Structural Analysis of Carbohydrates by Mass Spectrometry Chiharu Konda Purdue University

Purdue University Purdue e-Pubs

Open Access Dissertations Theses and Dissertations

Fall 2013 Structural Analysis of Carbohydrates by Mass Spectrometry Chiharu Konda Purdue University

Follow this and additional works at: https://docs.lib.purdue.edu/open_access_dissertations Part of the Analytical Chemistry Commons

Recommended Citation Konda, Chiharu, "Structural Analysis of Carbohydrates by Mass Spectrometry" (2013). Open Access Dissertations. 141. https://docs.lib.purdue.edu/open_access_dissertations/141

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for additional information. Graduate School ETD Form 9 (Revised 12/07) PURDUE UNIVERSITY GRADUATE SCHOOL Thesis/Dissertation Acceptance

This is to certify that the thesis/dissertation prepared

By Chiharu Konda

Entitled Structural Analysis of Carbohydrates by Mass Spectrometry

Doctor of Philosophy For the degree of

Is approved by the final examining committee:

Yu Xia Chair Peter T. Kissinger

Nikolai R. Skrynnikov

Hilkka I. Kenttamaa

To the best of my knowledge and as understood by the student in the Research Integrity and Copyright Disclaimer (Graduate School Form 20), this thesis/dissertation adheres to the provisions of Purdue University’s “Policy on Integrity in Research” and the use of copyrighted material.

Approved by Major Professor(s): ______Yu Xia ______

Approved by: Robert E. Wild 10/25/2013 Head of the Graduate Program Date i

STRUCTURAL ANALYSIS OF CARBOHYDRATES BY MASS SPECTROMETRY

A Dissertation

Submitted to the Faculty

Purdue University

Chiharu Konda

In Partial Fulfillment of the

Requirements for the Degree

Doctor of Philosophy

December 2013

Purdue University

West Lafayette, Indiana LL

7RP\IDPLO\DQGIULHQGV LQFOXGLQJKDLU\RQHV

:LWKRXW\RXUVXSSRUWWKLVZRXOGQRWKDYHEHHQSRVVLEOH LLL

$&.12:/('*0(176

)LUVW , ZRXOG OLNH WR DFNQRZOHGJH P\ SDUHQWV

EURWKHU+LURIXPL0RPDQGGDG\RXKDYHDOZD\VEHHQHQFRXUDJLQJPHDQGVXSSRUWLQJ

PH LQ PDQ\ ZD\V 'DG \RX GHILQLWHO\ VKRZHG PH D JUHDW DWWLWXGH WRZDUGV OLIH

DOZD\V ORYH WR UHDG DQG OHDUQ :KHQ \RX SDVVHG RQH RI WKH PRVW GLIILFXOW -DSDQHVH

QDWLRQDOOLFHQVHVDIWHU\HDUVRIVWXG\\RXZHUH\HDUVROG$WWKDWWLPH,KHVLWDWHGWR

JRWRXQLYHUVLW\EHFDXVH,WKRXJKW,ZDVWRRROGWRVWXG\EXW\RXSURYHGWKDWWKHDJHGRHV

QRW PDWWHU WR DFFRPSOLVK VRPHWKLQJ 0RP DQG GDG , DP WKDQNIXO WR KDYH VXFK D

ZRQGHUIXOSDUHQWVOLNH\RX+LURIXPL,UHDOO\DSSUHFLDWH\RXUVXSSRUWDQGWDNLQJFDUHRI

RXUSDUHQWVZKLOH,DPLQWKH86

,DPGHHSO\JUDWHIXOWRP\PDMRUDGYLVRU3URIHVVRU

KDYLQJDYHU\WRXJKWLPHDQGZDVGHSUHVVHGZKHQ,FDPHWR3XUGXH

UHDGSHRSOH¶VPLQGEHFDXVH\RXVKRZHGPHDQH[FLWLQJSDUWRIFRQGXFWLQJUHVHDUFKDW

HDUO\VWDJHRIP\JUDGXDWHFDUULHU7KLVH[SHULHQFHHQJDJHGPHLQWRUHVHDUFKDQGPDGH

PH IHHO FRQILGHQW DERXW ZRUNLQJ LQ \RXU JURXS :KHQHYHU , ZHQW WR \RXU RIILFH ZLWK

XQZDQWHG UHVXOWV \RX JDYH PH YDOXDEOH VXJJHVWLRQV DQG VROXWLRQV $OVR \RXU SDVVLRQ

DQG FRQILGHQFH WRZDUGV UHVHDUFK HQFRXUDJHG PH WR NHHS PRYLQJ IRUZDUG

HQWKXVLDVWLFDERXWWHDFKLQJDQGDOZD\VNQRZKRZWREUHDNGRZQFRPSOLFDWHGWKLQJVLQWR

VLPSOH DQG HDV\ IRU XV WR XQGHUVWDQG , UHDOO\ DSSUHFLDWH \RXU KLJK VWDQGDUGV RQ WKH LY

TXDOLW\RISUHVHQWDWLRQVOLGHVDQGSXEOLFDWLRQVZKLFKJDYHPHDFKDQFHWRLPSURYHWKRVH

VNLOOVZKLFKDUHH[WUHPHO\XVHIXOLQDGGLWLRQWRWKHUHVHDUFKVNLOOV$OVRWKDQN\RXIRU

OHWWLQJPHDWWHQGWR$606FRQIHUHQFHVHYHU\\HDUDQG,06&LQ.\RWR-DSDQ7KURXJK

WKHVHFRQIHUHQFHV ,FRXOGPHHW JUHDWVFLHQWLVWVDURXQGWKHZRUOG DQGWKHVHH[SHULHQFHV

NHSWPHPRWLYDWHG

,DPWKDQNIXOWR3URIHVVRU%UDG%HQGLDNDW8QLYHUVLW\RI&RORUDGR:HKDYHEHHQ

FROODERUDWLQJVLQFHWKHYHU\EHJLQQLQJRIP\JUDGXDWHFDUHHU

DQGVKDUHGH[SHUWLVHZLWKPH:KHQHYHU,DVNHG\RXTXHVWLRQVWKURXJKHPDLOVHQGQHZ

GDWD RU PDQXVFULSWV \RX DOZD\V JDYH PH DQVZHUV ZLWK D IXOO RI H[SODQDWLRQV ZKLFK

KHOSHGPHDORWWRXQGHUVWDQGLQJWKHLPSRUWDQWFRQFHSWRIWKHPDWHULDO

,DPDOVRWKDQNIXOWR3URIHVVRU=KHQJ2X\DQJDQGVWXGHQWVLQKLVODEHVSHFLDOO\

0HORGLH 'X IRU WKH FROODERUDWLRQ RQ WKH DOJRULWKP GHYHORSPHQW IRU VSHFWUDOPDWFKLQJ

DQGFRQGXFWLQJVSHFWUDO VLPLODULW\FDOFXODWLRQV IRU VRPHRIP\SDSHUV 7R 'U *UDKDP

&RRNV'U6FRWW0F/XFNH\'U+LONND.HQWWlPDD'U0DU\:LUWKDQG'U$QG\7DR

, DSSUHFLDWH \RXUNLQGQHVVHVIRURIIHULQJPHDQDFFHVVWR \RXUHTXLSPHQWVLQVWUXPHQWV

DQGDOVRKHOSIURPVWXGHQWVLQ\RXUJURXS

,ZRXOGOLNHWRWKDQN'U-LP+DJHU)UDQN/RQGU\DQG'U/DUU\&DPSEHOOIURP

$%6FLH[IRUDYDULHW\RIKHOSVVXFKDVLQVWDOOLQJWKHVFDQGHVLJQHUPRGLILFDWLRQRQRXU

LQVWUXPHQWDQGJLYLQJVXJJHVWLRQVWRLPSURYHP\SDSHU

7R DOO RI P\ FXUUHQW ODE PDWHV WKDQN \RX IRU RIIHULQJ VXFK D QLFH DQG KRPH\

DWPRVSKHUH , WUXO\ HQMR\HG RXU PXOWLFXOWXUDO HQYLURQPHQW &KDVLW\ /RYH ZH DUH

3URIHVVRU;LD¶VYHU\ILUVWVWXGHQWVDQGVKDUHGPDQ\H[SHULHQFHVJRRGWLPHVDVZHOODV Y

GLIILFXOW WLPHV , DP UHDOO\ JODG WKDW , KDYH EHHQ KDYLQJ \RX DQG , FRXOG QRW PDNH LW

WKURXJKSDVWIRXU\HDUVZLWKRXW\RXIRUVXUH

\RXEHVWOXFNLQWKHIXWXUH;LDR[LDR0D\RXZHUHDOVRZLWKPHDQG&KDVLW\ZKHQ;LD

ODEMXVWJRW ODXQFKHG 7KDQN \RX IRUHQFRXUDJLQJPHGXULQJP\ILUVW \HDURI JUDGXDWH

FDUHHUDQG,DOVRHQMR\HGKDYLQJ\RXEDFNKHUHDVDSRVWGRF.LUW'XUDQGWKDQN\RXIRU

EHLQJ WKH PRVW XQGHUVWDQGLQJ SHUVRQ LQ WKH JURXS DQG DOVR KHOSHG PH LPSURYLQJ P\

FRPSXWHU VNLOOV /HL 7DQ WKDQN \RX IRU EHLQJ IXQQ\ PDNLQJ DOO RI XV ODXJK GXULQJ

FKDWWLQJ LQ WKH RIILFH 2Q WKH RWKHU KDQG \RX EHFDPH UHDOO\ VHULRXV DERXW UHVHDUFK ,

DSSUHFLDWH \RXUKHOSIXO FRPPHQWVDQG DGYLVHRQP\ UHVHDUFKSURMHFWV &UDLQJ6WLQVRQ

WKDQN\RXIRUSURRIUHDGLQJP\PDQXVFULSWVDQGDOVRIRUVKDULQJ\RXUNQRZOHGJHDERXW

LQVWUXPHQWDWLRQ -LD 5HQ WKDQN \RX IRU KHOSLQJ PH DQG WDNLQJ RYHU P\ FDUERK\GUDWH

SURMHFWV,UHDOO\HQMR\HGZRUNLQJZLWK\RX HVSHFLDOO\IRUWKHQLJKWVDQGZHHNHQGV IRU

SDVWRQH\HDU+DQIHQJ+XWKDQN\RXIRUODEFKDWWLQJDQGVKDULQJ\RXUFRRNLHVGXULQJ

JURXSPHHWLQJ,ZLVKDOORI\RXDJRRGOXFNDQG,DPVXUH\RXZLOOGRJUHDW

$OVRWKDQNVWR'U-DPHOOH:LOOLDPVIURP:HQWKROGJURXS'U/HLGG\$OYDUDGR

IURP5DIWHU\JURXSDQG'U0DULMD0HQWLQRYDIURP0F/XFNH\JURXS'U%UHWW6FKLOOLQJ

IURP +U\F\QD JURXS 'U &KDULW\ :D\XD DQG 'U 6FRWW 3RK IURP /RZ JURXS IRU

VXSSRUWLQJPHDWHDUO\VWDJHRIP\JUDGXDWHFDUULHU7KDQNVWR2VFDU0RUDOHVDQG0DWW

&DVVHOPDQ IURP :HL JURXS :DWDUX 0LWVXKDVKL IURP 5DPDFKDGUDQ JURXS +LURIXPL

.DPDGD IURP 1HJLVKL JURXS IRU KHOSIXO GLVFXVVLRQ RQ P\ VXJDU V\QWKHVLV SURMHFW

(VSHFLDOO\+LURIXPL.DPDGDZLWKRXW\RXUVXSSRUW,FRXOGQRWKDYHILQLVKHGWKHSURMHFW

ZLWKLQWKHOLPLWHGDPRXQWRIWLPH YL

,DOZD\V JRWPDQ\HQFRXUDJHPHQWVDQGVXSSRUWVIURPSURIHVVRUV DQGIULHQGVDW

3XUGXH DQG RXWVLGH RI 3XUGXH , DP YHU\ IRUWXQDWH WR KDYH VXFK D ZRQGHUIXO SHRSOH

DURXQGPH6SHFLDOWKDQNVWR'U.LKDUDDQGKLVZLIH1DKRDQGWKHLUGDXJKWHU0HL

WUHDWHGPHOLNHDSDUWRI\RXUIDPLO\'U0L]XNDPL DQG\RXUORYHO\FDWV WKDQN\RXIRU

\RXUFRPSDQ\DVZHOODVVFLHQWLILFDQGQRQVFLHQWLILFDGYLFHVZKHQ,QHHGHG7KDQNVWR

DOORIP\SURIHVVRUVLQXQGHUJUDGXDWHLQVWLWXWH9DOGRVWD6WDWH8QLYHUVLW\IRUVXJJHVWLQJ

PHWRJRWRJUDGXDWHVFKRRODQGVXSSRUWLQJPHGXULQJP\JUDGXDWHFDUULHU

)LQDOO\,ZRXOGOLNHWR WKDQN'U3HWHU.LVVLQJHU'U1LNRODL6NU\QQLNRYDQG

'U+LONND.HQWWlPDDIRUVHUYLQJDVP\WKHVLVFRPPLWWHHPHPEHUV7KDQN\RXIRU\RXU

WLPHDQGDOORI\RXUKHOSWKURXJKRXWP\JUDGXDWHFDUHHU

YLL

7$%/(2)&217(176

3DJH

/,672)7$%/(6[

/,672)),*85(6[L

/,672)6&+(0(6[YLL

/,672)6<0%2/6$1'$%%5(9,$7,216[YLLL

$%675$&7[[

&+$37(5,1752'8&7,21

&DUERK\GUDWHV *O\FRV\ODWLRQ 2YHUYLHZRI*O\FDQ6WUXFWXUDO$QDO\VLV 6WUXFWXUDO$QDO\VLVRI*O\FDQVE\0DVV6SHFWURPHWU\ ,RQL]DWLRQ 'HULYDWL]DWLRQ7HFKQLTXHVIRU*O\FDQ$QDO\VLV 1RPHQFODWXUHIRUWKH*DV3KDVH)UDJPHQWDWLRQRI2OLJRVDFFKDULGHV 6WUXFWXUDO$QDO\VLVRI2OLJRVDFFKDULGHVE\0DVV6SHFWURPHWU\ 06YV06Q $1HZ06Q$SSURDFKIRU+LJK/HYHO6WUXFWXUDO$QDO\VLVRI 2OLJRVDFFKDULGHV &RQFOXVLRQV 5HIHUHQFHV

&+$37(5',))(5(17,$7,212)7+(67(5(2&+(0,675<$1' $120(5,&&21),*85$7,21)25/,1.(' ',6$&&+$5,'(6

