Armet is an effector protein mediatingaphid–plant interactions
Wei Wang,*,† Huaien Dai,‡ Yi Zhang,§ Raman Chandrasekar,‡ Lan Luo,* Yasuaki Hiromasa,‡
Changzhong Sheng,‡ Gongxin Peng,§ Shaoliang Chen,† John M. Tomich,‡ John Reese,{
Owain Edwards,k Le Kang,* Gerald Reeck,‡,1 and Feng Cui*,‡,1
*State Key Laboratory of Integrated Management of Pest Insects and Rodents, Institute of Zoology,Chinese Academy of Sciences, Beijing, China; †College of Biological Sciences and Technology, BeijingForestry University, Beijing, China; ‡Department of Biochemistry and Molecular Biophysics, Kansas StateUniversity, Manhattan, Kansas, USA; §Department of Mathematics, Hebei University of Science andTechnology/Hebei Laboratory of Pharmaceutic Molecular Chemistry, Shijiazhuang, Hebei, China;{Department of Entomology, Kansas State University, Manhattan, Kansas, USA; and kCommonwealthScientific and Industrial Research Organisation Ecosystem Sciences, Centre for Environment and LifeSciences, Floreat, Australia
ABSTRACT Aphid saliva is predicted to contain pro-teins that modulate plant defenses and facilitate feeding.Armet is a well-characterized bifunctional protein in mam-malian systems. Here we report a new role of Armet,namely as an effector protein in the pea aphid, Acyrtho-siphon pisum. Pea aphid Armet’s physical and chemicalproperties and its intracellular role are comparable tothose reported for mammalian Armets. Uniquely, wedetected Armet in aphid watery saliva and in the phloemsap of fava beans fed on by aphids. Armet’s transcriptlevel is several times higher in the salivary gland whenaphids feed on bean plants than when they feed on anartificial diet. Knockdown of the Armet transcript byRNA interference disturbs aphid feeding behavior onfava beans measured by the electrical penetration graphtechnique and leads to a shortened life span. Inocula-tion of pea aphid Armet protein into tobacco leavesinduceda transcriptional response that includedpathogen-responsive genes. The data suggest that Armet is an effec-tor protein mediating aphid–plant interactions.—Wang,W., Dai, Zhang, Y., Chandrasekar, R., Luo, L., Hiromasa,Y., Sheng, C., Peng, G., Chen, S., Tomich, J. M., Reese, J.,Edwards, O., Kang, L., Reeck, G., Cui, F. Armet is an ef-fector proteinmediating aphid–plant interactions.FASEB J.29, 000–000 (2015). www.fasebj.org
NUMEROUS APHID SPECIES aremajor crop pests throughdirectdamage or as vectors of plant viruses (1). Aphids feed onphloem sap from sieve elements while keeping these cells
alive and their sieve-plate pores open (2). Continuousproduction of watery saliva occurs as aphids probe sieveelements and ingest phloem sap from individual sieveelements for hours or even days (3). Saliva is predicted tocontain numerous proteins that overcome plant defenses(4–6) and is thought to deliver effector proteins in analogyto plant pathogens (6–8). However, the functions of in-dividual protein effectors in aphid saliva remains poorlyunderstood.
The identification of individual proteins of aphid salivahas progressed considerably in recent years. Nine proteinssecreted in the saliva of the pea aphid,Acyrthosiphon pisum,5 enzymes in the saliva of the green peach aphid Myzuspersicae, 12 and 7 proteins in the saliva of 2 cereal aphidspecies, Sitobion avenae and Metopolophium dirhodum, re-spectively, havebeen identifiedusingproteomics (9–11).Acatalog of candidate effector proteins from the salivaryglands of the pea aphid has been generated using a com-bined proteomics and transcriptomics approach, 42 ofwhose transcripts were enriched in salivary glands (6).When feddifferent diets, theRussianwheat aphid,Diuraphisnoxia, secretes saliva with qualitative and quantitative dif-ferences in soluble proteins (12).
Several studies on the effects of individual proteins ofsalivahave appeared. ProteinC002 (aphidbase:ACYPI008617;National Center for Biotechnology Information [NCBI]reference sequence [RefSeq]: XP_001948358) was de-tected in pea aphid–infested host plants, and knock-down of its transcript affected the foraging and feedingbehavior of aphids (13). The role of protein C002 as an
1 Correspondence: Feng Cui, State Key Laboratory of In-tegrated Management of Pest Insects & Rodents, Institute ofZoology, Chinese Academy of Sciences, Beijing 100101,China. E-mail: [email protected]. Gerald Reeck, Department ofBiochemistry and Molecular Biophysics, Kansas State Univer-sity, Manhattan, KS 66506, USA. E-mail: [email protected]: 10.1096/fj.14-266023This article includes supplemental data. Please visit http://
effector has been supported by overexpression of greenpeach aphid C002 (RefSeq: EC389283) in Nicotianabenthamiana or Arabidopsis thaliana, which enhancedgreen peach aphid colonization (14, 15), and by feedingaphids on transformed plants producing double-strandedRNA (dsRNA) of protein C002 that induced a contraryeffect (16). Two candidate effectors, Me10 (RefSeq:GAAF01000080) and Me23 (RefSeq: GAAF01000028)from the potato aphid, Macrosiphum euphorbiae, increaseaphid fecundity with their overexpression in N. ben-thamiana (17). A candidate effector, Mp10 (RefSeq:ES225905), from the green peach aphid, was found toinduce chlorosis and local cell death in planta (14). Anunknown salivary gland protein (ACYPI39568; RefSeq:NP_001232999) of A. pisum, belonging to an aphid-specific cysteine-rich protein family, was characterizedfor its physical and chemical properties andpossible rolesin the aphid–plant interactions (18).
Herewe report work on a protein, Armet, not previouslystudied in aphids. Armet is bifunctional in mammals, bothintracellularly, as a component of the unfolded proteinresponse in the lumen of the endoplasmic reticulum (ER)(19, 20), and extracellularly, as a neurotrophic factor. Inthe latter role, the protein is typically referred to as mes-encephalic astrocyte–derived neurotrophic factor (MANF)(21, 22). Mammalian Armet is widely distributed in manyorgans and tissues; it is most highly expressed in tissues ororgans that are characterized by especially active proteinsecretion (23, 24). In invertebrates, published work onArmet (MANF) has been limited to Drosophila (RefSeq:AAF55303), documenting both the neurotrophic role ofthe protein (22) and its function as a unfolded proteinresponse component (25) in that species. We note thatmammals have a MANF homolog called conserved dopa-mine neurotrophic factor (CDNF), which, in contrast toMANF, is not increased by ER stress (20).
Our attention was drawn to Armet for 2 reasons. One isthat the pea aphid salivary glandhas amuchhigher level oftheArmet transcript thandoes the rest of theorganism(6).The other is that pea aphid Armet was predicted to havecalcium-binding ability in an early round of genome an-notation. Calcium-binding proteins in aphid saliva arethought to undermine the calcium-triggered sieve-tubeocclusion mechanism by conversion of proteinaceous“forisomes” in sieve elements of legume plants to a con-tracted state, thuspreventing the forisomes fromoccludingthe sieve tubes (26). Suppression of sieve-tube occlusion byaphid saliva is a general phenomenon, independent ofaphid species, host plant species, and host family (27, 28).
In this study, we explored both the intracellular andextracellular roles of Armet (ACYPI008001, RefSeq:XP_001949541) in the pea aphid, with an emphasis on theextracellular role of Armet as a component of saliva. Ourresults suggest that Armet is an effector protein that medi-ates the compatible interaction between aphid and plant.
