Secondary Organic Aerosol-Forming Reactions of Glyoxal with Amino Acids

14 Pages·2009·1.13 MB·English

Subscriber access provided by University of Colorado at Boulder
Article
Secondary Organic Aerosol-Forming
Reactions of Glyoxal with Amino Acids
David O. De Haan, Ashley L. Corrigan, Kyle W. Smith, Daniel R. Stroik, Jacob J. Turley,
Frances E. Lee, Margaret A. Tolbert, Jose L. Jimenez, Kyle E. Cordova, and Grant R. Ferrell
Environ. Sci. Technol., 2009, 43 (8), 2818-2824• DOI: 10.1021/es803534f • Publication Date (Web): 12 March 2009
Downloaded from http://pubs.acs.org on April 14, 2009
More About This Article
Additional resources and features associated with this article are available within the HTML version:
• Supporting Information
• Access to high resolution figures
• Links to articles and content related to this article
• Copyright permission to reproduce figures and/or text from this article
Environmental Science & Technology is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Environ.Sci.Technol.2009,43,2818–2824
Secondary Organic Aerosol-Forming incomplete.Inpartsoftheatmosphereasdissimilarasthe
MexicoCityboundarylayer(7)andthefreetroposphere(8),
Reactions of Glyoxal with Amino
SOA formation rates and/or concentrations are much
higherthanmodelpredictions.Theuptakeofglyoxalonto
Acids
aerosolwasimplicatedinMexicoCityasbeingresponsible
foratleastpartofthemismatchbetweenmeasurements
andmodeling(9).Glyoxalisscavengedbyaqueous-phase
DAVID O. DE HAAN,*,†,‡
aerosol (10, 11) and even by solid aerosol with surface-
ASHLEY L. CORRIGAN,† KYLE W. SMITH,†
adsorbed water (12, 13).
DANIEL R. STROIK,† JACOB J. TURLEY,†
FRANCES E. LEE,† Cloudprocessingofsmallaldehydes,especiallyglyoxal
MARGARET A. TOLBERT,‡ andmethylglyoxal,hasalsobeenidentifiedasasourceof
JOSE L. JIMENEZ,‡ KYLE E. CORDOVA,‡ aerosol(14-16),supportedbyfieldmeasurementsofhigh
AND GRANT R. FERRELL‡ levelsofoxalicacid,theoxidationproductofglyoxal,inthe
aerosollayeraboveclouds(17).Modelrunsincludingglyoxal
DepartmentofChemistryandBiochemistry,UniversityofSan
andmethylglyoxaloxidationincloudsandaerosolsuggest
Diego,5998AlcalaPark,SanDiegoCalifornia92110,and
DepartmentofChemistryandBiochemistry,andCooperative thatthesetwocompoundsareresponsiblefor38%ofmodeled
InstituteforResearchinEnvironmentalSciences,Universityof globalSOA,largelyviacloudprocessing(18).
Colorado,BoulderColorado80309-0216 Theaqueous-phasechemistryofglyoxalissummarized
inScheme1.Glyoxaltakespartinreversiblehydrationwhen
Received December 12, 2008. Revised manuscript received itentersthecondensedphasewhereitcanbeoxidizedby
February17,2009.AcceptedFebruary19,2009. OHandformoxalicacid.However,attypicalatmospheric
oxidant concentrations, a large majority of glyoxal is not
oxidizedwithinthe∼15min.lifetimeofaclouddroplet(19).
What then is its fate? ATR-FTIR observations suggest that
Glyoxal,thesimplestandmostabundantR-dicarbonyl aftermostofthewaterhasevaporatedfromadroplet,the
compoundintheatmosphere,isscavengedbycloudsand glyoxaldihydrateconvertsintoanextremelyreactivemono-
aerosol,whereitreactswithnucleophilestoformlow-volatility hydrate form, which self-oligomerizes and remains in the
condensedphase(20).Computationsindicatethatdehydra-
products.Hereweexaminethereactionsofglyoxalwithfive
tionmayalsobetriggeredbynucleophilicattack(21).Amino
aminoacidscommoninclouds.Whenglyoxalandglycine,serine,
acids have been implicated in the formation of light-
asparticacidorornithinearepresentatconcentrationsas
absorbingoligomersviaaldolreactionswithacetaldehyde
lowas30µMinevaporatingaqueousdropletsorbulksolutions, (22, 23). As we show in this work, the presence of such
1,3-disubstitutedimidazolesareformedinirreversiblesecond- nucleophilesinanevaporatingdropletresultsintheforma-
orderreactionsdetectedbynuclearmagneticresonance tion of different nonvolatile products. The resulting dry
(NMR),aerosolmassspectrometry(AMS)andelectrospray aerosolthencontainstheseproductsalongwithanyother
ionizationmassspectrometry(ESI-MS).Incontrast,glyoxalreacts low-volatilitymaterialfromtheoriginaldroplet,suchasoxalic
withargininepreferentiallyatsidechainaminogroups, acid.
formingnonaromaticfive-memberedrings.Allreactionswere Inthisworkwedescribethereactionsbetweenglyoxal
andthemostcommonaminoacidsinclouds,fog,andaerosol
accompaniedbybrowning.Theuptakeof45ppbglyoxalbysolid-
(24-26), and we use electrospray ionization mass spec-
phaseglycineaerosolat50%RHwasalsostudiedandfound
trometry(ESI-MS)andnuclearmagneticresonance(NMR)
tocauseparticlegrowthandtheproductionofimidazole
to characterize the imidazole products and measure the
measuredbyscanningmobilityparticlesizingandAMS,
kinetics of their formation. Finally, we quantify aerosol
respectively,withaglyoxaluptakecoefficientR)0.0004. formation via quadrupole aerosol mass spectrometry (Q-
Comparisonofreactionkineticsinbulkandindryingdroplets AMS) and high-resolution time-of-flight (HR-ToF)-AMS
showsthatconversionofglyoxaldihydratetomonohydrate experimentsthatsimulateclouddropletdryingandthedirect
acceleratesthereactionbyover3ordersofmagnitude,allowing uptakeofglyoxalbyaerosol.
thesereactionstooccuratatmosphericconditions.
