Metabolites of 2,3-diketogulonate delay peroxidase action and induce non-enzymic H2O2 generation: Potential roles in the plant cell wall

Me aboli es o 2,3-dike ogulona e delay pe oxidase ac ion and induce non-enzymic H 2 O 2 gene a ion: Po en ial oles in he plan cell wall Anna K€ a k€ onen a , b , * , Rebecca A. Dewhi s b , 1 , C. Logan Mackay c , S ephen C. F y b a Depa men o Ag icul u al Sciences, Viikki Plan Science Cen e , Uni e si y o Helsinki, Finland b The Edinbu gh Cell Wall G oup, Ins i u e o Molecula Plan Sciences, The Uni e si y o Edinbu gh, Edinbu gh EH9 3BF, UK c Eas CHEM School o Chemis y, The Uni e si y o Edinbu gh, Edinbu gh EH9 3FJ, UK a icle in o A icle his o y: Recei ed 30 Sep embe 2016 Recei ed in e ised o m 22 Feb ua y 2017 Accep ed 12 Ma ch 2017 A ailable online 14 Ma ch 2017 Keywo ds: Asco ba e Dehyd oasco bic acid Dike ogulona e Apoplas Pe oxidase Hyd ogen pe oxide Hyd oxyl adical Plan cell wall abs ac A p opo ion o he plan 's L -asco ba e ( i amin C) occu s in he apoplas , whe e i and i s me aboli es may ac as p o-oxidan s and an i-oxidan s. One asco ba e me aboli e is 2,3-dike ogulona e (DKG), p epa a ions o which can non-enzymically gene a e H 2 O 2 and delay pe oxidase ac ion on a oma ic subs a es. As DKG i sel gene a es se e al by-p oduc s, we cha ac e ised hese and hei abili y o gene a e H 2 O 2 and delay pe oxidase ac ion. DKG p epa a ions apidly p oduced a by-p oduc , compound (1), wi h l max 271 and 251 nm a neu al and acidic pH espec i ely. On HPLC, (1) co-elu ed wi h he majo H 2 O 2 -gene a ing and pe oxidase- delaying p inciple. Compound (1) was slowly des oyed by asco ba e oxidase, and was less s able a pH 6 han a pH 1. Elec opho esis o an HPLC-en iched p epa a ion o (1) sugges ed a s ongly acidic (pK a z2.3) compound. Mass spec ome y sugges ed ha un-ionised (1) has he o mula C 6 H 6 O 5 , i.e. i is a educ ion p oduc o DKG (C 6 H 8 O 7 ). In conclusion, compound (1) is he majo H 2 O 2 -gene a ing, pe oxidase-delaying p inciple o med non- enzymically om DKG in he pa hway asco ba e /dehyd oasco bic acid /DKG /(1). We hypo hesise ha (1) gene a es apoplas ic H 2 O 2 (and consequen ly hyd oxyl adicals) and delays cell-wall c osslinking dbo h hese e ec s a ou ing wall loosening, and possibly playing a ole in pa hogen de ence. ©2017 The Au ho s. Published by Else ie Inc. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). 1. In oduc ion L -Asco bic acid (C 6 H 8 O 6 ; i amin C) is an impo an edox compound in all plan s and animals. In plan s, i is syn hesised in he p o oplas , and a p opo ion o i is eleased in o he apoplas (aqueous solu ion ha pe mea es he cell wall) [11,37], whe e some o i is enzymically and non-enzymically oxidised by O 2 o o m monodehyd oasco ba e, which apidly disp opo iona es in o asco ba e and dehyd o- L -asco bic acid (DHA; C 6 H 6 O 6 ). Asco ba e oxidase is well es ablished o be a wall-localised enzyme capable o modula ing he asco ba e:DHA a io [43]. Apoplas ic asco ba e and i s downs eam me aboli es ha e been widely discussed as impo an playe s in p o ec ing he plan agains en i onmen al oxida i e s esses, especially ul a iole i adia ion, a mosphe ic ozone pollu ion and pa hogen challenge [4,5,10,40,41,44,47,53]. Apoplas ic asco ba e me aboli es ha e also been p oposed o se e oles in modula ing he cell wall's biophysical p ope ies, especially influencing he so ening o ipening ui s and he ex ensibili y o he p ima y wall [1,9,12,15,25]. DHA is uns able in neu al aqueous solu ions and is easily de- lac onised o 2,3-dike o- L -gulonic acid (DKG; C 6 H 8 O 7 ) which i sel con e s non-enzymically o se e al u he deg ada ion p oduc s depending on he incuba ion condi ions [8,26,35,46,51]. In he apoplas o cul u ed ose cells, a po ion o he DHA is oxidised o oxala e and L - h eona e (and es e s he eo ), some o hese e- ac ions being p oposed o gene a e H 2 O 2 , and a u he po ion o he DHA is hyd olysed o DKG [17,18]. The balance be ween hese wo pa hways (oxida ion:hyd olysis a io o DHA) is dic a ed by he se e i y o he ambien oxidising condi ions [38,39]. Some Abb e ia ions used: AAO, asco ba e oxidase; ABTS, 2,2 0 -azino-bis(3- e hylbenz hiazoline-6-sulphonic acid; DHA, dehyd oasco ba e; DKG, 2,3-dike o- L - gulonic acid; 2,3-enediol-DKGL, he 2,3-enediol o m o 2,3-dike ogulono- d - lac one; 3,4-enediol-DKGL, he 3,4-enediol o m o 2,3-dike ogulono- d -lac one; ROS, eac i e oxygen species; XO, xylenol o ange. *Co esponding au ho . Cu en add ess: Na u al Resou ces Ins i u e Finland (Luke), G een Technology, La oka anonkaa i 9, Helsinki, Finland. E-mail add ess: anna.ka konen@luke.fi(A. K€ a k€ onen). 