Site-specificN-glycosylationofchickenserumIgGNorikoSuzuki1andYuanC.LeeDepartment
of Biology, Johns Hopkins University, Baltimore, MD
21218ReceivedonSeptember27,2003;revisedonNovember7,2003;acceptedonNovember10,2003Avianserumimmunoglobulin(IgGor
IgY) is functionallyequivalent to mammalian IgG but has one
additional constantregion domain (CH2) in its heavy (H) chain. In
chicken IgG,each H-chain contains two potential
N-glycosylationsiteslocated on CH2 and CH3 domains. To clarify
characteristicsof N-glycosylation on avian IgG, we analyze
N-glycans
fromchickenserumIgGbyderivatizationwith2-aminopyridine(PA) and
identified by HPLC and MALDI-TOF-MS.Thereweretwotypesof N-glycans:
(1) high-mannose-typeoligosaccharides (monoglucosylated 26.8%,
others
10.5%)and(2)biantennarycomplex-typeoligosaccharides(neutral,29.9%;
monosialyl, 29.3%; disialyl, 3.7%) on molar basis oftotal
N-glycans. Toinvestigatethesite-specificlocalizationof different
N-glycans, chicken serum IgG was digested
withpapainandseparatedintoFab[containingvariableregions(VHVL) CH1
CL] andFc (containingCH3 CH4)fragments. Con A stained only Fc (CH3
CH4) and RCA-Istained only Fabfractions, suggesting that
high-mannose-type oligosaccharides were located on Fc (CH3
CH4)fragments, andvariable regions of Fabcontains
complex-typeN-glycans. MSanalysisof
chickenIgG-glycopeptidesrevealed that chicken CH3 domain
(structurally equivalent tomammalian CH2 domain) contained only
high-mannose-typeoligosaccharides, whereas chicken CH2 domain
containedonlycomplex-typeN-glycans.
TheN-glycosylationpatternonavianIgGismoreanalogoustothatinmammalianIgEthanIgG,
presumablyreflectingthestructural similaritytomammalian
IgE.Keywords:Asn297/chickenserumIgG/IgG-Fc/monoglucosylatedhigh-mannose-type/N-glycan
processingIntroductionAvian IgG, the predominant
serumimmunoglobulin inbirds,
iscloselyrelatedtobothmammalianIgGandIgE,basedontheirfunctional
andstructural properties(Warret al., 1995). Incontrast
tomammalianIgG, avianIgGcontainsoneadditional
domainintheconstantregionofits heavy(H) chains (designatedupsilon,
u), but it lacksfunctional hinge regions found in mammalian
IgG(Figure1A)(Magoretal.,1992;Parvarietal., 1988;Warret al., 1995).
In short, the CH3 and CH4 domains
ofchicken/duckIgGresembletheCH2andCH3domainsofmammalianIgGinstructure,respectively,andtheequiva-lent
of the CH2 domain in avian IgG is absent in
mamma-lianIgG.BecauseofitsdistinctstructuraldifferencefrommammalianIgG,
avianIgGisalsocalledIgY(LeslieandClem,1969;Warretal.,1995).TheIgY-likemoleculesarealsofoundinreptilesandamphibians(Fellahetal.,1993;Warr
et al., 1995). Structural properties of IgY in birds andamphibians
are rather close to mammalian IgE with respectto the number of CH
domains as well as the organization
ofintradomainandinterchaindisulfide bonds (Figure 1A)(Fellahet
al.,1993; Parvariet al.,1988; Warret al.,1995).Itishypothesizedthat
g andegeneshavebeengeneratedfromarelativelyrecent
gene-duplicationevent andthatIgY-likemoleculewas
theimmediateprogenitor bothofIgG and IgE (Parvari et al., 1988;
Warr et al., 1995).Basedontheir aminoacidsequences,
chicken(Parvarietal.,1988)andduck(Magoretal.,1992)IgGssharetwopotential
N-glycosylationsites, predictablefromthecon-sensus sequence(sequon)
inconstant regions (Figure1Aand 1B). One of them is located in the
CH2 (Cu2) domain,which is absent in mammalian IgG. The other is
located inthe CH3 (Cu3) domain, whichcorresponds tothe CH2(Cg2)
domain of mammalian IgG (Asn297,
Eu-numbering).Figure1Bshowsthesequencealignment forH-chainsofavian
(chicken, duck) IgG (Cu-chains), mammalian(human) IgG1 (Cg1-chain),
mammalian(human, mouse,rat) IgE(Ce-chains), and amphibian (Xenopus,
axolotl)IgY(Cu-chains) aroundthe twosequons onavianIgGH-chains. The
sequence alignment indicates that thelocationof bothpotential
N-glycosylationsites of
avianIgGsarealsoconservedinmammalianIgE(except CH2inhumanIgE).
Incontrast, sequons of amphibianIgYare located at different
positions fromavian IgGandmammalianIgG/IgE, probablyreflectingthe
evolutionaldistances between mammals/avians and amphibians.We
recentlyfoundthat pigeonserumIgGhas
uniqueN-glycanfeatures,suchasthepresenceofalargequantityof
highlygalactosylatedtriantennaryoligosaccharides
aswellasmonoglucosylatedhigh-mannose-typeoligosaccha-rides
(monoGlc-high-Man) (Suzukiet al., 2003), which arenot
foundinmammaliannormal serumIgG. MonoGlc-high-Man is probably a
characteristic in avian IgG
becauseithadbeenalsofoundinchicken(Ohtaetal., 1991)andquail
(Matsuuraet al., 1993) eggyolkIgGs andchickenserum IgG (Raju et
al., 2000). Except avian IgGs,
however,themonoGlc-high-Manisrarelyfoundinsecretedmatureglycoproteins,andonlytransientlyexistsonglycoproteinsduringfoldingprocessintheendoplasmicreticulum(ER).The
currently proposedmechanismis that after
correctfoldingofglycoproteins,
themonoGlcresidueisremoved1Towhomcorrespondenceshouldbeaddressed;e-mail:[email protected]
vol.14 no.3#OxfordUniversityPress2004;allrightsreserved.
275Glycobiologyvol.14no.3pp.275292,2004DOI:10.1093/glycob/cwh031AdvanceAccesspublicationDecember23,2003
by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from by
a-glucosidase II (GII) (Helenius and Aebi, 2001; Parodi,2000). The
retained monoGlc-high-Man of avian IgG mighthave resultedfromthe
steric hindrance imposedby theuniqueconformational
structuresofavianIgG. However,the relationship between the protein
structure of avian IgGandits N-glycosylationpatternhas not
beenexamined.Because chicken IgGis a goodsource of
glycoproteinsthat contain monoGlc-high-Man and its peptide
sequencesareknown,webelieveditisworthwhileinvestigatinghowsuch
unique oligosaccharides exist in chicken IgG.Fig. 1. Structures of
Igs and their N-glycosylation sites. (A) Structures of avian IgG,
mammalian IgG, and mammalian IgE molecules. Avian(chicken, duck)
IgGhas five domains in the heavy chain, termed VH (in variable
region), CH1, CH2, CH3, and CH4 (in the constant region). The
domainsof the light chains are termed VL (in the variable region)
and CL (in constant region). The potential N-glycosylation sites on
avian IgG are shown ashexagons. N-glycosylation on variable regions
(gray hexagon) could occur sometimes, depending on the peptide
structures. Location of disulfide bondslinking the chains were
predicted from analogy to human IgE (Parvari et al., 1988; Wan et
al., 2002), indicated with dotted bold lines. Numbering forchicken
IgG H-chain (u-chain) is based on the deduced amino acid sequences
from cDNA, starting from the first methionine in the leader region
(Parvariet al., 1988; Reynaud et al., 1989). Position of actual
N-glycosylated sites on human IgG1 and IgE were shown in hexagons.
Interchain disulfide bonds arebased on human IgG1 and IgE
structures, respectively (Paul, 1999; Wan et al., 2002), and
indicated as bold lines. Numbering for mammalian Ig g-chain isbased
on Eu numbering (Edelman et al., 1969). Numbering for mammalian Ig
e-chain is modified from a reference (Dorrington and Bennich, 1978)
tocomply with an human immunoglobulin e-chain of 547 amino acids.
(B) Comparison of amino acid sequences around potential
N-glycosylation sites onchicken IgG H-chain and others. Potential
N-glycosylation sites were indicated as bold. Residue numbers on
Asn were given for chicken IgG H-chain.Homologies of these
sequences are not always high, but Trp and Cys residues (indicated
with arrows), which are hallmarks of Ig domains, are wellconserved
each other. Primary accession numbers for Entrez database are;
chicken upsilon chain, CAA30161; duck upsilon chain, CAA46322;
humangamma chain, CAC20454; human epsilon chain, AAB59424; mouse
epsilon chain, EPC_MOUSE; rat epsilon chain, AAA41364; Xenopus
upsilon chain,S04845; axolotl upsilon chain,
CAA49247.N.SuzukiandY.C.Lee276 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from Inthis study,
we first determine the
detailedN-glycanstructuresofchickenserumIgGforthecomparisonwiththoseofpigeonIgGandthendemonstratethesite-specificN-glycosylationonchickenIgG.
Ourdataindicatedthathigh-mannose-type N-glycans are exclusively
located onCH3 domains (as we found in pigeon IgG),
whereascomplex-typeN-glycansarepresent inCH2domainsandFab
regions(most likelyvariable regions).N-glycosylationpatterns,
including the oligosaccharide structures, onchicken IgGare more
similar to mammalian IgEthanto IgG, which is probably due to the
structural simi-larityinheritedinmolecular
evolutionfromIgYtoIgElineage.
Byanalogyofthe3DstructureofhumanIgE-Fc(containing Ce2-4 domains)
(Wan et al., 2002), we speculatethat protein folding and assembly
mechanisms of avian IgGenables retaining monoGlc-high-Man on the
secretedproteins.ResultsStructuralanalysisofPA-derivatizedoligosaccharidesfromchickenIgGToinvestigate
structural profile of N-glycans inchickenserum IgG, we utilized a
3D mapping technique (Takahashietal., 1995b),
inwhich2-aminopyridine(PA)derivatizedN-glycans were chromatographed
on high-performanceliquidchromatography(HPLC)
using(1)anionexchangewithaDEAEcolumn, (2) reversed-phase
withanocta-decylsilica(ODS)column,andthen(3)normalphasewithan
Amide-80 column. The elution positions of eachPA-oligosaccharide
were recordedas glucoseunits (GU)(TableI), andtheir structures
weredeterminedbasedontheelutionpositionsandmatrix-assistedlaserdesorption/ionization
time-of flight mass spectrometry (MALDI-TOFMS) data (Table I) by
comparing with reference PA-derivatized oligosaccharides (Figure
2).TotalPA-oligosaccharidesfromchickenIgGweresepa-ratedintoneutral,mono-,anddisialyloligosaccharidesonaDEAEcolumn(Figure3A).