,QWURGXFWLRQ ([SHULPHQWDO 0DWHULDOV 2ODEHOLQJRI5HGXFLQJ'LVDFFKDULGHV YLLL

3DJH 0DVV6SHFWURPHWU\ 5HVXOWVDQG'LVFXVVLRQV &,'RI'HSURWRQDWHG'LVDFFKDULGH,RQVIURP/LQNHG'LVDFFKDULGHV &,'RIm/z,RQV*HQHUDWHGIURP/LQNHG'LVDFFKDULGHV 7KH(IIHFWRI&,'&RQGLWLRQVRQWKH)RUPDWLRQRIm/z'LDJQRVWLF ,RQV 6WHUHRFKHPLVWU\'HWHUPLQDWLRQRIWKH1RQ5HGXFLQJ6XJDUZLWKLQ /LQNHG'LVDFFKDULGHV 6SHFWUDO0DWFKLQJE\6LPLODULW\6FRUHV &RQFOXVLRQV 5HIHUHQFHV

&+$37(52%7$,1,1*67(5(26758&785$/,1)250$7,21:,7+ 6,1*/(68*$55(62/87,21)520/,1($5 2/,*26$&&+$5,'(6

,QWURGXFWLRQ ([SHULPHQWDO 0DWHULDOV 5HGXFLQJ(QG0RGLILFDWLRQ 5HGXFWLYH$PLQDWLRQ 0DVV6SHFWURPHWU\ 06&,''DWDEDVHRI'$OGR+H[*$6\QWKHWLF6WDQGDUGV 6SHFWUDO0DWFKLQJE\6LPLODULW\6FRUHV 5HVXOWVDQG'LVFXVVLRQV 06Q$SSURDFK $+\EULG7ULSOH4XDGUXSROH/LQHDU,RQ7UDSIRU06Q 7KH(IIHFWRI&,'&RQGLWLRQVRQWKH)RUPDWLRQRI.H\3URGXFW,RQV 'HWHUPLQDWLRQRI,QGLYLGXDO6XJDU,GHQWLW\DQG$QRPHULF &RQILJXUDWLRQIURP3HQWDVDFFKDULGHV &RQFOXVLRQV 5HIHUHQFHV

&+$37(5=,216$6',$*1267,&,216)25/,1.$*( '(7(50,1$7,21

,QWURGXFWLRQ ([SHULPHQWDO 0DWHULDOV 5HGXFLQJ(QG0RGLILFDWLRQ $PLQDWLRQZLWKRXW5HGXFWLRQ 0DVV6SHFWURPHWU\ +3/& 5HVXOWVDQG'LVFXVVLRQV 'LVDFFKDULGH/LQNDJH$QDO\VLV%DVHGRQ06 &,'RI=,RQV L[

3DJH 7KH(IIHFWRI&,'&RQGLWLRQVRQWKH)RUPDWLRQRI=,RQV Q /LQNDJH'HWHUPLQDWLRQIRU2OLJRVDFFKDULGHVYLD06 &,'RI=,RQV /LQNDJH'HWHUPLQDWLRQIRUD%UDQFKHG2OLJRVDFFKDULGH $OJRULWKP$VVLVWHG/LQNDJH'HWHUPLQDWLRQ &RQFOXVLRQV 5HIHUHQFHV

&+$37(57+(())(&762)5('8&,1*(1'02',),&$7,2121*$6 3+$6(&+(0,675<2)60$//2/,*26$&&+$5,'(6

,QWURGXFWLRQ ([SHULPHQWDO 0DWHULDOV 2'HULYDWL]DWLRQWR5HGXFLQJ6XJDUV $FHW\ODWLRQRI+\GUR[\O*URXSVRQ6XJDUV 2'HULYDWL]DWLRQWR$FHW\ODWHG6XJDUV 'HDFHW\ODWLRQRI2'HULYDWL]HG$FHW\ODWHG6XJDUV 'HVDOWLQJRI2'HULYDWL]HG6XJDUVE\+3/& 0DVV6SHFWURPHWU\ 5HVXOWVDQG'LVFXVVLRQV *DV3KDVH)UDJPHQWDWLRQ6WXGLHVRI1'HULYDWL]HG'LVDFFKDULGHV *DV3KDVH)UDJPHQWDWLRQ6WXGLHVRI2'HULYDWL]HG'LVDFFKDULGHV &RPSDULVRQVEHWZHHQ1'HULYDWL]HG E\$PLQDWLRQ DQG2'HULYDWL]HG 7ULVDFFKDULGHV &RQFOXVLRQV 5HIHUHQFHV

9,7$

38%/,&$7,216 [

/,672)7$%/(6

7DEOH 3DJH

/LVWRIGLVDFFKDULGHVEHLQJVWXGLHG

6SHFWUDOVLPLODULW\VFRUHVIRUOLQNHGGLVDFFKDULGHVYVPRQRVDFFKDULGH*$ VWDQGDUGV

7KHWRSUDQNHG'DOGR+H[FDQGLGDWHVEDVHGRQVSHFWUDOVLPLODULW\VFRUHV LQGLFDWHGLQSDUHQWKHVLV

([SHULPHQWDOFRQGLWLRQVIRUFRQGXFWLQJ06QH[SHULPHQWV

6SHFWUDOVLPLODULW\VFRUHVIRUHDFKXQLWLQROLJRVDFFKDULGHVDJDLQVW' DOGR+H[*$VWDQGDUGV

/LVWRIGLVDFFKDULGHVDQGROLJRVDFFKDULGHVEHLQJVWXGLHG

Q 6SHFWUDOVLPLODULW\VFRUHVRI06 &,'RI=LRQVSHFWUDGHULYHGIURP ROLJRVDFFKDULGHVYVGLVDFFKDULGHVWDQGDUGV

/LVWRIUHGXFLQJHQGGHULYDWLYHV

6XPPDU\RIWKHQRUPDOL]HGSURGXFWLRQLQWHQVLW\RIFKDUDFWHULVWLFLRQV JHQHUDWHGIURPGHULYDWL]HGGLVDFFKDULGHVDQGWULVDFFKDULGHV

/,672)),*85(6

)LJXUH 3DJH

5LQJDQGRSHQULQJIRUPVRI'JOXFRVH

7KHQXPEHURISRVVLEOHVWUXFWXUDOLVRPHUVDVLQFUHDVLQJGHJUHHRI SRO\PHUL]DWLRQ FRQVLGHULQJRQO\DOGRKH[RVHVDQGOLQHDUVWUXFWXUH

&RPPRQFRUHVWUXFWXUHVIRU D 1DQG E 2JO\FDQV

6WUXFWXUDOLQIRUPDWLRQQHFHVVDU\WRIXOO\FKDUDFWHUL]HJO\FDQV

5HDFWLRQVFKHPHIRUDPLQDWLRQDQGUHGXFWLYHDPLQDWLRQ

1RPHQFODWXUHIRUJO\FRFRQMXJDWHSURGXFWLRQVJHQHUDWHGE\WDQGHP06

06&,'VWHSVIRUVXJDUXQLWLGHQWLW\DQGDQRPHULFFRQILJXUDWLRQ GHWHUPLQDWLRQ

06&,'VSHFWUDLQWKHQHJDWLYHLRQPRGHREWDLQHGIURPGHSURWRQDWHGȕ' *OFp '*OF m/z XQGHUORZHQHUJ\GLVVRFLDWLRQFRQGLWLRQV D EHDPW\SH&,' &( 9 DQG E LRQWUDS&,' $) ,QVHWVLQ D DQG E VKRZWKHLVRODWLRQRIm/zLRQVJHQHUDWHGIURPEHDPW\SH&,' LQMHFWLRQWLPH V DQGLRQWUDS&,' LQMHFWLRQWLPH V UHVSHFWLYHO\

06LRQWUDS&,'VSHFWUDRIm/zLRQVGHULYHGIURPV\QWKHWLFVWDQGDUGV D Į'*OFp*$$) DQG E ȕ'*OFp*$$) 06&,'VSHFWUD RIm/zLRQVGHULYHGIURPORZHQHUJ\EHDPW\SH&,'RIJOXFRVH FRQWDLQLQJGLVDFFKDULGHV F Į'*OFp *OF&( 9IRU06DQG$) IRU06 G ȕ'*OFp *OF&( 9IRU06DQG$) IRU 06 H Į'*OFp *OF&( 9IRU06DQG$) IRU06DQG I ȕ'*OFp *OF&( 9IRU06DQG$) IRU067KHHUURU EDUVLQWKHVSHFWUDVKRZWKHVWDQGDUGGHYLDWLRQRIWKHSHDNLQWHQVLW\EDVHGRQ VHYHQVSHFWUDFROOHFWHGRYHUD\SHULRG

[LL

)LJXUH 3DJH

06VSHFWUDRI2ODEHOHGĮ'*OFp '*OFXQGHU D UHODWLYHO\ORZ HQHUJ\EHDPW\SH&,' &( 9 DQG E LRQWUDS&,' $) ,QVHWVLQ D DQG E VKRZLVRODWLRQRIm/zDQGLRQVJHQHUDWHGIURPEHDPW\SH &,' LQMHFWLRQWLPH V DQGLRQWUDS&,' LQMHFWLRQWLPH V UHVSHFWLYHO\

06VSHFWUDRIm/zDQGLRQVGHULYHGIURP2ODEHOHGĮ'*OFp '*OFDQGȕ'*OFp '*OF D &,'RIm/zLRQV&( 9 06 $) 06 E &,'RIm/zLRQV&( 9 06 $) 06 IURP2ODEHOHGĮ'*OFp '*OFDQG F &,'RIm/zLRQV &( 9 06 $) 06 G &,'RIm/zLRQV&( 9 06 $) 06 IURP2ODEHOHGȕ'*OFp '*OF

,VRODWLRQRIm/zDQGm/zLRQVGHULYHGIURP2ODEHOHGȕ'*OFp '*OFXQGHUGLIIHUHQWFROOLVLRQDODFWLYDWLRQFRQGLWLRQV%HDPW\SH&,' D &( 9 E &( 9 F &( 9,RQWUDS&,'DWT I N+]H[FLWDWLRQWLPH PV G P9SS H P9SS I P9SS

06&,'RIm/zLRQVJHQHUDWHGYLDXVLQJKLJK&( &( WR9 IRU WKHGLVVRFLDWLRQRIGHSURWRQDWHGGLVDFFKDULGHLRQV m/z $ Į'*OFp '*OF&( 9 06 $) 06 % ȕ'*OFp '*OF &( 9 06 $) 06 & Į'*OFp ')UX&( 9 06 $) 06 ' Į'*DOp '*DO&( 9 06 $) 06 DQG ( Į'0DQp '0DQ&( 9 06 $) 06 7KHHUURUEDUVLQWKHVSHFWUDVKRZWKHVWDQGDUGGHYLDWLRQRISHDN LQWHQVLWLHVEDVHGRQVHYHQVSHFWUDFROOHFWHGRYHUD\SHULRG

&,'VSHFWUDRIm/zLRQVGHULYHGIURPĮ'*OFp ')UX&( 9 06 $) 06

&,'VSHFWUDRIm/zLRQVGHULYHGIURP D Į'*DOp*$$) 06 E Į'0DQp*$$) 06 F Į'*DO '*DO&( 9 06 $) 06 G Į'0DQ '0DQ&( 9 06 $) 06 7KHHUURUEDUVLQWKHVSHFWUD D DQG E VKRZWKHVWDQGDUG GHYLDWLRQRIWKHUHODWLYHLQWHQVLW\FDOFXODWHGEDVHGRQVHYHQVSHFWUDFROOHFWHG RYHUD\SHULRG

$YHUDJHG06WUDS&,'GDWDRIGHSURWRQDWHG'DOGR+H[*$VWDQGDUGV m/z 7KHPDMRUSHDNVZHUHODEHOOHGZLWKVWDQGDUGGHYLDWLRQEDVHGRQ DYHUDJHGVSHFWUDREWDLQHGRYHUD\SHULRG

[LLL

)LJXUH 3DJH

$VFKHPDWLFRIDPRGLILHGWULSOHTXDGUXSROHOLQHDULRQWUDSPDVVVSHFWURPHWHU 475$3 %HDPW\SH&,'FDQEHSHUIRUPHGWKURXJKHQHUJHWLFD[LDO LRQWUDQVIHUEHWZHHQTXDGUXSROHDUUD\V 4WR4 D FRQYHQWLRQDOPHWKRG 4WR4ZLWKPDVVLVRODWLRQLQ4E UHYHUVHGLRQWUDQVIHUIURP4WR4 ZLWKPDVVLVRODWLRQLQ4DQGF UHYHUVHGLRQWUDQVIHUIURP4WR4ZLWK PDVVLVRODWLRQLQ4

06 DQG06 H[SHULPHQWVRIDSHQWDVDFFKDULGH>Į'*OFp @'*OF $%$ D ,RQWUDQVIHUVFKHPHIRUFRQGXFWLQJHLWKHUEHDPW\SH&,'RULRQ UG WUDS&,'LQWKH VWDJH0606IRUWKHIRUPDWLRQRI&LRQV m/z YLD 06> ÆÆIUDJPHQWV@ E 06EHDPW\SH&,'IURP4WR4 &( 9 F 06LRQWUDS&,'LQ4T 9DQGPVRI DFWLYDWLRQ G ,RQWUDQVIHUVFKHPHIRUFRQGXFWLQJEHDPW\SHRULRQWUDS&,' LQWKHWKVWDJH0606IRUWKHIRUPDWLRQRIm/zGLDJQRVWLFLRQVYLD06 > ÆÆÆIUDJPHQWV@ H 06EHDPW\SH&,'IURP4WR4&( 9 I 06LRQWUDS&,'LQ4T 9DQGPVRIDFWLYDWLRQ 7KHVWUXFWXUHRI>Į'*OFp @'*OF$%$LVVKRZQLQWKHLQVHW

06 H[SHULPHQWVRISHQWDVDFFKDULGHV>Į'*OFp @'*OF$%$DQG>ȕ '*OFp @'*OF$%$ D ,RQWUDQVIHUVFKHPHIRUFRQGXFWLQJ06 &,' IRUERWKSHQWDVDFFKDULGHV06 LRQWUDS&,'RI E >Į'*OFp @'*OF $%$DQG F >ȕ'*OFp @'*OF$%$06 LRQWUDS&,'RIV\QWKHWLF VWDQGDUGV G Į'*OFp*$DQG H ȕ'*OFp*$

06Q Q RU IRUWKHGHWHUPLQDWLRQRIVWHUHRVWUXFWXUDOLQIRUPDWLRQRI VXJDUXQLWVIURP>Į'*OFp @'*OF$%$6SHFWUD D WR F VKRZ WKHVHTXHQWLDO06WR06GDWDEDVHGRQ> ÆÆÆIUDJPHQWV@WR GHWHUPLQHVXJDUXQLWVWUXFWXUDOLQIRUPDWLRQ6SHFWUD G WR I VKRZWKH VHTXHQWLDO06WR06GDWDEDVHGRQ> ÆÆÆÆIUDJPHQWV@ WRGHWHUPLQHVXJDUXQLW6SHFWUD J WR L VKRZWKHVHTXHQWLDO06WR06 GDWDEDVHGRQ> ÆÆÆÆIUDJPHQWV@WRGHWHUPLQHVXJDUXQLW 7KHH[SHULPHQWDOFRQGLWLRQVIRUHDFKVWHSFDQEHIRXQGLQ7DEOH