MATERIALS AND METHODS
Aphids
Pea aphids were collected from peas (Pisum sativum) in 2010 atYuxi, Yunnan Province, China. Offspring of several female adultswere reared on fava beans (Vicia fabae) in incubators at 216 1°C,606 5% relative humidity, and 16 hour light/8 hour dark pho-toperiod to form a stable clone named YYC. The experimentsconducted at Kansas State University used aphids maintained onfava beans, as described elsewhere (29), with a clone of aphidsderived from the LSR1 clonemaintained byD. Stern at Princetonand used in the pea aphid genome project. The YYC clone wasused in the experiments of injection of ER stress inducer, analysisof aphid survival and feedingbehavior after dsRNA injection.TheLSR1 clone was used in the rest of experiments. When necessary,the developmental stage of aphidswas synchronized by collectingfirst instar aphids, then rearing these aphids at 23°C for 3 days toreach the third instar and 7 days to reach adult stage.
Cloning of aphid Armet transcript, protein sequenceanalysis, homology modeling, and molecularevolution analysis
Total RNA was extracted from 16 pairs of salivary glands of adultaphids using Trizol (Invitrogen, Carlsbad, CA, USA) and thenreverse transcribed into cDNA with SuperScript III first-strandsynthesis system for RT-PCR (Invitrogen). A pair of primers,namely, Armet-ORF-F andArmet-ORF-R (Supplemental Data 1),was designed for amplifying Armet open reading frame (ORF)from the salivary gland cDNA library. The predicted proteinsequence was analyzed with SignalP (http://www.cbs.dtu.dk/services/SignalP) and PSORT (http://psort.ims.u-tokyo.ac.jp)servers for identifying signal peptide. Prediction of possibleO- and N-glycosylation sites was performed online at http://www.cbs.dtu.dk/services/NetOGlyc and http://www.cbs.dtu.dk/services/NetNGlyc, respectively. A phylogenetic tree containinghomologous proteins from 31 organisms with complete genomeswas constructed by the maximum likelihood method. Molecularevolution analysis was carried out using Datamonkey (http://www.datamonkey.org). Positive selection sites were identified with singlelikelihood ancestor counting andfixed effects likelihoodmethods.
Protein expression, purification, and antibody preparation
Thenucleotide sequencesencodingArmet lacking theN-terminalsecretion signal or its 2 domains individually (N-terminal and C-terminal domains) were constructed in the pET28a vector, witha hexa-histidine-tag added at the 39 terminus for purificationconvenience. Recombinant plasmids were used to transformEscherichia coli BL21 (DE3) cells for expression after sequenceverification. The supernatant of the sonicated cells was used forpurification. Also, full-length Armet, including the signal peptide,was inserted in vector pFastBac1 and expressed in the Bac-to-Bacbaculovirus system (Invitrogen) with or without a hexa histidine-tag at the 39 terminus. The resulting plasmids, after sequenceverification, plus the blank vector, were used to generaterecombinant baculoviruses. Spodoptera frugiperda Sf9 cells wereinfested with the recombinant baculovirus. The cell-free mediumand the supernatant of sonicated cells were kept for purification.The primers for our constructions are presented in SupplementalData 1. The expressed recombinant proteins were purified se-quentially with Ni-NTA agarose column (Qiagen, Germantown,MD, USA) and Q-Sepharose Fast Flow column (Sigma-Aldrich,St. Louis, MO, USA), and then treated with Chelex-100 resin(Sigma-Aldrich) to remove the remaining calcium ion. Purified
(continued from previous page)MANF, mesencephalic astrocyte-derived neurotrophic factor;MS, mass spectrometry; NCBI, National Center for Bio-technology Information; ORF, open reading frame; qPCR,quantitative PCR; RefSeq, reference sequence; TCEP, Tris (2-carboxyethyl) phosphine
2 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
protein was served as antigen to produce rabbit anti-Armet poly-clonal antibody inCocalicoBiologicals Inc. (Reamstown,PA,USA).
Molecular weight measurement and determination ofdisulfide bonding pairs by mass spectrometry
Molecular weights of recombinantly expressed proteins weremeasured by matrix-assisted laser desorption-ionization time-of-flight mass spectrometry (MALDI-TOF-MS) with a BrukerUltraflex II (Bruker Daltonics, Fremont, CA, USA) using Brukerprotein molecular weight standard I for molecular weight cali-bration. Armet was subjected to SDS-PAGE in the absence ofa reducing agent. After subjected to alkylation of free cysteineresidues using 100 mM iodoacetamide, the gel sections weretreated with 100 ng of sequencing-grade trypsin or chymotrypsin.Tomeasure thedigestedpeptideprofile for each sample, 1.0ml ofthe peptide solution was dispensed on 1.0 ml of 30 mg/ml 2,5-dihydroxylbenzonic acid (Sigma-Aldrich) onto a stainless steeltarget plate. MALDI-TOF spectra were acquired with a BrukerUltraflex II and selected precursor ions were subjected to TOF/TOF (MS/MS) analysis with laser-induced fragmentation. Thespectra were processed by FlexAnalysis 3.0 and Biotools software3.0. Thegeneratedpeak lists were transferred to ProteinScape1.3software for protein identification. Tris (2-carboxyethyl) phos-phine (TCEP) was applied to do on-target reduction of disulfidebonded cysteines, which were confirmed by the +2 Da shift ofpeptide molecular mass in the spectra.
Measurement of circular dichroism spectra
Circulardichroism spectrawere recorded from190nmto260nmat 25°C on a J-815 spectropolarimeter (Jasco, Tokyo, Japan). Thespectra were an average of 5 consecutive scans at a scan rate of50 nm/min. All spectra were corrected by subtracting the signalfrom the protein-free solutions recorded under the same con-ditions. Data were recorded as the mean residue ellipticity indegrees cm2 dmol21.
Analytical ultracentrifugation
Sedimentationvelocitywasconductedat20°CusinganOptimaXL-Iultracentrifuge (BeckmanCoulter, Fullerton, CA) with an An-60 Tirotor. Recombinant Armet at 0.4 mg/ml was buffered in 20 mMTris-HCland120mMNaClatpH8.0.Sedimentationwasmonitoredat A280 for 120 minutes at 5 minute intervals. Data were analyzedusing DCDT1 software 1.16 (http://www.jphilo.mailway.com).
Calcium binding assay of pea aphid Armet and its 2 domains
The binding of Ca2+ to Armet or its domains was examined byisothermal titration calorimetry using the MCS-ITC system(MicroCal,Northampton,MA) at 30°C.RecombinantArmet, itsN-and C- terminal domains, and CaCl2 were buffered in 20 mMTris-HCl and 120mMNaCl at pH 8.0. Armet (48mM) was titrated with3 mM CaCl2. The N-terminal domain (50 mM) was titrated with3.5 mM CaCl2, and the C-terminal domain (100 mM) was titratedwith15mMCaCl2.Datawerefitted topossiblebindingmodels, andthermodynamicparametersweredetermined fromnonlinear least-squares fitting using Origin ITC software supplied by MicroCal.