MSanalysis revealedthatneutralfractions eluted between 5 and
15minon theODScolumn(Figure3B) wereHex711HexNAc2-PA(datanotshown),
suggesting that these are high-mannose-type oligo-saccharides. The
three major peaks (n-4, n-5, n-6) wereisolated, digestedwith
a-mannosidase, andfurtherexam-ined by 2D HPLC mapping (Tomiya et
al., 1988). Fractionn-4 showed the same elution position as
Man9GlcNAc2-PAfrombovineRNaseB(Figure2),
andbothmovedtothepositionofMan1GlcNAc2-PAafter
a-mannosidasediges-tion (Figure 4A), suggesting that n-4 has the
same structureas Man9GlcNAc2-PA. This was supported by the MS
dataof n-4before(m/z1984.86; Hex9HexNAc2-PA)
andafter(m/z687.97;Hex1HexNAc2-PA) a-mannosidasedigestion.Onthe
other hand, neither n-5(m/z 1984.81; Hex9Hex-NAc2-PA) nor n-6 (m/z
2147.06; Hex10HexNAc2-PA)could be digested to Hex1HexNAc2-PA under
the same con-ditions used for n-4 (Figure 4A). Mild a-mannosidase
diges-tion(50mU/100pmolePA-N-glycans,37
C,overnight)ofn-6 yielded a major peak on either ODS or
Amide-80column (GUODS7.1; GUAmide6.9; m/z 1498.60,Hex6HexNAc2-PA),
which was transformed to a peak(GUODS6.5; GUAmide5.9; m/z 1336.26,
Hex5HexNAc2-PA) on exhaustive digestion with a-mannosidase (200
mU/100 pmole PA-N-glycans, 37
C, two overnight). This mightbe a reflection of difficulty of
removal of the a1-6 mannoseresidue by jack bean
a-mannosidase.Althoughthe final product of
a-mannosidase-digestedn-6hadthesamemassvalueasMan5GlcNAc2-PAfrombovineRNaseB(GUODS7.3;GUAmide6.1),itelutedatadistinctly
different position on an ODS column (Figure 4A).These results
suggest that one of the
a-mannosylatedbranchesonn-6areblockedatthenonreducingterminus,so
that a-mannosidase could not trim it to
Man1GlcNAc2-PA.Theblockingismostlikelybyaglucosylationonthenonreducing
terminal of Mana1-2Mana1-2Mana-
branch,becausetherelativeelutionpositionof n-6onbothODSandAmide-80
columns were coincidental with those ofGlc1Man9GlcNAc2-PA(Tomiya et
al., 1988). Fractionn-5, smaller by one hexose than n-6, yielded
the sameproduct asn-6aftera-mannosidasedigestion,
suggestingthatthisalsohasamonoglucosylatedbranch, butshorterby
onea-mannoside residue thann-6. Because n-5 waseluted earlier than
n-6 on the ODS column, one a1-2-mannoside residue on the
Mana1-3Mana1-6Manb1-4GlcNAc arm is absent in n-5 (Tomiya et al.,
1991; Tomiyaand Takahashi,1998).Thusthestructuresofn-4,n-5,andn-6
were deduced as shown in Table I.Elution positions of the remaining
eight fractions of neu-tral oligosaccharides fromchickenIgG, n-9,
n-10, n-11,n-12, n-14, n-15, n-16, and n-17, were coincidental
onODSandAmide-80 columns withthose of humanIgGN-glycans F, J, H, L,
M, N, O, and P, respectively (Figure
3B,Figure4B,andTableI).MALDI-TOFMS(TableI)gavegoodagreement
withthese results. The structure assign-ments were also supported
by digestion with b-galactosidaseand/or a-fucosidase (Figure 4B).
After b-galactosidasedigestion, the elution positions of the
products on the ODSand Amide-80 columns were shifted as follows:
(1) both n-9andn-11yieldedthe same
GlcNAc-terminatedstructure(N-glycanEinFigure2),whichhascore
a1-6fucoseresi-dues andnobisecting GlcNAc; (2) bothn-10
andn-12yieldedthesameproduct(N-glycanIinFigure2),
whichhasbisectingGlcNAcandno a1-6corefucose; (3) n-15,n-16, and
n-17 all yielded the same product (N-glycan M inFigure2),
whichhasbothcore a1-6fucoseandbisectingGlcNAc. After a-fucosidase
digestion, n-9, n-11, n-14, n-15,n-16,
andn-17yieldedtheexpectedrespectivefucose-lessstructures (N-glycans
B, D, I, J, K, and L, respectively, as inFigure 2). Before the
exoglycosidase-digestion, n-9 (GUODS13.5; GUAmide6.3)
andn-10(GUODS13.5; GUAmide6.2)exhibit very close GUvalues. However
they are distin-guished by m/z values (Table I) as well as by the
sensitivityto a-fucosidase. Fractions n-15 (GUODS 19.5; GUAmide
6.5)andn-16(GUODS19.6; GUAmide6.6) alsoexhibit similarGU values,
but a-fucosidase-treated n-15 and n-16 (i.e.,
N-glycansJandK,respectively)wereclearlydistinguishablebytheir
elutionpositions ontheODScolumn. Thus thestructures of n-9, n-10,
n-11, n-12, n-14, n-15, n-16, andn-17 were firmly assigned (Table
I).Three fractions of monosialylatedPA-oligosaccharides(ms-5, ms-7,
andms-8) andone fractionof
disialylatedGlc1Man9GlcNAc2-AsnonchickenIgG-CH3domains277 by guest
on September 5, 2012http://glycob.oxfordjournals.org/Downloaded
from TableI. Assignment of the major PA-oligosaccharides from
chicken IgG based on HPLC and MSN.SuzukiandY.C.Lee278 by guest on
September 5, 2012http://glycob.oxfordjournals.org/Downloaded from
PA-oligosaccharides (ds-7) were isolated, and theirstructures were
analyzed with MALDI-TOF MS andHPLC(Figure4C). Compositions
deducedfromtheMSdata were ms-5, NeuAcHex5HexNAc4dHex-PA;
ms-7,NeuAcHex4HexNAc5dHex-PA; ms-8, NeuAcHex5Hex-NAc5dHex-PA; and
ds-7, NeuAc2Hex5HexNAc5dHex-PA. After a-sialidasedigestion,
elutionpositionsofms-5,ms-7, ms-8,
andds-7shiftedaspredictedandwereindis-tinguishablefromthoseofhumanIgGN-glycansH,O,P,andP,
respectively(Figure4C). TheirGUAmidedecreasedabout 0.30.4 after
desialylation, suggesting that theirsialylationis a2-6linkageandnot
a2-3(Nakagawaetal.,1995; Takahashi et al., 1995a,b;
TomiyaandTakahashi,1998). The monosialylated branching position on
ms-5,ms-7, andms-8 were determinedas follows: whenms-5and ms-8 were
digested witha-fucosidase, followed byb-galactosidase,
andthena-sialidase, the products wereidentical withN-glycansCandK,
respectively(Figures4and 5). b-Galactosidase-treated ms-8 was
identical to ms-7.When ms-7 was digested with a-fucosidase and
a-sialidase,theproducts
wereidenticalwithN-glycanK.Theseresultssuggest that monosialylated
site in ms-5, ms-7, and ms-8 areontheMana1-3Manarm. Accordingly,
thestructuresofthese monosialylated oligosaccharides were deduced
asshowninTableI. Desialylatedds-7was further
digestedwitha-fucosidase toyieldaN-glycanwithGal onbothterminals,
having bisecting GlcNAc and no core a1-6fucose (N-glycanLinFigure
2). Therefore, structure ofds-7 was elucidated (Table I).Our
HPLCstudyclearlyindicatedthat chickenserumIgGhas high-mannose-type
(37.2%) and complex-type(62.8%) oligosaccharides. This ratio is in
a range of that ofpigeonserumIgG(high-mannose-type, 33.3%;
complex-type, 66.7%) (Suzuki et al., 2003). Thepresenceof
high-mannose-type set avian IgG apart from human (Figure 3B,Figure
2) and other mammalian IgGs that possess complex-type
oligosaccharides exclusively. MonoGlc-high-Man(Glc1Man89GlcNAc2)
was71.2%ofthetotalhigh-mannose-typeN-glycans.Thispercentisalsointherangeofthatofpigeon
IgG (61.7% of the total high-mannose
type).Isolationandlectin-blottingsofchickenIgG-FabandFcChicken
serum IgG was known to be cleaved into Fab andFc fragments by
papain digestion (Dreesman and
Benedict,1965;KuboandBenedict,1969),althoughtheexactcleav-agesiteshavenotbeenelucidatedyet.
ToisolatechickenIgG-FabandFcfragments, papain-digestedchickenIgGwas
applied to a DEAE-Sepharose column and eluted withagradient of NaCl
toyieldthreepeaksfr. 1, fr. 2, andfr. 3, asshowninFigure6A.
Immunoblottingswithanti-chickenIgG-Fcantibodyrevealedthatfr.1andfr.3wereFab
and Fc, respectively (Figure 6B). Fr. 2 is mostly Fab butwas
slightly contaminated by Fc fragments, which could beremoved with
an affinity column using anti-chicken IgG-Fcantibody as affinant.
The pass-through fraction of the affin-itycolumnwas designatedas
fr.20. Fabwas detectedasbroad bands on sodium dodecyl
sulfatepolyacrylamide gelelectrophoresis (SDSPAGE), perhaps due to
peptide hetero-geneity of the variable regions (Figure 6B). Fc
appeared as asharp band under the reducing condition (data not
shown),but under the nonreducing condition it separated into
threeaRelative quantity was calculated based on CLC-ODS elution
profiles for neutral (67.1%), monosialyl (29.3%), anddisialyl
(3.65%) PA-oligosaccharides, including minor peaks ( 51%/each
peak). The sumof PA-oligosaccharides on thistable is 85.7%.