06Q Q RU IRUWKHGHWHUPLQDWLRQRIVWHUHRVWUXFWXUDOLQIRUPDWLRQRI VXJDUXQLWVIURP>ȕ'*OFp @'*OF$%$6SHFWUD D WR F VKRZ WKHVHTXHQWLDO06WR06GDWDEDVHGRQ> ÆÆÆIUDJPHQWV@WR GHWHUPLQHVXJDUXQLWVWUXFWXUDOLQIRUPDWLRQ6SHFWUD G WR I VKRZWKH VHTXHQWLDO06WR06GDWDEDVHGRQ> Æ ÆÆÆIUDJPHQWV@WRGHWHUPLQHVXJDUXQLW6SHFWUD J WR L VKRZ WKHVHTXHQWLDO06WR06GDWDEDVHGRQ> ÆÆÆÆIUDJPHQWV@WRGHWHUPLQHVXJDUXQLW7KHH[SHULPHQWDO FRQGLWLRQVIRUHDFKVWHSFDQEHIRXQGLQ7DEOH

[LY

)LJXUH 3DJH

06EHDPW\SH&,'RI2ODEHOHGĮ'JOXFRVHKRPRGLPHUVm/z >0 +@ ZLWKGLIIHUHQWOLQNDJHSRVLWLRQV D &(9 E &(9 F &(9DQG G &(9,QVHWVLQHDFKRIWKHVSHFWUDVKRZWKH H[SDQGHGUHJLRQFRYHULQJm/zDQG

06LRQWUDS&,'RIm/zSURGXFWLRQVGHULYHGIURP2ODEHOHG GLVDFFKDULGHV DDQRPHULFFRQILJXUDWLRQ ZLWKGLIIHUHQWVXJDULGHQWLWLHVDQG OLQNDJHSRVLWLRQV D Į'*OFp '*OF E Į'*DOp '*DO F Į '*OFp '*OF G Į'0DQp '0DQ H Į'*OFp '*OF I ȕ'*DOp '0DQ J Į'*OFp '*OFDQG K Į'*DOp '*OF

$YHUDJHG06 LRQWUDS&,'GDWDRI=LRQV m/z GHULYHGIURP 2 ODEHOHGJOXFRVHKRPRGLPHUVZLWKGLIIHUHQWOLQNDJHSRVLWLRQV D E F DQG G 7KHHUURUEDUVUHSUHVHQWVWDQGDUGGHYLDWLRQVRISHDN LQWHQVLWLHVEDVHGRQDYHUDJHGVSHFWUD VSHFWUDHDFKIURPDDQGE DQRPHUV REWDLQHGRYHUDPRQWKSHULRG

06&,'VSHFWUDRIm/zSURGXFWLRQVIURP D Į'*OFp '*OF2 DQG E ȕ'*OFp '*OF2$GGLWLRQDOSHDNV m/zDQG ZHUH IRXQGLQWKHȕFRQILJXUDWLRQ

06LRQWUDS&,'RIm/z SURGXFWLRQVIURP D Į'0DQp '0DQ 2 E ȕ'*DOp '*OF2 F Į'*DOp '*DO2 G ȕ' *DOp '0DQ2 H Į'0DQp '*OF2 I Į'*DOp '*DO2DQG J Į'*DOp '*OF2

7KHHIIHFWRI&,'FRQGLWLRQVRQWKHIRUPDWLRQRI= m/z YVLWVVWUXFWXUDO LVRPHUV m/z 'DWDZHUHREWDLQHGIURP2ODEHOHGJOXFRVHKRPRGLPHUV D Į'*OFp '*OF E ȕ'*OFp '*OF F ȕ'*OFp ' *OFDQG G Į'*OFp '*OF

$YHUDJHG06 LRQWUDS&,'GDWDRI=SURGXFWLRQV m/z GHULYHGIURP nativeXQODEHOHGJOXFRVHKRPRGLPHUVZLWKGLIIHUHQWOLQNDJHSRVLWLRQV D E F DQG G 7KHHUURUEDUVVKRZVWDQGDUGGHYLDWLRQRISHDN LQWHQVLWLHVEDVHGRQDYHUDJHGVSHFWUD VSHFWUDHDFKIURPDDQGE DQRPHUV REWDLQHGRYHUDPRQWKSHULRG,RQWUDS&,'ZLWKP9SSZDV XVHGIRU06&,'RIm/zVWHSIRU D WKURXJK G

)LJXUH 3DJH

06 LRQWUDS&,'RIWKH=LRQIURP 2ODEHOHGIUXFWRVHFRQWDLQLQJ GLVDFFKDULGHVDQG06 LRQWUDS&,'RIWKH=LRQJHQHUDWHGE\LRQWUDS&,' ZLWKORZHQHUJ\IURPnative XQODEHOHGIUXFWRVHFRQWDLQLQJGLVDFFKDULGHV D Į'*OFp ')UX2 E Į'*OFp ')UX F ȕ'*DOp ' )UX2DQG G ȕ'*DOp ')UX

06Q&,'VSHFWUDIRUWKHGHWHUPLQDWLRQRI OLQNDJHW\SHVZLWKLQWULVDFFKDULGHV Į'*OFp Į'*OFp '*OF2 D 06EHDPW\SH&,' >ÆIUDJPHQWV@ E 06LRQWUDS&,'>ÆÆIUDJPHQWV@ F 06LRQ WUDS&,'>ÆÆIUDJPHQWV@ G 06LRQWUDS&,' >ÆÆÆIUDJPHQWV@Į'*DOp ȕ'*DOp '*DO2 H 06LRQWUDS&,'>ÆÆIUDJPHQWV@DQG I 06LRQWUDS&,' >ÆÆÆIUDJPHQWV@

06DQG06&,'VSHFWUDIURPĮ'*DOp ȕ'*DOp '*DO2 D 06EHDPW\SH&,'>ÆIUDJPHQWV@DQG E 06LRQWUDS&,' >ÆÆIUDJPHQWV@

06 DQG06 &,'VSHFWUDIURP>ȕ'*OFp @'*OF 2([SHULPHQWDO VWHSIRUOLQNDJHDQDO\VLVLV>ÆÆIUDJPHQWV@ D 06EHDPW\SH &,'>ÆIUDJPHQWV@ E 06LRQWUDS&,'>ÆÆIUDJPHQWV@ ([SHULPHQWDOVWHSIRUOLQNDJHDQDO\VLVLV>ÆÆÆIUDJPHQWV@ F 06LRQWUDS&,'>ÆÆIUDJPHQWV@DQG G 06LRQWUDS&,' >ÆÆÆIUDJPHQWV@

06 WR06 &,'VSHFWUDIURP$%((PRGLILHG>ȕ'*OFp @'*OF ([SHULPHQWDOVWHSIRUOLQNDJHDQDO\VLVLV> ÆÆÆÆIUDJPHQWV@ D 06EHDPW\SH&,'> ÆIUDJPHQWV@ E 06LRQWUDS&,'> ÆÆIUDJPHQWV@ F 06LRQ WUDS&,'> ÆÆÆIUDJPHQWV@DQG G 06LRQWUDS&,' > ÆÆÆÆIUDJPHQWV@([SHULPHQWDOVWHSIRUOLQNDJH DQDO\VLVLV> Æ ÆÆÆIUDJPHQWV@ H 06LRQWUDS&,' > Æ ÆIUDJPHQWV@ I 06LRQWUDS&,'> Æ ÆÆIUDJPHQWV@DQG J 06LRQWUDS&,'> Æ ÆÆÆIUDJPHQWV@

+3/&FKURPDWRJUDPVIURP D VWHSDQG E VWHSVHSDUDWLRQV

06Q&,'VSHFWUDIRUWKHGHWHUPLQDWLRQRI OLQNDJHW\SHVZLWKLQWULVDFFKDULGHV GLJHVWHGIURP0DQĮ'0DQp Į'0DQp '0DQ2 D 06 EHDPW\SH&,'>ÆIUDJPHQWV@ E 06LRQWUDS&,' >ÆÆIUDJPHQWV@ F 06LRQWUDS&,'>ÆÆIUDJPHQWV@ G 06LRQWUDS&,'>ÆÆÆIUDJPHQWV@ [YL

)LJXUH 3DJH

06EHDPW\SHYVLRQWUDS&,'VIURP D ȕ'*OFp '*OF2 E ȕ '*OFp '*OF15$%$ F ȕ'*OFp '*OF5$%$ G ȕ' *OFp '*OF15$%((DQG H ȕ'*OFp '*OF5$%((

06EHDPW\SHYVLRQWUDS&,'IURP D ȕ'*OFp '*OF+%(( E ȕ '*OFp '*OF+%$ F ȕ'*OFp '*OF+%6$ G ȕ'*OFp '*OF+1$DQG H ȕ'*OFp '*OF+1$

06EHDPW\SH&,'IURP D Į'*OFp Į'*OFp '*OF15 $%$DQG E Į'*OFp Į'*OFp '*OF15$%((

06 EHDPW\SHYVLRQWUDS&,'IURP D >Į'*OFp @'*OF 2 E >Į'*OFp @'*OF+%(( F >Į'*OFp @'*OF+%$DQG G >Į'*OFp @'*OF+%6$ [YLL

/,672)6&+(0(6

6FKHPH 3DJH

7KH06Q Q RU DSSURDFKIRUGHWDLOHGVWUXFWXUDODQDO\VLVRIDOLQHDU ROLJRVDFFKDULGH&KDUJHVDUHQRWLQGLFDWHGRQWKHVWUXFWXUH³0´VWDQGVIRU UHGXFLQJHQGPRGLILFDWLRQ

)RUPDWLRQRIJO\FRV\O*$DQLRQVDWm/zIURP&,'RIGHSURWRQDWHGOLQNHG GLVDFFKDULGHV

$VFKHPDWLFGLDJUDPRIWKH475$3DK\EULGWULSOHTXDGUXSROHOLQHDULRQ WUDSPDVVVSHFWURPHWHU

7KH06Q Q RU DSSURDFKIRUVWHUHRVWUXFWXUDODQDO\VLVRIROLJRVDFFKDULGHV ³0´VWDQGVIRUUHGXFLQJHQGPRGLILFDWLRQ6XJDUXQLWQXPEHULVDVVLJQHGIURP WKHUHGXFLQJHQG

7KH06QDSSURDFKIRUOLQNDJHGHWHUPLQDWLRQRIDQROLJRVDFFKDULGH³0´VWDQGV IRUUHGXFLQJHQGPRGLILFDWLRQ

6DPSOHSUHSDUDWLRQVWHSVIRUEUDQFKHGROLJRVDFFKDULGHEHIRUH06DQDO\VLV

5HDFWLRQVWHSVIRU2GHULYDWL]DWLRQDWUHGXFLQJVXJDUV

[YLLL

/,672)6<0%2/6$1'$%%5(9,$7,216

K\GUR[\QDSWKRLFDFLG +1$

DPLQREHQ]RLFDFLG $%$

K\GUR[\EHQ]HQHVXOIRQLFDFLG +%6$

K\GUR[\EHQ]RLFDFLG +%$

$FWLYDWLRQHQHUJ\IRULRQWUDS&,' $)

$OWHUQDWLQJ&XUUHQW $&

%RURQWULIOXRULGHGLHWK\OHWKHUDWH %)ā(W2

&ROOLVLRQ(QHUJ\IRUEHDPW\SH&,' &(

&ROOLVLRQ,QGXFHG'LVVRFLDWLRQ &,'

'LFKORURPHWKDQH '&0

'LPHWK\OVXOIR[LGH '062

'LUHFW&XUUHQW '&

(OHFWURQ&DSWXUH'LVVRFLDWLRQ (&'

(OHFWURQ'HWDFKPHQW'LVVRFLDWLRQ (''

(OHFWURQ([FLWHG'LVVRFLDWLRQ (('

(OHFWURQ,PSDFW (, [L[

(OHFWURQ,QGXFHG'LVVRFLDWLRQ (,'

(OHFWURVSUD\,RQL]DWLRQ (6,

(WK\ODPLQREHQ]RDWH $%((

(WK\OK\GUR[\EHQ]RDWH +%((

)DVW$WRP%RPEDUGPHQW )$%

)RXULHUWUDQVIRUPLRQF\FORWURQUHVRQDQFH )7,&5

*DV&KURPDWRJUDSK\ *&

*O\FRODOGHK\GH *$

+LJK3HUIRUPDQFH/LTXLG&KURPDWRJUDSK\ +3/&

+\GURSKLOLF,QWHUDFWLRQ/LTXLG&KURPDWRJUDSK\ +,/,&

,QFOXGLQJ,QIUDUHG0XOWLSKRWRQ'LVVRFLDWLRQ ,503'

0DVV6SHFWURPHWU\ 06

0DWUL[$VVLVWHG/DVHU'HVRUSWLRQ,RQL]DWLRQ 0$/',

1DFHW\OJDODFWRVDPLQH *DO1$F

1DFHW\OJOXFRVDPLQH *OF1$F

1DFHW\OQHXUDPLQLFDFLG 1HX$F

1RQUHGXFHGIRUP 15

1XFOHDU0DJQHWLF5HVRQDQFH 105 pDPLQRSKHQ\ODPPRQLXPFKORULGH 70$3$

5DGLR)UHTXHQF\ 5)

