable properties regarding resistance to proteolysis or ability to accumulate in specific tissues in plants and, therefore, merit testing as transgenes. Certainly design and testing of AMPs offer substantial opportunity for better understanding interactions between AMPs and hosts or targets as affected by amino acid sequence.
Most AMPs were very similar in effectiveness across the three species, showing growth suppression at either identical MICs or within one tested concentration increment (AMPs were tested at concentrations increasing ≈3-fold in each adjoining category). However, some bacterial species × AMP interactions were observed. Most notably, apidaecin IA and drosocin were more effective against SM than AT or XCC, showing no suppression of AT or XCC at any concentration tested (the maximum tested was 30 μM) but suppressing SM at 1 to 3 μM. α-purothionin demonstrated much greater effectiveness against XCC than AT or SM. Such species × AMP interactions likely relate to interactions between bacterial membrane structures and motifs in the AMPs displaying specificity (Huang, 2000), and further study may ultimately permit design of AMPs with considerable specificity against targeted organisms.
Although in vitro assays of AMPs are much faster than tests of transgenic genes in planta for species where transgenesis is inefficient, there are also significant limitations. Activity of a gene in a transformed plant can be substantially different from performance in vitro; therefore, in vitro data can provide some guidance on AMP selection but must be considered with caution. Transgenically expressed spinach defensin (SoD2) has shown substantial activity against the diseases ACC and HLB in greenhouse and field trials of citrus (E. Mirkov, unpublished data). However, in our in vitro system, there was no suppression of growth on any of the three test organisms at the highest concentration tested. SoD2 has multiple cysteine disulfide bridges (Segura et al., 1998), which stabilize the secondary structure that is likely critical to antimicrobial activity (Broekaert et al., 1995), and the peptide may not have undergone proper folding and crosslinking when chemically synthesized. Caution should be used in testing synthesized peptides that may require post-translational processing to achieve functional conformations, although many of the AMPs that were quite effective in our assays also include cysteine disulfide bridges (e.g., Kawano et al., 1990). In vitro assessment may provide an initial screen to determine promising AMPs but it may also substantially underestimate activity of peptide sequences requiring post-translational modifications. In addition, very frequently amphipathic AMPs are more effective in planta than in vitro and there is evidence that AMPs in vitro form large aggregates that hinder function, whereas in planta, AMPs likely exist as discrete molecules or much smaller aggregates (J. Jaynes, unpublished data).
D4E1, a 17 amino acid synthetic AMP, which forms a beta sheet (Lucca et al., 1998), was among the most active AMPs in our in vitro assays. D4E1 was active against AT expressed as a transgene in hybrid poplar [Populus tremula × P. alba (Mentag et al., 2003)] and has been used extensively in initial efforts to create HLB-resistant transgenic citrus in the citrus improvement program at the U.S. Horticultural Research Laboratory, Fort Pierce, FL. Despite significant AMP design modifications intended to enhance activity, the lack of increased AMP effectiveness in this report suggests that other approaches may be desirable for enhancing efficacy. There are reports of synergistic effects in combining diverse AMPs (e.g., Yan and Hancock, 2001) and transgenic tobacco (Nicotiana tabacum) simultaneously expressing two different spinach defensins appear to be more effective in suppressing Candidatus Liberibacter solanacearum than either defensin expressed alone (E. Mirkov, unpublished data). In our studies as part of the current report, combining diverse AMPs in in vitro tests was always strictly additive in their effect (data not shown). A combination of AMPs may provide the additional benefit of helping suppress development of pathogens resistant to a single AMP.
Improved transgene expression is sometimes realized by inclusion of a transit sequence to concentrate proteins in critical tissues for activity of the transgenic product and/or with greater protection from proteolysis (summary in Frank et al., 2011). However, for HLB, the causal pathogen is located exclusively in phloem sieve elements (Bové, 2006) and a suitable transit peptide to target these cells is not apparent.
Chimeral peptides that combine AMP motifs with other sequences having high affinity to bacterial membranes may permit markedly greater AMP effectiveness at similar expression levels. Effectiveness of such a chimeral peptide was recently demonstrated to enhance resistance to pierce’s disease (caused by X. fastidiosa) in genetically engineered grapevine [Vitis vinifera (Dandekar et al., 2012)].
These approaches are being implemented in the next generation of citrus AMP transgenics.
Contributor Notes
Thanks to AgroMed and Southern Gardens Citrus of U.S. Sugar for granting permission to publish this article covering material owned and provided by them, to Dr. William Dawson of the University of Florida for providing several custom synthesized AMPs used in the initial screening and to Dr. Robert Shatters, Jr. of USDA/ARS for providing the Sinorhizobium meliloti isolate Sm 1021. Dr. Jaynes designed the peptides for this study while a consultant for AgroMed, LLC.
Corresponding author. E-mail: ed.stover@ars.usda.gov.