Structures of minor peaks were not assigned.bm/z of neutral and
sialylated oligosaccharides were detected as [MNa]
and [MH], respectively.cPA-derivatized reference N-glycans whose
elution positions on both CLC-ODS and Amide-80 columns and
[MNa]
coincide with the chicken IgG PA-oligosaccharides were indicated
on the table. Because coincident reference oligosac-charides for
n-5, n-6, ms-5, ms-7, ms-8, and ds-7 were not available, their
relative elution positions on the HPLCcolumns were compared with
those of the related reference N-glycans (see text). Structures of
the reference compoundswere shown in Figure 2.dMonosaccharides were
denoted by F, fucose; M, mannose; Glc, glucose; GN,
N-acetylglucosamine; G, galactose; NA,N-acetylneuraminic acid.eN/A,
not available.Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains279 by
guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from bands, all
stained with anti-chicken IgG-Fc antibody(Figure 6B, Coomassie
brilliant blue [CBB] andanti-Fcantibodystaining).
ThedifferentmobilityofthethreeFcfragmentsonthegelisattributabletopartialreductionofdisulfide
bonds and/or multiple or alternative cleavagesduring papain
digestion, as was found in mammalian IgG-Fc (Coligan,
1991).Onemajor N-terminal aminoacidsequenceof
theiso-latedFcfragmentwasshowntobeSCSPIQL---(startingfrom Ser346),
located near the initial part of CH3 domain.MALDI-TOF MS revealed
that the [MH]
values of thewhole chicken IgG, Fab, and Fc were 167,940.56,
44,771.61,and53,925.79, respectively. ThevalueforFcwasclosetothe
theoretical molecular mass of dimerizedCH3 CH4regions, apparently
lacking CH2 domains. Therefore itwas designatedas Fc (CH3 CH4). The
mass value
forthewholechickenIgGmoleculewerealmostthesameaspreviouslyreported,
andthevaluesforFabwereclosetothose for
Fab0preparedbypepsindigestion(Sunet al.,Fig. 2. Structures of
reference oligosaccharides isolated from humanIgG and bovine RNase
B. All the oligosaccharides were derivatized withAP for HPLC and
MALDI-TOF MS analysis. Monosaccharides weredenoted by F, fucose; M,
mannose; GN, N-acetylglucosamine;G, galactose.Fig. 3. HPLC
separation of PA-oligosaccharides from chicken serumIgG. (A) Total
PA-oligosaccharides from chicken IgG were separatedinto neutral,
mono-, and disialyl oligosaccharides on a DEAEcolumn. (B) Elution
profiles of the neutral, mono-, and disialyl PA-oligosaccharides
from chicken serum IgG on an ODS column. Structuresof human IgG
N-glycans (AP) were shown in Figure 2. HumanIgG N-glycans I, J, and
K were prepared with a-fucosidase-digestion ofhuman IgG N-glycans,
and their elution positions on an ODS columnwere indicated. Each
peak was collected and analyzed with MS. Fractionnumbers indicated
for chicken IgG N-glycans correspond to those inTable
I.N.SuzukiandY.C.Lee280 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from 2001). The
molecular mass value of Fab agrees withtheoretical values of
{(VLCL) (VHCH1)}but not{(VLCL) (VHCH1 CH2)}, suggesting that
theFabfragmentalsolacksCH2domains.TheCH2domainsmighthavebeenlostbythepapaindigestion,
presumablyby excessive
fragmentation.ConcanavalinA(ConA)andanti-chickenIgG-Fcanti-bodystainedonlyFc(CH3
CH4) fractions (fr. 3), butRCA-IstainedFabfractions(fr. 1andfr.
20exclusively;Figure 6B). These data strongly suggest the gross
differenceof glycans between chicken IgG-Fc (CH3 CH4) and Fab.Based
on the lectin specificities, it is most likely
thathigh-mannose-typeoligosaccharidesarelocatedontheFc(CH3 CH4)
region, whereas galactosylated glycans are inFab regions.
Oligosaccharides containing
b-galactosidesonchickenIgG-FabwereN-glycans (i.e.,
complex-type),becauseglycoamidaseF(GAF)-treatedchickenIgG-Fabcouldnolonger
be stainedwithRCA-I lectin(datanotshown). The N-glycosylation on
Fab fragments most likelyoccurredonvariableregions,thatis,VLand
VHdomains(Figure 1),as seeninmammalianserumIgG.N-glycosyla-tions on
VL and VH occur only when the peptide
sequencespossesstheN-glycosylationsignals,andthelocationofN-glycosylation
sites are varied among polyclonal serum IgG.Fig. 4. Structural
analysis of N-glycans from chicken IgG by a 2Dmapping technique.
(A) Elution position of high-mannose-type PA-oligosaccharides from
chicken IgG (n-4, triangle; n-5, square; n-6, opencircle) and
bovine RNase B (solid circle) on ODS and Amide-80
columns.Structures of bovine RNase B N-glycans (M9 and M5) are
shown inFigure 2. Arrows with dot-dot-dashed lines indicate the
changes of thecoordinates of N-glycans after digestion with
a-mannosidase. (B) Elutionpositions of neutral complex-type
PA-oligosaccharides from chicken(open circle) and human IgG (solid
circle). Arrows with dashed lines andwith dotted lines indicate the
changes of the coordinates of N-glycansafter digestion with
b-galactosidase and a-fucosidase, respectively. Alsosee Figure 2
for structures. (C) Elution positions of mono- (ms-5, square;ms-7,
triangle; ms-8, reversed triangle), and disialylated (ds-7,
opencircle) PA-oligosaccharides from chicken IgG and human IgG
(solidcircle). Arrows with solid lines, dashed lines, and dotted
lines indicate thechanges of the coordinates of N-glycans after
digestion with a-sialidase,b-galactosidase, and a-fucosidase,
respectively.Fig. 5. Sequential exoglycosidase digestions for
determination of themonosialylated branch of biantennary N-glycans
from chicken serumIgG. Products of exoglycosidase-treated ms-5,
ms-7, and ms-8 wereanalyzed with ODS and Amide-80 columns and
compared with referenceN-glycans (Figure 2). Monosaccharides were
denoted by F, fucose;M, mannose; GN, N-acetylglucosamine; G,
galactose; NA,N-acetylneuraminic
acid.Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains281 by guest on
September 5, 2012http://glycob.oxfordjournals.org/Downloaded from
Therefore, N-glycosylation sites on Fab fragments were
notdetermined by isolating glycopeptides originating from VLand
VH.N-glycosylationonchickenIgG-CH3domainThe known amino acid
sequences of chickenIg upsilon (u)chains (H-chains)indicate
theexistenceof onlyonepoten-tial
N-linkedglycosylationsitelocatedontheCH3-CH4domains(Figure1A)(Parvari
etal., 1988). Thiswascon-firmed by GAF treatment and SDSPAGE (data
notshown). Wealsoconfirmedthepresenceof
N-glycansontheCH3domainwithMALDI-TOFMS. MSof trypticglycopeptides
fromthe reducedformof chickenIgG-Fc(CH3
CH4)fragmentareasshowninFigure7,andthe[MH]
molecular ions areassignedinTableII. Peakswith m/z higher than
5000 were affected by GAF treatment,producing two new peaks of m/z
4036.66 and 4294.34. Thelargerofthesetwopeakshavingadditional
EandKresi-dues on N-terminal side could have arisen from
incompletetrypsin digestion (Table II). The N-glycans on the
glycopep-tides were assignedtobe exclusively
high-mannose-typeglycans, because the masses of the glycopeptides
were com-pletelyshiftedaftertreatment
withendo-b-N-acetylgluco-saminidase H(Endo H). The intact
glycopeptides wereassigned as Hex610HexNAc2-peptide (Table II).
Aftera-mannosidasedigestion,someoftheintactglycopeptideschangedtoHex1HexNAc2peptide,butalargeamountofHex56HexNAc2-peptide
was also produced (Figure 7).Concomitant
withthedecreaseinthepeaksof Hex6Hex-NAc2peptides after
exhaustivea-mannosidasedigestion,peaksof
Hex5HexNAc2peptidesincreased.
NopeaksforHex24HexNAc2-peptideweredetected,
evenwhenexcessa-mannosidase was used in the digestion. This
suggests thepresenceof Glc-cappedoligomannosyl branch.
ThereforeHex56HexNAc2peptides must have been derived fromlarger
monoglucosylatedoligosaccharides.
Hex1HexNAc2peptidecanbeassignedasManb1-4GlcNAcb1-4GlcNAc-peptide,whichcanbederivedfromglycopeptidesofhigh-mannose
type without monoglucosylation. These resultssuggest that
bothmonoglucosylatedandnonglucosylatedhigh-mannose-typeoligosaccharideswereonthesamesiteof
CH3 domain.N-glycosylationonchickenIgGCH2domainOneofthetwopotential
N-glycosylationsitesonchickenIgG CH domains is located on the CH2
domain. To demon-strate the actual glycosylation at this site, the
correspondingglycopeptides were isolated. Tryptic peptides of
wholechickenIgGwerepreparedasdescribedinMaterialsandmethods,
andtheelutionprofilesonaC18columnbeforeandafter GAFtreatment were
compared. Three of thepeaks, designatedfr. A, B, andC,
shiftedtheirpositionsafter GAFtreatment (datanot shown). Theactual
peptidesequence analysis of the pooled fr. A indicated its
completeagreement withthe predictedpeptide sequence
withtheN-glycosylationsite onthe CH2domain(Table II). Likewise,MS
and peptide sequencing data indicated that fr. B and fr.Cwere
shorter andlonger glycopeptides fromthe CH3domain, respectively.
TheCH2glycopeptides beforeandafter GAF digestion or sequential
exoglycosidase digestionswere analyzedwithMALDI-TOFMS(Figure 8).
Peaksaroundm/z28003300(Figure8A) wereeliminatedafterGAF digestion,
resulting in a peak of m/z 1035 (Figure 8B).The m/z value of the
GAF-treated glycopeptide corre-sponded to the expected [MH]
molecular ion for de-N-glycosylated CH2 glycopeptides prepared
with trypsindigestion(TableII). Thesignalsat about m/z17001800were
not shifted by GAF digestion, suggesting that they arenot
N-glycosylated peptides. When CH2 glycopeptidewasdigestedwith
a-sialidase, onepeak, atm/z3298, dis-appeared, and
theintensityofthem/z3005peakincreased(Figure 8C). The mass
difference indicates that a singleNeuAcwasremovedbya-sialidase.