Experimental Methods
Introduction
15N-enrichedglycine(CambridgeIsotopes),aminoacids,and
Aerosols have a strong effect on visibility (1) and human glyoxaltrimerdihydrate(Sigma-Aldrich)wereusedwithout
health(2),andcanbothdirectlyandindirectlyinfluencethe further purification. Glyoxal-amino acid solutions in the
climate (3). Secondary organic aerosol (SOA) material concentrationrange0.1-1.0Mweremadein18MΩwater.
represents an important fraction of the particle mass One µL droplets of these solutions were dried at room
throughoutthetroposphere(4,5).Agreatamountofeffort temperatureonattenuatedtotalreflectance(ATR)9-reflec-
has been expended to chemically speciate this organic tiondiamondcrystals(DuraSamplIR,SmithsDetection)and
materialandunderstandtheprocessesbywhichitforms(6). analyzedbyFouriertransforminfrared(FTIR)spectroscopy
Recentcomparisonsbetweenfieldmeasurementsandmod- (JASCO4800)inordertoobservereactionsduringthedrying
elswherebothgas-phaseaerosolprecursorconcentrations process(20).Larger150µLsampleswerealsodriedatroom
andaerosolformationratesweremeasured,however,have temperatureinvialstobrownsolids,andredissolvedin18
shownthatourunderstandingofaerosolformationremains MΩ water at 1 mg/mL for ESI-MS analysis or in D2O
(Cambridge Isotopes) for NMR analysis. Solutions were
injected into the ion trap ESI-MS (Thermo-Finnigan LCQ
*Correspondingauthore-mail:[email protected].
†UniversityofSanDiego. Advantage)viasyringepumpat2.5µL/minwithESInozzle
‡UniversityofColorado. voltage ) +4.3 kV, medical N2 sheath gas flow ) 25, and
28189ENVIRONMENTALSCIENCE&TECHNOLOGY / VOL.43,NO.8,2009 10.1021/es803534fCCC:$40.75 2009AmericanChemicalSociety
PublishedonWeb03/12/2009
SCHEME 1. Cloud Processing of Glyoxal to Form Imidazole Derivatives, Oligomers, And Oxalic Acid
capillarytemperature)200or250°C.1H,13C,15N,and2-D Results and Discussion
NMRspectrawerecollectedatroomtemperatureonVarian Products.Analysisofdropscontaining0.1Mglyoxal+glycine
400or500MHzinstruments.Reactionkineticsweremoni- by ATR-FTIR shows that the peaks assigned to the amine
toredby1HNMRinD2Osolutionsat295Kcontaining1M groupat3174(asymNsHstr),1508,1133,and1188cm-1(all
glyoxal and either 1 M glycine, 0.86 M arginine, or 0.5 M NH+deformations)(31)arealleithereliminatedorsignifi-
3
serine. The latter two amino acids have limited water cantlyreducedinintensityupondropletdrying,relativeto
solubility. peaksassignedtoglycine’sCH orcarboxylategroup(Figure
2
Thedryingofclouddropletswassimulatedbyatomizing 1andSupportingInformationFigureS1).Inaddition,the
0.03-10mMaqueousmixturesofglyoxalandanaminoacid glyoxal oligomer peaks at 950 and 980 (CsOsC oligomer
inHPLC-gradewater(Sigma-Aldrich)usingeitherpneumatic ring stretch), and 1070 cm-1 (CsOH stretch) that appear
nebulization(TSImodel9302,1.5µmmodewetdiameter) when glyoxal solutions are dried (20) become much less
or ultrasonic nebulization (∼5 µm). The latter was ac- pronouncedinthepresenceofaminoacids.Theseobserva-
complished using an inexpensive, commercially available tionssuggestthatglycine’saminegroupreactswithglyoxal
mistgeneratorthatwasmodifiedbyexposingitsconductivity andlargelypreemptsglyoxalself-reaction.