Dual luciferase assay
A 427 base region containing an endoplasmic reticulum stressresponse element (ERSE) motif CCAATN9CCAAG was cloned
from the genomic DNA using primers Armet-pro-F and Armet-pro-R (SupplementalData 1). The ERSE region was then deletedor replaced with a 19 base “random” sequence CAGCAGTAA-TACATTTGAA to generate mutant promoters, which were se-quenced for verification. The intact or mutant promoters werecloned into the psiCHECKTM-2 vector (Promega, Madison, WI,USA), allowing simultaneous expression of Renilla and fireflyluciferases, and then the vector was transferred to Drosophila S2cells. The Armet promoter replaced the SV40 early promoter ofthe vector to regulate the expression of Renilla luciferase. Thefirefly luciferase served as control. ER stress inducer tunicamycin(2 and 5 mg/ml) or DTT (2 and 3 mM) was added to the cellcultures. Cells were collected after 24 hours of treatment, and theactivities offirefly andRenilla luciferases weremeasuredusing thedual-luciferase reporter assay system (Promega). Light emissionwas determined by a modulus microplate multimode reader(TurnerBiosystems, Sunnyvale,CA,USA). Fifteen replicateswereprepared and assayed for each treatment. Differences were eval-uated statistically by SPSS 17.0 (IBM, Armonk, NY), either by t testto compare 2 means, or by 1-way ANOVA followed by a Tukey’stest for multiple comparisons.
Injection of ER stress inducer into pea aphids
TheER stress inducer tunicamycinwas injected in the third instarstageof theaphids.Exactly23nlof tunicamycin(about0.5mg/mlin water) was injected in each aphid, which caused 30% to 40%mortality over 3 days of treatment. The mortality of controlaphids that were injected with 23 nl of water was approximately20% over 3 days. Ten surviving aphids as one biologic replicatewere collected on the third day; 5 replicates were conducted.Armet transcript levels in thetreatedaphidswerecomparedwiththose in the aphids injected with water using quantitative PCR(qPCR). RibosomalproteinL27transcript (RefSeq:CN584974)wasamplified as internal control (Supplemental Data 1). qPCRwas per-formedonaRocheLightCycler480(Roche,Mannheim,Germany).Differences were evaluated by t test statistically by SPSS 17.0.
Treating insect cells with ER stress inducers
S2 and Sf9 cells were treated with 2 mM or 5mMDTT (dissolvedin water) and 5 mg/ml or 10 mg/ml tunicamycin (dissolved inDMSO) for 12, 24, or 48 hours. The transcript levels of the Dro-sophila melanogaster Armet and S. frugiperda Armet (RefSeq:DY898745) in S2 and Sf9 cells, respectively, were quantified usingqPCR. The ribosomal protein L27 transcripts of D. melanogaster(RefSeq: NM_143160) and S. frugiperda (RefSeq: AF400191), asinternal controls, were quantified (Supplemental Data 1). Dif-ferences were statistically evaluated by t test by SPSS 17.0. Threebiologic replicates of the cells were collected for assay.
Protein preparation from phloem sap of plantsand aphid saliva
Favabeanplants (13daysold)were infestedwithaphids for2days,thenwashed 3 times inwater. Ten leaves with petioles were cut offandeach immediatelyplaced in2.5mlof20mMEDTA,pH8.0, inonewell of a 12-well tissue culture plate in the dark. After 4 hours,the buffer-containing phloem sap was collected and passedthrough a 0.2 mm filter. Phloem sap from plants not infested byaphids was collected in the same way, as the control. Cooled tri-chloroacetic acid was added to the collected phloem sap to attain15% final concentration and kept at 220°C for at least30 minutes, and then at 4°C overnight. After centrifugation for10minutes at 12,0003 g, the precipitate was washed 3 times withcold acetone and then dissolved in a 1:1 (v/v)mixture of solution
A (phenol, pH8.0) and B (0.1M Tris-HCl, 2% SDS, 30% sucrose,5% b-mercaptoethanol). The supernatant (phenol phase) wasretained after centrifugation, mixed with 5 volumes of 0.1 Mammonium acetate in methanol, and incubated for 30 min at220°C and then at 4°C overnight. The precipitate was collectedby centrifugation, washed 3 times with cold acetone, and solubi-lized in SDS-PAGE loading buffer.
To collect aphid saliva, about 1000 aphids were reared on arti-ficial diet (30) for 2 days at room temperature under a 16 hourphotoperiod. The diet was sealed between 2 layers of Parafilm inapetriplate (150315mm). Saliva-containingdiet from8000aphids(from 8 petri plates) was collected after feeding, passed through0.2 mm filter, and dialyzed against 20 mM Tris-HCl, pH 8.0, over-night. Protein was precipitated with 20% trichloroacetic acid. Afterwashing 3 times with cold acetone, the precipitate was dissolved inSDS-PAGE loading buffer. Diet that had not been fed on was pro-cessed in the same way as control. The gelling saliva was collectedandproteins in the gelling salivawere viewed in a silver-stainedSDS-PAGE gel following the method presented by Will et al. (31).
Western blot analysis and immunohistochemistry
InvitrodetectionofArmetprotein in thephloemsap, aphidsaliva,and aphid heads (containing salivary glands) was performedwithWestern blot analysis using purified anti-Armet polyclonal anti-body and X-ray exposure method. Immunohistochemistry wascarried out to localize in vivo the expression of Armet in aphidsalivary glands. Detailed experimental steps have been describedin a previous study (13).
Comparison of Armet or C002 transcript in diet-fedand plant-fed pea aphids
Adult pea aphids were placed on a healthy fava bean leaf insertedinto sterilized 2% agar [supplemented with 0.1%miracle growthfertilizer (Miracle-Gro, Marysville, OH, USA) and 0.03% methyl4-hydroxybenzoate] in a petri plate for feeding for 24 hours.Another group of adult aphids were transferred to an artificialdiet (30) in a sachet for feeding for 24 hours. The transcript levelsofArmet orC002 in theheads (containing salivary glands) of diet-fed and plant-fed pea aphids were compared through qPCR.Actin transcript (RefSeq: NM_001142636) was amplified as in-ternal control (Supplemental Data 1). Forty-five to 50 aphidheads were used for a measurement, and 3 replicates were per-formed. Differences were evaluated by t test using SPSS 17.0.
dsRNA synthesis and delivery
A 286 and 439 bp dsRNAof Armet and greenfluorescent protein(GFP), respectively, were generated using T7 RiboMAX ExpressRNAi System (Promega) and purified using Wizard SV Gel andPCR Clean-Up System (Promega) following the manufacturer’sprotocol. Injection of 23 nl of dsRNAs at 6 mg/ml was performedon the third instar stage of aphids. The dsRNAs were deliveredinto hemolymph from dorsal abdomens by microinjectionthrougha glass needle at slow speed usingNanoliter 2000 (WorldPrecision Instruments, Sarasota, Florida, USA). Six groups ofaphids and 15 individuals in each group were injected and thenreared on fava bean plants or an artificial diet for 9 days (31).Survival rates were recorded every day and reported as means6SE. The survival curves of interference and control groups werecompared for statistical differences using the log-rank (Mantel-Cox) test (SPSS 14.0). Four groups of aphids and 5 aphids in eachgroup were collected on the first and third days after injection ofdsRNA for testing inhibition efficiency of Armet transcription inaphid whole bodies or in heads (containing salivary glands) and
guts using qPCR. The change in protein level of Armet in heads(containing salivary glands) at the third day after injection ofdsRNA was examined by Western blot analysis.
Analysis of aphid feeding behavior
The electrical penetration graph (EPG) technique was used toanalyze the feeding behavior of individual aphids on fava beanplants according to a procedure previously described (13). Thefeeding behavior of dsGFP-RNA-injected aphids and dsArmet-RNA-injected aphids was monitored continuously for 8 h on thethird day after injection. Valid data sets from 23 and 29 aphidswere collected for the dsGFP-RNA and dsArmet-RNA groups,respectively. The EPG waveforms were recorded by a Giga am-plifier series, GIGA-8 model (EPG-Systems, Wageningen, TheNetherlands) and analyzed with the software Stylet+ according tothe manufacturer’s manual and the information from Tjallingii(5). Differences between the 2 groups were analyzed statisticallyby independent-sample t test using SPSS 17.0.