Thisagreeswiththefact that predominant sialylatedN-glycans
derivedfromFig. 6. Separation of chicken IgG Fab and Fc by
DEAE-Sepharose.(A) Elution profile of papain-digested chicken IgG
on DEAE-Sepharose.One milligramof papain-digested chicken IgGwas
loaded onto a DEAE-Sepharose column (1 ml) and eluted by a linear
gradient of NaCl in10 mM TrisHCl (pH 8.0). (B) Localization of
N-glycans in chicken IgG.Lectin- and immunoblottings of chicken IgG
Fab and Fc (CH3 CH4).Fractions from the DEAE-Sepharose (fr. 1 and
3) and affinity column(fr. 20) were heat-denatured with sample
buffer containing 3% SDSwithout reducing, separated by SDSPAGE
(12.5% gel, 1 mg/lane),transferred to polyvinyl difluoride
membranes, and stained with CBB,Con A, RCA-I, or anti-chicken
IgG-Fc antibodies.N.SuzukiandY.C.Lee282 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from
chickenserumIgGis
monosialylatedbiantennaryoligo-saccharidesasshowninTableI.
a-Sialidase-treatedCH2glycopeptides were further digested
sequentially withb-galactosidase (Figure 8D),
b-N-acetylhexosaminidase(Figure 8E), and a-fucosidase (Figure 8F),
and the [MH]
oftheproductateachstepwasrecorded.Newlyproducedpeaks of
glycopeptides were unambiguously assignedasshown in Table II, and
the results indicated that N-glycanson the CH2 domain are
exclusively
complex-type.HomologymodelingofchickenIgG-Fc(CH2CH4)Basedontheassumptionthatthe3DstructureofchickenIgGissimilartothat
of humanIgE,
the3DstructureofchickenIgG-Fc(CH2CH4)waspredictedbyhomologymodeling(Figure9).Thecrystalstructureofrecombinanthuman
IgE-Fc (CH2 CH4) (PDB ID, 1o0v) was utilized asatemplate. Like the
template structure, chickenIgG-Fc(CH2 CH4) was built to form an
asymmetric homodimerwith highly bent CH2CH3 junctions. One of the
N-glyco-sylation sites on chicken IgG u-chains, Asn407 on the
CH3domain corresponded well to Asn394 on human IgEe-chains (Figure
1A, 1B), and the same orientation asfoundinthetemplates was
adopted.
High-mannose-typeoligosaccharidesatAsn407arelocatedinthecaveformedbytwoCH3
CH4regionchains,andpartiallyburiedbytheCH2domains,suggestingthatthisCH2domainsmayconfer
steric hindrance to access oligosaccharides atAsn407. The model
alsopredicts that without FabandCH2 regions, two CH3 CH4 region
chains can onlyformaframewithacavebut cannot confer strict
sterichindrancetoretainmonoglucosylatedhigh-mannose-typeoligosaccharidesatAsn407intheframe.
IfentirechickenIgG molecule formed Y-shaped structure as seen
inmammalian IgG(Harris et al., 1997), terminal
glucoseresiduesonGlc1Man89GlcNAc2-oligosaccharides, whichoccupy
wider space thanMan5GlcNAc2-, wouldreadilyprotruded fromthe (CH3
CH4) frame and would beaccessible toa-glucosidase II in the ER. On
the otherhand, N-glycosylationsites at Asn308onCH2domainsare
exposedonthe surface of the molecule tobe easilyaccessible by
processing enzymes. Thus this model supportsour experimental
results showing the site-specific N-glycanson chicken IgG and can
give a rational explanation for theexpression mechanismof
monoGlc-high-Man on avian IgG.DiscussionCarbohydrate chains on
glycoproteins can have diverseroles, such as mediators of protein
foldings, tags forintracellular and extracellular trafficking,
protein stabilizer,conferringhydrophilicproperties, protectors
against pro-teolyticdigestion, ligandsof
carbohydratebindingrecep-tors, and so on (Varki, 1993). Glycans can
assume
differentstructuresdependingontheirbiological,biochemical,andstructural
properties of the individual glycoproteins as
wellastheirlocationincellsand/orinbodies. Suchstructuraland
functional complexity of glycans often makes it difficultto
understand their biological function immediately.However,
comparison of oligosaccharide structures andglycosylation patterns
among glycoproteins with similarFig. 7. MALDI-TOF MS analysis of
the glycopeptides fromchicken IgG-Fc (CH3 CH4). Tryptic digest of
chicken IgG-Fc(CH3 CH4) was analyzed with MALDI-TOF MS before
andafter digestion with GAF, Endo H, or a-mannosidase. Assignment
ofthe [MH]
molecular ions detected were listed in Table II.Asterisks on the
m/z values indicate the peaks of glycopeptidescontaining two
additional amino acid residues (Glu-Lys, 257.29 Da)produced by
alternative trypsin digestion. Because all trypticpeptide fragments
are 54000 Da, they are not shown in the
selectedwindows.Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains283 by
guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from structures
andfunctions canprovide useful
informationaboutthestructuralrelationshipbetweentheoligosaccha-ride
chains and the core proteins. Mammalian IgG is one ofthe best
studied glycoproteins with regard to their carbohy-drate structures
andfunctions, as well as the entire
3Dstructures(Deisenhofer,1981;Harrisetal.,1997;Kobata,1990;
Matsudaet al., 1990; Mimuraet al., 2000; RadaevandSun, 2001).
Indeed, it has beenusedas astructuralmodel
forotherclassesofmammalianIgs(Burton, 1987;RuddandDwek, 1997).
DetailedstudiesofcarbohydratechainsonIgsare, however,
mostlylimitedtothosefrommammals,andlessattentionhasbeenpaidtootherverte-brate
immunoglobulins.Our previous work for structural analysis of pigeon
serumIgGN-glycans by HPLC, MS, and tandemMS haverevealed that the
prominent N-glycans are triantennaryTableII. Assignment of the
[MH]
molecular ion signals afforded by glycopeptides from chicken
IgGaaData were compiled from the result of MALDI-TOF MS (Figures 7
and 8) analysis. All [MH]
values given were the averaged masses.bMonosaccharide
composition indicated were deduced from the results of
exoglycosidase digestions.cThese exoglycosidase digestions were
sequentially performed as described in Materials and methods.dAfter
deglycosylation with GAF, the glycosylated Asn residues were
converted into Asp.eIncomplete trypsin digestion produced an
alternative glycopeptide fragment containing Glu-Lys at the
N-terminal site. [MH]
values given from thealternative were indicated in
parentheses.N.SuzukiandY.C.Lee284 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from
complextypeaswellashigh-mannose-type,bothofwhichare rarely found in
mammalian normal serum IgGs(Hamako et al., 1993). When we compared
the pigeon IgGN-glycanswiththosefromchickenIgGreportedbyothergroups
(Ohta et al., 1991; Raju et al., 2000), we realized thatconfirming
the N-glycan structures of chicken serum IgG isnecessary by our
conventional methods (Suzuki et al.,
2001;Takahashietal.,2001)forseveralreasons.First,althoughegg yolk
IgG, which arises fromtransport of maternalantibodies by
receptor-specific process (Pattersonet
al.,1962),hasidenticalbiophysicalpropertieswithserumIgG(LoekenandRoth,1983),itisalsoreportedthateggyolkIgGdoes
not mediateanaphylacticreactionwhichavianserumIgGdoes(FaithandClem,
1973). Inthisregard, apossibility that the transport of IgG to egg
yolk is restrictedby certain carbohydrate structures have not been
excluded.Second, Rajuet al. (2000) reportedchickenserumIgGN-glycans
analyzed by MALDI-TOF MS, but the
proposedstructuresareonlybasedonmassvaluesanddeducedbyanalogy to
mammalian IgGs. Although simple MS is a con-venient tool for
oligosaccharide analysis, it cannot distinguishisoforms, such as
bisected biantennary and nonbisectedtriantennary oligosaccharides.
We have demonstratedearlier that N-glycan structures and
N-glycosylation patternof pigeon IgG are quite different from those
of mammalianIgGs, so it was necessary to investigate the
differencesbetween chicken and pigeon IgG N-glycans carefully.Thus
we analyzed the N-glycan structures of chickenserumIgGbyHPLC,
whichcandistinguishtriantennaryandbisectedbiantennarystructures(Tomiyaetal.,
1988).Fig. 8. MALDI-TOF MS analysis for glycopeptides from chicken
IgG CH2. The isolated chicken IgG CH2 glycopeptide (A) was treated
withGAF (B) or sequentially digested with a-sialidase (C),
b-galactosidase (D), b-N-acetyl-D-hexosaminidase (E), and
a-fucosidase (F) and analyzed withMALDI-TOF MS. Assignment of the
[MH]
molecular ions detected were listed in Table
II.Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains285 by guest on
September 5, 2012http://glycob.oxfordjournals.org/Downloaded from
Moreover, we also determined the preferred galactosylationand
sialylation branches in complex-type as well as isoformsof
high-mannose-type oligosaccharides, which have notbeen reported for
chicken serum
IgG.Ourresultsconfirmedthatstructuralpropertiesofcom-plex-type
N-glycans from chicken serum IgG is more
similartothoseofhumanIgGthantothoseofpigeonserumIgG(Suzukietal.,2003),becausenotriantennaryorextendedbranches
with b- and a-galactosylation were
detected.Thisfactpointsoutthatstructuralpropertiesofcomplex-type
N-glycan are somehow conserved between chicken andmammalianIgGs
regardless of thedifferent N-glycosyla-tionsites but nolonger
maintainedinpigeonIgG. Bothhuman andchicken IgGs possess
biantennary complex-type oligosaccharides with and without core
a1-6 Fucand/or bisecting GlcNAc. Both IgGs have a
mono-galactosylatedbranchpredominantly onthe
GlcNAcb1-2Mana1-6Manarm(N-glycansn-9, n-10, andn-15fromchickenIgG;
showninTable I, N-glycans B, F, andNfrom human IgG; shown in
Figures 2 and 3B). Monosialy-lation in chicken IgG(ms-5, ms-7, ms-8
in Table I)exclusively occurs on the
Galb1-4GlcNAcb1-2Mana1-3-Manarm, whichis alsothe case innormal
humanIgG(Takahashi et al., 1995b). There are, however, some
notabledifference in structural patterns of complex-type
N-glycansbetween human and chicken IgG. More than half ofthe
complex-type N-glycans fromchicken IgGcontainbisecting GlcNAc and
core a1-6 Fuc and are fully b-galac-tosylated (n-17, ms-8, and ds-7
in Table I). Althoughhuman IgGalso has the same structure,
predominantoligosaccharides are core a1-6 fucosylated
biantennaryN-glycans without bisecting GlcNAc (Figure 3B).