1 Cu en add ess: wildFIRE Lab, Ha he ly Labo a o ies, Uni e si y o Exe e , P ince o Wales Road, Exe e EX4 4PS, UK. Con en s lis s a ailable a ScienceDi ec A chi es o Biochemis y and Biophysics jou nal homepage: www.else ie .com/loca e/yabbi h p://dx.doi.o g/10.1016/j.abb.2017.03.006 0003-9861/©2017 The Au ho s. Published by Else ie Inc. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). A chi es o Biochemis y and Biophysics 620 (2017) 12e22 asco ba e deg ada ion p oduc s a e o in e es in cell-wall physi- ology and in pa hogen de ence because o hei unique edox p ope ies. DKG is a highly uns able compound, eadily o ming a wide ange o by-p oduc s. Some o he many epo ed asco ba e deg ada ion p oduc s, p obably o med ia DKG, include 2,3- enediol-DKGL (C 6 H 6 O 6 ; he 2,3-enediol o m o 2,3-dike o- gulono- d -lac one), 3,4-enediol-DKGL (also C 6 H 6 O 6 ; he 3,4-enediol o m o 2,3-dike o-gulono- d -lac one) and L -e y h oasco bic acid (C 5 H 6 O 5 ), ha e educing ac i i y and migh unc ion as educing agen s in a simila way o asco ba e in i o [26,27,35,48]. The pa hway om DKG o e y h oasco ba e was sugges ed o p oceed ia 2,3-enediol- L -lyxose, wi h O 2 (o DHA i p esen ) se ing as oxidan in he con e sion o 2,3-enediol- L -lyxose o e y h- oasco ba e [19]. Howe e , he physiological significance o e y h- oasco ba e o ma ion om DKG may be limi ed, as he p ocess was only apid in he p esence o cyanide, 0.5 M phospha e, and a pH o 8. De-lac onisa ion o DHA o DKG is o en conside ed o be i e- e sible [35]. Howe e , some o ma ion o DHA by e-lac onisa ion o DKG has been demons a ed, especially a lowe pH alues [31]. DHA can be educed o asco ba e when a sui able educ an is p esen ; 3,4-enediol-DKGL o med om DKG was conside ed o be his educ an [48]. Indeed, he concen a ion o 3,4-enediol-DKGL ose o 10% o ha o DKG a e 30 min incuba ion o DKG a neu al pH in a ni ogen a mosphe e [48]. DKG has also been epo ed o unde go deca boxyla ion o L - xylosulose (¼ L -xylosone) [19,32,55], om which e y h oasco ba e may be o med [21]. L -Xylosulose may also gi e ise o se e al s ongly acidic edox compounds e.g. 2- u oic acid and he de- lac onised o m o 5-me hyl-3,4-dihyd oxy e one, al hough mos o hese we e only o med unde highly unphysiological condi ions such as 0.5 M H 2 SO 4 a 90  C[26]. Non-acidic dioxo p oduc s may be o med om DKG a e he loss o oxala e by hyd olysis o o m L -e y h ulose and subsequen ly 3-deoxy- L - h eosulose (¼3-deoxy- L - h eosone) o a e he loss o oxala e by an oxida i e pa hway o o m L - h eosulose dall o which a e compounds discussed as being o ele ance o he ageing o animal lens p o eins [34]. Howe e , we a e no awa e o any o hese pa icula dioxo p od- uc s being epo ed o delay pe oxidase ac ion o o educe O 2 o H 2 O 2 . Addi ional p oduc s o DKG deg ada ion, o med in he plan apoplas , include compounds ‘C’and ‘E’[17], p o isionally iden i- fied [39] as C¼2-ca boxy- L -xylonolac one plus 2-ca boxy- L -lyx- onolac one and E¼ hei de-lac onised p oduc ; Cand Ea e in e con e ible, bu o he wise ela i ely s able bo h in i o and in i o. ‘DKG’p epa a ions ha e biologically in e es ing edox p ope - ies, p obably due o he o ma ion o DKG deg ada ion p oduc s. An eneediol g oup o some o he unc ional g oup ha is as easily oxidisable as an eneediol is hough o be a common ea u e o he asco ba e deg ada ion p oduc s ha ing educing p ope ies [51]. Fo example, ‘DKG’(120 m M) has been epo ed o be an an i- oxidan agains he oxida i e modifica ion o yolk lipop o ein in a coppe -con aining solu ion, whe eas i has a p o-oxida i e e ec a lowe concen a ions (75 m M; [29]). Fu he mo e, ‘DKG’a mic omola concen a ions delays coppe -induced oxida i e o - ma ion o conjuga ed dienes in yolk lipop o ein, he lag ime leng hening wi h inc easing ‘DKG’concen a ions [29]. Since DKG i sel has no educing ac i i y, i was sugges ed ha 3,4-enediol- DKGL, he mos p e alen b eakdown p oduc de ec ed, was esponsible o he an i-oxida i e e ec [29]. 2,3-Enediol-DKGL was also p esen and possibly con ibu ed o he an i-oxida i e unc- ion. Likewise, [42] hypo hesised ha he p o ec i e e ec o DHA on coppe -induced oxida i e modifica ion o human low-densi y lipop o ein was due o s able modifica ion o he p o ein by DHA o i s b eakdown p oduc (s). ‘DKG’has p e iously been obse ed o accele a e he pe oxi- da ion o linoleic acid in neu al bu no in sligh ly acidic solu ions [49]. A supe oxide-sca enging agen , Ti on, supp essed linolea e pe oxida ion whe eas ca alase had no inhibi o y e ec , sugges ing ha supe oxide was he eac i e oxygen species (ROS) gene a ed du ing incuba ion wi h DKG. Al hough H 2 O 2 was he ROS de ec ed in he p esen s udy, he possibili y emains ha supe oxide was he o iginal ROS gene a ed, o ming H 2 O 2 by dismu a ion. Asco ba e induces clea non-enzymic H 2 O 2 gene a ion when added in o a solu ion con aining a ansi ion me al [7,12]. Also DHA, and especially a DKG p epa a ion p epa ed om comme cial DHA, led o H 2 O 2 gene a ion when added in o a solu ion con aining a ace o coppe ions [22]. To de ec he H 2 O 2 gene a ed a e DKG addi ion we used wo sepa a e assays: he xylenol o ange (XO) assay [3,16] and an indi ec pe oxidase ac i i y assay in which o- dianisidine was used as a pe oxidase subs a e. Di e ences in he esul s ob ained by hese wo me hods led o he disco e y ha he DKG p epa a ion con ained a compound ha inhibi s pe oxidase ac i i y. Since he DKG p epa a ion con ained se e al b eakdown p oduc s, as obse ed by pape elec opho esis ollowed by sil e s aining (Fig. 9 in Re . [22], we wan ed o esol e which o hese was he majo ac i e componen inducing non-enzymic H 2 O 2 gene a- ion and delaying pe oxidase ac ion. 2. Ma e ials and me hods 2.1. Chemicals Asco ba e oxidase (AAO), 2,2 0 -azino-bis(3-e hylbenz hiazoline- 6-sulphonic acid) (ABTS), ca alase, dehyd oasco bic acid and dehyd o- L -asco bic acid dime , o-dianisidine dihyd ochlo ide and ho se adish pe oxidase ype II we e ob ained om Sigma-Ald ich. AAO was dissol ed as a s ock a 1000 U ml 1 in 50 mM succina e (Na þ ) bu e , pH 5.6, supplemen ed wi h 0.05% bo ine se um al- bumin. Pe oxidase was dissol ed (1 m g m l 1 ) and u he dilu ed in he same bu e . DKG was p epa ed om he comme cial DHA by alkali ea - men [56]. A s ock o DHA (50 mM) was p epa ed in wa e (i ook a leas 30 min o dissol e DHA). A sligh mola excess o NaOH (1.3 ) was added and he mix u e incuba ed a 20  C o 6 min. Rou inely, he hyd olysis was hen s opped wi h 1 M L - a a ic acid and he pH was checked by pH pape (~3.5e4.0). Howe e , o samples o be ac iona ed by HPLC, hyd olysis was s opped wi h 1MH 2 SO 4 o a final pH o ~1 o ~6. F eshly-made DHA and DKG solu ions we e s o ed on ice be o e he assays. DKG was p epa ed also by an ioda e me hod [20]. A solu ion o asco bic acid (0.12 M) was incuba ed wi h po assium ioda e (0.36 M) o 5 min. KOH (1 M) was hen added d opwise un il he solu ion became colou less. Cold e hanol (8 ol, 20  C) was added, and he p ecipi a ed DKG was acuum fil e ed, insed in 70% e hanol, d ied and s o ed a 80  C. 2.2. In- i o pe oxidase ac i i y assays The e ec s o a ious asco ba e b eakdown p oduc s on pe oxidase ac i i y we e es ed in i o. DHA and DKG s ock solu- ions we e eshly p epa ed and added a a ious concen a ions o a eac ion mix u e ( o al olume 1.0 ml) ha con ained ei he 550 m M ABTS o 800 m Mo-dianisidine, and 250 o 500 m MH 2 O 2 , 3.13 o 6.25 ng/ml ho se adish pe oxidase ype II and he com- pound o in e es , in 44 mM succina e (Na þ ) bu e , pH 5.6. [All concen a ions quo ed a e final, in he comple e eac ion mix u e, unless o he wise s a ed.] The eac ion was ini ia ed by he addi ion A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e22 13 o he enzyme and ollowed a 420 and 405 nm o ABTS and o- dianisidine espec i ely. When an asco ba e oxidase (AAO) p e- ea men was included, he compound o in e es (~0.5 mM) was p e-incuba ed in 4 U/ml AAO and 44 mM succina e (Na þ ) bu e , pH 5.6, o 10, 15 o 60 min a 20  C be o e addi ion o he o he assay componen s. 2.3. Sea ch o he ac i e componen (s) in he DKG p epa a ion ha s imula es H 2 O 2 p oduc ion and inhibi s pe oxidases As he DKG p epa a ion con ained se e al compounds in addi- ion o DKG (Fig. 9 in Re . [22], hese ‘me aboli es’we e sepa a ed by p epa a i e high- ol age pape elec opho esis a pH 2.0, 3.5 and 6.5 acco ding o [14]. Each elec opho e og am was cu in o s ips, and he compounds we e elu ed om he pape in wa e , concen a ed in acuo (SpeedVac, Sa an ) and s o ed a 75  C. The e ec o elu ed compounds on non-enzymic H 2 O 2 p oduc ion was es ed in i o:10 m l o each ac ion, supplemen ed wi h 1 m M CuSO 4 , was es ed o H 2 O 2 o ma ion by he xylenol o ange (XO) me hod [3]. The XO me hod de ec s hyd ope oxides ha oxidise Fe 2þ in an acidic solu ion, and he amoun o e ic p oduc is measu ed as a XO complex [16]. Also he e ec o each ac ion on pe oxidase ac i i y in i o was es ed. To confi m he iden i y and s abili y o compounds used in he assays, we e-elec opho esed each ac ion a he o iginal pH, and s ained he solu es wi h AgNO 3 [13]. 