sensortoallowittoworkinhighlydiluteaqueoussolutions. ESI-MSandNMRdatashowthatglyoxalreactswiththe
Wet drop sizes were calculated from scanning mobility freeaminoacidsglycine,serine,asparticacid,andornithine
particlesizing(SMPS)distributionsfordriedNaClaerosol. inaqueoussolutionstoformextremelystable,zwitterionic
Theultrasonicallyproducedmistwaspushedbyprepurified 1,3-disubstitutedimidazolesinirreversiblereactions(Scheme
nitrogen at flow rates of 7-10 L min-1 through a charge 2).ESI-MSspectraofdriedandredissolvedglyoxal+amino
neutralizer (TSI 3077 or NRD model P-2021-1000) into a acid bulk solutions are shown in Figure 2 with imidazole
collapsible300LTeflonchamber(NewStarEnvironmental, peakslabeled.1Hand13CNMRpeakassignmentsofproducts
LLC)containingnitrogenhumidifiedto>70%RH.Further formed in glyoxal + amino acid reactions are listed in
evaporationfromdropletsenteringthechamberraisedthe SupportingInformationTableS1.Forthereactionsofglyoxal
relativehumiditybeyond80%,asmeasuredbyahumidity withglycineandserine,theimidazolesweretheonlyproducts
sensor(VaisalaHMT337)atthechamberoutlet.After∼20 detectedbyESI-MSor13Cor2-DNMRexperiments.However,
minoffilltime,theparticlesinthechamberweresampled the formation of imidazoles was always accompanied by
byQ-AMS,(AerodyneResearch)(27),HR-ToF-AMS(Aero- browningofthesolution.Thecompoundscausingthebrown
dyne)(28)operatinginthehighest-resolution“W”mode,or color are melanoidins, a term used by food chemists to
both.BothAMSinstrumentswereoperatedwithvaporizer describecomplex,light-absorbing,nitrogen-containingoli-
temperature)600°Cand70eVelectronimpactionization. gomercompoundsthatareendproductsofMaillardreactions
AMSsamplinglineswereblackconductivetubing(TSI)and/ betweensugarsandaminoacidswhereR-dicarbonylssuch
or1/4”o.d.copperand2-4minlength.Allotheraerosol asglyoxalarereactiveintermediates(32).Melanoidinsmay
flowswerethroughcoppertubing. havebeenresponsibleforseveralsmallpeaksobservedby
Direct uptake of glyoxal by solid-phase glycine aerosol 1HNMR.
was also probed by Q-AMS and SMPS. A 3 mM glycine Additionalproductswereobservedforthereactionsof
solutionwasultrasonicallynebulizedandthedropletswere glyoxal with aspartic acid and ornithine. In an ESI mass
driedto<44%RH,belowtheefflorescencepoint(29),inthe spectrum (Figure 2), H+ and Na+ adducts of unreacted
chamberdescribedabove.Watervaporwasthenaddedto
reach50%RH,stayingwellbelowthedeliquescencepoint
of95%RH(29).Gas-phaseglyoxalwasgeneratedbyheating
equal masses of glyoxal trimer dihydrate and phosphorus
pentoxide to 150 °C and trapping the yellow-green gas
producedat-105°C.Glyoxalwaspurifiedbytrap-to-trap
distillation,retainingthemiddlefraction(30).A1.5mLflow-
throughmetalvesselwaspumpedoutandpressurizedwith
afewmbarofglyoxal(measuredbyconvectiongauge),and
thevesselcontentsweresweptintothechamberwithdry
nitrogen. Glyoxal concentrations in the chamber were
calculatedtobe45ppbbasedondilutionintothecurrent FIGURE 1. Low wavenumber ATR-FTIR peaks observed at the
chambervolume.SMPStotalaerosolvolumeswerecorrected endofdryingofaqueousdropletsondiamondcrystal.Red:1uL
for volume dilution and for wall losses by fitting the total droplet of 0.095 M glycine. Blue dashed: 5 uL of 0.01 M glyoxal
aerosolvolumesbeforeglyoxaladditiontoameasuredfirst- (×2).Black:1uLdropletcontaining0.095Mglycineand0.05M
orderdecayconstant,0.0034min-1. glyoxal.
VOL.43,NO.8,2009 / ENVIRONMENTALSCIENCE&TECHNOLOGY92819
SCHEME 2. Formation Mechanism for 1,3-Disubstituted Imidazole and Diimine from Glyoxal and the Amino Acids Glycine (RdH),
Serine (RdCHOH), Ornithine (RdCHCHCHNH), and Aspartic Acid (RdCHCOO-); Odd Carbon Designated by *
2 2 2 2 2 2
aspartic acid are dominant at m/z 134 and 156, and the ThepreferencefortheR-aminepositionislikelyduetoits
imidazolepeakatm/z301isonlyoneofseveralsignificant higherprobabilityofdeprotonation;itspK is8.69compared
a
productpeaks.Whiletheglyoxal-ornithineimidazolepeak to10.76forthesidechainaminegroup.
atm/z299isdominantinESI-MS(Figure2),multipleisomers Theirreversibleformationof1,3-disubstitutedimidazoles
are observed by NMR because of ornithine’s two reactive has been previously observed for the room temperature
amine groups. Integration of 1H NMR data suggests that reaction of glyoxal and glycine (33), the reaction between
glyoxalreactsatornithine’sR-aminegroup80%ofthetime, glyoxal,formaldehyde,andseveralaminoacidsat80°C(34),
andonly20%ofthetimeatthesidechainaminegroup.The and the reaction of glyoxal with lysine side chains, which
reactionwithglyoxalrequiresadeprotonatedaminogroup. causesproteincross-linkingunderphysiologicalconditions
(35). The imidazole yield is highest under mildly acidic
conditions(pH3-5);however,melanoidinformation(brown
discoloration)wasmaximizedatpH8fornonacidicamino
acidsandpH5forglutamicacid(34).Wehypothesizethat
imidazoleformationintheabsenceofformaldehydebegins
withanucleophilicattackbyanunprotonatedaminegroup
on glyoxal monohydrate’s carbonyl moiety (Scheme 2).
Proton transfer followed by water loss produces an imine
intermediate. Reaction at glyoxal’s second carbonyl (after
dehydration)producesadiimine,aminorproducttowhich
weassignthe8.1ppm1HNMRpeak(SupportingInformation
TableS1)basedonotherdiiminescharacterizedbyNielsen
etal.(36).In1Mglyoxal+glycineexperiments,iminesreact
witheachother80%ofthetime(basedonrelative1HNMR
signalsfortheproducts),formingazwitterionintermediate,
whichcyclizesandgainsanH+toformthefive-membered
ring(36).DehydrationthentriggerstheCsCbondrupture
andlossofformicacid,formingthedisubstitutedimidazole
withan“oddcarbon”notbondedtoanothercarbon(marked
with*inScheme2).Severallinesofevidenceinsupportof
thismechanismarepresentedbelow.