Theobserved difference is consistent with an earlier report(Raju et
al., 2000).We also confirmed that complex-type N-glycan
structuresof chickenserumIgGwere mostlythe same as
chickeneggyolkIgG(Ohtaet al., 1991) but different fromquailegg yolk
IgG (Matsuura et al., 1993) or pigeon serum IgG.In contrast,
high-mannose-type N-glycans includingmonoGlc-high-Manarewell
conservedamongthem. Ourprevious data suggested that pigeon serum
IgG-CH3domain also possess exclusively high-mannose-type (Suzukiet
al., 2003). Because the same site-specific location of
high-mannose-type was found eveninthe distantly
removedavianorderssuchasGalliformes(chicken)andColumbi-formes
(pigeon), this feature may be widely occurringamong avian
IgGs.CH3domainofavianIgGisequivalenttoCH2domainofmammalianIgG.
InmammalianIgG, theN-glycosyla-tion site is located at Asn297 on
CH2 domains (Figure 1A),and is well conserved among mammals
(Burton, 1987). TheN-glycosylation at Asn297 is essential because
it influencesthermal stability of IgG, recognition by Fc
receptors,associationwithcompliment component C1q, andinduc-tion of
antigen-dependent cellular cytotoxicity (Kobata,1990; Mimura et
al., 2000; Radaev and Sun,
2001).N-glycansatthissiteareexclusivelybiantennarycomplextype.
ItisreportedthatrecombinantIgGpossessingonlyhigh-mannose-typeoligosaccharides
is defectivefor com-plement activation(Wright andMorrison, 1994).
CrystalstructuresofhumanandmouseIgGstudiedbyX-raydif-fraction
revealed that their Fc region form a cavity betweenthetwog-chains,
andN-glycans
onAsn297canbeseenoccludingthecavityatthecenteroftheFc(Harrisetal.,1997).1Hnuclear
magneticresonanceprovidedevidencethatconformationalchangesinthesugarchainscanaffectthe
structure of the Fc (Matsuda et al., 1990). By analogy
oftheN-glycansatAsn297onmammalianIgG,andbecauseof the structural
similarity of the complex-type N-glycans ofchickenandmammalianIgG,
one mayassume that theFig. 9. Ribbon representations of predicted
3D structures ofchicken IgG-Fc (CH2 CH4) by homology modeling. The
two chains(A-chain, green and B-chain, orange) are indicated in two
orthogonalviews. Backbones and side chains of Asn308, Asn407 (light
blue), andconserved Cys residues (at 252, 264, 322, 340, 372, 431,
477, and 546,yellow) forming intradomain and interchain disulfide
bridges areindicated. Man5GlcNAc2- linking to Asn407 are visualized
bysuperimpose from the template structure (PDB ID: 1o0v). The
atomcolors for oligosaccharide chains are carbon, white; oxygen,
red;nitrogen, blue.N.SuzukiandY.C.Lee286 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from glycans at the
corresponding position on chicken IgG(Asn407) are also complex
type.However, we have demonstrated that
N-glycosylationpatternofavianIgGisclosertothoseofmammalianIgEthantomammalianIgG.
Althoughthenumberandposi-tions of N-glycosylationsites onavianIgG,
mammalianIgG, and IgEare varied, one site (Asn297 on CH2
ofmammalianIgG,Asn394 on CH3ofIgE, Asn407 on CH3of avianIgG) iswell
conservedamongthem(Figure1A,1B). The crystal structure of human
IgE-Fc (including Ce2-4domains)(Wanetal.,
2002)revealedthatN-glycansonAsn394(seeFigure1A)areburiedinacavitybetweenthetwo
heavy chains, as seen in N-glycans on Asn297 of mam-malianIgG.
UnlikemammalianIgG, however, N-glycansat Asn394 of IgE is
exclusively high-mannose-typeoligosaccharides (Baenziger et al.,
1974; DorringtonandBennich, 1978), although the precise structures
of N-glycansatthispositionderivedfromtheentireIgEmolecule(i.e.,not
truncated mutants) remain to be elucidated. OtherN-glycosylation
sites on IgE are complex-type oligosacchar-ides. The presence of
high-mannose-type N-glycans atAsn394 in IgE can be understood from
the crystal structure,which shows that the gap between two Ce3
domains of thetwochainsarenarrowerthanthoseof twoCg2domainsand
covered with Ce2 domains formed by highly bentstructureat
Ce2Ce3junctions (Wanet al., 2002; Zhenget al., 1992). In contrast,
mammalian IgG has a wider cavityin the Fc region, and its hinge is
flexible enough to make theFccavitynot coveredbytheFabregions
(Harris et al.,1997; Zheng et al., 1992). Enzymes for N-glycan
processingapparentlycanaccess the N-glycans inCg2cavitymorereadily,
leading to complex-type structure.Structural properties of
chickenIgG based on the
aminoacidsequencesareknowntobeclosetomammalianIgEinterms of the
number of CHdomains as well as theorganization of intradomain and
interchain disulfidebonds (Magor et al., 1992; Parvari et al.,
1988; Warr et
al.,1995)(Figure1A).Inaddition,bothofthemmediateana-phylactic
reaction (Faith and Clem, 1973; Warr et al.,
1995),anditisbelievedthattheyarecloserelativesinmolecularevolution
(Warr et al., 1995). Moreover, as we have demon-strated in this
study, the site-specific presence of
high-mannose-typeN-glycanatAsn407onchickenIgGisalsosimilar tothat
inhumanIgE(Asn394). Basedonthesesequencesandwithfunctional
similarityinmind, wecon-structeda3Dstructuremodel
ofchickenIgG-Fc(CH2CH4) (Figure 9). The model suggests that unless
FabandCH2regions conferredsterichindranceagainst
GII,oligosaccharides inthe cave of Fc (CH3 CH4) regionswere
relatively exposed and could not retain
monoglucosy-latedforms.AlthoughthebentstructureofentirechickenIgGwithFabregions
shouldbe confirmedpreciselybyother biochemical and biophysical
approaches, such a uniquestructureismost
likelytoimposethesterichindrancetoN-glycan processing enzymes to
retain Glc1Man89GlcNAc2structures at
Asn407.GlycoproteinsbearingmonoGlc-high-Manoligosaccha-ridesareusuallyfoundintheERasanearlyintermediateinthebiosynthesis
of N-glycans
andaresupposedtoberecognizedbylectin-likechaperones,calnexin(CNX)and/orcalreticulin(CRT),
toaidinthefoldingprocess. Afterthe correct folding, the terminal
glucose is released by GII,followed bya-mannosidase action for
further N-glycanprocessing (Helenius and Aebi, 2001). Although it
isreportedthatCRTcanbindtochickenIgGinvitro(Patilet al., 2000;
Saito et al., 1999), the role of CNX or CRT inprotein folding for
IgG in ER is not clear. In contrast, it isreportedthat
partiallyfoldedmammalianIgGH-chainsform complex with BiP
(immunoglobulin heavy-chain bind-ingprotein)
andsomeotherchaperones,
includingUDP-Glc:glycoproteinglucosyltransferase, but not with
CNXnor CRT (Meunier et al., 2002).Acurrentlyproposedmodel
fortheproteinfoldingandassemblyof mammalianIgG(HaasandWabl, 1983;
Leeetal.,1999)isthatVH,CH2,andCH3domainsonmam-malian IgG are folded
first, then folding of the CH1 domainis accomplished by assembly
with light (L) chains. Weexpected that this two-step protein
folding process of mam-malian IgG and formation of the IgE-like
highly bent struc-ture can account for the presence of
monoGlc-high-Man onavianIgG-CH3domain. If thetwo-stepfoldingmodel
isalso applicable to avian IgG, it follows that VH, CH3, andCH4
domains of the H-chains are folded first, thenassembledwithLchains
toallowthe completionof theentireIgmolecule(Figure10).
Afterthefoldingof
avianCH3domain,deglucosylationonthisdomaincanonlybepartially
carriedout, becauseavianIgG-CH3 domainpos-sesses
bothmonoglucosylatedandnonglucosylatedhigh-mannose-type N-glycans.
However, concurrent with butindependent of deglucosylation, the
H-chains are assembledwith L-chains, and the fully folded molecule
no longerconfersaccessibilitybyGII, sothattheIgG-CH3domaincan
retain monoglucosylated N-glycans (Figure 10). On
theotherhand,N-glycosylationsitesontheCH2domainandvariable regions
are more exposed to allow
advancedN-glycanprocessing,sothattheycanbepossessedtofullyb-galactosylated,
bisected, andcore fucosylatedcomplextype N-glycans. CH3 domains of
chicken IgGcontainsexclusively high-mannose-type oligosaccharides,
whereasCH2 domains only contains complex-type, suggestingthat the
presence of monoGlc-high-Man N-glycans
ismainlyduetostrictsterichindrance,ratherthanrandom,incomplete
cleavages by the processing enzymes.The reasonfor the
N-glycosylationsite inacavityofFc region to be conserved among
mammalian IgG, IgE, andavianIgGandyet withdifferent N-glycantypes
remainsto be elucidated. Basu et al. (1993) had reported that
humanIgElackingN-glycosylationattheAsn394bypointmuta-tions tended
to self-aggregate, although the loss of N-glyco-sylation did not
influence its binding to Fce receptors.Therefore the
N-glycosylation is probably involved at leastin stabilization of
the protein structures by
conferringsuitablehydrophilicityinthecavityof IgE, aswell asinIgG.
If the hypothesis that IgY evolved to mammalian IgGis correct,
mammalian IgGmight have gained a hinge regionthat gives higher
segmental flexibilitytoaccommodateamore diverse range of antigens,
such as adjusting for cross-linkingof epitopesontwolargeantigens.