Armet protein injection from leaf petiole
Approximately 2.5 mg of purified pea aphid Armet, recombi-nantly expressed inE. coli, in50ml buffer (20mMTris-HCl, 120mMNaCl, pH 8.0) was injected through leaf petiole in the abaxialside of the leaves of 1moold tobaccoN. benthamianausing a 1mlsterile syringe and a 0.4 3 13 RWLB needle (Shanghai MisawaMedical Industry, Shanghai,China). Anequal volumeofpurifiedproduct from pET28a empty vector was injected as control. Theleaves from the treatment group and the control group werecollected for RNA extraction after 60 hours.
Transcriptomic analysis of N. benthamiana
One-month-old tobacco N. benthamiana was used for infiltrationof purified pea aphid Armet protein that has been expressedrecombinantly in E. coli. Five micrograms of pea aphid Armet in100 ml of buffer (20 mM Tris-HCl, 120 mM NaCl, pH 8.0) wasinfiltrated in 1 leaf, and 2 leaves were treated in each plant.Control leaves were infiltrated with 100 ml of buffer. Eight leavesfrom the treatment group and the control group were collectedfor RNA extraction after 60 h. Total RNA was sent to BGIShenzhen, China, for RNA-Seq analysis using the Illumina HiSeq2000 sequencer. At least 10 million 49 bp clean reads wereobtained for each sample. High-quality reads were mapped toN. benthamiana gene set using SOAP2. The reads per kilobase oftranscript per million reads mapped values are used for com-paring the difference in transcript levels between samples. A falsediscovery rate of #0.001 and the absolute value of log2 ratio $1were used as a threshold to judge the significance of the geneexpression difference. Unigenes were assigned to Kyoto Encyclo-pedia of Genes and Genomes pathways. Twenty up-regulated genes(with initials of Nb in the N. benthamiana gene set) of the plant–pathogen interaction pathway were verified with qPCR in thepetiole injection groups of Armet or the pET28a empty vectorcontrol (Supplemental Data 1). The elongation factor 1a(NbS00007372g0013) was quantified as an internal control. Sixto 8 replicates and 8 leaves in each replicate were prepared forqPCR verification.
RESULTS
Armet gene, transcript, and predicted protein inpea aphid
The Armet gene is located in scaffold 17420 of the peaaphid genome. By aligning with Armet expressed
4 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
sequence tags, the gene is found to have 4 exons (of 94,122, 142, and 167 bases) partitioned by 3 introns (of 114,156, and 79 bases). A 58-base intron is observed in the 59UTR, separating this region into 130- and 44-base sections(Fig. 1A). A putative cis-acting ERSE, CCAATN9CCAAG,occurs 76 bases upstreamof the transcription start site (Fig.1A), and differs only at one nucleotide (the underlined A)from the canonical ERSE sequence inmammals (32). TheORF of the Armet transcript contains 522 bases and en-codes a protein of 174 amino acid residues, 8 of which arecysteines (Fig. 1B). SignalP and PSORT predict an N-terminal signal secretion peptide with cleavage predictedeitherbetweenAla20 andGln21 (SignalP)orbetweenSer22
and Arg23 (PSORT). The protein’s sequence includesKEEL at its C-terminus, a putative ER retention signal, andno predicted O- or N-glycosylation sites.
Molecular evolution of Armet
Using BLAST analysis, transcripts homologous to peaaphid Armet were identified inmany animal species, from
invertebrates to mammals. A related protein, CDNF,occurs in vertebrates, whereas only Armet exists in inver-tebrates (Fig. 2). Common features of Armet acrossinvertebrates and vertebrates include a secretory signalpeptide at theN-terminus, a KDEL-like ER retentionmotifat the C-terminus, and 8 cysteines (Supplemental Data 2).For Armet, no positive selection sites with P , 0.05 wereidentified, but we found positive selection in the codonencodingLeu62 inCDNF(SupplementalData2), inwhichthe dN-dS value was 2.36 (P = 0.08) based on the singlelikelihood ancestor counting method and 2.24 (P = 0.02)based on the fixed effects likelihood method.
Characterization of recombinantly expressed peaaphid Armet
Armet (excluding the signal peptide) was expressed inE. coli andpurified, as were the protein’s N- andC-terminaldomains individually (Supplemental Data 3A, B). The cir-cular dichroism spectrum of pea aphid Armet exhibitedminima at 208 and 222 nm (characteristic of high helix
Figure 1. Nucleotide sequence of pea aphidArmet gene region. A) 59-Upstream region. Thetranscription start site is marked with an arrow;ERSE is enclosed in a box. The 58 base intron isunderlined. B) The 522-base ORF with theencoded amino acid sequence. The amino acidresidues of the signal secretion peptide are inred. The KDEL-like ER retention motif isunderlined by double strands. The linker be-tween domains is enclosed in brackets. Eight con-served cysteines are marked in green.
content) and was unaffected by the presence of calcium(Fig. 3A). Analysis of sedimentation velocity revealed thatpea aphidArmet exists as amonomerwith a sedimentationcoefficient of 1.70 S (Fig. 3B). The frictional coefficientratio (f/f0 = 1.45) fits a somewhat extended globularstructure (Fig. 3C).
Thermodynamic properties of calcium binding to peaaphid Armet were determined by isothermal titration cal-orimetry. The titration data fit best to a 1-site model—thatis, tobinding at a 1:1molar ratio (Fig. 3D). Thedissociationconstant Kd was estimated to be 240 6 12 mM. Armet’sbinding of Ca2+ was driven mostly by a change in enthalpy(DH° =25.96 0.2 kcal/mol) rather than entropy (TDS° =0.9kcal/mol).Wheneither theN-orC-terminaldomainorthe 2 domains, mixed, were titrated with Ca2+ under thesame conditions, no binding was observed.
No suggestion of intermolecular disulfide bonding be-tween Armet molecules was observed from the SDS-PAGEperformed in theabsenceofmercaptoethanol, andno freecysteine residue was detected by MS analysis. Three disul-fide bonds, namely Cys82Cys11, Cys822Cys93, andCys1272Cys130, were identified after digestion of theprotein with trypsin, and a fourth disulfide bond, Cys392Cys51, was identified in chymotrypsin digests (Fig. 3E–H).
Response of Armet from 3 insect species to ER stress
Transcription driven by the promoter of the pea aphidArmetgenewasexamined inDrosophilaS2cellsusingadualluciferase assay. Transcription was significantly increasedby the treatmentof cellswitheitherDTT(a reducingagent)or tunicamycin (an N-glycosylation inhibitor), whereas nosignificant increase in transcriptional activity was observedafter deletion of ERSE or its replacement with a randomsequence (Fig. 4A).
An aphid cell line is not available, so we used insect celllines S2 and Sf9 to examine their Armet transcript levelswith and without ER stress. Normalized levels of DrosophilaArmet transcript in S2 cells ranged from about 5-fold toover 15-fold higher in the presence of DTT (Fig. 4B), andfrom about 2-fold to 3-fold higher in the presence oftunicamycin (Fig. 4C). In Sf9 cells, the increase in nor-malized Spodoptera Armet transcript levels was somewhatless than 4-fold in the presence of DTT (Fig. 4D) and up to10-fold in the presence of tunicamycin (Fig. 4E).
To impose ER stress throughout aphids, tunicamycinwas injected into pea aphids at a level that caused30% to 40% mortality over 3 days of treatment. After 3days, the normalized transcript level of the pea aphid
Figure 2. Phylogenetic tree for Armet and CDNF from invertebrates and vertebrates. The tree was constructed using themaximum likelihood method based on the alignment of amino acid sequences. The bar indicates branch scale and geneticdistance. The RefSeq numbers in NCBI are indicated.