Previousworkershaveattributedtheoddcarboninthe
imidazoleringtotheadditionofformaldehydeproducedin
the Strecker degradation of glycine (33). This mechanism
wouldleaveathirdaminoacidsidechainRattachedtothe
oddcarbonoftheimidazolering.Whiletheproductisthe
sameregardlessofthesourceoftheoddcarbonatomfor
the reaction with glycine (RdH), other amino acids will
FIGURE 2. Electrospray ionization mass spectra of the products
ofbulkphasereactionsofglyoxalwithaminecompounds,after producedifferentproducts.Sinceproductsmassescontaining
drying to a solid residue and redissolving residue in water to athirdRgroupwerenotobserved(Figure2)forserine(m/z
reach 1 mg/mL. Absolute abundances (all × 106): Glycine, 2.7 275), aspartic acid (m/z 359), or ornithine (m/z 356), we
Serine, 16.1. Aspartic acid, 0.88. Ornithine, 4.0. Arginine, 14.5. * concludethattheoddcarboninallourexperimentsmust
denotesmassof1,3-disubstitutedimidazoleproductshown. befromasecondglyoxalmolecule.
28209ENVIRONMENTALSCIENCE&TECHNOLOGY / VOL.43,NO.8,2009
TABLE 1. Room-Temperature 1H-NMR Kinetics Data for the Reactions of Glyoxal with Amino Acids (AA) in D2O in the Absence of
Drying
aminoacid [glyoxal](M) [AA](M) f a f AAinitiallossrate(mMd-1) 2ndorderrateconstant,kb(M-1s-1)
glyoxal AA
glycine 1.0 1.0 7×10-4 1.7×10-4c 1.9(0.6 0.12(0.04
serine 1.0 0.48 7×10-4 7.1×10-4c 2.0(0.2 0.09(0.01
arginine 1.0 0.86 7×10-4 1 24(3 (4.5(0.6)×10-4
aFractionofglyoxalinmonohydrateformestimatedusingthermodynamicparametersfromabinitiocomputations(21).
bExpressed in terms of concentrations of reactive forms of glyoxal and amino acids. cFraction of amino acid with
deprotonatedaminegroupatpH6.
SCHEME 3. Products Formed in Glyoxal + Arginine Reaction
FIGURE 3. Kinetics of amino acid loss during reactions with
glyoxal at room temperature in DO, without drying. Two
2
reactions per amino acid were performed; all data is shown
Similar quantities of N-derivatized imidazoles were together. Amino acids are quantified by relative peak areas of
producedevenifbulkglyoxal+aminoacidmixtureswere major 1H NMR signals, normalized to 1 at t ) 0. Initial
driedinthedarkorundernitrogen.Whileitisunusualfor concentrations: glyoxal and glycine ) 1.0 M, serine ) 0.48 M,
CsCbondstorupturewithoutphotolysisoroxidantchem- and arginine ) 0.86 M. Reaction rates are calculated from
istry,theveryhighthermodynamicstabilityofthearomatic initial loss rates, taken to be a linear fit to all data for glycine
imidazoleringispivotal.Wedonotbelievethattheimidazole and serine (as shown) and to first 7 days of data for arginine
peaksobservedbyESI-MSaretheresultoffragmentation (linear fit not shown). Curve shown for arginine is to guide the
becauseNMRdataclearlyindicatesahydrogenatomattached eye.
attheoddcarbon.Furthermore,thereisnoNMRevidence
fboereeixtpheecrtaendaitftatchheedseaclodnehdygdleyooxraglemca-rdbioonlgraotoumpthwaetrweosutildll -d[AA] )k [Glyoxal] f [AA] f (1)
dt AA tot glyoxal tot AA
attachedtotheimidazolering.
The reaction between glyoxal and arginine is different whereAAisanyaminoacid,k istherateconstantinM-1
AA
fromthoseoftheotheraminoacidsbecausearginine’sside s-1expressedintermsofthereactiveformsofeachreactant,
chaincontainsthreeaminegroups,twoofwhichcanreact [Glyoxal] and[AA] aretheconcentrationsofallformsof
tot tot
evenwhenoneisprotonated.Thus,glyoxalreactstwiceat thereactantsinMintheaqueousphase,andfisthefraction
the arginine side chain, forming 1:1 adducts rather than ofeachreactantinreactiveform:
imidazoles.Themajorproducts1and2(Scheme3),both
withprotonatedm/z233(Figure2),wereidentifiedbyNMR f ) [GlyoxalMonohydrate] (2)
andwereproducedinaratioof3:1.Productsfromthereaction glyoxal [Glyoxal]
ofglyoxalatthearginineR-aminegroupareatleast30times tot
lessabundant.Theimidazolidinederivativeof1,produced
-
[NH CHRCOO ]
byasingledehydrationfollowedbyhydridetransfer(37),is f ) 2 (3)
observedwhenglyoxalisincubatedat110°Cwithproteins AA [AA]
tot
(38).