2.4. Sea ch o he AAO- esponsi e ‘me aboli e’in he DKG p epa a ion by HPLC DKG [4.7 mM, in 45 mM succina e (Na þ ) bu e , pH 5.6] was ea ed wi h AAO (12 U/ml) o dena u ed AAO (10 min boiling) a 25  C o 15 min wi h gen le mixing, hen he enzymic eac ion was e mina ed by addi ion o H 2 SO 4 o pH ~1. Reac ion p oduc s, Table 1 Ul a iole abso p ion p ope ies o asco ba e and some o i s deg ada ion p oduc s. Compound l max a acidic pH (nm) l max a neu al pH (nm) Re e ences Cmpd (1) 251 271 p esen wo k L -Asco ba e 245 265 [19] Dehyd o- L -asco bic acid a <195 [54] 223 [2,54,59] 225 w 2,3-Dike o- L -gulona e <195 <225 [36,54] L -E y h oasco ba e 245 265 [19] 2,3-Enediol-DKGL 210, 300 225, 345 [29,51] 3,4-Enediol-DKGL 245 265 [29,36] 2-Fu oic acid 255 245 [26,60] 252 5-Me hyl-3,4-dihyd oxy e one 245 265 [26] L -E y h ulose 279 279 [33] Oxala e <205 <205 Succina e b <205 <205 w¼Weak abso bance maximum. a Fig. 3ao [21] shows ha esh dehyd oasco bic acid has almos no abso bance a ~300 nm, bu acqui es abso bance a o nea ha wa eleng h a e leng hy s o age o he solu ion. b No a p oduc o asco ba e ca abolism, bu used in he p esen wo k as a bu e . Fig. 1. E ec o dehyd oasco ba e and a dike ogulona e p epa a ion on he pe oxidase eac ion wi h o-dianisidine as subs a e. The e ec o he DKG p epa a ion p e- ea ed wi h AAO is also shown (þAAO). DHA, dehyd oasco bic acid; DKG, dike ogulona e. A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e2214 analysed by HPLC, we e compa ed wi h hose in an un ea ed DKG aliquo . 2.5. Semi-pu ifica ion o he ac i e compound(s) in DKG p epa a ion by HPLC HPLC was used o pu i y he compound(s) ha inhibi s pe oxi- dases and gene a es H 2 O 2 when added in o 1 m MCu 2þ .DKG p epa a ions (~46 mM, pH ~1 and ~6) we e fil e ed (0.4 m m, Ch omacol), and 40 m l was ac iona ed on a Phenomenex Rezex ROA column, un (0.5 ml min 1 )a 35  C, ou inely wi h 47 mM H 2 SO 4 as mobile phase. In some expe imen s, 13 mM TFA [0.1% ( / )] was used when a ola ile mobile phase was equi ed. Deg ada ion p oduc s we e de ec ed by UV abso bance a a ious wa eleng hs. The majo peak o cmpd (1) was collec ed and s o ed ozen p io o analysis. Mass spec ome y measu emen s we e pe - o med by elec osp ay on a 12T Sola iX Fou ie ans o m mass spec ome e (B uke Dal onics) equipped wi h an infini y cell and ope a ing in posi i e mode. Spec a we e he sum o 20 mass an- alyses and collec ed wi h a da a size o 4 Mwo d. Agilen une mix was used o ex e nal calib a ion. Analysis was achie ed wi h Da a Analysis e sion 4.4 (B uke Dal onics). 2.6. Analysis o edox p ope ies o HPLC- ac iona ed me aboli es Fo p epa a i e pu poses, HPLC ac ions (0.5 ml) we e collec ed and used in pe oxidase o H 2 O 2 assays ei he immedia ely o a e s o age. In he la e case he ac ions we e ozen in liquid ni- ogen be o e s o age a 75  C. Since asco ba e deg ada ion p oduc s we e elu ed om he HPLC column in 47 mM H 2 SO 4 (pH ~1), he assays we e modified as ollows. The pe oxidase ac i i y assay mix u e (1.0 ml) con ained (added in he ollowing o de ; final concen a ions a e gi en): 25 mM Na 2 -succina e, 37 mM succina e (Na þ ) bu e (pH 5.6), 225 m l o he HPLC ac ion (i <225 m l, he emaining olume was added as 47 mM H 2 SO 4 ), 550 m M ABTS, 250 m MH 2 O 2 , and 3.13 ng/ml pe oxidase. When an AAO ea men was included, AAO (1 U/ml; ac i e o boiled) was added a e he HPLC ac ion had been mixed wi h he succina e; hen, a e 10 min incuba ion a 20  C, ABTS and H 2 O 2 we e added and he assay was s a ed by addi ion o pe oxidase. The assay mix u e o non-enzymic H 2 O 2 gene a ion con ained (final olume 3.0 ml): 8.3 mM Na 2 -succina e and 225 m l o HPLC ac ion (i <225 m l, he emaining olume was added as 47 mM H 2 SO 4 ) and 1 m M CuSO 4 (added las ). When an AAO p e- ea men was included, he enzyme (0.7 U/ml) was added a e he HPLC ac ion had been mixed wi h he succina e, and he ials we e incuba ed o 10 min on a shake (100 pm) a 20  C be o e addi ion o CuSO 4 ( o 1 m M). As a con ol o he AAO ea men , he HPLC ac ion was ea ed o 10 min wi h dena u ed AAO (10 min boiling). This ea men was also impo an o show whe he in- cuba ion a an inc eased pH was enough o alle ia e he e ec o he compound, i.e. whe he he compound was mo e labile a pH~5 han a pH~1. H 2 O 2 gene a ed was measu ed by he XO assay [3,16,24]. Ali- quo s (100 m l) o he eac ion mix u e we e sampled a ime poin s and immedia ely added o 1 ml o XO mix u e (con aining 125 m M XO, 100 mM D -so bi ol, 250 m M FeSO 4 , 250 m M (NH 4 ) 2 SO 4 and 25 mM H 2 SO 4 ). All sample-XO mix u es we e incuba ed o 40 min a oom empe a u e be o e measu emen o A 560 agains a Fig. 2. HPLC o dike ogulona e and i s by-p oduc s. (a) The DKG p epa a ion was ac iona ed by HPLC wi h 47 mM H 2 SO 4 as eluen , e ealing se e al me aboli es. The elua e was moni o ed simul aneously o abso bance a 210, 250 and 300 nm, and ac ions (labelled 1 o 13) we e collec ed. A possible s uc u e o cmpd (1), based on i s mass spec um (Supplemen al Fig. 5), is shown. (b) UV spec um o he peak con aining cmpd (1), elu ing a 11.01 min. (c) Asco ba e oxidase (AAO; 12 U/ml), o dena u ed enzyme as a con ol,was applied o a new p epa a ion o DKG o 15 min, hen he p oduc s we e ac iona ed as in (a). Abso bance a 250 nm is shown. Solid a ows, nomencla u e o significan peaks; open a ows wi h names in b acke s, expec ed elu ion posi ions o he named compounds. A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e22 15 blank p epa ed wi h 100 m l solu ion con aining 1 m M CuSO 4 and 8.3 mM Na 2 -succina e þ1 ml o XO mix u e. CuSO 4 was obse ed no o in e e e wi h he XO assay. A s anda d cu e was p epa ed wi h di e en concen a ions o H 2 O 2 in 1 m M CuSO 4 . A dilu ion se ies o H 2 O 2 was p epa ed in wa e and a po ion o each solu ion was adjus ed o 1 m M CuSO 4 immedia ely be o e addi ion o XO eagen . 