Accompanyingwiththe changes inproteinstructures, N-glycans
inthecavityof Fcregionof
mammalianIgGbecamecomplex-typebutareconservedatthesamesite(Asn297)toconferthe
protein stabilization.Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains287
by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from
MaterialsandmethodsMaterialsPapain (2crystallized), iodoacetamide,
and alkaline phos-phataseconjugated ExtrAvidin were purchased
fromSigma(St.Louis, MO). GAFisalso known
asPNGaseF,glycopeptideN-glycosidaseF,orN-glycanase.OneunitofGAFactivity
is definedas the amount of enzyme thatcatalyzes the release of
N-linked oligosaccharides from1nmol denaturedribonuclease Bin1minat
37
C, pH7.5)wasfromProzyme(SanLeandro, CA), andEndoHfrom
Streptomyces plicatus was a gift from Dr. C. E. Ballou(Berkeley,
CA). L-(Tosylamido-2-phenyl) ethyl chloro-methyl ketone
(TPCK)-treated trypsin (3 crystallized)was from Worthington
Biochemical (Lakewood, NJ). Alka-line phosphataseconjugated Con A
and RCA-I lectin werepurchased from EY Labs (San Mateo, CA).
Biotin-conjugated anti-chicken IgG-Fc was from Biotrend Chemi-cals
(Destin, FL). 5-Bromo-4-chloro-3-indoyl phosphate/nitro blue
tetrazolium kit for use with alkaline
phosphatasewaspurchasedfromZymedLaboratories(SanFrancisco,CA).
DEAE-Sepharose Fast Flow column (HiTrap, 1 mL)was from Amersham
Pharmacia Biotech (Piscataway, NJ).TSKgel DEAE-5PWcolumn(7.5 75mm)
andTSKgelAmido-80 column(4.6 250 mm) were
fromTosoHaas(Montgomeryville, PA). Shim-Pack CLC-ODS column(6.0
150mm)wasfromShimadzu(Kyoto,Japan).Poly-vinylidene difluoride
membranes for blotting and CentriconYM 10 were from Millipore
(Bedford, MA). Chicken serumIgGwas fromPel-Freez Biologicals
(Rogers, AR). BCAProtein Assay reagents and immobilized avidin (on
6%cross-linkedbeadedagarose)werefromPierce(Rockford,IL).
Neuraminidase fromArthrobacter ureafaciens was agenerousgift
fromDr. TsukadaandDr. Ohtaof KyotoResearch Institute (Uji, Japan).
Other exoglycosidases usedwere b-galactosidase (from jack beans
SeikagakuAmerica),b-N-acetyl-D-hexosaminidase(fromjackbean, Sigma),
a-mannosidase (fromjackbean, Glyko), anda-fucosidase(frombeef
kidney, Roche). The matrices for MALDI-TOFMS,
3,5-dimethoxy-4-hydroxycinnamicacid(sinapi-nic acid, SA),
a-cyano-4-hydroxycinnamic acid
(ACH),2,5-dihydroxybenzoicacid(DHB),and20,40,60-trihydroxy-acetophenonemonohydrate(THAP)werepurchasedfromAldrich
(Milwaukee, WI).BuffersandstandardproceduresTris-buffered saline
(TBS) contains 50 mMTrisHCl(pH7.4)and150mMNaCl.
TBSTcontains0.1%Tween20 in TBS. Digestion buffer for papain
treatment wasFig. 10. Diagram of hypothesis of folding and assembly
of avian IgG. In the ER, nascent H-chains of avian IgG possesses
Glc3Man9GlcNAc2 onboth CH2 and CH3 domains. Partially folded
H-chains with Glc1Man9GlcNAc2 can be produced by concerted actions
of a-glucosidase I, II (GI, GII),UDP-Glc:glycoprotein
glucosyltransferase (GT) (Parodi, 2000), and some ER chaperones.
When folding of VH, CH3, and CH4 domains anddimerization of
H-chains are proceeded in analogy to mammalian g-chains (Lee et
al., 1999), N-glycan processing enzymes such as GII and
a-mannosidase I might be able to partially process Glc1Man9GlcNAc2
on CH3 domain. However, concurrent with but independent on
deglucosylation,L-chains are assembled with the H-chains mediated
by BiP and other ER chaperones, then the CH3 domains became
sterically unaccessible to theprocessing enzymes after full folding
and assembly of the avian IgG molecules. Although timing of the
folding of CH2 domain is unknown, folded CH2domains might confer
highly bent structure between Fab and Fc regions in analogy to
mammalian IgE (Wan et al., 2002). N-glycans on CH2 domains,however,
are amenable to processing enzymes and can become complex-type
eventually.N.SuzukiandY.C.Lee288 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from 50
mMsodiumphosphate (pH7.0) containing 1 mMEDTAand10mMcysteine.
ProceduresforSDSPAGE,lectin- and immunoblottings, and N-terminal
sequenceanalyseswereasdescribedpreviously (Suzuki etal.,
2001).ProteinconcentrationsweremeasuredbytheBCAassay(Smith et al.,
1985) using bovine serumalbumin as
astandard.PapaindigestionofchickenserumIgGPapainsuspension(28
mg/mL) was dilutedindigestionbuffertobe1mg/mLandincubatedat 37
Cfor10minfor activation.
ChickenserumIgGdissolvedindigestionbuffer(2mg/ml)wasincubatedwiththeactivatedpapain(enzyme:substrate
ratio of 1:100) at 37
Cfor 4 h. Thereactionwas terminatedby adding iodoacetamide
(finalconcentration 30 mM) and incubation at room
temperaturefor30mininthedark. Themixturewasdialyzedagainst10 mM
TrisHCl (pH 8.0) for the following
anion-exchangechromatography.IsolationofFabandFcfragmentsofchickenserumIgGAcolumnofDEAE-SepharoseFastFlow(HiTrap,1mL)waswashedwith1MNaClin10mMTrisHCl(pH8.0)andequilibratedwith10
mMTrisHCl (pH8.0). Afterpapain-digested chicken IgG (1 mg) was
loaded, the columnwas washed with 10 mM TrisHCl (pH 8.0) for 20 min
(flowrate 1mL/min), andthenconcentrationof NaCl
intheelutionwaslinearlyincreasedupto0.3Mwithin120min(flow rate 0.5
mL/min). The major peaks detected by A280nmwere collected and
concentrated with a Centricon YM-10 at4
C. One of the chicken IgG Fab fractions was
incompletelyseparatedfromtheFcfraction,anditwasfurtherisolatedwith
an affinity column. Avidin-agarose (250 mL) in a
1-mLsyringecolumnwas washedwith5ml water
andequili-bratedwith5mLbindingbuffer (20mMsodiumphos-phate, pH7.4,
with 500 mMNaCl). Biotin-conjugatedanti-chickenIgGFcantibodies(100
mg)wereloadedontothe columnandallowed tostand for 30 min,
thenthecolumn was washed with 5 mL binding buffer. The samplewas
loaded into the column and incubated for 30 min; thenthe column was
washed with 2 mL binding buffer. The pass-throughfractionwas
collectedandconcentratedwithaCentriconYM-10. All
theaffinitypurificationproceduresdescribed were performed at 4
C.GAFandEndoHtreatmentofglycoproteinsForGAFdigestion,
glycoproteins(15mg) weredissolvedwith30mL50 mMNaHPO4(pH7.5)
containing 0.1%SDSand100mM2-mercaptoethanol andheatedat90
Cfor3minfordenaturation.
Afterthesolutionwascooledtoroomtemperature, 1%(v/v) NP-40was
addedtotheheat-denaturedglycoproteins. Themixturewasincubatedwith
GAF(40 U/mg substrates) at 37
Cfor 16 h forcomplete de-N-glycosylation, and heated at 100
C
for10mintoinactivateGAF.Forpartialde-N-glycosylation,glycoproteins
were incubated with GAF(1 U/mg sub-strates)for10min, 40min,
and3handheatedat100
Cfor 10
min.EndoHdigestionforglycoproteinswasperformedsimi-larly,but50mMsodiumacetate(pH5.5)wasusedasthereaction
buffer. The glycoproteins before and after thedigestions were
analyzed with SDSPAGE.ReductionandalkylationofglycoproteinsThe
formations of intramolecular disulfide bonds on
glyco-proteinswerereducedandblockedasfollows.
Glycopro-teins(0.2mg)weredissolvedwith120 mL8Mguanidine-HCl
in0.2MTrisHCl (pH8.0)andreducedwith60 mL0.18 M dithiothreitol in
the 8 M guanidine solution at roomtemperaturefor1h. Forthiol
alkylation,
240mL0.18Miodoacetamideintheguanidinesolutionwasaddedtothemixtureandincubatedatroomtemperaturefor30mininthedark.
ThereactionmixturewasdialyzedagainstH2Oand
lyophilized.PreparationandisolationofglycopeptidesOne-third of the
reduced and alkylated glycoproteinsweresuspendedin50
mL50mMNH4HCO3,pH7.8,andincubatedwithTPCK-treatedtrypsinat 37
Covernight.Trypsinwas inactivatedbyheatingat 100
Cfor 5min.Aportion of the reaction mixture was further
treatedwithGAF(1U/10mL) at 37
C, overnight. The
peptidefragmentsbeforeandaftertreatmentwithGAFwereana-lyzedwithreversed-phase
HPLConaShim-packCLC-ODScolumn(6.0 150mm). Themobilephasewas
(A)0.05%trifluoroacetic acid (TFA) and (B) 90%CH3CNwith0.05%TFA.
Elution(1ml/min)wasconductedbyalinear gradient of 050%of (B) in (A)
developed over100min. EachpeakdetectedbyA210nmwascollectedandkept
at 4
C.PreparationandsepalationofPA-derivatizedoligosaccharidesforstructuralanalysesChicken
serum IgG (reductive alkylated, 1 mg) weredigested with trypsin and
chymotrypsin in 50 mMNH4HCO3, pH7.8, at 37
Covernight, andtheenzymeswereinactivatedbyheatingat100
Cfor5min.Oligosac-charides were released with GAFtreatment in 50
mMNH4HCO3, pH7.8, at37
Covernight. AfterinactivatingGAF by heating at 100
C for 5 min, the digest was lyophi-lized. Cations and peptides
were removed with Dowex50W 2(H
form,50100mesh,250 mL,Sigma)packedin a 1-mL syringe. The column
was washed with 1 mL
H2O,andthecollectedeffluentwaslyophilized.Thesamplewasreconstituted
with 100 mL H2O, loaded onto a
Carbographtube(25mg,Alltech),andwashedwith400
mLH2O,thenelutedwith400mL50%CH3CNcontaining0.05%TFA.ForthePAderivatizationbyreductiveamination,lyophi-lizedoligosaccharidefractionsweredissolvedin40
mLPAsolution(1g/580mLinconcentratedHCl, pH6.8), andheatedat 90
Cfor 15 minwithheating block.