Kinetics. Figure 3 and Supporting Information Figures These fractions were estimated using the results of ab
S2andS3showthetimedependenceofthepeakareasof initiocalculationsforglyoxal(21),pK valuesfortheamino
a
1H NMR peaks assigned to amino acids, diimine and acids glycine and serine, and experimental pH 6. Since
imidazoleproducts(SupportingInformationTableS1)for [Glyoxal] ≈[GlyoxalDihydrate]insolution,f represents
tot glyoxal
the reactions between glyoxal and arginine, serine, and an equilibrium constant for the dehydration reaction of
glycine.Agraphof1/[reactant]vstime(SupportingInfor- glyoxaldihydrate.Becausetheargininesidechainappears
mationFigureS2)ismostlinearforasecondorderreaction tobereactiveevenwhenprotonated,allargininemolecules
withreactantspresentinequimolarquantities,andthiswas are considered to be active-form reactants, f ) 1. Rate
Arg
observedforallthreereactions.Becauseglyoxal-aminoacid constantswerethencalculatedintermsofthesefractions
reactions must occur between glyoxal monohydrate and and measured initial reaction rates (Table 1). Glyoxal loss
aminoacidswithunprotonatedaminegroups,weexpress rates were not measured, but should be the same as the
theratelawas aminoacidlossrateslisted.
VOL.43,NO.8,2009 / ENVIRONMENTALSCIENCE&TECHNOLOGY92821
Time-dependentyieldsofdiimineandimidazoleproducts
(producedinglycineandserinesystems)remainedpropor-
tionalthroughouteachexperiment(SupportingInformation
FigureS3),suggestingthatdiiminesareside-productsrather
thanintermediates,consistentwithScheme1.Iminesignals
werenotobserved,consistentwithashort-livedintermediate
whoseformationisrate-determining.
Reactionstoichiometryandyieldscanbeestimatedby
comparing growth rates of NMR product peaks with the
aminoacidlossrates.Forthereactionbetweenglyoxaland
glycine, the production rate of the imidazole product
(measuredatthe7.4ppmpeakduetotwoCHgroupsonthe
ring)is60(20%ofthelossrateforglycine(measuredatthe
CH grouppeakat3.4ppmandpropagating1σuncertainties
2
in the slopes). Because each of the signals is due to 2 H
FIGURE4. Q-AMSpeakheightchangesforpeaksexhibitingthe
atoms,theratesaredirectlycomparable,andsuggestthat
∼1.2(0.4imidazoleproductmoleculesformforeachtwo largest relative change upon exposure of solid glycine aerosol
at 50% RH to 45 ppb gas-phase glyoxal. Changes are expressed
glycinemoleculesthatreact.Similarly,theproductionrate
as ratios of average signals postexposure divided by
forthediimineintermediate(measuredatthe8.1ppmpeak
pre-exposure. The total AMS aerosol-phase signal declined by
duetotwoCHiminegroups)is15(5%oftheglycineloss 41% due to wall losses. All bars are referenced to this
rate,indicatingthat0.3(0.1diimineproductmoleculesare benchmark, so that any peaks that declined by more than this
formedpertwoglycinereactants.Thetotalobservedproduct are shown as negative, whereas peaks that increased in size
formation stoichiometry is 1.5 ( 0.5 molecules per two extend above 1. HR-ToF-AMS peak identifications are shown in
glycine,withinuncertaintyoftheexpectedstoichiometryof figure. Other peaks: m/z 30 (CHN+ glycine fragment), 47 (formic
4
1:2,andsuggestingthatnootherproductswereformedin acid or H3CO2+ fragment), 86 (product fragment C3H4NO2+), 97
significantamounts.Asimilaranalysisofthereactionwith (product fragment C4H3NO2+). Full Q-AMS spectra are shown in
arginineconfirmsnetstiochiometricyieldsforproducts1 SupportingInformationFigureS4.
and2(SupportingInformationTableS1).
ConcentrationDependenceandCloudDropletDrying. bothreactantstoactiveformsbythetimethewatermole
We simulated the drying of cloud droplets by generating fractiondeclinesto0.7,buttodifferentextents.
aqueous∼5µmdropletscontaining4mMglycineandglyoxal Glyoxal Uptake by Aerosol. Since reactions between
(butnoinorganicsalts)andintroducingthemintoachamber glyoxalandaminoacidsaccelerateduringdropletevapora-
tion,theymayoccuratthehighestratesinaerosolparticles
at90%RH.Whentheresultingaerosolwassampled10min
(13). We examined this process by exposing solid-phase
laterbyHR-ToF-AMS,1,3-disubstitutedimidazoleproduct
glycineaerosol(modediameter)84nm)at50%RHto45
wasobservedatm/z185,alongwithlargerfragmentpeaks
suchasm/z86,CHNO+.Therelativeabundancesofthese ppbofgas-phaseglyoxal,whileanalyzingtheresultantaerosol
3 4 2 by SMPS and Q-AMS. (All fragment identifications were
peakschangedonlyslightlywithfurthersamplingordrying,
confirmedbyHR-ToF-AMSindryingdropletstudies.)Total
suggestingthatthereactionwasnearlycompletebythefirst
aerosolvolumeatthetimeofglyoxaladditionwas1.2×103
sampling. When the experiment was repeated at atmo-
µm3cm-3,whichmeansthatglyoxalandglycinewerepresent
sphericallyrelevantreactantconcentrationsof32µM(125
in equimolar quantities in the chamber. The addition of
times less concentrated), the molecular ion peak was no
glyoxaltriggered30min.ofparticlegrowth,increasingaerosol
longervisible,butthem/z86fragmentwasstilldetectable
volumeby20%(afterwalllossandvolumedilutioncorrec-
withanabundance2000timeslower.Otherproductfragment
tions)andincreasingparticle-phaseglyoxalsignalsatm/z
ionsobservedbybothQ-AMSandHR-ToF-AMS(containing
58 (Figure 4 and Supporting Information Figure S4). Fur-
bothCandN,m/z97,CHNO+andm/z126,CHNO+),
4 3 2 5 6 2 2 thermore,significantsignalincreasesoccurredatm/z184
respondedproportionallytochangesinreactantconcentra-
and 185 (imidazole product), 86 and 97 (major product
tions, suggesting that the glycine-glyoxal reaction has an fragments), and 47 (HCO+), which is either formic acid
overallreactionorderof∼1.6,roughlyconsistentwithNMR 3 2
producedinScheme1,ormorelikelyanionizationfragment
kineticsresultsdescribedabove.