3. Resul s 3.1. Dike ogulona e gene a es a by-p oduc ha educes O 2 o H 2 O 2 and delays pe oxidase ac ion on model subs a es DKG, p epa ed by hyd olysis o DHA, exhibi ed a p ominen peak o UV abso bance a pH 5.6 ( l max 271e272 nm; [22].In con as , he DHA had negligible abso bance a his wa eleng h o a 265 nm (which is he l max o asco ba e, e y h oasco ba e and 3,4-enediol-DKGL in neu al solu ion; Table 1). Since eshly-made aqueous DKG is s a ed o ha e no s ong abso p ion abo e 225 nm [36], he obse ed abso bance a 271e272 nm was p obably a ibu able o uniden ified DKG deg ada ion p oduc s, he majo one o which is he e e med cmpd (1). Simila DKG solu ions had been shown o educe O 2 o H 2 O 2 non-enzymically in he p esence o a ace o Cu 2þ [22]. We now gi e e idence ha his is due o he p esence o cmpd (1). In addi ion, we ha e ound ha in in- i o pe oxidase assays, he DKG p epa a ion caused a concen a ion-dependen lag, p esumably also due o cmpd (1), be o e he oxida ion o a model subs a e, o- dianisidine, became isible (Fig. 1). A e his lag, he eac ion e- loci y was simila o (o , in he case o high ‘DKG’concen a ions, sligh ly slowe han) ha o he non-inhibi ed pe oxidase, and he du a ion o he lag co ela ed wi h he amoun o ‘DKG’added. These e ec s, which seem unlikely o be due o DKG i sel , a e simila o hose exe ed by asco ba e [50]. DHA, on he con a y, showed li le inhibi ion o pe oxidase ac i i y. 3.2. Asco ba e oxidase pa ially inac i a es cmpd (1) Since he UV spec um o he DKG p epa a ion ( l max 271 nm a pH 5.6; Fig. 8 o [22] was eminiscen o ha o asco ba e ( l max 265 nm a pH 5.6), we explo ed he possible p esence o asco ba e- like subs ances. P e- ea men o he DKG p epa a ion wi h asco - ba e oxidase (AAO) diminished bu did no abolish he lag pe iod (Fig. 1). The A 271 o he DKG p epa a ion dec eased slowly wi hou any enzyme addi ion; addi ion o AAO accele a ed his eac ion (Fig. 8 o [22], bu i ook se e al minu es be o e he abso bance alue eached a minimum, and e en hen some A 271 pe sis ed. In con as , he A 265 o a solu ion o pu e asco ba e eaches ze o a ew seconds a e AAO addi ion (Supplemen al Fig. 1), confi ming ha cmpd (1) is no asco ba e. In a mix u e o he DKG p epa a ion and pu e asco ba e, AAO caused a apid dec ease in A 265 (due o asco ba e oxida ion) ollowed by a slowe dec ease due o cmpd (1) oxida ion (Supplemen al Fig. 1); hus i canno be a gued ha he DKG was inhibi ing he AAO. Ne e heless, cmpd (1) appea s o be asco ba e- ela ed since AAO has a high specifici y owa ds L - asco ba e and ela ed compounds ha ha e a lac one ing wi h an adjacen eneediol g oup such as e y h oasco ba e [6]. 3.3. High- ol age elec opho esis o he c ude DKG p epa a ion As he alkali-gene a ed DKG p epa a ion con ained se e al by- p oduc s [22], we a emp ed o sepa a e hese by elec opho esis and es hem indi idually o pe oxidase ac ion delay (Supplemen al Fig. 2) and non-enzymic H 2 O 2 gene a ion (da a no shown). A e elec opho esis a pH 2.0 (Supplemen al Fig. 2a), only ac ion 2 (con aining neu al and weakly acidic ma e ial) exe ed hese e ec s, and only o a low deg ee. Thus he only ac i e p in- ciple de ec ed had clea ly sepa a ed om DKG i sel (a ela i ely s ong acid, ound in ac ions 3 and 4). Howe e , he o al eco - e ed zones had a less H 2 O 2 -gene a ing and pe oxidase-delaying capaci y han he c ude DKG ha had been applied o he elec opho e og am. A e elec opho esis a pH 3.5 (Supplemen al Fig. 2b) o 6.5 (Supplemen al Fig. 2c), he only ac i e p inciple de ec ed was ound o ha e co-mig a ed wi h DKG. This could indica e ha (i) cmpd (1) co-mig a ed wi h DKG a hese pH alues, and/o (ii) he cmpd (1) o iginally p esen was deg aded du ing he elec opho esis and subsequen elu ion bu new cmpd (1) was o med om he elu ed DKG i sel . 