5HGXFHGIRUP 5

7DQGHP0DVV6SHFWURPHWU\ 0606 [[

$%675$&7

.RQGD &KLKDUX 3K ' 3XUGXH 8QLYHUVLW\ 'HFHPEHU 6WUXFWXUDO $QDO\VLV RI &DUERK\GUDWHVE\0DVV6SHFWURPHWU\0DMRU3URIHVVRU

*O\FRV\ODWLRQLVDFWLYHO\LQYROYHGLQLQWHUPROHFXODUDQGLQWHUFHOOXODUELQGLQJHYHQWVWKDW

DUHLPSRUWDQWWRDZLGHUDQJHRIELRORJLFDOIXQFWLRQVVXFKDVLPPXQLW\DQGIHUWLOLW\7KH

XQLTXH IXQFWLRQV RI JO\FDQV DUH GLUHFWO\ UHODWHG WR WKHLU VWUXFWXUHV ,Q RUGHU WR IXOO\

FKDUDFWHUL]H WKH VWUXFWXUH RI DQ XQNQRZQ JO\FDQ OHYHOV RI VWUXFWXUDO LQIRUPDWLRQ LV

QHHGHG VXJDU XQLW LGHQWLW\ DQRPHULF FRQILJXUDWLRQ OLQNDJH W\SHV VHTXHQFH DQG

EUDQFKLQJORFDWLRQ2QHRIWKHJUDQGFKDOOHQJHVOLPLWLQJWKHDGYDQFHRIJO\FRVFLHQFHVLV

WKH ODFN RI VHQVLWLYH WRROV IRU GHWDLOHG JO\FDQ VWUXFWXUDO DQDO\VLV 7DQGHP PDVV

VSHFWURPHWU\ 0606 LVDSRZHUIXOWRROIRUVWUXFWXUDODQDO\VLVRIFDUERK\GUDWHVZLWKLWV

KLJKVHQVLWLYLW\DQGVHOHFWLYLW\06 WZRVWDJHWDQGHPPDVVVSHFWURPHWU\ LVPRVWRIWHQ

DGRSWHG GXH WR WKH ZLGH DYDLODELOLW\ RI FRPPHUFLDO LQVWUXPHQWV KRZHYHU JO\FDQ

VWUXFWXUDOGHWHUPLQDWLRQW\SLFDOO\VWRSVDWWKHOHYHORIWRSRORJ\ UHODWLYHORFDWLRQRIVXJDU

XQLWHV 7KH JRDO RI WKLV UHVHDUFK LV WR GHYHORS D PXOWLSOH VWDJH WDQGHP PDVV

VSHFWURPHWU\ DSSURDFK WRZDUG D VHQVLWLYH DQG IXOO OHYHO VWUXFWXUDO DQDO\VLV RI OLQHDU

ROLJRVDFFKDULGHV E\ REWDLQLQJ LQIRUPDWLRQ RI LQGLYLGXDO VXJDU XQLW LGHQWLW\ DQRPHULF

FRQILJXUDWLRQ OLQNDJH W\SHV DQG VHTXHQFH 7KLV DSSURDFK URRWV LQ WKH GLVFRYHU\ RI [[L

GLDJQRVWLFLRQVZKLFKDUHVXEVWUXFWXUHVRIGLVDFFKDULGHVDQGWKHIUDJPHQWDWLRQSDWWHUQV

RIZKLFKFDQEHOLQNHGWRDVSHFLILFOHYHORIVWUXFWXUDOLQIRUPDWLRQ$Q06VFUHHQLQJ

PHWKRG ZDV GHYHORSHG DQG D QHZ GLDJQRVWLF LRQ = IRU OLQNDJH GHWHUPLQDWLRQ ZDV

HVWDEOLVKHGIRUWKHILUVWWLPH:LWKWKHXVHRIDQRWKHUGLDJQRVWLFLRQ QRQUHGXFLQJVXJDU

JO\FRVLGLFDOO\OLQNHGWRJO\FRODOGHK\GHm/z GLVFRYHUHGE\%HQGLDNJURXSLGHQWLW\

DQRPHULF LQIRUPDWLRQ DQG OLQNDJH SRVLWLRQ RI WKH QRQUHGXFLQJ VXJDU ZLWKLQ D

GLVDFFKDULGHFDQEHGHWHUPLQHGIURP06H[SHULPHQWV%\XWLOL]LQJSDWWHUQFRPSDULVRQ

DOJRULWKP VSHFWUDO VLPLODULW\ VFRUH KLJKO\ FRQILGHQW VWUXFWXUDO LQIRUPDWLRQ FDQ EH

REWDLQHG DXWRPDWLFDOO\ 7KH VWUDWHJ\ RI GLDJQRVWLF LRQV ZDV IXUWKHU GHPRQVWUDWHG RQ

PRGHOOLQHDUROLJRVDFFKDULGHVXVLQJ06Q Q IRUIXOOOHYHOVWUXFWXUDOGHWHUPLQDWLRQ

CHAPTER 1 INTRODUCTION

1.1 Carbohydrates

Carbohydrates (also called sugars, oligosaccharides, glycans) are defined as polyhydroxyaldehydes, polyhydroxyketones and their simple derivatives, or larger compounds that can be hydrolyzed into such units. A monosaccharide is the smallest unit of carbohydrates which cannot be hydrolyzed into a simpler form. Free monosaccharides can exist in ring or open-ring forms as shown in Figure 1.1. Changes in the orientation of hydroxyl groups around specific carbon atoms generate new sugar molecules. For example, mannose which has hydroxyl group on C2 is sticking up instead of down for glucose and mannose is the C2 epimer of glucose. There are 3 chiral centers (C2-C4) and

8 (= 23) possible structures for D-monosaccharides and 16 possible structures (4 chiral centers, C2-C5) if L-monosaccharides are included. The ring form of a monosaccharide generates a chiral anomeric center at C1 for aldo sugars or at C2 for keto sugars. There are two types of anomeric configurations, α and β, depending on the hydroxyl group is sticking down or up, respectively.

Ring formOpen-ring form Ring form

1 6 6 2 5 5 3 4 1 4 1 2 4 2 3 5 3 6 α-D-Glucose β-D-Glucose

Figure 1.1 Ring and open-ring forms of D-glucose.

Any sugar that has aldehyde group or be able to form one in solution through isomerism is called “reducing sugars” and can attach to another monosaccharide via the hydroxyl group of anomeric center. A sugar which two monosaccharides connected through the newly formed bond (“glycosidic bond”) is called a disaccharide. Due to the enormous number of combinatory ways to construct an oligomer, even a small subunit of carbohydrates, disaccharide, can possibly have more than 104 structural isomers

(estimated by Laine1 as shown in Figure 1.2). Unlike oligonucleotides and proteins which are expressed in a linear fashion, carbohydrates can form branched structures. In addition, the hydroxyl groups are subjected to various modifications (i.e. acetylamine, sulfate, or sialic acid groups). This structural complexity imposes challenges to any existing analytical methods for complete structural characterization. The heterogeneity of many carbohydrate samples also requires separation tools such as high performance liquid chromatography (HPLC) to be coupled with analysis methods including mass spectrometry (MS) and nuclear magnetic resonance (NMR). 3

1.E+12

1.E+10

1.E+08

1.E+06

1.E+04

# of isomers 1.E+02

1.E+00 012345 Degree of polymerization

Figure 1.2 The number of possible structural isomers as increasing degree of polymerization (considering only aldohexoses and linear structure).

1.2 Glycosylation

Glycosylation is one of the most frequent post-translational modifications and more than 50% of known proteins as well as 80% of membrane proteins are estimated to be modified with glycans.2,3 Proteins and lipids modified with glycans are called glycoproteins and glycolipids, respectively. Those glycan modifications are widely involved in intermolecular and intercellular binding events from fertility to immunity.2 In eukaryotic cells, there are two major classes of glycans according to the nature of the linkage regions to the proteins: N- and O-linked glycans. An N-linked glycan (N-glycan) is covalently linked to an asparagine residue of a peptide chain where the peptide sequence matches with Asn-X-Ser or Asn-X-Thr (X could be any amino acid except proline). N-glycans share a common core structure (Man3GlcNAc2) and can be divided into three classes: high-mannose type, complex type, and hybrid type (Figure 1.3a).4 An

O-linked glycan (O-glycan) is typically linked to serine or threonine residue via N- acetylgalactosamine (GalNAc). Unlike N-glycans, O-glycans have a variety of different 4 structural core classes. The most commonly occurring O-glycan structures are shown in

Figure 1.3b.4 Since the discovery of glycoproteins in bacteria and their pathogenicity relates to the glycan structure, bacterial glycoproteins are attractive targets for therapeutic intervention.5,6 However, prokaryotic glycans consist of unusual monosaccharide units and their analysis is much more challenging than already complex eukaryotic glycans.5,7

The development of sensitive analytical method for structural characterization which does not depend on the previous knowledge of biological glycan synthesis is highly desirable.

(a) N-glycans N-Acetylneuraminic Galactose (Gal) acid (Neu5Ac) N-Acetylgalactosamine Glucose (Glc) (GalNAc) N-Acetylglucosamine Mannose (Man) (GlcNAc)

(b) O-glycans

Complex type Hybrid type High-mannose type Core 1 Core 2 Core 3 Core 4

Figure 1.3 Common core structures for (a) N- and (b) O-glycans.

1.3 Overview of Glycan Structural Analysis

In order to understand the biological function of glycans, full-level of structural information is necessary. Full-level of structural characterization of a glycan include the following five levels: the identity (stereochemistry and modifications) of each monosaccharide unit, the anomeric configuration of the glycosidic bonds, linkage positions, branching location, and the sequence of the individual monosaccharides in the oligomer (Figure 1.4).

Identity

Anomeric Configuration Linkage

Branching location Sequence

Figure 1.4 Structural information necessary to fully characterize glycans.

NMR is the most powerful technique for structural analysis of glycans. The advantage of

NMR over other methods is that it offers the full-level of structural information in a non- destructive way.8,9 The main drawback of this technique is that it typically requires large amount of samples (μg to mg quantities) to characterize unknowns.9,10 In many cases, such as the analysis of N- and O-linked glycans from glycoproteins, there is rarely enough material available. The preferred techniques which are capable of smaller quantity of samples (pg to ng) are enzymatic analysis or lectin affinity chromatography 6 coupled with MS or HPLC. Enzymatic analysis of glycans uses exoglycosidases which are known to cleave bonds that are specific to the type of linkage, anomeric configuration, and stereochemistry of monosaccharide. Selection of exoglycosidase largely relies on the knowledge of biological synthesis. Sialidase (cleave sialic acid which has linkage of α2-3, -6, and -8), β-galactosidase (β1-4 galactose), α-fucosidase

(α1-2, -3, -4, -6 fucose), β-N-acetylhexosaminidase (β1-2, -3, -4, -6 GlcNAc), α- mannosidase (α1-2, -3, -6 mannose) are some examples of exoglycosidases.11

Susceptibility to cleavage by different kinds of exoglycosidases is assessed by analysis of the reaction products by MS or HPLC.12-14 This approach gives a lot of structural information if prior knowledge about the glycan is available (e.g. the glycan is an N- glycan).15,16 However, the variation of exoglycosidases which have been isolated so far is very limited to N-glycans. Especially most of exoglycosidases cleave multiple linkages, thus linkage information cannot be obtained for completely unknown molecule. Lectins are carbohydrate-binding proteins which recognize specific sugar moieties.17 Based on the relative affinity of glycans to different lectins, each glycan is eluted differently in lectin affinity chromatography and their structures can be identified.18,19 A variety of lectins has been discovered and a series of lectin columns as well as lectin array is available for glycan analysis. These methods allow high-throughput analysis. However, the detail glycan structure is difficult to obtain since lectins recognize whole sugar moieties instead of a single sugar unit resolution.

1.4 Structural Analysis of Glycans by Mass Spectrometry

Mass spectrometry was first applied to carbohydrate analysis in 195820 with electron impact (EI) ionization.21,22 EI is said to be “hard” ionization method since it induces extensive fragmentation and the molecular ions are typically not observed. Since

EI is only suitable for volatile organic molecules, hydroxyl groups in non-volatile carbohydrates had to be protected by permethylation,23 or peracetylation24 to increase the volatility prior to analysis.25 Since then, many researches for structural analysis of glycans using mass spectrometry have been reported with the improvement of analytical techniques, mass spectrometric instrumentation, as well as development of “soft” ionization methods such as fast atom bombardment (FAB),26 electrospray ionization

(ESI)27,28 and matrix-assisted laser desorption ionization (MALDI)29,30 which enable native carbohydrates to be ionized with molecular ion information to be obtained.

1.4.1 Ionization

FAB has been introduced in 198126 and was the first ionization method used in

MS experiment of carbohydrates. A mechanism of FAB is similar to EI, instead of impacting the sample by electrons, FAB uses beam of atoms and the sample has to be in the liquid matrix.26 The intact non-volatile compounds were ionized by FAB, especially acidic oligosaccharides (containing Neu5Ac or sulfate) in negative ion mode worked relatively well. However, the ion signal was not strong enough and the size of the molecular weight applicable was typically low.31 Currently, FAB has been replaced by later developed “soft” ionization methods: ESI and MALDI. 8

ESI and MALDI are two common methods for the analysis of carbohydrates.

Both ionization techniques were introduced in the late 1980’s and enabled non-volatile and thermally labile compounds to be ionized. In ESI, the sample solution is highly charged and sprayed through a capillary into a strong electric field to form a fine mist of charged droplets. The solvent evaporates and produces gas-phase ions or solvated ions. In order to be efficiently ionized, analytes are necessary to stay on the surface of the droplets. However, the hydrophilicity of carbohydrates limits their surface activity in ESI droplets, and thus ionization efficiency of carbohydrates is significantly lower than those of peptides and proteins.32,33 NanoESI,34 which uses much lower flow rates down to some tens of nL/min as compared to 1 to 10 μL/min for regular ESI, produces smaller charged droplet sizes and enhances the ionization efficiency for carbohydrates.35

MALDI uses a matrix which contains small organic molecules that have strong absorption at the laser wavelength for ionization. Before the analysis, analytes are dissolved in the matrix, which are subsequently dried to form crystals. Intense laser pulse ablates the surface of the dried crystal in vacuum. This ablation excites and sublimates the matrix molecule into gas phase.36 Although the exact mechanism of the MALDI process is still under debate,37,38 the widely accepted theory for ion formation involves proton transfer in the solid phase before desorption. Ions in the gas phase are then accelerated by the electric field towards the mass analyzer.38 Both ESI and MALDI ionization methods have advantages and disadvantages for carbohydrate analysis. The advantage of MALDI is that the ionization efficiency for neutral carbohydrates is relatively constant as the size of the molecule increases, in contrast to ESI, where the ionization efficiency decreases with an increasing molecular weight.39 The limitation of 9

MALDI comes from extensive fragmentation due to the higher internal energies deposited to the ions by laser ablation as compared to ESI.

1.4.2 Derivatization Techniques for Glycan Analysis

Derivatization of sugars, such as permethylation of hydroxyl groups23,40 and reducing end modification,41,42 reduces hydrophilicity of the sugar molecules and increases the ionization efficiency significantly.43,44 Permethylation is the most widely used modification for glycan analysis in mass spectrometry, especially after the introduction of the solid phase permethylation method developed by Novotny group.40,45,46 Solid phase permethylation allows simple and efficient permethylation especially for small quantity of samples. The aldehyde group at the reducing end of glycans can also react with alkylamines (typically aromatic amines, which also function as chromophore for UV detection upon separation). Two approaches (amination and reductive amination) of reducing end derivatization are summarized in Figure 1.5. Schiff base formation by amination produces open-ring and closed-ring structures in equilibrium. It has been confirmed by NMR and other techniques that closed-ring

(glycosylamine) structure is prevalent in solution with aromatic substitutions.47,48 The

Schiff base can be further reduced by sodium borohydride to form reduced Schiff base which can only exist in the open-ring structure. While reduced Schiff base is more stable than a glycosylamine, glycosylamine structure has been reported with analytical advantages such as better HPLC separation and enhancement of cross-ring cleavages by

CID.47,49 A number of amine groups have been used as reducing end derivatives containing cationic and anionic charges such as p-aminophenyl ammonium chloride 10

(TMAPA)50 and 5-aminosalicylic acid.33 TMAPA was shown to increase sensitivity

5000-fold under ESI condition relative to the native form oligosaccharides.50 These charge-containing derivatizations do not only offer improved ionization efficiency but also show positive effects on the fragmentation pattern of CID.51 Reducing end derivatization (including isotopic labeling, i.e. 18O) can also be used to simplify the complex fragmentation spectrum from CID by introducing mass differences of product ions derived from the reducing end vs. non-reducing end. 41,42

Native Glycan Labeled Glycan Non-Reduced (NR) form Reduced (R) form by amination by reductive amination

H2N-R2 (-H2O)

reduction

open-ring open-ring open-ring (Schiff base) (reduced Schiff base)

closed-ring closed-ring (hemiacetal) (glycosylamine)

Figure 1.5 Reaction scheme for amination and reductive amination.