6 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
Figure 3. Biophysical and biochemical characterizations of pea aphid Armet. A) Far-ultraviolet circular dichroism spectra of peaaphid Armet in 20 mM Tris-HCl, 120 mM NaCl (pH 8.0), and EDTA and CaCl2 as indicated. B) Sedimentation velocity profiles ofanalytical ultracentrifugation of pea aphid Armet in 20 mM Tris-HCl, 120 mM NaCl (pH 8.0). Scans at A280 across the cell at5 min intervals are shown. Direction of sedimentation is to the right. These are the primary data from the measurement. C) Time-derivative analysis of the sedimentation velocity profiles using DCDT1 software (see Methods). The y-axis is the time derivative ofthe protein concentration, in units of ABS/s, as a function of sedimentation coefficient, s*, in units of svedbergs. The solid curveis the result of a least-squares fit to the data for a single, homogenous species with a sedimentation coefficient of 1.70 S. Theexcellent fit to the time-derivative data indicates that the protein is monodisperse, without tendency to aggregate. D) Isothermaltitration calorimetry of pea aphid Armet with Ca2+. At top is heat release with incremental addition of CaCl2 to recombinant peaaphid Armet; at bottom are data fitted with best-fit single-site binding model. E) MS profile of TFTEEDCPVCVLTIDK peptide bytryptic digestion. F) MS profile of CLSTKIDKEKRLCY peptide by chymotryptic digestion. G) MS profile of ICERLKKMDAQVCDIK
(continued on next page)
ARMET MEDIATES APHID–PLANT INTERACTIONS 7
Armet had increased from 0.040 6 0.003 (SE) to 0.077 60.008 (SE) (P, 0.05).
Pea aphid Armet is secreted
When full-length Armet, including the signal secretionpeptide, was expressed in Sf9 cells, with or without a His tagat theC-terminus,Armetwasdetected inboth cell lysate andthe culture medium, indicating that some Armet protein issecreted (Fig. 5A). The molecular mass of the secretedArmet (carrying aHis tag at theC-terminus) in themediumof Sf9 cells was found to be 18,556 Da (Supplemental Data3C,D),which is consistentwithcleavageof the signalpeptidebetweenSer22 andArg23.MALDI-TOF-MSanalysis of trypticpeptides revealed the peptide RTFTEEDCPVCVLTIDK,thus confirming that cleavage site (Fig. 5B).
Immunohistochemistry using anti–pea aphid Armetpolyclonal antibody revealed that Armet occurs preferen-tially in a symmetrically disposed pair of large secretorycells in middle of the 2 lobes of principal salivary glands
(Fig. 5C). Knockdown of Armet transcription by dsRNAgreatly reduced the immune signal in those middle largesecretory cells (Fig. 5C).Using immunoblotting, a 17.7 kDaband, the size of secreted Armet, appeared in watery saliva(800 aphids), head extract (10 aphids), and phloem sap offava beans fed on by aphids (800 aphids), but not inphloem sap from fava beans not exposed to aphids orgelling saliva (800 aphids) (Fig. 5D).
We conclude that pea aphid Armet is a component ofaphid watery saliva and is secreted into fava beans duringthe feeding process.
Armet promotes pea aphid feeding on plants
The transcript levels of Armet and of protein C002,a protein that is required for the feeding of pea aphid onfava bean plants (13), were measured in RNA isolatedfrom aphid heads (containing salivary glands). The nor-malized level of the Armet transcript was about 3.5-foldhigher in plant-feeding aphids compared with the level
peptide by tryptic digestion. H) MS profile of ILDNWGEICDGCLEK peptide by tryptic digestion. The proteolytic peptide ofArmet was subjected to MALDI-TOF-MS analysis before (top) and after (bottom) reduction with TCEP. The shift of 2 amu massafter reduction with TCEP indicates addition of 2-H atoms. Cysteines are underlined.
Figure 4. Response of Armet from 3 insects to ER stress. A) Dual luciferase reporter gene for the analysis of transcription drivenby pea aphid Armet promoter with intact, deleted, or replaced ERSE region in S2 cells. The activity of the Renilla luciferase isdivided by the activity of the firefly luciferase. This ratio of luciferase activities is obtained at different DTT and tunicamycinconcentrations as indicated, then divided by the same ratio in the absence of DTT and tunicamycain (i.e., the control ratio ofluciferase activities). This value, compared with the control value, is reported as mean 6 SE. *P , 0.05 for the difference betweenthe treatment and the control. Letters above bars indicate results of multiple comparisons among the 3 types of promoters. B–E)Transcript level ratios of Drosophila Armet and Spodoptera Armet in S2 cells or Sf9 cells after the treatment of cells from 12 to 48 hwith DTT or tunicamycin relative to control (in DTT-free cells or in DMSO-treated cells). Drosophila and Spodoptera Armettranscript levels were measured by real-time PCR and normalized with the transcript level of respective ribosomal protein L27.Ratios are reported as mean 6 SE. *P , 0.05, **P , 0.01 for the difference between the treatment and the control.
8 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
in diet-fed aphids (Fig. 6A). The normalized level ofthe protein C002 transcript was 6-fold higher in plant-feeding aphids comparedwith that in diet-fed aphids (Fig.6A). Both values suggest a major increase in synthesis ofproteins of saliva when aphids are moved from liquid dietto plants.
When dsArmet-RNA was injected into pea aphids, theArmet transcript level in the aphid whole bodies was re-duced by 39 6 10% and 34 6 6% at 1 and 3 days afterinjection, respectively (Fig. 6B). The knockdown was ashigh as 48% in aphid heads (containing salivary glands),while no change was observed in aphid guts 3 d after in-jection (Fig. 6C).Adecrease inArmetproteinwasobservedin theheads (containing salivary glands) of theknockdowninsects (Fig. 6D). Injection of dsArmet-RNA significantly
reduced the life span of aphids feeding onhost plants (P,0.001) (Fig. 6E) but did not influence that of aphidsfeeding on artificial diet (P = 0.08) (Fig. 6F).
Knockdown of the Armet transcript led to several sta-tistically significant changes in the feeding behavior of peaaphids on fava bean plants as monitored by the EPGtechnique (Fig. 6G). Most notably, the total (average)length of E2 (passive ingestion from sieve elements) wasreduced to 50 min from 190 min in control insects (Sup-plemental Data 4). Furthermore, the dsArmet-injectedaphids showed ahigher frequency of E1 (watery salivation)resurgence and E2/E1 transition than the dsGFP-injectedaphids, as illustrated in Fig. 6G, where 4 E2/E1 transitionswereobserved in thedsArmet-injectedaphid.This suggeststhat aphids have to secretemore saliva for phloem feeding
Figure 5. Pea aphid Armet is secreted. A) Western blot assay of pea aphid Armet expressed in insect Sf9 cells using anti–pea aphidArmet polyclonal antibody. B) MALDI-TOF-MS spectrum of Sf9-expressed pea aphid Armet in cell medium after trypsindigestion. The peptides in red are confirmed further with MALDI-TOF-TOF. C) Immunohistochemical localization of Armet inpea aphid salivary glands (YYC clone) using purified anti–pea aphid Armet antibody before or after dsArmet-RNA injection onthe third day. Red is the positive signal. Nuclei are shown in blue. PG, principal gland; AG, accessory gland. Negative control isdepleting the anti-Armet antibody by pretreatment with recombinant Armet for the immunohistochemistry. D) Western blotanalysis of Armet in aphid saliva (LSR1 and YYC clones) and plant phloem sap fed by aphids using anti–pea aphid Armetpolyclonal antibody. 1, aphids’ Armet expressed in the medium of insect Sf9 cells; 2, aphids’ gelling saliva collected from artificialdiet fed by aphids; 3, aphids’ watery saliva collected from artificial diet fed by aphids; 4, aphids’ heads; 5, phloem sap of fava beanfed by aphids; 6, phloem sap of fava bean without aphids’ infestation; 7, artificial diet without aphids’ infestation.
with decreased levels of Armet in saliva. Thus, Armet isneeded to sustain phloem feeding to normal duration.Correspondingly (and necessarily, given the time-limiteddesign of the experiments), the combined lengths of thenonprobing and pathway components were increased inthe knockdown insects (Supplemental Data 4).