3.4. HPLC o he DKG p epa a ion On HPLC, eshly p epa ed c ude DKG e ealed se e al peaks o abso bance a 210 nm (A 210 peaks), sugges ing ca boxylic acids, es e s o lac ones, and a leas h ee A 250 peaks (sugges ing con- juga ed double-bonds; labelled 1,2and 3in Fig. 2). Cmpd (1)was elu ed sho ly a e he DKG peak, only pa ially sepa a ed om i ( e en ion imes 11.02 and 10.64 min espec i ely; Fig. 2). We p opose ha cmpd (1) showed a pH-dependen abso bance shi , Fig. 3. Selec ed HPLC ac ions om a dike ogulona e p epa a ion delay pe oxidase ac ion. ABTS was used as pe oxidase subs a e. Volume o HPLC ac ion added o he assay: 150 m l. The e ec o p e ea men o he ac ions wi h AAO (1 U/assay; þAAO) o wi h boiled AAO (þboiled enz) a pH ~5.2e5.6 o 10 min be o e he pe oxidase ac i i y assay is also shown. A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e2216 wi h l max 251 (Fig. 2b) and 271 nm a acidic and neu al pH espec i ely ( he HPLC eluen was 47 mM H 2 SO 4 ). This shi would mi o he beha iou o asco ba e and 3,4-enediol-DKGL, which ha e l max 245 and 265 nm in acidic and neu al solu ions espec- i ely [36,48]. Cmpd (1) was again confi med no o be asco ba e i sel , as his elu es a 13.25 min in his sys em. 3.5. Abili y o HPLC ac ions o gene a e H 2 O 2 and delay pe oxidase ac ion F ac ion 5 (Fig. 2a), which con ained he majo i y o cmpd (1), was he mos e ec i e ac ion a delaying pe oxidase ac ion on a model subs a e (ABTS; Fig. 3). O he delaying agen s we e also de ec ed, e.g. in ac ions 8 and 9. A mode a e delaying e ec was also obse ed in ac ion 4, which con ained mos o he DKG (Figs. 2 and 3). Howe e , since 24% o cmpd (1) elu ed in ac ion 4, and because some o he DKG may be u he deg aded o cmpd (1) a e elu ion om he column, we conclude ha he pe oxidase delaying agen in ac ion 4 was cmpd (1), no he DKG i sel . F ac ions 4 and 5 also caused non-enzymic H 2 O 2 p oduc ion in he p esence o O 2 and a ace o Cu 2þ , ac ion 5 again being mo e e ec i e (Supplemen al Fig. 3). The o he HPLC ac ions es ed, e en hose ha caused a sligh delay in pe oxidase ac i i y assays, did no gene a e H 2 O 2 . 3.6. Asco ba e oxidase diminishes he abili y o HPLC ac ions o delay pe oxidase ac ion and gene a e H 2 O 2 AAO p e- ea men o he pe oxidase- e a ding HPLC ac ions (4, 5, 8 and 9; Fig. 2a) diminished hei abili y o delay pe oxidase ac ion (Fig. 3). T ea men wi h hea -dena u ed AAO (i.e., ‘ageing’ he ac ions a ele a ed pH (5.2e5.6) in he absence o ac i e AAO) also sligh ly educed he lag caused by ac ions 4, 5, 8 and 9. Howe e , AAO did no comple ely des oy he pe oxidase delaying e ec s o any o hese ac ions, sugges ing ei he ha se e al agen s we e p esen in each ac ion, only some o hem being AAO- oxidisable, o ha he AAO gene a ed new p oduc s whose pe oxidase-delaying p ope ies we e weake han hose o he ini ial compounds. AAO-p e ea men o ac ions 4 and 5 also diminished hei abili y o non-enzymically gene a e H 2 O 2 (Supplemen al Fig. 3). 3.7. UV-de ec able compounds (1), (2) and (3) can be oxidised by asco ba e oxidase Since AAO a ec ed he HPLC ac ions' e ec s on H 2 O 2 Fig. 4. E ec o ea men a pH 1 o 6 on he HPLC p ofile o dike ogulona e and i s by-p oduc s. The DKG, p epa ed by NaOH ea men o DHA, was adjus ed o pH 1 o pH 6 wi h H 2 SO 4 and, a e s o age o 0.5e3.0 h a 0 C, ac ioned by HPLC. Du ing each un, abso bances a (a) 210, (b) 250 and (c) 300 nm we e simul aneously moni o ed. Solid a ows, nomencla u e o significan peaks; open a ows wi h names in b acke s, expec ed elu ion posi ions o he named compounds. A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e22 17 gene a ion and pe oxidase ac ion, we a emp ed o de e mine which UV-de ec able compounds we e a ec ed by he AAO ea - men . The c ude DKG p epa a ion was ea ed wi h AAO (ac i e o dena u ed) o 15 min and e- un by HPLC. The mos ema kable changes we e obse ed in compounds abso bing a 250 nm (Fig. 2c). In pa icula , ac i e AAO s ongly diminished he cmpd (1) peak. A sligh dec ease was also no iced in he A 210 o his peak (da a no shown), sugges ing ha cmpd (1) is a ca boxylic acid o es e as well as possessing conjuga ed double bonds. Compounds 2 and 3also diminished (Fig. 2c; he peak elu ing a 15.8 min is p obably he succina e bu e ). 3.8. Cmpd (1) is less s able a pH 6 han a pH 1 Since ea men e en wi h dena u ed AAO (a pH ~5.6) modified he HPLC p ofile (Fig. 2c) and he ac ions' abili y o delay pe ox- idase ac ion (Fig. 3) and o p omo e H 2 O 2 p oduc ion (Supplemen al Fig. 3), we es ed he e ec o pH on he s abili y o DKG and i s by-p oduc s. Fo his wo k, DKG was p epa ed by alkaline hyd olysis o DHA and he eac ion was s opped wi h H 2 SO 4 ei he o pH 6 o o pH 1 (Fig. 4). DKG i sel , de ec ed a 210 nm, was almos una ec ed by s o age on ice o 0.5e3 h a pH 1 o 6 (Fig. 4a). When p e- ea ed a pH 6, ins ead o pH 1, compounds (1), (2) and (3) we e diminished, li le a ec ed and inc eased espec i ely (Fig. 4b). The e o e, i cmpd (1) was he educ an ha delays pe oxidase ac ion, hen i s e ec should be weakened by s o age a pH 6. This p edic ion was es ed on samples s o ed a pH 1 o 6 be o e HPLC. P e- ea men o he un ac iona ed p epa a ion only sligh ly diminished i s abili y o delay pe oxidase ac ion ega dless o he pH o which