1.4.3 Nomenclature for the Gas-phase Fragmentation of Oligosaccharides

The nomenclature for oligosaccharide fragmentation commonly used in the mass spectrometry field has been introduced by Domon and Costello as shown in Figure 1.6.52

Fragment ions that contain a non-reducing end are labeled with uppercase letters A, B, 11 and C, and those contain the reducing-end of the oligosaccharide or the aglycon are labeled with X, Y, and Z; subscripts indicate the location of cleavage within the oligosaccharide ion. B and Y ions are cleaved at the non-reducing side of a glycosidic oxygen and C and Z ions are cleaved at the reducing side of a glycosidic oxygen. B, C, Y, and Z ions are classified as “glycosidic bond cleavages.” A and X ions resulted from cleavages across the glycosidic ring are termed as “cross-ring cleavages.” They are labeled with superscript of two broken ring bond numbers that are assigned as shown in

Figure 1.6.

Y2 Z2 Y1 Z1 Y0 Z0 1,5 X0

5 5 4 4 Non-reducing end 0 0 Reducing end 3 3 1 1 2 2

0,2 A1 B C B C B C 1 1 2 2 3 3

Figure 1.6. Nomenclature for glycoconjugate product ions generated by tandem MS.52

1.4.4 Structural Analysis of Oligosaccharides by Mass Spectrometry

Mass spectrometry is a widely applied method in structural analysis of carbohydrates, due to its high sensitivity and capability of providing detailed molecular information.53 However, obtaining full-level of structural information (sugar identity, anomeric configuration, linkage position, sequence, and branching location) has not been 12 possible by using only MS. The molecular weight information of carbohydrates is readily obtained from soft ionization methods such as ESI33,54 and MALDI.29,39,55 Tandem mass spectrometry (MS/MS) based on collision-induced dissociation (CID) is heavily relied on to obtain structural information for carbohydrates. Glycosidic bond cleavages and/or cross-ring cleavages are typically observed from collisional activation. Glycosidic bonds are labile and can be easily cleaved under low-energy collisions. These fragments are useful for sequencing and identifying branching location. On the other hand, cross-ring cleavages need relatively higher-energy to induce and these fragments are useful for identifying linkage positions. Sequence and branching location can be readily obtained from MS2 CID of permethylated33,56-59 or peracetylated59,60 oligosaccharides in the positive ion mode and from native oligosaccharide in the negative ion mode.61

Assignment of linkage positions between disaccharides can be obtained based on the product ions generated by MS2 CID in negative ion mode49,62-66 and positive ion mode with metal cation adducts.60,67-73 Linkage determination by MS2 CID has been applied to oligosaccharides. As the molecular gets larger, it is more difficult for linkage determination due to possible ion suppression of the diagnostic product ions for structural determination.63,74 Stereochemistry of monosaccharides and anomeric configuration are the most difficult structural information to obtain by MS. Notable examples to obtain stereo-structure information include tandem mass spectrometry (MS2) of transition metal chelated75 and amino acid complexed monosaccharides.76

1.4.5 MS2 vs MSn

Tandem mass spectrometry (MS/MS) is the key technique for structural analysis by MS.77-81 MS2, the simplest MS/MS, has been routinely used for sequencing and identifying branching locations by a series of product ions from glycosidic bond cleavages. In order to identify linkage positions unambiguously, obtaining all of the possible product ions from cross-ring cleavages is the key. Besides CID, different activation/dissociation methods have been investigated, including infrared multiphoton dissociation (IRMPD),82 electron capture dissociation (ECD),83,84 electron-induced dissociation (EID),85 electron detachment dissociation (EDD),86 and electron excited dissociation (EED).87 These methods have been shown to increase cross-ring cleavages and are complementary to CID. However, the application of the above methods has been limited to research groups which have these instrument capabilities. Overall, MS2 has advantages of good sensitivity (pmol to fmol of sample quantiteis),88 high throughput, and simplicity in performance and instrumentation. Depending on the nature of analytes and the dissociation method used, key diagnostic ions (i.e. A type ions) may be missing in MS2, leading to either miss-assigned or un-assigned linkage positions or other structural information.63,74,89 Higher stages of tandem mass spectrometry (MSn, where n>2), enables a sequential disassembly of the intact molecule, which can be back-tracked based on the precursor-product relationships at each step. In general, MSn leads to higher confidence in structural characterization, and is extremely useful in probing subtle structure changes between isomers.57,90-92 Viseux et al. demonstrated the use of MS3 to

MS5 CID for structural characterization (sequence, linkage, and branching determination) of linear and branched oligosaccharides. In their work, an unknown oligosaccharide was 14 first fragmented into disaccharide and/or trisaccharide substructures via CID. These substructures were further subjected to CID and the fragmentation spectra were compared to that from known reference molecules for linkage determination. By using the substructures generated in MSn CID as opposed to the whole oligosaccharide, it significantly reduced the size of reference molecules needed for comparisons and it overcame the low mass cut off issue for detecting key ions with the use of electrodynamic ion trap mass spectrometer.58 Furthermore, determining the identity of low mass ions through comparison to a much more limited number of possible structural isomers provides greater confidence in their structural determination, particularly in discriminating between their unique stereochemical or regiochemical isomers.

1.4.6 A New MSn Approach for High Level Structural Analysis of Oligosaccharides

One of the grand challenges of oligosaccharide analysis is to develop a complete approach based on MS toward a full level structural analysis of linear oligosaccharides by obtaining information of sugar unit identity, anomeric configuration, linkage types and sequence. We propose a new MSn approach to achieve this goal by 4 steps: (1) find a diagnostic ion which gives any of the structural information listed above in the product ions from MS3 CID of disaccharide ions (shown in the blue square in Scheme 1.1), (2) make a standard spectral library by collecting MSn (n = 2 or 3) CID of the diagnostic ions from all of the possible isomeric structures which was either synthesized or fragmented in gas-phase from disaccharide ions, (3) find the optimized pathway to fragment down oligosaccharide ions into overlapping disaccharide substructures (shown in the red square in Scheme 1.1), and (4) employ spectral matching algorithm for easy and fair comparison. 15

I. Oligosaccharide Ion Æ Disaccharide Substructure Ions

Y3 Y2 MS1 4 O 3 OOM2 1 C2 MS2 MS2 MS2 HO 3 OOM2 1 MS3

DisaccharideDisac SubstrSubstructurecture FFormationn HO 4 OOHO 3 HO 3 OOOH2 HO 2 OM1

MS3 MS4 MS3

Diagnostic Ion MS4 MS5 MS4

Identity Anomeric Config. Location Linkage

II. Disaccharide Ion Æ Diagnostic Ion Æ Structural Information

Scheme 1.1 The MSn (n = 4 or 5) approach for detailed structural analysis of a linear oligosaccharide. Charges are not indicated on the structure. “M” stands for reducing end modification.

Sugar Identity Diagnostic Ion Anomericity Unknown Disaccharide Sugar-GA (m/z 221) 131131 - - -H -H MS2 CID MS3 CID 101 8787 113113 161161 221221 220303

Figure 1.7 MS3 CID steps for sugar unit identity and anomeric configuration determination.93 16

Finding the diagnostic ion which is a substructure of disaccharide ion is the key concept of this approach since the number of structural isomers can be minimized and significantly reduce the cost of library construction. Previous study by our collaborator,

Bendiak group, has demonstrated that the CID patterns of m/z 221 product ions (C8H13O7, non-reducing sugar glycosidically linked to a glycolaldehyde (sugar-GA)) from disaccharide anions can be used to determine the non-reducing sugar identity and anomeric configuration (Figure 1.7).93,94 Using this m/z 221 diagnostic ion successfully reduces the size of synthetic standards from 11520 possible structures (disaccharides) to

24 (monosaccharides) as shown in Figure 1.2. By following this concept, further discovery of diagnostic ion for linkage type is expected. Next step is the application of this structural analysis method to oligosaccharides. Oligosaccharides need to be broken down into overlapping disaccharide substructures. This can be done by MSn (n = 2 or 3) via fragmenting the oligosaccharide ions into a ladder of Yn-1 to Y2 ions, then, cleaving

n off disaccharide substructures (C2 ion) from non-reducing end of each of the Y ions. MS

(n>2) is typically performed using ion trap (Paul trap, linear ion trap, and Fourier transform ion cyclotron resonance (FTICR)) in a tandem-in-time fashion. In practice, the number of stages (n) of MS/MS that can be executed on a certain instrument is limited by the number of ions remained after each step of ion activation and isolation.95 The loss of the low abundance fragment ion could be significant for ion traps due to the space charge effect.96,97 Due to the above reason, a systematic study of fragmentation chemistry of oligosaccharides and the way to enhance the formation of desired ions are important as well to accomplish multiple stages of MSn (n = 4 or 5) experiment.

Previously, Bendiak group successfully demonstrated the stereo-structure analysis by the m/z 221 diagnostic ion for 1-2, 1-4, and 1-6 linked disaccharides. However, 1-3 linked disaccharides produced very few m/z 221 ions using ion trap CID, making it difficult to perform MS3 analysis on the stand-alone ion trap instrument. We investigated the unimolecular dissociation chemistry of 1-3 linked disaccharide anions under different

CID conditions to understand fundamental gas-phase ion chemistry for the formation of m/z 221 diagnostic ion in Chapter 2. In Chapter 3, the MSn approach is discussed to generate m/z 221 within linear oligosaccharides to obtain the stereo-structure information of each sugar unit. Furthermore, we explored and found the new diagnostic ion for linkage determination which is smaller than a monosaccharide, Z1 ion, and this Z1 ion was used to obtain linkage information in linear and branched oligosaccharides in

Chapter 4. Finally, a variety of reducing end modification groups and their significance on the fragmentation chemistry of oligosaccharides were studied in Chapter 5 in order to enhance selective bond cleavages, which is critical in improving the sensitivity of the overall MSn analysis.

1.5 Conclusions

MS has been extensively used in glycan analysis both independently and coupled with other analytical techniques. Due to its high sensitivity and simplicity, it is highly expected that the full-level characterization can be obtained only using MS. The current limitation of MS method is that linkage position cannot be obtained from oligosaccharides unambiguously under most of circumstances, as well as obtaining stereo-structure (stereochemistry and anomeric configuration) information. As opposed to the trend in this field using MS2, we focused on the development of MSn approach which overcomes the current weakness of MS method on glycan analysis, obtaining unambiguous information of stereochemistry, anomeric configuration, and linkage position from oligosaccharides.

1.6 References

(1) Laine, R. A., Glycobiology 1994, 4, 759-767.

(2) Perkel, J. M., Science 2011, 331, 95-97.

(3) Apweiler, R.; Hermjakob, H.; Sharon, N., Biochim. Biophys. Acta 1999, 1, 4-8.

(4) Varki, A.; Cummings, R. D.; Esko, J. D.; Freeze, H. H.; Stanley, P.; Bertozzi, C. R.; Hart, G. W.; Etzler, M. E., Essentials of glycobiology. Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 2009.

(5) Dube, D. H.; Champasa, K.; Wang, B., Chem. Commun. 2011, 47, 87-101.

(6) Balonova, L.; Hernychova, L.; Bilkova, Z., Expert Rev. Proteomics 2009, 6, 75-85.

(7) Maki, M.; Renkonen, R., Glycobiology 2004, 14, R1-R15.

(8) Vliegenthart, J. F. G.; Dorland, L.; Halbeek, H. v., High-Resolution, 1H-Nuclear Magnetic Resonance Spectroscopy as a Tool in the Structural Analysis of Carbohydrates Related to Glycoproteins. In Advances in Carbohydrate Chemistry and Biochemistry, Tipson, R. S.; Derek, H., Eds. Academic Press: 1983; Vol. 41, pp 209-374.

(9) Duus, J. Ø.; Gotfredsen, C. H.; Bock, K., Chem. Rev. 2000, 100, 4589-4614.

(10) Allen, S.; Richardson, J. M.; Mehlert, A.; Ferguson, M. A. J., J. Biol. Chem. 2013, 288, 11093-11105.

(11) Hoja-Łukowicz, D.; Link-Lenczowski, P.; Carpentieri, A.; Amoresano, A.; Pocheć, E.; Artemenko, K.; Bergquist, J.; Lityńska, A., Glycoconj J 2013, 30, 205-225.

(12) Guile, G. R.; Rudd, P. M.; Wing, D. R.; Prime, S. B.; Dwek, R. A., Anal. Biochem. 1996, 233, 205-211.

(13) Rudd, P. M.; Gulle, G. R.; Kuster, B.; Harvey, D. J.; Opdenakker, G.; Dwek, R. A., Nature 1997, 388, 205-207.

(14) Edge, C. J.; Rademacher, T. W.; Wormald, M. R.; Parekh, R. B.; Butters, T. D.; Wing, D. R.; Dwek, R. A., Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 6338-6342. 20

(15) Mechref, Y.; Novotny, M. V., Anal. Chem. 1998, 70, 455-463.

(16) Geyer, H.; Schmitt, S.; Wuhrer, M.; Geyer, R., Anal. Chem. 1998, 71, 476-482.

(17) Sharon, N.; Lis, H., Science 1989, 246, 227-234.

(18) Cummings, R. D., Methods Enzymol. 1994, 230, 66-86.

(19) Endo, T., J. Chromatogr. A 1996, 720, 251-261.

(20) Finan, P. A.; Reed, R. I.; Snedden, W., Chem. Ind. (London) 1958, 36, 1172.

(21) Bleakney, W., Phys. Rev. 1929, 34, 157-160.

(22) Nier, A. O., Rev. Sci. Instrum. 1947, 18, 398-411.

(23) Ciucanu, I.; Kerek, F., Carbohydr. Res. 1984, 131, 209-217.

(24) Bourne, E. J.; Stacey, M.; Tatlow, J. C.; Tedder, J. M., J. Chem. Soc. 1949, 2976- 2979.

(25) Finan, P. A.; Reed, R. I., Nature 1959, 184, 1866.

(26) Barber, M.; Bordoli, R. S.; Sedgwick, R. D.; Tyler, A. N., Nature 1981, 293, 270- 275.

(27) Meng, C. K.; Mann, M.; Fenn, J. B., Z. Phys. D. 1988, 10, 361-368.

(28) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F., Mass Spectrom. Rev. 1990, 9, 37- 70.