Pea aphid Armet induces antipathogen reaction intobacco plants
We infiltratedE. coli recombinantly expressed andpurifiedpea aphid Armet into leaves of tobacco, N. benthamiana,a nonhost plant but the species that is best established for
Figure 6. Armet promotes aphid feeding on plants. A) Armet and protein C002 transcript levels in artificial diet-fed and favabean-fed pea aphids (LSR1 clone) that were normalized with actin transcript. Values are reported as mean 6 SE. ***P , 0.001. B,C) The Armet transcript levels in aphid whole bodies, in heads that contain salivary glands, or in guts at the third day after dsRNAinjection (YYC clone). *P , 0.05. Interference ratios are indicated above the bars. D) Western blot analysis of Armet in aphidheads (YYC clone) that contain salivary glands at the third d after dsRNA injection using anti–pea aphid Armet polyclonalantibody. E, F) Survival graphs of aphids (YYC clone) on fava beans or on artificial liquid diet after injection of dsArmet-RNA ordsGFP-RNA. G) Representative EPG waveforms of dsArmet-injected or dsGFP-injected (as control) aphids (YYC clone) on favabeans. np, nonprobing; pd, potential drop; C, pathway phase; E1, watery salivation; E2, passive ingestion.
10 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
infiltration experiments (33). The transcriptome of to-baccowas comparedbetweenArmet-infiltratedandbuffer-infiltrated leaves. Compared with the latter, 144 genes inthe plant–pathogen interaction pathwaywere up-regulatedin the Armet-infiltrated plants. These genes were mainlycategorized as kinases (45 genes), WRKY transcriptionfactors (28 genes), calcium-binding proteins (11 genes),anddisease resistanceproteins (6 genes), alongwith a largegroup of uncharacterized proteins (31 genes). qPCR wasperformed to verify the expression levels of 20 genes in theplant–pathogen interaction pathway when the purifiedproduct from pET28a empty vector was used as negativecontrol. Among the 20 genes, 18 were confirmed to be up-regulated in the Armet-delivered group relative to thepET28a empty vector group (Fig. 7).
DISCUSSION
Aphids deliver proteins in their saliva to their host plantsduring feeding. Although numerous proteins of aphid sa-liva have now been identified (6, 9, 10), we are still in theearly stages of studying individual proteins to unravel theirmodes of action. Many such studies on individual proteinswill be needed in order to developmeaningful insight intothe molecular basis of aphid–plant interactions.
Here we report work on the intracellular and extracel-lular roles of the protein Armet, which occurs in saliva ofthe pea aphid. In a broad sense, pea aphid Armet appearssimilar tomammalianArmet in that theprotein servesbothintracellular and extracellular roles in the aphid and inmammals. We present a physicochemical characterizationof the aphid protein and demonstrate that its promoter isresponsive to ER stress. Our greater emphasis, however, ison a novel extracellular role of the protein as a secretedeffector protein that facilitates successful aphid feeding onhost plants. Interference with Armet expression in theaphid undermined the compatible interaction betweenaphid and plant.
Our focus here on the role of Armet in saliva is notintended to imply that the extracellular, neurotrophic roleof Armet is absent in aphids. It is clear from gene knockoutstudies that the protein (under its alternative name,MANF) acts on the nervous system in another insect,namely D. melanogaster (22). Although we have no directevidence for a neurotrophic role of Armet in the peaaphid, it can be assumed that such a role exists. Extrap-olations from the results on Drosophila to the pea aphidsystem would be hazardous at best, not only because of theevolutionary distance between flies and aphids, but alsobecause of the very different methods used (gene knock-out in Drosophila, transcript knockdown in pea aphid) anddifferent developmental stages studied (embryogenesis inDrosophila, adult aphids in our work). We believe the mostlikely interpretation of our results is that the effects ofArmet transcript knockdown are due to decreased pro-duction of Armet as a component of saliva, based partly oncomparable results (in altered feeding and reduced lifespan) obtained by transcript knockdown for another ef-fector protein in pea aphid, protein C002 (13, 29).
Protein C002 is perhaps the most clearly identifiedeffector-protein in aphids (13–16, 29), and it is thereforeinteresting to compare some of the results reported here
on Armet with earlier results on protein C002. Both pro-teins occur in subsets of secretory cells of the principalsalivary glands, but those subsets are different for proteinC002 and Armet. Armet occurs in a pair of large secretorycells in middle of the 2 lobes of principal salivary glands,while protein C002 is present in several other secretorycells (13). We suggest that a preferential expression ofa protein in a subset of secretory cells of the salivary glandmay be a hallmark of saliva proteins. As regards the effectsof transcript knockdown, in each case, the life span ofknockdown aphids feeding on fava beans is reducedcompared to control insects, but the effect is lessmarked inthe case of Armet. Correspondingly, there are differencesin theeffects of transcript knockdownon feedingbehavior,as shown by EPG. Both the total duration of passive in-gestion and the total duration of pathway phase are af-fected in the Armet-knockdown insects, whereas only thetotal duration of passive ingestion is affected in proteinC002–knockdown insects, leading essentially, in that case,to a great reduction in sustained phloem feeding in theknockdown insects. Thus, the role of Armet as effectorappears to be significantly different from the role of pro-tein C002 as effector.
An important tool in this work is knockdown of theArmet transcript in the pea aphid salivary gland by in-jection of dsArmet RNA. We have previously obtained ef-ficient knockdown of the transcripts encoding proteinC002 and the cysteine-rich protein ACYPI39568 by thesamemethod (13, 18, 29). It appears that dsRNA injectioninto the hemolymph of the pea aphid is quite effective intranscript knockdown in the salivary gland and consider-ably less effective in other organs—for instance the gut,which also has relatively high Armet transcript levels (Fig.6C) but did not incur a statistically significant knockdown.Our interpretation of the phenotype resulting fromknockdown of the Armet transcript (i.e., altered EPG pat-terns, reduced life span) is that these aremost likely due totranscript knockdown in the salivary glands.
As do other plant pathogens or parasites, aphids mustsecrete bioactive effector proteins into host plants tomodulate the plant defense response and allow successfulfeeding (14). These effectors are almost certainly in-troduced into the plant as a part of the aphid saliva, andplant responses typical of pathogen effectors (e.g., chloro-sis, necrosis) have been associated with some proteins ofaphid saliva (14, 17). Plant defense responses to aphidfeeding have been characterized in several systems andoften involve induction of elements of the salicylic acid,jasmonic acid, and ethylene defense response pathways(34–36). Jasmonic acid appears to be most important inregulating effective defense responses in incompatibleaphid–plant interactions (37, 38). It has been demon-strated that some components of aphid saliva elicit localdefenses even in compatible interactions (39), and henceother components of aphid saliva must act to effectivelyneutralize these plant defenses (40). Infiltration of peaaphid Armet into tobacco, a nonhost, induced a transcrip-tional response that included many genes associated withpathogen infection, suggesting that Armet’s function mayinduce rather than suppress defenses. One possible way ofinterpreting these results is to imagine Armet (in a hostplant) as being a double-edged sword—on the one hand,an agent that is useful or required to sustain feeding from
ARMET MEDIATES APHID–PLANT INTERACTIONS 11
sieve elements, but on the other hand, an agent that alsostimulates plant defense mechanisms. However, it is alsopossible that Armet might elicit plant defenses in a non-host even if its function in a compatible plant is to suppress
host defenses. Additional functional work is required toclarify Armet’s effector function.