hey we e adjus ed (Fig. 5a), possibly because o he educ an s e.g. compound (3) inc eased a e ea men a pH 6 (Fig. 4). Howe e , in HPLC-pu ified p epa a ions ( ac ions 4 and 5), s o age a pH 6 did weaken he e ec compa ed wi h s o age a pH 1, app oxima ely hal ing he lag pe iod obse ed be o e pe oxidase ac ion began (Fig. 5b) [ he pH o he pe oxidase eac ions was adjus ed o >5 wi h he addi ion o Na 2 -succina e o HPLC ac- ions]. These da a s ongly suppo he idea ha cmpd (1), he main 250-nm-abso bing compound in ac ions 4 and 5, was he majo educ an ha delayed pe oxidase ac ion on i s model subs a e. 3.9. Analysis o cmpd (1) by high- ol age pape elec opho esis HVPE has p o ed e y use ul o esol ing asco ba e me aboli es [17,18,38,39]. Pa ially pu ified cmpd (1) ob ained by p epa a i e HPLC wi h 13 mM TFA as eluen (chosen because i is eadily ol- a ile) ga e s ainable spo s on analy ical elec opho e og ams (Fig. 6). The HPLC p ofile wi h TFA as eluen (Supplemen al Fig. 4a) was b oadly simila o ha wi h H 2 SO 4 (Fig. 2a), DKG elu ing sligh ly be o e cmpd (1). The g ea es amoun s o cmpd (1) we e ound be ween 10.5 and 11.25 min (Supplemen al Fig. 4b and c). This 0.75-min window o ac ions was pooled, d ied in acuo and e-dissol ed in wa e ; po ions we e elec opho esed a pH 2.0 and 6.5 and s ained wi h AgNO 3 (Fig. 6), e ealing: DKG, which is only pa ially esol ed om cmpd (1) by he HPLC sys em used; compounds Cand E(a ace), which a e p oposed [39] o be C¼2-ca boxy- L -xylonolac one plus 2-ca boxy- L -lyxonolac one; E¼ hei de-lac onised p oduc (a dianionic ca boxypen ona e); and a spo whose m DKG alues we e 1.22 a pH 2.0 and 1.05 a pH 6.5, which s ained a sligh ly yellowish b own a he han he g eyish b own usually p oduced by AgNO 3 . O hese possible iden i ies, we ha e al eady shown ha cmpd (1) is no DKG i sel . Fu he mo e, we ound ha cmpd (1) is no C o Esince pu ified Cand Ega e peaks clea ly esol ed om cmpd (1) on HPLC (Fig. 7). This was demons a ed when samples o Cand E(elu ed om pape a e p epa a i e elec opho esis [17];we e analysed by HPLC wi h 13 mM TFA as eluen (Fig. 7b), and a sample con aining cmpd (1) was un immedia ely he ea e (Fig. 7a). Cand Eshowed majo peaks o A 210 a 10.1 and 9.2 min espec i ely, and only small peaks o A 250 (Fig. 7b); in con as , cmpd (1) showed g ea e abso bance a 250 nm and elu ed a 10.5 min (Fig. 7a). Thus cmpd (1) is clea ly dis inguished om cmpds Cand E. Compounds Cand E o m om DKG in aqueous condi ions [17,39], and he spo s o hem seen in Fig. 6 would ha e o med om he DKG a e being elu ed om he column. We he e o e sugges ha , o he spo s seen in Fig. 6, one wi h m DKG alues 1.22 and 1.05 a pH 2.0 and 6.5 espec i ely is likely o be cmpd (1). Al hough i s s uc u e emains unknown, some o i s ionic p ope ies can be deduced om he elec opho e ic mobil- i ies. A pH 6.5, all eCOOH g oups a e almos ully ionised, so he Fig. 5. E ec o pH on he abili y o dike ogulona e and/o i s by-p oduc s o delay pe oxidase ac ion. The samples es ed we e (a) he whole DKG p epa a ion was s o ed a 0 C o 0.5e3.0 h a pH ~1 o a pH ~6; and (b) HPLC ac ions he eo ha had been collec ed in 47 mM H 2 SO 4 (pH ~1) as in Fig. 2a. In each case, he samples we e hen es ed o e ec on in- i o pe oxidase ac ion wi h ABTS as subs a e. In (a), he lag imes caused by he whole p epa a ion we e e y long, so he samples we e dilu ed o he assay ( o ~23 and 46 m Mfinal). In (b), wi h HPLC ac ions 4 and 5, undilu ed 100- m l po ions we e included in 1-ml pe oxidase eac ion mix u es wi h Na 2 -succina e addi ion o bu e he pH abo e 5. A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e2218 compound's p oximi y o DKG and Ca ha pH (Fig. 6b) indica es ha i has a cha ge:mass a io simila o hei s, i.e. 1 nega i e cha ge pe ~6 ca bon a oms. Fu he mo e, since he compound appea s o be anionic e en a pH 2.0, wi h a mobili y exceeding ha o DKG (m DKG ¼1.22; Fig. 6a), i is p obably a s ong acid wi h a pK a e en lowe han ha o DKG (p edic ed pK a z2.38; h p://www. hmdb.ca/me aboli es/HMDB05971). 3.10. Mass spec ome y o cmpd (1) A u he sample o cmpd (1), pa ially pu ified by HPLC as in Supplemen al Fig. 4, was analysed by MS in posi i e mode (Supplemen al Fig. 5). Ion peaks a m/z159.03164 and 181.01351 we e obse ed, which we e absen om he blank. These ions a e in e p e ed as C 6 H 6 O 5 $H þ and C 6 H 6 O 5 $Na þ (m/z alues espec i ely 17 and 12 ppm de ia ion om heo e ical, which is accep able since he nea es calib a ion poin was a m/zz332). 