(29) Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F., Int. J. Mass Spectrom. 1987, 78, 53-68.

(30) Karas, M.; Hillenkamp, F., Anal. Chem. 1988, 60, 2229-2231.

(31) Egge, H.; Peter-Katalinić, J., Mass Spectrom. Rev. 1987, 6, 331-393.

(32) Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J., Anal. Chem. 1994, 66, 634R-683R.

(33) Reinhold, V. N.; Reinhold, B. B.; Costello, C. E., Anal. Chem. 1995, 67, 1772-1784.

(34) Wilm, M. S.; Mann, M., Int. J. Mass Spectrom. Ion Processes 1994, 136, 167-180.

(35) Bahr, U.; Pfenninger, A.; karas, M.; Stahl, B., Anal. Chem. 1997, 69, 4530-4535.

(36) Dreisewerd, K., Chem. Rev. 2003, 103, 395-425.

(37) Zenobi, R.; Knochenmuss, R., Mass Spectrom. Rev. 1998, 17, 337-366.

(38) Knochenmuss, R.; Zenobi, R., Chem. Rev. 2003, 103, 441-452.

(39) Harvey, D. J., Rapid Commun. Mass Spectrom. 1993, 7, 614-619.

(40) Mechref, Y.; Kang, P.; Novotny, M. V., Methods in Mol. Biol. 2009, 534, 53-64.

(41) Harvey, D. J., J. Am. Soc. Mass Spectrom. 2000, 11, 900-915.

(42) Harvey, D. J., J. Chromatogr. B 2011, 879, 1196-1225.

(43) Karas, M.; Bahr, U.; Dulcks, T., Fresenius J. Anal. Chem. 2000, 366, 669-676.

(44) Dell, A., Methods Enzymol. 1990, 193, 647-660.

(45) Kang, P.; Mechref, Y.; Klouckova, I.; Novotny, M. V., Rapid Commun. Mass Spectrom. 2005, 19, 3421-3428.

(46) Kang, P.; Mechref, Y.; Novotny, M. V., Rapid Commun. Mass Spectrom. 2008, 22, 721-734.

(47) Her, G. R.; Santikarn, S.; Reinhold, V. N.; Williams, J. C., 1987, 6, 129-139.

(48) Nishikaze, T.; Kaneshiro, K.; Kawabata, S.; Tanaka, K., Anal. Chem. 2012, 84, 9453-9461.

(49) Li, D. T.; Her, G. R., Anal. Biochem. 1993, 211, 250-257. 22

(50) Hansson, G. C.; Karlsson, H., Methods Molec. Biol. Humana Press: Totowa, 1993.

(51) Chen, S.-T.; Her, G.-R., J. Am. Soc. Mass Spectrom. 2012, 23, 1408-1418.

(52) Domon, B.; Costello, C. E., Glycoconjugate J. 1988, 5, 397-409.

(53) Zaia, J., Mass Spectrom. Rev. 2004, 23, 161-227.

(54) Fenn, J.; Mann, M.; Meng, C.; Wong, S.; Whitehouse, C., Science 1989, 246, 64-71.

(55) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T.; Matsuo, T., Rapid Commun. Mass Spectrom. 1988, 2, 151-153.

(56) Reinhold, V. N.; Sheeley, D. M., Anal. Biochem. 1998, 259, 28-33.

(57) Sheeley, D. M.; Reinhold, V. N., Anal. Chem. 1998, 70, 3053-3059.

(58) Viseux, N.; de Hoffmann, E.; Domon, B., Anal. Chem. 1998, 70, 4951-4959.

(59) Perreault, H.; Costello, C. E., J. Mass Spectrom. 1999, 34, 184-197.

(60) Domon, B.; Müller, D. R.; Richter, W. J., Biol. Mass Spectrom. 1990, 19, 390-392.

(61) Chai, W.; Lawson, A.; Piskarev, V., J. Am. Soc. Mass Spectrom. 2002, 13, 670-679.

(62) Ballistreri, A.; Montaudo, G.; Garozzo, D.; Giuffrida, M.; Impallomeni, G.; Daolio, S., Rapid Commun. Mass Spectrom. 1989, 3, 302-304.

(63) Garozzo, D.; Giuffrida, M.; Impallomeni, G.; Ballistreri, A.; Montaudo, G., Anal. Chem. 1990, 62, 279-286.

(64) Lamb, D. J.; Wang, H. M.; Mallis, L. M.; Linhardt, R. J., J. Am. Soc. Mass Spectrom. 1992, 3, 797-803.

(65) Carroll, J. A.; Willard, D.; Lebrilla, C. B., Anal. Chim. Acta 1995, 307, 431-447.

(66) Mulroney, B.; Traeger, J. C.; Stone, B. A., J. Mass Spectrom. 1995, 30, 1277-1283.

(67) Domon, B.; Müller, D. R.; Richter, W. J., Org. Mass Spectrom. 1989, 24, 357-359.

(68) Domon, B.; Muller, D. R.; Richter, W. J., 1990, 100, 301-311.

(69) Zhou, Z.; Ogden, S.; Leary, J. A., J. Org. Chem. 1990, 55, 5444-5446.

(70) Hofmeister, G. E.; Zhou, Z.; Leary, J. A., J. Am. Chem. Soc. 1991, 113, 5964-5970.

(71) Lemoine, J.; Fournet, B.; Despeyroux, D.; Jennings, K. R.; Rosenberg, R.; de Hoffmann, E., J. Am. Soc. Mass Spectrom. 1993, 4, 197-203.

(72) König, S.; Leary, J., J. Am. Soc. Mass Spectrom. 1998, 9, 1125-1134.

(73) Xue, J.; Song, L.; Khaja, S. D.; Locke, R. D.; West, C. M.; Laine, R. A.; Matta, K. L., Rapid Commun. Mass Spectrom. 2004, 18, 1947-1955.

(74) Dallinga, J. W.; Heerma, W., Biol. Mass Spectrom. 1991, 20, 215-231.

(75) Gaucher, S. P.; Leary, J. A., Anal. Chem. 1998, 70, 3009-3014.

(76) Augusti, D. V.; Carazza, F.; Augusti, R.; Tao, W. A.; Cooks, R. G., Anal. Chem. 2002, 74, 3458-3462.

(77) Cooks, R. G.; Beynon, J. H.; Caprioli, R. M., Metastable Ions. Elsevier: Amsterdam, 1973.

(78) Burinsky, D. J.; Cooks, R. G.; Chess, E. K.; Gross, M. L., Anal. Chem. 1982, 54, 295-299.

(79) McLafferty, F. W., Tandem Mass Spectrometry. John Wiley: New York, 1983.

(80) Busch, K. L.; Glish, G. L.; McLuckey, S. A., Mass Spectrometry / Mass Spectrometry: Techniques and Applications in Tandem Mass Spectrometry. VCH: New York, 1988.

(81) Burinsky, D. J.; Cooks, R. G.; Chess, E. K.; Gross, M. L., Anal. Chem. 1982, 54, 295-299.

(82) Xie, Y.; Lebrilla, C. B., Anal. Chem. 2003, 75, 1590-1598. 24

(83) Adamson, J. T.; Håkansson, K., Anal. Chem. 2007, 79, 2901-2910.

(84) Zhao, C.; Xie, B.; Chan, S.-Y.; Costello, C.; O’Connor, P., J. Am. Soc. Mass Spectrom. 2008, 19, 138-150.

(85) Wolff, J.; Laremore, T.; Aslam, H.; Linhardt, R.; Amster, I. J., J. Am. Soc. Mass Spectrom. 2008, 19, 1449-1458.

(86) Wolff, J.; Amster, I. J.; Chi, L.; Linhardt, R., J. Am. Soc. Mass Spectrom. 2007, 18, 234-244.

(87) Yu, X.; Huang, Y.; Lin, C.; Costello, C. E., Anal. Chem. 2012, 84, 7487-7494.

(88) Geyer, H.; Geyer, R., Biochim. Biophys. Acta 2006, 1764, 1853-1869.

(89) Dongre, A. R.; Wysocki, V. H., Org. Mass Spectrom. 1994, 29, 700-702.

(90) Li, B.; An, H.; Hedrick, J.; Lebrilla, C., Methods Mol. Biol. 2009, 534, 133-145.

(91) Ashline, D. J.; Lapadula, A. J.; Liu, Y.-H.; Lin, M.; Grace, M.; Pramanik, B.; Reinhold, V. N., Anal. Chem. 2007, 79, 3830-3842.

(92) Weiskopf, A. S.; Vouros, P.; Harvey, D. J., Anal. Chem. 1998, 70, 4441-4447.

(93) Fang, T. T.; Bendiak, B., J. Am. Chem. Soc. 2007, 129, 9721-9736.

(94) Fang, T. T.; Zirrolli, J.; Bendiak, B., Carbohydr. Res. 2007, 342, 217-235.

(95) McLuckey, S. A.; Glish, G. L.; Vanberkel, G. J., Int. J. Mass Spectrom., Ion Processes 1991, 106, 213-235.

(96) Cox, K. A.; Cleven, C. D.; Cooks, R. G., Int. J. Mass Spectrom., Ion Processes 1995, 144, 47-65.

(97) March, R. E., J. Mass Spectrom. 1997, 32, 351-369. 25

CHAPTER 2 DIFFERENTIATION OF THE STEREOCHEMISTRY AND ANOMERIC CONFIGURATION FOR 1-3 LINKED DISACCHARIDES

2.1 Introduction

A tandem mass spectrometry approach has been developed to differentiate the stereochemistry and anomeric configuration for the non-reducing unit of hexose- containing disaccharides having any of the 16 possible stereochemical variants.1,2 In this method, diagnostic ions at m/z 221 were formed from CID of deprotonated disaccharide ions (m/z 341). It was established that the m/z 221 ions consisted of the intact non- reducing sugar glycosidically linked to glycolaldehyde, as indicated in Scheme 2.1

(where GA abbreviates glycolaldehyde). Note that an open-chain form for the reducing sugar is indicated in Scheme 2.1 and also for other disaccharides discussed later. This is based on the observation of absorbance in the carbonyl stretch region in variable wavelength infrared radiation photo-dissociation of deprotonated monosaccharide anions in the gas phase.3 When m/z 221 ions were further dissociated by collisional activation, disaccharides having different non-reducing sugar units and anomeric configurations showed distinct fragmentation patterns that matched synthetic glycosyl-GAs. This method was shown to be useful for assigning the stereochemistry as well as the anomeric configuration of the glycosidic bond for the non-reducing sugar in disaccharides having

1-2, 1-4, and 1-6 linkages.1,2 However, due to the low abundance of m/z 221 ions 26 produced from 1-3 linked disaccharides, MS3 CID of m/z 221 ions could not be performed, and it was unclear whether the fragmentation patterns could be used for assigning either their stereochemistry or anomeric configuration.

-H -

-H - CID CID

α-D-Glcp-(1-4)-Glc, m/z 341 α-D-Glcp-GA, m/z 221

Scheme 2.1 Formation of glycosyl-GA anions at m/z 221 from CID of deprotonated 1-4 linked disaccharides.

In this chapter, a series of 1-3 linked disaccharides were studied on a hybrid triple quadrupole-linear ion trap mass spectrometer (QTRAP 4000). MS3 CID data of m/z 221 ions from the 1-3 linked disaccharides were obtained for the first time. The formation of m/z 221 ions was examined using different collisional activation methods, i.e. beam-type

CID and on-resonance ion trap CID of the deprotonated disaccharides. 18O-labeling of the reducing carbonyl oxygen in 1-3 linked disaccharides was used to enable mass- discrimination of structural isomers of the (usually) m/z 221 ions. By choosing the proper

CID conditions, the diagnostic m/z 221 ions (the glycosyl-GAs) could be formed as the dominant isomer. Their CID fragmentation patterns could be used to establish the stereochemistry and anomeric configuration of the non-reducing sugar unit from 1-3 linked disaccharides. 27

2.2 Experimental

2.2.1 Materials

A list of 7 disaccharides (6 reducing sugars and 1 non-reducing sugars) is shown in Table 2.1. All samples were purchased from commercial sources (indicated by the

18 superscripts in Table 2.1) and used without further purification. H2 O was purchased from Sigma-Aldrich, Inc. (St. Louis, MO). α- and β-D-monosaccharide-glycolaldehyde standards, glucopyranosyl-glycolaldehydes (Glcp-GA), galactopyranosyl- glycolaldehydes (Galp-GA), and mannopyranosyl-glycolaldehydes (Manp-GA) were synthesized as previously described.1 Disaccharides and synthetic standards were dissolved in methanol to a final concentration of 0.01 mg/mL and NH4OH was added to a final concentration of 1 % immediately before use.

Table 2.1 List of disaccharides being studied.

Type Linkage 1-2 α-D-Glcp-D-Glca β-D-Glcp-D-Glca α-D-Glcp-D-Glca β-D-Glcp-D-Glcb Disaccharides 1-3 α-D-Glcp-D-Frua α-D-Galp-D-Galb α-D-Manp-D-Manb

Analytes were purchased from: aSigma-Aldrich, Inc. (St. Louis, MO, USA) and bCarbosynth, Ltd. (Berkshire, UK).

2.2.2 18O-labeling of Reducing Disaccharides

Carbonyl oxygen of the reducing sugar was 18O-labeled by dissolving 0.5 mg of a

18 disaccharide or oligosaccharide in 100 μL of H2 O and heated up to 60 °C for 6 - 24 h.

The progress of the reaction was monitored by MS. Once the reaction completed; the above solution was further diluted to 0.01 mg/mL (30 μM) with methanol. 1% (volume) of triethylamine was added to the diluted solution right before MSn analysis.

2.2.3 Mass Spectrometry

All samples were analyzed in the negative-ion mode on a QTRAP 4000 mass spectrometer (Applied Biosystems/Sciex, Toronto, Canada) equipped with a home-built nanoelectrospray ionization (nanoESI) source. A schematic diagram of the instrument ion optics is shown in the Scheme 2.2. Two types of low energy collisional activation methods were accessible on this instrument, i.e. beam-type CID and ion trap CID. In beam-type CID, the precursor ions (m/z 341 or 343) were isolated in Q1, accelerated in the Q2 collision cell for collisional activation, and all products were analyzed in the Q3 linear ion trap. Collision energy (CE) was defined by the potential difference (absolute value) between Q0 and Q2. In ion trap CID, the precursor ions were isolated in the Q3 linear ion trap via the RF/DC mode and a dipolar excitation was used for collisional activation. In order to perform ion trap CID at different Mathieu q-parameters, an AC

(alternating current) generated from an external waveform generator (Agilent

Technologies, Santa Clara, CA, USA) was used for resonance excitation. Frequency and the low mass cut-off were calculated by SxStability (Pan Galactic Scientific, Omemee,

Ontario, Canada). MS3 CID experiments were carried out by first performing beam-type 29

CID of precursor ions in Q2. The fragment ions of interest were isolated in Q3 and then subjected to ion trap CID. Analyst 1.5 software was used for instrument control, data acquisition, and processing. The typical parameters of the mass spectrometer used in this study were set as follows: spray voltage, -1.1 to -1.5 kV; curtain gas, 10; declustering potential, 50 V; beam-type CID collision energy (CE), 5 to 30 V; ion trap CID activation energy (AF2), 5 to 60 (arbitrary units); scan rate, 1000 m/z/s; pressure in Q2, 6.9x10-3

Torr, and in Q3, 3.6x10-5 Torr. Ion injection time was controlled to keep a similar parent ion intensity: typically 3x106 counts per second (cps) for MS2 CID experiments and

1x106 cps for MS3 CID experiments. Activation time was kept constant at 200 ms for all ion trap CID experiments. Seven spectra were collected for CID of m/z 221 ions from synthesized monosaccharide-GA standards (deprotonated molecules) and disaccharides over a one year period. Standard deviations of peak heights were calculated for major fragments such as m/z 87, 99, 101, 113, 129, 131, 159, 161, 203, and 221, which were observed from all the standards and disaccharides studied here except β-D-Glcp-GA and

β-D-Glcp-(1-2)-D-Glc, which showed no peaks at m/z 99.