One of the important phloem-related mechanisms ofplants that aphids have to overcome is an occlusion
Figure 7. Real-time qPCR verification of the transcript levels of 20 plant–pathogen interaction genes of Nicotiana benthamianawhen Armet was delivered by petiole injection. The purified product from pET28a empty vector was used as negative control. *P ,0.05; **P , 0.01.
12 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
response to injury in the sieve tubes. It has been shown thatcalcium binding by components of aphid saliva may un-dermine a calcium-triggered occlusion mechanism (26).We have shown here that pea aphid Armet binds calcium,despite the absence of any recognized calcium-bindingmotif in its sequence. Thus, we should consider the possi-bility that Armet might conceivably take part in counter-acting the calcium-triggered occlusion mechanism inplants. However, any physiologically significant effect ofcalciumbinding byArmet (or, for thatmatter, by any othercalcium-binding protein in phloem sap) would have to behighly localized because the molar concentrations of in-dividual aphid proteins in phloem sap are undoubtedlyvery low, and thus the effect of calcium binding by sucha protein on global calcium concentration in phloem sapwould be negligible. Such localization would result fromspecific interactions betweena calcium-bindingprotein andother components (presumably proteins) in sieve elements.
A potentially important aspect of our studies on peaaphid Armet is the pairing of cysteine residues in the pro-tein’s disulfide bonds. Actually, this is an issue that is al-ready represented, but has never been discussed, in theliteratureon the structureofmammalianArmet. Studies ofthe3-dimensional structuresofhuman(RefSeq:NP_006001)andmammalianArmets (41–43) report disulfidebonds inthe following pairings: Cys8-Cys93, Cys11-Cys82, Cys39-Cys51, and Cys127-Cys130. (For convenience, we use thenumbering for the pea aphid Armet sequence, withArg ofthe sequence RTFTEED taken as residue 1 of the matureprotein.) It is interesting tonote, however, that aMS-basedapproach on mouse Armet (RefSeq: NP_083379) (19)revealed 2 differences from the pairings listed above—namely, the existence of Cys8-Cys11 and Cys82-Cys93.These 2 pairs are in fact what we have found here forpea aphid Armet, also using an MS approach. Bothapproaches—the elucidation of Armet’s 3-dimensionalstructure and MS of Armet peptides—are valid; neithersupplants or invalidates the other as regards the disulfidebonding pattern. We therefore hypothesize that Armet,whether mammalian or insect in origin, has alternativedisulfide arrangements in a portion of the N-terminaldomain, both of which exist simultaneously in a pop-ulation of Armetmolecules. Cys8, Cys11, Cys82, andCys93are all clustered in the 3-dimensional structure of Armet,and both disulfide pairings that have been reported areentirely feasible spatially. It seems that the differentmethods of identifying the disulfidepairings in this regionpreferentially detect one of the 2 forms. The structuredeterminations would emphasize the major pairings, andthe MS approach could detect the minor pairings prefer-entially. The hypothesis of alternative disulfide pairingsclustered in the N-terminal domain, if validated, could beimportant in understanding the function of Armet in-tracellularly, extracellularly, or both.
This work was supported by grants from the StrategicPriority Research Program of the Chinese Academy ofSciences (Grant XDB11040200), the Major State BasicResearch Development Program of China (973 Program;Grant 2012CB114102), the Cooperative Research Centre forNational Plant Biosecurity (Australia), the Grains Researchand Development Corporation (Australia), and the KansasAgricultural Experiment Station (publication 15-035-J).
REFERENCES
1. Blackman, R. L., and Eastop, V. F. (1984) Aphids on the World’sCrops: An Identification and Information Guide, John Wiley & Sons,New York
2. Tjallingii, W. F., and Esch, T. H. (1993) Fine structure of thestylet route in plant tissues by some aphids. Physiol. Entomol. 18,317–328
3. Cherqui, A., and Tjallingii, W. F. (2000) Salivary proteins ofaphids, a pilot study on identification, separation and immu-nolocalisation. J. Insect Physiol. 46, 1177–1186
4. Miles, P. W. (1999) Aphid saliva. Biol. Rev. Camb. Philos. Soc. 74,41–85
5. Tjallingii, W. F. (2006) Salivary secretions by aphids interactingwith proteins of phloem wound responses. J. Exp. Bot. 57,739–745
6. Carolan, J. C., Caragea, D., Reardon, K. T., Mutti, N. S., Dittmer,N., Pappan, K., Cui, F., Castaneto, M., Poulain, J., Dossat, C.,Tagu, D., Reese, J. C., Reeck, G. R., Wilkinson, T. L., andEdwards, O. R. (2011) Predicted effector molecules in thesalivary secretome of the pea aphid (Acyrthosiphon pisum): a dualtranscriptomic/proteomic approach. J. Proteome Res. 10,1505–1518
7. Jones, J. D. G., and Dangl, J. L. (2006) The plant immune system.Nature 444, 323–329
8. Hogenhout, S. A., Van der Hoorn, R. A. L., Terauchi, R., andKamoun, S. (2009) Emerging concepts in effector biology ofplant-associated organisms. Mol. Plant Microbe Interact. 22,115–122
9. Harmel, N., Letocart, E., Cherqui, A., Giordanengo, P.,Mazzucchelli, G., Guillonneau, F., De Pauw, E., Haubruge, E.,and Francis, F. (2008) Identification of aphid salivary proteins:a proteomic investigation of Myzus persicae. Insect Mol. Biol. 17,165–174
10. Carolan, J. C., Fitzroy, C. I. J., Ashton, P. D., Douglas, A. E., andWilkinson, T. L. (2009) The secreted salivary proteome of thepea aphid Acyrthosiphon pisum characterised by mass spectrometry.Proteomics 9, 2457–2467
11. Rao, S. A. K., Carolan, J. C., and Wilkinson, T. L. (2013)Proteomic profiling of cereal aphid saliva reveals both ubiquitousand adaptive secreted proteins. PLoS ONE 8, e57413
12. Cooper, W. R., Dillwith, J. W., and Puterka, G. J. (2010) Salivaryproteins of Russian wheat aphid (Hemiptera: Aphididae).Environ. Entomol. 39, 223–231
13. Mutti, N. S., Louis, J., Pappan, L. K., Pappan, K., Begum, K.,Chen, M.-S., Park, Y., Dittmer, N., Marshall, J., Reese, J. C., andReeck, G. R. (2008) A protein from the salivary glands of the peaaphid, Acyrthosiphon pisum, is essential in feeding on a host plant.Proc. Natl. Acad. Sci. USA 105, 9965–9969
14. Bos, J. I. B., Prince, D., Pitino, M., Maffei, M. E., Win, J., andHogenhout, S. A. (2010) A functional genomics approachidentifies candidate effectors from the aphid species Myzuspersicae (green peach aphid). PLoS Genet. 6, e1001216