FreshlypreparedNaCNBH3solution(7mg/4 mL)wasaddedintothe reaction
mixture, then heated at 90
Cfor 1 h. PAoligosaccharides were fractionatedbygel
filtrationonaSephadex G-15 column (1.040 cm, in 10 mM NH4-
HCO3),Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains289 by guest on
September 5, 2012http://glycob.oxfordjournals.org/Downloaded from
and the effluent was monitored with a fluorometer (excita-tion 300
nm, emission 360 nm) and lyophilized.The mixture of
PA-oligosaccharides was separatedbyHPLC with three different
columns as described previously(Nakagawa et al., 1995). In the
first stage, the
PA-oligosac-charideswereseparatedonaTSKgelDEAE-5PWcolumn(7.5 75
mm), and the neutral, monosialyl, and disialylfractions
(monitoredbyfluorescence, excitation320nm,emission 400 nm) were
collected separately and lyophilized.In the second stage, neutral,
mono-, and disialylatedoligosaccharide fractions were individually
dissolved inH2Oandseparatedona Shim-PackCLC-ODScolumn(6.0 150 mm),
monitoring effluent with fluorescence(excitation 320 nm, emission
400 nm). Elution wasperformed at a flow rate of 1.0 mL/min at
55
C using eluentA (0.005% TFA) and eluent B (0.5 % 1-butanol in
eluent A).The columnwas equilibratedwitha mixture of
eluentsA:B90:10(v/v), andaftersampleinjection,
theratiooftheeluentswaschangedlinearlytoA:B60:40in60min.Eachpeakwas
collected, lyophilized, andanalyzedwithMALDI-TOF
MS.Inthethirdstage, major peaks fromtheODScolumnwere dissolved with
50 mL eluent C
(CH3CN:3%CH3COOH-trietylamine,pH7.3,65:35),andseparatedonaTSKgel
Amide-80column(4.6 250mm), monitoringeffluent with fluorescence,
excitation 300 nm, emission360nm. Elutionpositionof
eachPA-oligosaccharidesonCLC-ODSandAmide-80columnswasexpressedinGUs,basedontheelutionpositionofisomaltoseseries.Fortheassignment
of GUODS, analysis on a CLC-ODS column wasperformed using 10
mMsodium phosphate, pH 3.8(Tomiyaet al., 1988) insteadof 0.005%TFA.
ReferencePA-derivatized oligosaccharides fromasialo-human
IgGandbovineRNaseBwerepreparedbythesamemethod.Structures of the
reference compounds were shown inFigure2. PA-derivatizedN-glycans
A, B, C, D, I, J, K,and L (Figure 2) were obtained from
a-fucosidase-digestedPA-oligosaccharides from human
IgG.MALDI-TOFMSMALDI-TOFMS was performed on a Kompact SEQ(Kratos
Analytical, Manchester, England), equippedwith a 337-nmnitrogen
laser and set at 20 kVextrac-tionvoltage. Eachspectrumwas
theaverageof 50lasershots. Glycoproteins, glycopeptides, and
neutral
oligo-saccharideswereanalyzedinthelinearpositive-ionmode,andsialylatedoligosaccharideswereanalyzedinthelinearnegative-ionmode.
SAandACHwere usedas matricesin the analysis of glycoproteins and
glycopeptides,respectively. Thesematrices
weredissolvedtobe10mg/mLwith50%CH3CNwith0.05%TFA. Forneutral
andsialylatedPA-derivatizedoligosaccharides, DHB(10 mg/mLin5
mMNaCl) andTHAP(2
mg/mLin25%CH3CNwith10mMdibasicammoniumcitrate)wereused,respec-tively
(Papac et al., 1998). Samples (0.5 mL) were applied toatarget
andmixedwithmatrix(0.5mL),
andallowedtodryunderambientcondition(forSAorACH)orvacuum(forDHBorTHAP)
at
roomtemperaturepriortomassanalysis.Endo-andexoglycosidasedigestionforglycopeptidesandoligosaccharidesGlycopeptides(about90pmol/2.5mL)weredigestedwithGAF
(1 U) in 50 mM NH4HCO3 (pH 7.8) or with Endo
H(2.5mg)in20mMsodiumacetate(pH5.6)at37
Cover-night. For exoglycosidase digestion, glycopeptides were
dis-solved with 20 mM sodium acetate (pH 4.5) to be 40 pmol/mL.
Glycopeptides containing high-mannose-type oligosac-charides (90
pmol) were digested with a-mannosidase(250 mU/100 pmol
glycopeptides) at 37
C overnight.Glycopeptides containing complex-type
oligosaccharides(250 pmol) were sequentially digested with
a-sialidase(0.63 mU/100 pmol of glycopeptides),
b-galactosidase(0.38mU), b-N-acetyl-D-hexosaminidase (1.11mU),
anda-fucosidase (0.5 mU). After the incubation at 37
Covernight andheat-inactivationat 90
Cfor5minat theeach step, a portion of the digestion products was
analyzedwithMALDI-TOFMS. PA-derivatizedoligosaccharideswere
digestedwith exoglycosidases in the same mannerand analyzed with
HPLC as described.HomologymodelingAmodel of the3Dstructureof
chickenIgG-Fc(CH2CH4)(fromSer250toGly567)wasconstructedusingtheSWISS-MODELprogram(Schwedeetal.,2003),whichismade
public by the Swiss Institute of
Bioinformatics(UniversityofBasel,Switzerland).ThecrystalstructureofhumanIgE-Fc(CH2CH4)(Wanetal.,2002)(PDBID:1o0v)wasusedasatemplateforthehomologymodelingbased
on their 32% amino acid sequence identity as well astheiroverall
structural similarity. Thesequencealignmentwas generated by
T-Coffee (Notredame et al., 2000),which correctly aligned all Cys
and Trp residues (hallmarksof Ig-like domains) conserved between
the target andtemplate. The optimize mode for oligomer modeling in
theSWISS-MODELserver was chosen for the appropriatemodeling. The
predicted 3D structure was visualizedwithDeepViewprogram.
Oligosaccharide chains
(Man5-GlcNAc2-)atAsn407onchickenIgGwasimposedfromthe
template.AcknowledgmentsTheauthorsaregratefulforDr.NoboruTomiyafortech-nicaladviceconcerningoligosaccharideanalysisbyHPLCand
Dr. Hao-Chia Chen for peptide sequencing. This workwas supported by
NIH Research Grant DK09970.AbbreviationsACH,
a-cyano-4-hydroxycycinnamicacid;CBB,Coomas-siebrilliantblue;CNX,calnexin;ConA,concanavalinA;CRT,
calreticulin; ER, endoplasmic reticulum; GII, a-glucosidaseII; GAF,
glycoamidaseF; GU, glucoseunit;MALDI,
matrix-assistedlaserdesorption/ionization; MS,mass spectrometry;
ODS, octadecylsilica; PA,
2-aminopyr-idine;PBS,phosphatebufferedsaline;SA,3,5-dimethoxy-4-hydroxycinnamic
acid (sinapinic acid); SDSPAGE,sodium dodecyl sulfatepolyacrylamide
gel electrophoresis;N.SuzukiandY.C.Lee290 by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from TBS,
Tris-buffered saline; TFA, trifluoroacetic acid;THAP,
20,40,60-trihydroxyaetophenone
monohydrate;TOF,timeofflightReferencesBaenziger, J., Kornfeld, S.,
and Kochwa, S. (1974) Structure of thecarbohydrateunitsof
IgEimmunoglobulin. I. Over-all composition,glycopeptide isolation,
andstructureof thehighmannose
oligosac-charideunit.J.Biol.Chem.,249,18891896.Basu, M., Hakimi,
J., Dharm, E., Kondas, J.A., Tsien, W.H., Pilson, R.S.,Lin, P.,
Gilfillan, A., Haring, P., Braswell, E.H., andothers.
(1993)Purification and characterization of human recombinant
IgE-FcfragmentsthatbindtothehumanhighaffinityIgEreceptor. J.
Biol.Chem.,268,1311813127.Burton, D.R.
(1987)Structureandfunctionofantibodies. InCalabi, F.andNeuberger,
M.S. (Eds.), Molecular genetics of
immunoglobulin.Elsevier,Amsterdam,NewYork,Oxford,pp.150.Coligan,J.E.(1991)Currentprotocolsinimmunology.Wiley,NewYork.Deisenhofer,
J. (1981) Crystallographic refinement and atomic models of
ahumanFcfragment andits complexwithfragment Bof proteinAfrom
Staphylococcus aureus at 2.9- and 2.8-Aresolution.
Biochemistry,20,23612370.Dorrington, K.J. andBennich, H.H. (1978)
Structure-functionrelation-shipsinhumanimmunoglobulinE.Immunol.Rev.,41,325.Dreesman,
G. andBenedict, A.A. (1965) Properties of
papain-digestedchicken7Sgamma-globulin.J.Immunol.,95,855866.Edelman,
G.M., Cunningham, B.A., Gall, W.E., Gottlieb, P.D.,Rutishauser, U.,
andWaxdal, M.J. (1969)
ThecovalentstructureofanentiregammaGimmunoglobulinmolecule. Proc.
Natl Acad.
Sci.USA,63,7885.Faith,R.E.andClem,L.W.(1973)Passivecutaneousanaphylaxisinthechicken.
Biological fractionationof
themediatingantibodypopula-tion.Immunology,25,151164.Fellah, J.S.,
Kerfourn, F., Wiles, M.V., Schwager, J., and Charlemagne, J.(1993)
Phylogeny of immunoglobulin heavy chain isotypes:structure of the
constant region of Ambystoma mexicanumupsilon chain deduced
fromcDNAsequence. Immunogenetics, 38,311317.Haas, I.G. andWabl, M.
(1983)
Immunoglobulinheavychainbindingprotein.Nature,306,387389.Hamako,
J., Matsui, T., Ozeki, Y., Mizuochi, T., andTitani, K.
(1993)Comparative studies of asparagine-linked sugar chains of
immunoglo-bulinGfromelevenmammalianspecies.Comp.
Biochem.Physiol.B,106,949954.Harris, L.J., Larson, S.B., Hasel,
K.W., and McPherson, A.