Dipolar AC Q2 Q0 Q1 Q3 nanoESI 5 mTorr 2.5x10-5 Torr

Beam-type Ion trap CID CID Scheme 2.2 A schematic diagram of the QTRAP 4000, a hybrid triple-quadrupole/linear ion trap mass spectrometer.

2.3 Results and Discussion

2.3.1 CID of Deprotonated Disaccharide Ions from 1-3 Linked Disaccharides

Ion trap CID of deprotonated 1-3 linked disaccharides (m/z 341) typically generates ions at m/z 221 in trace abundance on a Paul trap instrument, and isolation or further CID of m/z 221 ions have not been achieved before.2,4-6 The 4000QTRAP mass spectrometer used in this study has a unique triple quadrupole-linear ion trap configuration, offering high sensitivity due to the large capacity of the linear ion trap, and allowing either beam-type or ion trap collisional activation. In beam-type CID, the precursor ions were isolated in Q1 and accelerated in Q2 for collisional activation, while ion trap CID was conducted in Q3 with a dipolar excitation for collisional activation.

Since CID fragmentation patterns can be sensitive to the means of activation, the formation of m/z 221 ions from five 1-3 linked disaccharides was investigated via both beam-type and ion trap CID. Figure 2.1 compares the MS2 beam-type and ion trap CID of deprotonated β-D-Glcp-(1-3)-D-Glc (m/z 341) using low energy CID conditions. A relatively low CE (6 V) was used for beam-type CID; in ion trap CID, the AF2 for an AC dipolar excitation was set to 25 (arbitrary units) for 200 ms. Under either activation condition, the absolute intensities of m/z 221 ions (indicated by an arrow in Figure 2.1) were very low and their relative intensities were less than 1% (normalized to the base peak in the spectrum). This phenomenon was generally observed for all 1-3 linked disaccharides studied herein. The insets in Figure 2.1 demonstrate the isolated m/z 221 ions (with a 2 m/z isolation window) from each set of dissociation conditions. For beam- type CID, 1x106 cps of m/z 221 ions could be accumulated with an injection time of 1 s, 31 which was sufficient for performing the next stage of tandem mass spectrometry (MS3 in this case) with reasonable ion statistics and sensitivity. Far lower abundance of the m/z

221 ions (4.6x104 cps) could be isolated from ion trap CID of m/z 341, even after doubling the injection time to 2 s. As a result, it was not feasible to obtain MS3 CID for m/z 221 ions generated from m/z 341 precursor ions initially isolated within the trap. In experiments described below, beam-type CID was used to dissociate disaccharide precursor anions in the Q2 collision cell thereby generating m/z 221 product ions in high enough abundance to acquire their spectra in the linear trap reproducibly.

341 100 221 (a) 1.05e6

Intensity, cps Intensity, 222

161179 Relative Intensity, % Intensity, Relative 113 143

060 100 140 180 220 260 300 340 m/z 179 100 (b) 161 4.6e4 221

341 Intensity, cps Intensity,

113 Relative Intensity,Relative % 143

060 100 140 180 220 260 300 340 m/z

Figure 2.1 MS2 CID spectra in the negative ion mode obtained from deprotonated β-D- Glcp-(1-3)-D-Glc (m/z 341) under low energy dissociation conditions: (a) beam-type CID (CE = 6 V), and (b) ion trap CID (AF2 = 25). Insets in (a) and (b) show the isolation of m/z 221 ions generated from beam-type CID (injection time = 1 s) and ion trap CID (injection time = 2 s), respectively. 32

2.3.2 CID of m/z 221 Ions Generated from 1-3 Linked Disaccharides

Previous studies have demonstrated that m/z 221 product ions formed from collisional activation of disaccharide anions typically consist of an intact non-reducing sugar with a 2-carbon aglycon derived from the reducing sugar.2 Three dominant fragment peaks are commonly observed from CID of m/z 221: m/z 101, 131 and 161.

The relative intensities of these peaks together with some other fragment ions can be used to establish the fragmentation patterns and to distinguish the stereochemistry and anomeric configuration of the non-reducing sugar. Given that the CID patterns of m/z

221 ions will be used for structural identification, spectral reproducibility is an important issue. Similar to the findings from a Paul trap instrument,1,2 we noticed that the number of ions (m/z 221) in the linear ion trap and the energy input into an ion were among the most important parameters affecting spectral reproducibility. To ensure reasonable ion statistics and avoid adverse space charge effects, the intensity of the m/z 221 ions was kept at 1x106 cps before MS3 CID. Based on previous studies, the CID energies were tuned so that the ratio of remaining precursor ion to the most abundant product ion was kept around 18 ± 3%.1 Figures 2.2, a, b, e, and f were the averaged spectra from seven repetitions collected over a one year period and they were further used to make spectral comparisons in later discussion. Error bars in the spectra indicate the standard deviation of the peak intensity for 10 major fragment ions which were frequently observed for all the disaccharides studied herein (m/z 87, 99, 101, 113, 129, 131, 159, 161, 203, and 221).

The standard deviations for these peaks were less than 5% in most cases, indicating high reproducibility of the spectra from day to day by controlling the ion counts in the trap before CID and the energy input to the ions. Since the CID patterns upon dissociation of 33 m/z 221 ions can differ to some extent from instrumentto instrument7, CID spectra of the synthetic monosaccharide-GA were collected as standards for comparisons. Figures 2.2, a and b show the CID data of D- and β-D-Glcp-GA, respectively. The abundant peaks at m/z 101 and 131 in Figure 2.2a are a signature of a non-reducing glucose with an D anomeric configuration. Note that a distinct fragmentation pattern is observed for the β configuration (Figure 2.2b), where m/z 131 and 161 ions are dominant. The same collisional activation conditions were applied to m/z 221 ions derived from α-D-Glcp-(1-

3)-D-Glc and β-D-Glcp-(1-3)-D-Glc, anomeric isomers containing a non-reducing glucose. It is obvious that the spectra from the two anomeric isomers (Figures 2.2, c and d) were drastically different from their corresponding D-Glcp-GA standards, however, were similar to each other. This indicates that the m/z 221 ions generated using low collision energies from m/z 341 precursors have different structures from the D-Glcp-GA standards, and that their CID patterns cannot be used to assign either the stereochemistry or anomeric configurations of the ions. Note that beam-type CID was used to generate the m/z 221 ions from disaccharides shown in Figure 2.2, a condition differing from previous studies where ion trap CID had been used.1 This difference in activation could have contributed to the formation of structural isomers observed for the m/z 221 product ions. In order to test this hypothesis, m/z 221 ions of 1-2 linked disaccharides, α-D-Glcp-

(1-2)-D-Glc and β-D-Glcp-(1-2)-D-Glc, were formed using similar beam-type CID conditions and further subjected to MS3 CID (Figures 2.2, e and f). Except for a larger fluctuation in peak intensity for m/z 203, almost identical fragmentation patterns to the standards were observed (compare Figure 2.2, a to e and b to f), strongly indicating that the expected D-Glcp-GA structures were formed. We further investigated a wide variety 34 of disaccharides and found that the CID patterns of m/z 221 ions matched with their corresponding monosaccharide-GA standards with the exception of 1-3 linked disaccharides when low collision energy beam-type CID conditions were used to dissociate the disaccharides.

α-D-Glcp-(1-3)-Glc, m/z 341 α-D-Glcp-(1-2)-Glc, m/z 341 α-D-Glcp-GA, m/z 221 -H - -H - -H - m/z 221 m/z 221

131 161 131 100 100 100 (a) (c) (e)

101 203 101 99 99 113 131 113 113 87 221 87 129 159161 221 129 159161 221 Relative Intensity, % Intensity, Relative 203 % Intensity, Relative Relative Intensity, % Intensity, Relative 203 0 0 60 100 140 180 220 060 100 140 180 220 60 100 140 180 220 m/z m/z m/z β-D-Glcp-(1-3)-Glc, m/z 341 β-D-Glcp-(1-2)-Glc, m/z 341 β-D-Glcp-GA, m/z 221 -H - -H - -H - m/z 221 m/z 221

131 131 100 161 100 (b) 100 (d) (f)

161 161 113 131 203

221 221 221 129 87 129 Relative Intensity, % Intensity, Relative

Relative Intensity, % Intensity, Relative 87 Relative Intensity, % Intensity, Relative 101 113 159 203 101113 159 0 0 0 60 100 140 180 220 60 100 140 180 220 60 100 140 180 220 m/z m/z m/z

Figure 2.2 MS2 ion trap CID spectra of m/z 221 ions derived from synthetic standards (a) α-D-Glcp-GA, AF2 = 25 and (b) β-D-Glcp-GA, AF2 = 18. MS3 CID spectra of m/z 221 ions derived from low energy beam-type CID of glucose-containing disaccharides: (c) α- D-Glcp-(1-3)-Glc, CE = 6 V for MS2 and AF2 = 25 for MS3, (d) β-D-Glcp-(1-3)-Glc, CE = 6 V for MS2 and AF2 = 25 for MS3, (e) α-D-Glcp-(1-2)-Glc, CE = 5 V for MS2 and AF2 = 27 for MS3, and (f) β-D-Glcp-(1-2)-Glc, CE = 5 V for MS2 and AF2 = 25 for MS3. The error bars in the spectra show the standard deviation of the peak intensity based on seven spectra collected over a 1 y period.

18O-labeling at the carbonyl position of the reducing sugar was used to mass- discriminate the “sidedness” of dissociation events to either side of the glycosidic linkage and thus the origins of the m/z 221 and/or potential 223 product ions. Figure 2.3 compares relatively low energy beam-type and ion trap CID of 18O-labeled deprotonated

α-D-Glcp-(1-3)-D-Glc, m/z 343. Similar fragments were observed for both conditions; however, the ion abundance for m/z 283 (loss of 60 Da, C2H4O2) was much higher in beam-type CID relative to ion trap CID. It is possible that this fragmentation channel requires higher activation energy and is promoted, even in lower-energy beam-type CID, since higher collision energies (several eV) may have been obtained as compared to ion trap CID (hundreds of meV). The inset in Figure 2.3a shows data collected using a wide isolation window (6 m/z units) around m/z 221 after CID of m/z 343 precursor ions. A peak at m/z 223, due to incorporation of 18O, appeared with much higher abundance than m/z 221 ions for both low-energy beam-type and ion trap CID. Note that if the expected

D-Glcp-GA structure were formed, it should consist of the intact reducing sugar unit with a 2-carbon aglycon derived from the reducing sugar (C-2 and C-3 or C-3 and C-4). In this case, 18O should not be incorporated into the product ion and it should still appear at m/z 221. Therefore, the observation of abundant m/z 223 ions indicated that under relatively low energy CID conditions, most of the m/z 221 ions formed from α-D-Glcp-

(1-3)-D-Glc do not have the D-Glcp-GA structure which is the structural isomer needed to distinguish the stereochemistry and anomeric configuration of the non-reducing sugar.

Ions at m/z 221 and m/z 223 were further subjected to ion trap CID. Figure 2.4 compares the CID spectra of the isolated m/z 221 ions and the m/z 223 ions generated from 18O-labeled α-D-Glcp-(1-3)-D-Glc and β-D-Glcp-(1-3)-D-Glc. The CID spectra of 36 the m/z 221 ions (Figures 2.4, a and c) from the two anomeric isomers are distinct from each other and almost identical to those of the corresponding α and β-D-Glcp-GA standards (Figures 2.4, a and b). The fragmentation patterns of m/z 223 (Figures 2.4, b and d), however, were similar to each other yet were very different from the synthetic glucosyl-GA standards.

343 100 223 (a) 2.9e5

221 Intensity, cps Intensity, 0 283

Relative % Intensity, 163 179

060 100 140 180 220 260 300 340 m/z 343 100 223 (b) 4.0e4

221 Intensity, cps Intensity, 0 179 Relative Intensity, % Intensity, Relative 163

060 100 140 180 220 260 300 340 m/z

Figure 2.3 MS2 spectra of 18O-labeled α-D-Glcp-(1-3)-D-Glc under (a) relatively low- energy beam-type CID (CE = 6 V), and (b) ion trap CID (AF2 = 25). Insets in (a) and (b) show isolation of m/z 221 and 223 ions generated from beam-type CID (injection time = 1 s) and ion trap CID (injection time = 2 s), respectively.

The major fragments that resulted from CID of the m/z 223 ions included product ions at m/z 205, 163, 131, and 113. These ions are likely due to losses of water (-18 Da), a 2-

18 18 carbon piece, C2H4O2 (-60 Da), a 3-carbon piece including O, C3H6O2 + O (-92 Da), and sequential or concerted losses of a 3-carbon piece plus water including 18O (-110

Da), respectively. Interestingly, neither the loss of water nor the loss of 60 Da significantly involves loss of the 18O oxygen. The m/z 223 ions are hypothesized to have a structure in which the reducing sugar is connected to a 2-carbon piece from the non- reducing sugar as shown in the scheme above Figures 2.4b and 2.4d. Note that C-1 is no longer chiral on the piece from the (former) non-reducing sugar, which also explains the similarity in the CID data of the m/z 223 ion derived from the two anomeric isomers

(Figure 2.4, b and d).We also noticed some subtle differences between Figure 2.4b and

2.4d. For example, the relative intensities of m/z 205 and m/z 159 are higher (more than

10%) in Figure 2.4b than in 2.4d. These differences may be due to the existence of a small fraction of structural isomers other than that hypothesized for the m/z 223 ions.