15. Pitino, M., and Hogenhout, S. A. (2013) Aphid protein effectorspromote aphid colonization in a plant species-specific manner.Mol. Plant Microbe Interact. 26, 130–139
16. Pitino, M., Coleman, A. D., Maffei, M. E., Ridout, C. J., andHogenhout, S. A. (2011) Silencing of aphid genes by dsRNAfeeding from plants. PLoS ONE 6, e25709
17. Atamian, H. S., Chaudhary, R., Cin, V. D., Bao, E., Girke, T., andKaloshian, I. (2013) In planta expression or delivery of potatoaphid Macrosiphum euphorbiae effectors Me10 and Me23 enhancesaphid fecundity. Mol. Plant Microbe Interact. 26, 67–74
18. Guo, K., Wang, W., Luo, L., Chen, J., Guo, Y., and Cui, F. (2014)Characterization of an aphid-specific, cysteine-rich proteinenriched in salivary glands. Biophys. Chem. 189, 25–32
19. Mizobuchi, N., Hoseki, J., Kubota, H., Toyokuni, S., Nozaki, J.,Naitoh, M., Koizumi, A., and Nagata, K. (2007) ARMET is asoluble ER protein induced by the unfolded protein response viaERSE-II element. Cell Struct. Funct. 32, 41–50
20. Apostolou, A., Shen, Y., Liang, Y., Luo, J., and Fang, S. (2008)Armet, a UPR-upregulated protein, inhibits cell proliferation andER stress-induced cell death. Exp. Cell Res. 314, 2454–2467
21. Petrova, P., Raibekas, A., Pevsner, J., Vigo, N., Anafi, M., Moore,M. K., Peaire, A. E., Shridhar, V., Smith, D. I., Kelly, J., Durocher,Y., and Commissiong, J. W. (2003) MANF: a new mesencephalic,
astrocyte-derived neurotrophic factor with selectivity for dopa-minergic neurons. J. Mol. Neurosci. 20, 173–188
22. Palgi, M., Lindstrom, R., Peranen, J., Piepponen, T. P., Saarma,M., and Heino, T. I. (2009) Evidence that DmMANF is aninvertebrate neurotrophic factor supporting dopaminergicneurons. Proc. Natl. Acad. Sci. USA 106, 2429–2434
23. Lindholm, P., Voutilainen, M. H., Lauren, J., Peranen, J.,Leppanen, V. M., Andressoo, J. O., Lindahl, M., Janhunen, S.,Kalkkinen, N., Timmusk, T., Tuominen, R. K., and Saarma, M.(2007) Novel neurotrophic factor CDNF protects and rescuesmidbrain dopamine neurons in vivo. Nature 448, 73–77
24. Lindholm, P., Peranen, J., Andressoo, J. O., Kalkkinen, N.,Kokaia, Z., Lindvall, O., Timmusk, T., and Saarma, M. (2008)MANF is widely expressed in mammalian tissues and differentlyregulated after ischemic and epileptic insults in rodent brain.Mol. Cell. Neurosci. 39, 356–371
25. Palgi, M., Greco, D., Lindstrom, R., Auvinen, P., and Heino, T. I.(2012) Gene expression analysis of Drosophila Manf mutantsreveals perturbations in membrane traffic and major metabolicchanges. BMC Genomics 13, 134
26. Will, T., Tjallingii, W. F., Thonnessen, A., and van Bel, A. J.(2007) Molecular sabotage of plant defense by aphid saliva. Proc.Natl. Acad. Sci. USA 104, 10536–10541
27. Will, T., Kornemann, S. R., Furch, A. C., Tjallingii, W. F., andvan Bel, A. J. (2009) Aphid watery saliva counteracts sieve-tubeocclusion: a universal phenomenon? J. Exp. Biol. 212, 3305–3312
28. Medina-Ortega, K. J., and Walker, G. P. (2013) Does aphidsalivation affect phloem sieve element occlusion in vivo? J. Exp.Bot. 64, 5525–5535
29. Mutti, N. S., Park, Y., Reese, J. C., and Reeck, G. R. (2006) RNAiknockdown of a salivary transcript leading to lethality in the peaaphid, Acyrthosiphon pisum. J. Insect Sci. 6, 38
30. Akey, D. H., and Beck, S. D. (1971) Continuous rearing of thepea aphid, Acyrthosiphon pisum, on a holidic diet. Ann. Entomol.Soc. Am. 64, 353–356
31. Will, T., Steckbauer, K., Hardt, M., and van Bel, A. J. E. (2012)Aphid gel saliva: sheath structure, protein composition andsecretory dependence on stylet-tip milieu. PLoS ONE 7, e46903
32. Yoshida, H., Haze, K., Yanagi, H., Yura, T., and Mori, K. (1998)Identification of the cis-acting endoplasmic reticulum stressresponse element responsible for transcriptional induction ofmammalian glucose-regulated proteins. Involvement of basicleucine zipper transcription factors. J. Biol. Chem. 273,33741–33749
33. Hellens, R. P., Allan, A. C., Friel, E. N., Bolitho, K., Grafton, K.,Templeton, M. D., Karunairetnam, S., Gleave, A. P., and Laing,W. A. (2005) Transient expression vectors for functionalgenomics, quantification of promoter activity and RNAsilencing in plants. Plant Methods 1, 13
34. Moran, P. J., and Thompson, G. A. (2001) Molecular responsesto aphid feeding in Arabidopsis in relation to plant defensepathways. Plant Physiol. 125, 1074–1085
35. Cao, H. H., Wang, S. H., and Liu, T. X. (2014) Jasmonate- andsalicylate-induced defenses in wheat affect host preference andprobing behavior but not performance of the grain aphid, Sito-bion avenae. Insect Sci. 21, 47–55
36. Louis, J., and Shah, J. (2013) Arabidopsis thaliana–Myzus persicaeinteraction: shaping the understanding of plant defense againstphloem-feeding aphids. Front. Plant Sci. 4, 213
37. Cooper, W. C., Jia, L., and Goggin, F. L. (2004) Acquired and R-gene-mediated resistance against the potato aphid in tomato.J. Chem. Ecol. 30, 2527–2542
38. Gao, L. L., Kamphuis, L. G., Kakar, K., Edwards, O. R., Udvardi,M. K., and Singh, K. B. (2010) Identification of potential earlyregulators of aphid resistance in Medicago truncatula viatranscription factor analysis. New Phytol. 186, 980–994
39. De Vos, M., and Jander, G. (2009) Myzus persicae (green peachaphid) salivary components induce defence responses inArabidopsis thaliana. Plant Cell Environ. 32, 1548–1560
40. Giordanengo, P., Brunissen, L., Rusterucci, C., Vincent, C.,van Bel, A., Dinant, S., Girousse, C., Faucher, M., andBonnemain, J.-L. (2010) Compatible plant–aphid interactions:how aphids manipulate plant responses. C. R. Biol. 333, 516–523
41. Parkash, V., Lindholm, P., Peranen, J., Kalkkinen, N., Oksanen,E., Saarma, M., Leppanen, V. M., and Goldman, A. (2009) Thestructure of the conserved neurotrophic factors MANF andCDNF explains why they are bifunctional. Protein Eng. Des. Sel. 22,233–241
42. Hellman, M., Arumae, U., Yu, L. Y., Lindholm, P., Peranen, J.,Saarma, M., and Permi, P. (2011) Mesencephalic astrocyte-derived neurotrophic factor (MANF) has a unique mechanism torescue apoptotic neurons. J. Biol. Chem. 286, 2675–2680
43. Hoseki, J., Sasakawa, H., Yamaguchi, Y., Maeda, M., Kubota, H.,Kato, K., and Nagata, K. (2010) Solution structure and dynamicsof mouse ARMET. FEBS Lett. 584, 1536–1542
Received for publication October 14, 2014.Accepted for publication December 23, 2014.
14 Vol. 29 May 2015 WANG ET AL.The FASEB Journal x www.fasebj.org