(1997)RefinedstructureofanintactIgG2amonoclonalantibody.Biochem-istry,36,15811597.Helenius,
A. and Aebi, M. (2001) Intracellular functions of
N-linkedglycans.Science,291,23642369.Kobata,A.(1990)Functionandpathology
ofthesugarchainsofhumanimmunoglobulinG.Glycobiology,1,58.Kubo,R.T.andBenedict,A.A.(1969)Comparisonofvariousavianandmammalian
IgG immunoglobulins for salt-induced
aggregation.J.Immunol.,103,10221028.Lee,Y.K.,Brewer,J.W.,Hellman,R.,andHendershot,L.M.(1999)BiPandimmunoglobulinlight
chaincooperatetocontrol thefoldingofheavy chain and ensure the
fidelity of immunoglobulin assembly.
Mol.Biol.Cell,10,22092219.Leslie, G.A. and Clem, L.W. (1969)
Phylogeny of immunoglobulinstructureandfunction. 3. Immunoglobulins
of thechicken. J. Exp.Med.,130,13371352.Loeken, M.R. and Roth, T.F.
(1983) Analysis of maternal IgGsubpopulations which are transported
into the chicken oocyte.Immunology,49,2128.Magor, K.E., Warr, G.W.,
Middleton, D., Wilson, M.R.,
andHiggins,D.A.(1992)StructuralrelationshipbetweenthetwoIgYoftheduck,Anas
platyrhynchos: molecular genetic evidence. J. Immunol.,
149,26272633.Matsuda, H., Nakamura, S., Ichikawa, Y., Kozai, K.,
Takano, R., Nose, M.,Endo, S., Nishimura, Y., and Arata, Y. (1990)
Proton nuclearmagnetic resonance studies of the structure of the Fc
fragment ofhumanimmunoglobulinG1:
comparisonsofnativeandrecombinantproteins.Mol.Immunol.,27,571579.Matsuura,
F., Ohta, M., Murakami, K., and Matsuki, Y.
(1993)Structuresofasparaginelinkedoligosaccharidesofimmunoglobulins(IgY)
isolated fromegg-yolk of Japanese quail. Glycoconj. J.,
10,202213.Meunier, L., Usherwood, Y.K., Chung, K.T., and
Hendershot, L.M.(2002) A subset of chaperones and folding enzymes
form multiproteincomplexes inendoplasmic reticulumtobindnascent
proteins. Mol.Biol.Cell,13,44564469.Mimura, Y., Church, S.,
Ghirlando, R., Ashton, P.R., Dong, S., Goodall,
M.,Lund,J.,andJefferis,R.(2000)Theinfluenceofglycosylationonthethermal
stability and effector function expressionof
humanIgG1-Fc:propertiesofaseriesoftruncatedglycoforms.Mol.Immunol.,37,697706.Nakagawa,
H., Kawamura, Y., Kato, K., Shimada, I., Arata, Y., andTakahashi,
N. (1995) Identification of neutral and sialyl
N-linkedoligosaccharide structures fromhuman serumglycoproteins
usingthree kinds of high-performance liquid chromatography.
Anal.Biochem.,226,130138.Notredame, C.,Higgins,D.G., andHeringa,J.
(2000)T-Coffee: anovelmethodfor fast andaccurate multiple sequence
alignment. J. Mol.Biol.,302,205217.Ohta, M., Hamako, J., Yamamoto,
S., Hatta, H., Kim, M.,Yamamoto, T., Oka, S., Mizuochi, T., and
Matsuura, F. (1991)Structures of asparagine-linkedoligosaccharides
fromhenegg-yolkantibody (IgY). Occurrence of unusual
glucosylatedoligo-mannosetype oligosaccharides in a mature
glycoprotein. Glycoconj. J., 8,400413.Papac, D.I., Briggs, J.B.,
Chin, E.T., andJones, A.J. (1998) Ahigh-throughput microscale
methodtorelease N-linkedoligosaccharidesfromglycoproteins for
matrix-assisted laser
desorption/ionizationtime-of-flightmassspectrometricanalysis.Glycobiology,8,445454.Parodi,
A.J.
(2000)Proteinglucosylationanditsroleinproteinfolding.Annu.Rev.Biochem.,69,6993.Parvari,R.,Avivi,A.,Lentner,F.,
Ziv,E., Tel-Or,S., Burstein,Y.,andSchechter, I. (1988) Chicken
immunoglobulin g-heavy chains:limitedVHgenerepertoire,
combinatorial diversificationbyDgenesegments and evolution of the
heavy chain locus. EMBOJ., 7,739744.Patil, A.R., Thomas, C.J., and
Surolia, A. (2000) Kinetics and themechanismof interactionof the
endoplasmic reticulumchaperone,calreticulin, with monoglucosylated
(Glc1Man9GlcNAc2) substrate.J.Biol.Chem.,275,2434824356.Patterson,
R., Younger, J.S., Weigle, W.O., and Dixon, F. (1962)Antibody
production and transfer to egg yolk in chickens. J.
Immunol.,89,272278.Paul, W.E. (1999) Fundamental immunology.
Lippincott-Raven,Philadelphia.Radaev,S.andSun,P.D.(2001)RecognitionofIgGbyFcg.TheroleofFc
glycosylation and the binding ofpeptide inhibitors. J.Biol.
Chem.,276,1647816483.Raju, T.S., Briggs, J.B., Borge, S.M.,
andJones, A.J. (2000) Species-specific variationinglycosylationof
IgG: evidence for the species-specific sialylation and
branch-specific galactosylation and
importanceforengineeringrecombinant glycoproteintherapeutics.
Glycobiology,10,477486.Reynaud, C.A., Dahan, A., Anquez, V.,
andWeill, J.C.
(1989)SomatichyperconversiondiversifiesthesingleVhgeneof
thechickenwithahighincidenceintheDregion.Cell,59,171183.Rudd, P.M.
and Dwek, R.A. (1997) Glycosylation: heterogeneity and
the3Dstructureofproteins.Crit.Rev.Biochem.Mol.Biol.,32,1100.Saito,
Y., Ihara, Y., Leach, M.R., Cohen-Doyle, M.F., and Williams,
D.B.(1999) Calreticulin functions in vitro as a molecular
chaperonefor bothglycosylatedandnon-glycosylatedproteins. EMBOJ.,
18,67186729.Schwede, T., Kopp, J., Guex, N., and Peitsch, M.C.
(2003) SWISS-MODEL: anautomatedproteinhomology-modelingserver.
NucleicAcidsRes.,31,33813385.Glc1Man9GlcNAc2-AsnonchickenIgG-CH3domains291
by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from Smith, P.K.,
Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner,
F.H.,Provenzano, M.D., Fujimoto, E.K., Goeke, N.M., Olson, B.J.,
andKlenk,D.C.(1985)Measurementofproteinusingbicinchoninicacid.Anal.Biochem.,150,7685.Sun,
S., Mo, W., Ji, Y., and Liu, S. (2001) Preparation and
massspectrometricstudyof eggyolkantibody(IgY) against
rabiesvirus.RapidCommun.MassSpectrom.,15,708712.Suzuki, N., Khoo,
K.H., Chen, H.C., Johnson, J.R., and Lee, Y.C.
(2001)Isolationandcharacterizaionof major glycoproteins of
pigeoneggwhite: ubiquitous presence of unique N-glycans containing
Gala1-4Gal.J.Biol.Chem.,276,2322123229.Suzuki,N.,Khoo,K.H.,Chen,C.M.,Chen,H.C.,andLee,Y.C.(2003)N-glycan
structures of pegeon IgG: a major
serumglycoproteincontainingGala1-4Galtermini.J.Biol.Chem.,278,4629346306.Takahashi,
N., Lee, K.B., Nakagawa, H., Tsukamoto, Y., Kawamura, Y.,Li, Y.T.,
andLee, Y.C. (1995a) Enzymatic sialylationof
N-linkedoligosaccharides using an a-(2-3)-specific trans-sialicase
fromTrypanosomacruzi: structual
identificationusingathree-dimensionalelutionmappingtechnique.Anal.Biochem.,230,333342.Takahashi,
N., Nakagawa, H., Fujikawa, K., Kawamura, Y.,and Tomiya, N. (1995b)
Three-dimensional elution mapping ofpyridylaminatedN-linkedneutral
andsialyl oligosaccharides. Anal.Biochem.,226,139146.Takahashi, N.,
Khoo, K.H., Suzuki, N., Johnson, J.R., andLee, Y.C.(2001) N-glycan
structures from the major glycoproteins of pigeon eggwhite:
predominanceof terminal Gala(1-4)Gal. J. Biol. Chem.,
276,2323023239.Tomiya, N. and Takahashi, N. (1998) Contribution of
componentmonosaccharidestothecoordinatesofneutral andsialyl
pyridylami-nated N-glycans on a two-dimensional sugar map. Anal.
Biochem., 264,204210.Tomiya, N., Awaya, J., Kurono, M., Endo, S.,
Arata, Y., andTakahashi, N.(1988) Analyses ofN-linked
oligosaccharidesusing a two-dimensionalmapping technique. Anal.
Biochem., 171, 7390.Tomiya,N., Lee,Y.C.,Yoshida,T.,
Wada,Y.,Awaya,J.,Kurono,M.,andTakahashi, N. (1991)
Calculatedtwo-dimensional
sugarmapofpyridylaminatedoligosaccharides:elucidationofthejackbeanalpha-mannosidase
digestionpathwayof Man9GlcNAc2. Anal. Biochem.,193,90100.Varki, A.
(1993) Biological roles of oligosaccharides: all of the theories
arecorrect.Glycobiology,3,97130.Wan, T., Beavil, R.L., Fabiane,
S.M., Beavil, A.J., Sohi, M.K., Keown,
M.,Young,R.J.,Henry,A.J.,Owens,R.J.,Gould,H.J.,andSutton,B.J.(2002)ThecrystalstructureofIgEFcrevealsanasymmetricallybentconformation.
Nat. Immunol., 3, 681686.Warr, G.W., Magor, K.E., andHiggins, D.A.
(1995) IgY: clues
totheoriginsofmodernantibodies.Immunol.Today,16,392398.Wright, A.
andMorrison, S.L. (1994) Effect of
alteredCH2-associatedcarbohydratestructureonthefunctional
properties andinvivofateof chimeric mouse-humanimmunoglobulinG1. J.
Exp. Med., 180,10871096.Zheng, Y., Shopes, B., Holowka, D., and
Baird, B. (1992)
DynamicconformationscomparedforIgEandIgG1insolutionandboundtoreceptors.Biochemistry,31,74467456.N.SuzukiandY.C.Lee292
by guest on September 5,
2012http://glycob.oxfordjournals.org/Downloaded from