of larger aerosol-phase molecules. Glyoxal addition also
TheAMSresultsprovidedirectevidencethatimidazole causedglycineQ-AMSpeaksatm/z30and73-75todecline
formation can occur on a time scale of minutes in an
by half (after normalizing for the general decline in total
evaporatingdropletthatcontainsatmosphericconcentra-
aerosol-phasesignal)(Figure4andSupportingInformation
tionsofglyoxalandaminoacid(2-200µM(39,41)and1-100
Figure S4). If half of the glycine reacts to form imidazole
µM(24,25),respectively).Ifweassumethatthedryingprocess productviaa1:1reactionwithglyoxal,ifbyproductwater
in our experiment results in aerosol with water/glycine/ andformicacidevaporateintheAMSinletasexpected,and
glyoxalpresentata5:1:1molarratioat90%RH(consistent ifthedensityofthe1,3-disubstitutedimidazoleproductis
withelectrodynamicbalancemeasurementsforwaterand thesameasthatofimidazole(1.03g/cm3),thentheexpected
glycine(29)),soluteconcentrationswouldbe∼5M.Using volumegrowthwouldbe19%,matchingourSMPSaerosol
theactiveformfractionsfandrateconstantlistedinTable growthobservation.
1, we would predict that the reaction would take months Theseobservationssuggestthatglyoxaluptakeontosolid
unlessweincludethedrying-inducedconversionofglyoxal amino acid aerosol is a reactive process and not simply a
dihydratetoitsreactivemonohydrateform.Withfglyoxal)1, function of the interaction of glyoxal with aerosol water
predictedreactiontimesarewithinafactorof6ofthe∼30 content.Thecessationofparticlegrowthafter30minislikely
minreactiontimesobservedbytheAMS,butstilltooslow. causedbyproductbuildupatthesolidaerosolsurface,which
Theremainingdifferenceislikelyduetoenhanceddepro- physicallyseparatesthereactants.Anuptakecoefficientfor
tonationoftheglycineaminegroupinthemixedorganic/ glyoxal on solid-phase glycine, defined as the fraction of
water aerosol phase making f > 1.7 × 10-4. Thus, the molecularcollisionswiththeaerosolsurfacethatresultin
glycine
rapidityoftheobservedreactionindryingdropletsrelative glyoxaluptake,canbeestimated(12)basedonthisexperi-
to bulk solutions is further evidence that drying converts menttobeR)4×10-4.Thisisaboutanorderofmagnitude
28229ENVIRONMENTALSCIENCE&TECHNOLOGY / VOL.43,NO.8,2009
lessthantheuptakecoefficientinferredforglyoxalonMexico (8) Heald,C.L.;etal.Alargeorganicaerosolsourceinthefree
Cityaerosol(9). tropospheremissingfromcurrentmodels.Geophys.Res.Lett.
Atmospheric Significance. If a 5 µm cloud droplet 2005,32,L18809.
(9) Volkamer,R.;etal.Amissingsinkforgas-phaseglyoxalinMexico
contained 30 µM of both glyoxal and amino acids, and a
City:formationofsecondaryorganicaerosol.Geophys.Res.Lett.
complete conversion to imidazole product ensued, the 2007,34,L19807.
amount of resulting imidazole product would make up at (10) Liggio,J.;Li,S.-M.;McLaren,R.Reactiveuptakeofglyoxalby
least2%byvolumeofatypicalcloudprocessedaerosolwith particulatematter.J.Geophys.Res.2005,110,D10304.
diameter)130nm.Thus,wecanestimatethatthereactants (11) Kroll,J.H.;etal.Chamberstudiesofsecondaryorganicaerosol
glyoxal and amino acids in an evaporating cloud droplet growthbyreactiveuptakeofsimplecarbonylcompoundsJ.
Geophys.Res2005,110,D23207,DOI:10.1029/2005JD006004.
wouldbedilutedbyaboutafactorof50byothernonvolatile
(12) Hastings,W.P.;etal.Secondaryorganicaerosolformationby
compoundscomparedtooursimulatedclouddropletdrying glyoxalhydrationandoligomerformation:humidityeffectsand
experiments. Under these circumstances, using aerosol- equilibriumshiftsduringanalysis.Environ.Sci.Technol.2005,
phasereactantconcentrationsof5M/50)0.1M,f ) 39,8728–8735.
glyoxal
1andf )6×1.7×10-4,andtherateconstantforthe (13) Corrigan,A.L.;Hanley,S.W.;DeHaan,D.O.Uptakeofglyoxal
glycine
glyoxal+glycinereaction,conversionto1,3-disubstituted byorganicandinorganicaerosol.Environ.Sci.Technol.2008,
imidazoleswouldtake∼1day,muchlessthantheaverage 42,4428–4433.
(14) Lim,H.-J.;Carlton,A.G.;Turpin,B.J.Isopreneformssecondary
atmosphericlifetimeofaerosolparticles.Thetemperature organicaerosolthroughcloudprocessing:Modelsimulations.
dependenceofglyoxal-aminoacidreactionsisunknown, Environ.Sci.Technol.2005,39,4441–4446.
however. (15) Altieri,K.E.;etal.Evidenceforoligomerformationinclouds:
Glyoxalconcentrationsrangefrom10to40pptovermost reactionsofisopreneoxidationproducts.Environ.Sci.Technol.