4. Discussion F eshly p epa ed DKG is epo ed o ha e li le UV abso bance a wa eleng hs abo e abou 225 nm [36]. Howe e , ou DKG p epa- a ions apidly p oduced a p opo ion o compound (1), wi h l max 271 and 251 nm a neu al and acidic pH espec i ely. These l max alues may be compa ed wi h hose o asco ba e and some o i s p e iously epo ed deg ada ion p oduc s (Table 1). Cmpd (1)’s l max alues, and i s ba hoch omic shi when he pH is adjus ed om acidic o neu al, sugges some chemical simila i ies o asco ba e. Fu he mo e, like asco ba e, DHA and 3,4-enediol-DKGL [48], cmpd (1) was mo e s able a acidic han neu al pH. Fig. 6. Analysis o HPLC-en iched cmpd (1) by high- ol age elec opho esis. Cmpd (1) was pa ially pu ified by HPLC wi h 13 mM TFA as eluen (see Supplemen al Fig. 4), hen analysed by elec opho esis a pH 2.0 (a) o 6.5 (b). Each sample, and he ma ke s, con ained a ace o O ange G, which was ci cled in pencil be o e he o he compounds we e s ained in AgNO 3 . Abb e ia ions used: E yR, e y h a a e (¼meso- a a e); Th R, L- h ea a e (¼L- a a e); Th O, h eona e; OxT, oxalyl h eona e; cOxT, cyclic oxalyl h eona e; C,2- ca boxy-L-xylonolac one and/o 2-ca boxy-L-lyxonolac one; E, de-lac onised C; DHA, dehyd oasco bic acid; DKG, dike ogulona e (p epa ed by he ioda e me hod). Spo s p esen in he cmpd (1) p epa a ion a e labelled o he igh , ma ke s o he le o each elec opho e og am. Spo s labelled () a e con aminan s as hey a e no p ecisely in line wi h he o he spo s in he lane. A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e22 19 In ag eemen wi h he no ed esemblance be ween cmpd (1) and asco ba e, he delaying e ec o cmpd (1) on pe oxidase ac ion mimicked ha o asco ba e. Asco ba e se es as an an i-oxidan , sca enging pe oxidase-gene a ed phenolic adicals, so ha he oxida ion o a oma ic subs a es becomes isible only when all asco ba e has been oxidised [50]. Thus, cmpd (1) p obably a ec s pe oxidase ac ion in a simila manne . In he p esen wo k, se e al me aboli es ob ained om a DKG p epa a ion induced he non-enzymic p oduc ion o H 2 O 2 and delayed he onse o subs a e oxida ion in in- i o pe oxidase as- says. A scheme o asco ba e deg ada ion has been p oposed [17,18,38,39] in which he ini ial oxida ion p oduc , DHA, is ei he u he oxidised ( o oxalyl h eona e, cyclic oxalyl h eona e and oxala e þ h eona e) o hyd olysed ( o DKG and i s own down- s eam p oduc s Cand E). The oxidising b anch was p oposed o include, o lead o, h ee s eps ha migh gene a e H 2 O 2 . On he o he hand, he hyd oly ic pa hway was no p oposed o yield H 2 O 2 ; he e o e he disco e y epo ed he e ha DKG by-p oduc s, p incipally cmpd (1), do gene a e ROS, p obably mainly H 2 O 2 ,was o g ea in e es . Cmpd (1) has wo e ec s which supe ficially seem con adic- o y: (a) when p esen in pe oxidase assays i appea s o se e as an an i-oxidan , sca enging phenolic adicals such ha he oxida ion o a oma ic subs a es (o-dianisidine and ABTS) becomes isible only when all he cmpd (1) has been oxidised, and (b) i non- enzymically educes O 2 o H 2 O 2 , he oxidising subs a e o pe oxi- dase. E ec (a) would delay pe oxidase ac ion, whe eas e ec (b) would p omo e i . Bo h hese e ec s could ha e biological signifi- cance in he plan cell wall. E ec (a) would delay he pe oxidase- ca alysed c oss-linking o cell-wall phenolics [50], e.g. o e ula e o di e ula es and y osine o isodi y osine, hus po en ially p e en - ing wall igh ening. Con e sely, he H 2 O 2 gene a ed in e ec (b) can non-enzymically lead o he o ma ion o o he ROS. In pa icula , he hyd oxyl adical (  OH) is eadily o med om H 2 O 2 , especially in he p esence o some emaining asco ba e [12], he p ecu so o cmpd (1). I is known ha  OH causes non-enzymic scission o cell- wall polysaccha ides [12,45,52], po en ially loosening he p ima y cell wall. P e en ing wall igh ening and p omo ing wall loosening, caused by (a) and (b) espec i ely, a e bo h expec ed o lead o a mo e eadily ex ensible o so e wall. The e o e he wo appa en ly di e gen e ec s o cmpd (1) may ac in an equi alen di ec ion, bo h o hem acili a ing biological p ocesses ha depend on a ‘loose’cell wall such as cell expansion, ui so ening and abscis- sion. Addi ionally, simila ly o asco ba e, cmpd (1) may influence he oxida i e bu s occu ing du ing pa hogen a ack, and hence plan de ence esponses [41]. The only known sou ce o cmpd (1), po en ially exe ing such wall-loosening e ec s, is DKG dwhich is o med by he non- enzymic hyd olysis o apoplas ic DHA. DHA can i sel be o med om apoplas ic asco ba e, abou 50% by endogenous AAO ac ion Fig. 7. Cmpd (1) is no a ca boxypen ona e. (a) Cmpd (1), pu ified by HPLC elu ed in 13 mM TFA, was e- un by HPLC in wa e and he p oduc s we e de ec ed by A 210 and A 250 . (b) A sample con aining compounds Cand E(ca boxypen ona es, p epa ed by alkali ea men o DHA and subsequen elu ion om a p epa a i e elec opho e og am) was also un by HPLC in wa e and moni o ed by A 210 and A 250 . A. K€ a k€ onen e al. / A chi es o Biochemis y and Biophysics 620 (2017) 12e2220