2006,40,4956–4960.
oftheearth’ssurface,withenhancedlevelsobservedover
(16) Ervens,B.;etal.Secondaryorganicaerosolyieldsfromcloud-
tropical,oceanic,andurbanregions(42).Aminoacidsare
processingofisopreneoxidationproducts.Geophys.Res.Lett.
oftenenrichedinmarineprecipitationbyafactorof10over 2008,35,L02816.
terrestrialprecipitation(24),andby∼3ordersofmagnitude (17) Sorooshian,A.;etal.Onthesourceoforganicacidaerosollayers
over surface seawater (43). Measured organic N/C atomic aboveclouds.Environ.Sci.Technol.2007,41,4647–4654.
ratioshaverangedfrom∼0.02inMexicoCityaerosol(44)to (18) Fu, T.-M.; et al. Global budgets of atmospheric glyoxal and
∼0.2inCentralValleyfog(45).Imidazoleandmelanoidin methylglyoxal, and implications for formation of secondary
formationfromglyoxal-aminoacidreactionslikelyoccurs organicaerosols.J.Geophys.Res.2008,113,D15303.
(19) Ervens,B.;etal.Amodelingstudyofaqueousproductionof
overvastregionsoftheocean,overrainforests,andin“hot dicarboxylicacids:1.Chemicalpathwaysandspeciatedorganic
spots” where urban plumes mix with biologically derived massproduction.J.Geophys.Res.2004,109,D15205.
aerosol. (20) Loeffler,K.W.;etal.Oligomerformationinevaporatingaqueous
glyoxalandmethylglyoxalsolutions.Environ.Sci.Technol.2006,
Acknowledgments 40,6318–6323.
(21) Kua,J.;Hanley,S.W.;DeHaan,D.O.Thermodynamicsand
ThismaterialisbaseduponworksupportedbytheNational
kineticsofglyoxaldimerformation:acomputationalstudy.J.
ScienceFoundationunderGrantNo.ATM-0749145andby Phys.Chem.A2008,112,66–72.
ResearchCorporationforScienceAdvancement.Wethank (22) Nozie`re,B.;Dziedzic,P.;Co´rdova,A.Formationofsecondary
MichaelCubisonandDonnaSueperforassistancewithHR- light-abosrbing“fulvic-like”oligomers:acommonprocessin
ToF-AMShardwareandsoftwareandBrandonConnellyfor aqueousandionicatmosphericparticles?Geophys.Res.Lett.
2007,34,L21812.
synthesizinggas-phaseglyoxal.
(23) Nozie`re,B.;Co´rdova,A.Akineticandmechanisticstudyofthe
aminoacidcatalyzedaldolcondensationofacetaldehydein
Note Added after ASAP Publication
aqueousandsaltsolutions.J.Phys.Chem.A2008,112,2827–
TworeferenceswereaddedtotheversionpublishedonMarch 2837.
12,2009.ThecorrectedversionwaspublishedonMarch20, (24) Mopper,K.;Zika,R.G.Freeaminoacidsinmarinerains:evidence
2009. foroxidationandpotentialroleinnitrogencycling.Nature1987,
325,246–249.
Supporting Information Available (25) Zhang,Q.;Anastasio,C.Freeandcombinedaminocompounds
in atmospheric fine particles (PM ) and fog waters from
NMRpeakassignments,ATR-FTIRdata,NMRkineticsdata, NorthernCalifornia.Atmos.Environ2..52003,37,2247–2258.
andQ-AMSspectrabeforeandafterglyoxaluptakebyglycine (26) Matsumoto,K.;Uematsu,M.Freeaminoacidsinmarineaerosols
aerosol. This material is available free of charge via the overthewesternNorthPacificOcean.Atmos.Environ.2005,
Internetathttp://pubs.acs.org. 39,2163–2170.
(27) Canagaratna,M.R.;etal.Chemicalandmicrophysicalchar-
Literature Cited acterizationofambientaerosolswiththeAerodyneaerosolmass
spectrometer.MassSpectrom.Rev.2007,26,185–222.
(1) Watson, J. G. Visibility: science and regulation. J. Air Waste (28) DeCarlo,P.F.;etal.Field-deployable,high-resolution,time-
Manage.Assoc.2002,52,628–713. of-flightaerosolmassspectrometer.Anal.Chem.2006,78,8281–
(2) Harrison,R.M.;Yin,J.Particulatematterintheatmosphere: 8289.
whichparticlepropertiesareimportantforitseffectsonhealth. (29) Chan, M. N.; et al. Hygroscopicity of water-soluble organic
Sci.TotalEnviron.2000,249,85–101. compoundsinatmosphericaerosols:aminoacidsandbiomass
(3) Kanakidou, M.; et al. Organic aerosol and global climate derivedorganicspecies.Environ.Sci.Technol.2005,39,1555–
modelling:Areview.Atmos.Chem.Phys.2005,5,1053–1123. 1562.
(4) Zhang,Q.;etal.Ubiquityanddominanceofoxygenatedspecies (30) Volkamer,R.;Ziemann,P.J.;Molina,M.J.Secondaryorganic
inorganicaerosolsinanthropogenically-influencedNorthern aerosolformationfromacetylene(C2H2):seedeffectonSOA
Hemispheremidlatitudes.Geophys.Res.Lett.2007,34,L13801/ yieldsduetoorganicphotochemistryintheaerosolaqueous
1–6. phase.Atmos.Chem.Phys.Discuss.2008,8,14841–14892.
(5) Murphy, D. M.; et al. Single-particle mass spectrometry of (31) Fischer,G.;etal.TheFT-IRspectraofglycineandglycineglycine
troposphericaerosolparticlesJ.Geophys.Res2006,111,D23S32, zwitterionsisolatedinalkalihalidematrices.Chem.Phys.2005,
DOI: 10. 313,39–49.
1029/2006JD007340. (32) Cammerer,B.;Jalyschko,W.;Kroh,L.W.Intactcarbohydrate
(6) Hallquist,M.;etal.Theformation,propertiesandimpactof structuresaspartofthemelanoidinskeleton.J.Ag.FoodChem.
secondaryorganicaerosol:currentandemergingissues.Atmos. 2002,50,2083–2087.
Chem.Phys.Discuss.2009,9,3555–3762. (33) Velisek, J.; et al. New imidazoles formed in nonenzymatic
(7) Volkamer,R.;etal.Secondaryorganicaerosolformationfrom browningreactions.J.FoodSci.1989,54,1544–1546.
anthropogenicairpollution:Rapidandhigherthanexpected (34) Davidek, T.; et al. Amino acids derived 1,3-disubstituted
Geophys.Res.Lett.2006,33,L17811,DOI:10.1029/2006GL026899. imidazolesinnonenzymaticbrowningreactions.Sb.Vys.Sk.
VOL.43,NO.8,2009 / ENVIRONMENTALSCIENCE&TECHNOLOGY92823
Chem.-Technol.Praze,E:Potraviny1991,62,165–182. (41) Matsumoto,K.;Kawai,S.;Igawa,M.Dominantfactorscontrolling
(35) Chellan,P.;Nagaraj,R.H.ProteincrosslinkingbytheMaillard concentrationsofaldehydesinrain,fog,dewwater,andinthe
reaction:dicarbonyl-derivedimidazoliumcrosslinksinaging gasphase.Atmos.Environ.2005,39,7321–7329.
anddiabetes.Arch.Biochem.Biophys.1999,368,98–104. (42) Wittrock,F.;etal.Simultaneousglobalobservationsofglyoxal
(36) Nielsen,A.T.;etal.Polyazapolycyclicsbycondensationof and formaldehyde from space. Geophys. Res. Lett. 2006, 33,
aldehydeswithamines.2.Formationof2,4,6,8,10,12-hexa- L16804.
benzyl-2,4,6,8,10,12-hexaazatetracyclo[5.5.0.05.9.03,11]dode- (43) Milne,P.J.;Zika,R.G.Aminoacidnitrogeninatmospheric
canesfromglyoxalandbenzylamines.J.Org.Chem.1990,55, aerosols:occurrence,sourcesandphotochemicalmodification.
1459–1466. J.Atmos.Chem.1993,16,361–398.
(37) Chassonery,D.;etal.Reactionsdedismutationapartirduglyoxal (44) Aiken,A.C.;etal.O/CandOM/OCratiosofprimary,secondary,
etdemoleculesbasiquesdifonctionnelles.Bull.Soc.Chim.Fr. and ambient organic aerosols with high-resolution time-of-
1994,131,188–199. flightaerosolmassspectrometry.Environ.Sci.Technol.2008,
(38) Schwarzenbolz,U.;etal.Onthereactionofglyoxalwithproteins. 42,4478–4485.
Z.Lebensm.Unters.Forsch.A1997,205,121–124. (45) Zhang,Q.;Anastasio,C.ChemistryoffogwatersinCalifornia’s
(39) Igawa, M.; Munger, J. W.; Hoffmann, M. R. Analysis of Central Valley-Part 3: Concentrations and speciation of
aldehydesincloud-andfogwatersamplesbyHPLCwitha organic and inorganic nitrogen. Atmos. Environ. 2001, 35,
postcolumnreactiondetector.Environ.Sci.Technol.1989,23, 5629–5643.
556–561.
(40) Munger,J.W.;etal.Formaldehyde,glyoxal,andmethylglyoxal
inairandcloudwaterataruralmountainsiteincentralVirginia.
J.Geophys.Res.1995,100,9325–9333. ES803534F
28249ENVIRONMENTALSCIENCE&TECHNOLOGY / VOL.43,NO.8,2009
Secondary organic aerosol-forming reactions of
glyoxal with amino acids
Supplemental Information
David O. De Haan,1,2* Ashley L. Corrigan, 1 Kyle W. Smith, 1 Daniel R. Stroik, 1 Jacob J.
Turley, 1 Frances E. Lee,1 Margaret A. Tolbert, 2 Jose L. Jimenez,2 Kyle E. Cordova,2
Grant R. Ferrell2

Contents: 6 pp total
Table S1
Figures S1 – S4
S2
Table S1: 1H- and 13C-NMR peak assignments for the major (left side) and minor (right
side) products of the reactions of glyoxal with amino acids.
Product
Structures:

N-derivatized imidazole Diimine
A B C R-group D
Glycine 1H 8.6 7.3 4.6 na 8.1
13C 137 123 52 na
Serine 1H 8.8 7.4 4.0 4.1 8.2
Ornithine 1H 8.7 7.4 4.1 1.8 1.6 2.9 8.0
13C 136 123 63 25 23 39
Aspartic 1H 8.97 a 7.4 5.28 2.5 8.3
acid
13C 138.0 a 123.4 62.4 a 39.0 a
a
Product
Structures:

2
1
A B C R A B
Arginine 1H 5.0 5.0 3.3 1.7 1.8 3.6 5.2 5.2
13C 83 89 41 24 28 54 77 83
a: NMR data from Davidek, et al.,(32) synthesized including formaldehyde.