38 α-D-Glcp-(1-3)-Glc, m/z- 343 α-D-Glcp-(1-3)-Glc, m/z 343 -H -H - m/z 223 18 m/z 221 18 -H - -H -

100 131 100 163 (a) (b)

101 205 99 113 131 87 113 221 159 159 Relative Intensity, % Intensity, Relative

Relative Intensity, % Intensity, Relative 223 161 203 0 60 100 140 180 220 0 60 100 140 180 220 m/z m/z β-D-Glcp-(1-3)-Glc, m/z 343 β-D-Glcp-(1-3)-Glc, m/z 343 18 -H - m/z 221 18 -H - m/z 223 -H - -H -

131 163 100 100 (c) (d)

161 113 131 205 221

Relative Intensity, % Intensity, Relative 87 Relative Intensity, % Intensity, Relative 159 223 101 159 203 0 60 100 140 180 220 0 60 100 140 180 220 m/z m/z

Figure 2.4 MS3 spectra of m/z 221 and 223 ions derived from 18O-labeled α-D-Glcp-(1- 3)-D-Glc and β-D-Glcp-(1-3)-D-Glc. (a) CID of m/z 221 ions, CE = 15 V (MS2), AF2 = 15 (MS3), (b) CID of m/z 223 ions, CE = 5 V (MS2), AF2 = 14 (MS3) from 18O-labeled α- D-Glcp-(1-3)-D-Glc, and (c) CID of m/z 221 ions, CE = 15V (MS2) , AF2 = 28 (MS3), (d) CID of m/z 223 ions, CE = 5 V (MS2), AF2 = 24 (MS3) from 18O-labeled β-D-Glcp- (1-3)-D-Glc.

2.3.3 The Effect of CID Conditions on the Formation of m/z 221 Diagnostic Ions

As demonstrated in Figure 2.4, abundant structural isomers of m/z 221 product ions, (m/z 223 ions from the 18O-labeled disaccharides) were observed under relatively low-energy dissociation conditions of 1-3 linked disaccharides, either using beam-type or ion trap CID. This prevents the assignment of the stereochemistry or anomeric configuration of the non-reducing sugar in a typical scenario where either a disaccharide 39 is unlabeled or when it is isolated (unlabeled) from a larger oligosaccharide structure. It would be highly desirable to optimize CID conditions or, for that matter, to find any dissociation conditions whereby the relatively pure, structurally informative glycosyl-GA

(m/z 221) ions could be formed predominantly. Figure 2.5 shows the effect of collision energies on the formation of m/z 221 and m/z 223 ions under beam-type and ion-trap

CID, using 18O-labeled β-D-Glcp-(1-3)-D-Glc as an example. The data were collected using a wide isolation window around m/z 221 to observe both m/z 221 and m/z 223 ions.

It is clear from Figures 2.5a to c that the collision energy in beam-type CID affects the absolute and relative intensities of m/z 221 ions. When the CE was relatively low (CE = 5 V), m/z 223 ions were predominantly formed, with four times higher intensity than that of m/z 221. At a higher CE (CE = 10 V), m/z 221 and m/z 223 ions were seen at nearly equal intensities. Once the CE was increased to 15 V, m/z 221 ions became the dominant peak, accounting for 80% of the total intensities from m/z 221 and

223. Further increasing CE, however, resulted in a huge loss of ion abundance possibly due to competitive ion ejection thus the ratio was not improved.

Parameters that might affect the formation of m/z 221 ions versus m/z 223 ions were also examined for ion trap CID. When ion trap CID of m/z 343 was performed under the instrument default Mathieu q-parameter (q = 0.235), m/z 223 ions were formed exclusively independent of activation energies (data not shown). By changing the activation Mathieu q-parameter to a higher value, precursor ions are placed under a higher potential well depth, and higher activation energies can be applied. An AC generated from an external waveform generator was used for resonance excitation at q =

0.4. As shown in Figures 2.5d to f, the ratio of m/z 221 to m/z 223 ions was increased 40

from almost zero to about 1 as the activation amplitude was increased from 100 mVpp to

400 mVpp (activation time: 50 ms for all cases). Further increasing the activation amplitude resulted in a decrease in m/z 221 to 223 ratio as well as a huge ion loss. The data in Figure 2.5 suggest that m/z 221 ions, which have the desired monosaccharide-GA structures, are generated more favorably under relatively high collision energy conditions in both beam-type and ion trap CID. As compared to ion trap CID, beam-type CID provided more abundant and higher relative intensities of the m/z 221 ions that were wanted for discrimination of the stereochemistry and anomeric configuration of the non- reducing sugar. Evidently, a higher activation energy is needed for the formation of these m/z 221 product ions, and the pathway to generate the glycosyl-GAs is favored when the internal energies of the molecular ions increase. In beam-type CID, much higher collision energies can be applied (typically more than 10 V) as compared to ion trap CID

(hundreds of mV), which leads to a shift in the internal energy distribution of the molecular ions to the high energy direction.8 It is interesting to point out that the glycosyl-glycolaldehyde product ions are virtually the only isomeric species generated under ion trap or low-energy beam-type CID of the 1-2 linked disaccharide anions.2

Since much higher relative intensities of the glycosyl-glycolaldehyde product ions (10 –

40%, normalized to the most abundant peak) can be formed from 1-2 linkages as compared to that of 1-3 linkages (typically < 1% relative intensity) under ion trap CID conditions, it is reasonable to conclude that the formation of these ions from 1-2 linkages needs less energy than required for their generation from 1-3 linkages. Therefore, the formation of the glycosyl-glycolaldehyde product ions is a much lower energy dissociation channel for 1-2 linked disaccharides but a fragmentation pathway for this 41 isomeric species can only be promoted for 1-3 linked disaccharides when the collision energy is higher.

223 221 221 3.9e5 2.5e5 3.9e5 (a) (b) 223 (c)

221 Intensity, cps Intensity, Intensity, cps Intensity, Intensity, cps Intensity, 223

0 00 221 223 221 223 221 223 m/z m/z m/z 223 223 223 9.1e4 221 (d) 2.0e5 (e) 1.0e5 (f)

221 Intensity, cps Intensity, Intensity, cps Intensity, cps Intensity,

0 221 223 00221 223 221 223 m/z m/z m/z

Figure 2.5 Isolation of m/z 221 and m/z 223 ions derived from 18O-labeled β-D-Glcp-(1- 3)-D-Glc under different collisional activation conditions. Beam-type CID: (a) CE = 5 V, (b) CE = 10 V, (c) CE = 15 V. Ion trap CID at q = 0.4, f = 119.248 kHz, excitation time = 50 ms: (d) 100 mVpp, (e) 250 mVpp, (f) 400 mVpp.

Given the high pressure in the collision cell (~ 5 mTorr), multiple collisions happen in beam-type CID and the first-generation product ions may also be subjected to collisional activation once they are formed within the collision cell especially under higher CE conditions. In this sense, beam-type CID is less selective than ion trap CID, where fragment ions are not typically further activated. Indeed, MS3 CID studies in the ion trap showed that many fragment ions, including m/z 325, 323, 283, 281, 253, and 251 generated m/z 221 ions, which might contribute to the observation of higher intensity m/z

221 ions under beam-type CID due to secondary dissociation. 42

2.3.4 Stereochemistry Determination of the Non-Reducing Sugar within 1-3 Linked Disaccharides

Since relatively pure m/z 221 ions containing the intact non-reducing sugars could be formed using beam-type CID with high collision energies, it was possible to differentiate the stereochemistry and anomeric configuration of the non-reducing sugar in disaccharides without 18O-labeling. Figure 2.6 shows the MS3 CID spectra of m/z 221 ions generated by beam-type CID with relatively high CE (13 V to 22 V) from five 1-3 linked disaccharides. Each spectrum was an average of seven spectra and the error bars indicate standard deviations of the peak intensities. The standard deviations were found to be higher (0 - 12 %) for the disaccharide samples than those from standards (0 - 4 %).

This larger degree of spectral variation is likely contributed by the fluctuation in ion intensity of the low abundance m/z 221 isomers under slightly different instrument conditions, and these isomers fragment differently from the diagnostic and more abundant m/z 221 ions that have the monosaccharide-GA structures. Note that α-D-Glcp-

(1-3)-D-Glc, α-D-Glcp-(1-3)-D-Fru, and β-D-Glcp-(1-3)-Glc are disaccharide isomers containing a glucose as the non-reducing sugar; however, each has either a different anomeric configuration or reducing sugar. The characteristic fragmentation profile for disaccharides having glucose as the non-reducing sugar and an α-anomeric configuration can be clearly identified for Figure 2.6a (α-D-Glcp-(1-3)-D-Glc) and Figure 2.6c (α-D-

Glcp-(1-3)-D-Fru), which is distinct from the β-anomeric isomer as shown in Figure 2.6b

(β-D-Glcp-(1-3)-D-Glc, compare all three to the synthetic standards shown in Figure 2.2a and 2.2b). Figure 2.6d shows the characteristic fragmentation profile for the m/z 221 ion 43 of disaccharides having galactose as the non-reducing sugar and having an α-anomeric configuration (compare to Figure 2.8a, CID of m/z 221 from D-D-Galp-GA standard).

α-D-Glcp-(1-3)-Glc, m/z 341 β-D-Glcp-(1-3)-Glc, m/z 341 α-D-Glcp-(1-3)-Fru, m/z 341 -H - -H - -H -

m/z 221 m/z 221 m/z 221

131 131 131 100 100 100 (a) (b) (c)

161 101 99101 99 113 161 113 87 87 161 129 159 221 87 221 129 159 221 Relative Intensity, % Intensity, Relative

Relative Intensity, % Intensity, Relative 113 Relative Intensity, % Intensity, Relative 203 203 101 129 159 203 0 0 0 60 100 140 180 220 60 100 140 180 220 60 100 140 180 220 m/z m/z m/z

α-D-Galp-(1-3)-Gal, m/z 341 α-D-Manp-(1-3)-Man, m/z 341 -H - -H -

m/z 221 m/z 221

101 159 100 100 (d) (e) 131 161

161 131 99 221 101 129 221 113129 87 203 Relative Intensity, % Intensity, Relative Relative Intensity, % Intensity, Relative 87 159 203 99 113 0 0 60 100 140 180 220 60 100 140 180 220 m/z m/z

Figure 2.6 MS3 CID of m/z 221 ions generated via using high CE (CE = 13 to 22 V) for the dissociation of deprotonated disaccharide ions (m/z 341). (a) α-D-Glcp-(1-3)-D-Glc, CE = 15 V (MS2), AF2 = 26 (MS3), (b) β-D-Glcp-(1-3)-D-Glc, CE = 13 V (MS2), AF2 = 30 (MS3), (c) α-D-Glcp-(1-3)-D-Fru, CE = 18 V (MS2), AF2 = 25 (MS3), (d) α-D-Galp- (1-3)-D-Gal, CE = 22 V (MS2), AF2 = 35 (MS3), and (e) α-D-Manp-(1-3)-D-Man, CE = 20 V (MS2), AF2 = 36 (MS3). The error bars in the spectra show the standard deviation of peak intensities based on seven spectra collected over a 1 y period.

The MS3 CID of α-D-Manp-(1-3)-Man (Figure 2.6e) was similar to that of the D-D-

Manp-GA (Figure 2.8b). It is also important to note that under low-energy dissociation conditions, the spectra of the m/z 221 product ions derived from the disaccharides α-D- 44

Glcp-(1-3)-D-Fru, α-D-Galp-(1-3)-D-Gal and α-D-Manp-(1-3)-Man did not match those of their respective glycosyl-glycolaldehydes (Figures 2.7 and 2.8c and 2.8d). We conclude this is due to the presence of alternate isomers, possibly related in their origins to the hypothetical structures shown in Figures 2.4b and 2.4d but having different reducing monosaccharides.

α-D-Glcp-(1-3)-Fru, m/z 341 -H -

m/z 221

161 100 131

113

203 221 Relative % Intensity, 87 0 60 100 140 180 220 m/z

Figure 2.7 CID spectra of m/z 221 ions derived from α-D-Glcp-(1-3)-D-Fru, CE = 5 V (MS2), AF2 = 30 (MS3).

α-D-Galp-GA, m/z 221 α-D-Manp-GA, m/z 221 -H - -H -

101 159 100 100 (a) (b)

131 161 161

99 129131 203 221 221 113 Relative Intensity, % Relative Intensity, Relative Intensity, % Intensity, Relative 87 87 99 129 159 113 203 0 0 60 100 140 180 220 60 100 140 180 220 m/z m/z

α-D-Galp-(1-3)-Gal, m/z 341 α-D-Manp-(1-3)-Man, m/z-341 -H - -H -

m/z 221 m/z 221

113 189 100 100 (c) 131 (d) 161

131

203 161 163 113 203 221

101 221 Relative % Intensity, Relative % Intensity, 101 0 0 60 100 140 180 220 60 100 140 180 220 m/z m/z

Figure 2.8 CID spectra of m/z 221 ions derived from (a) α-D-Galp-GA, AF2 = 17 (MS2), (b) α-D-Manp-GA, AF2 = 13 (MS2), (c) α-D-Gal-(1-3)-D-Gal, CE = 5 V (MS2), AF2 = 20 (MS3), (d) α-D-Man-(1-3)-D-Man, CE = 5 V (MS2), AF2 = 30 (MS3). The error bars in the spectra (a) and (b) show the standard deviation of the relative intensity calculated based on seven spectra collected over a 1 y period.

2.3.5 Spectral Matching by Similarity Scores

The methodology for assigning the stereochemistry and anomeric configuration for the non-reducing sugar unit within a disaccharide is based on the comparison of the

CID patterns of m/z 221 ions to those of the synthetic monosaccharide-GA standards 1.

A high similarity between the compared spectra indicates a large likelihood of them sharing the same structure. Spectral similarity scores, which have been widely used in 46 mass spectral library search for both small molecules,9 peptides10-12 and oligosaccharides13 were chosen to facilitate these comparisons. The spectral similarity scores were calculated between each of the averaged spectra in Figure 6 and the averaged spectra from the monosaccharide-GA standards based on the following equation,

భȀమ σ൫௞ூభ ூమ ൯ σ ூమ Spectral similarity score ೘ ೘ , and k ೘ (Eq. 2.1) ൌ ೖ಺భ శ಺మ ൌ భ σ ೘ ೘ σ ூ೘ మ

1 2 where Im and Im are the normalized intensities of an ion at m/z = m for the two spectra.

Note that the spectral similarity score always has a value between 0 and 1. If two spectra are exactly the same, then, the spectral similarity score becomes 1. In general, a large similarity score indicates close similarity between the two spectra and a large degree of structural similarity. As shown in Table 2.2, the spectral similarity scores between a standard and a disaccharide having the same stereochemistry and anomeric configuration for the non-reducing side were the highest scores, ranging between 0.9838 and 0.9977.

When a disaccharide’s stereochemistry and anomeric configuration on the non-reducing side did not match with the standard, the spectral similarity score was significantly lower, between 0.6942 and 0.9178. Clearly, by comparing the spectral similarity scores, assigning the stereochemistry and anomeric configuration for the non-reducing side of the 1-3 linked disaccharides could be achieved with high confidence. Note that this was only possible under high-energy beam-type CID conditions where the m/z 221 product anions containing the intact non-reducing sugars were optimally generated from precursor disaccharides.

Table 2.2 Spectral similarity scores for 1-3 linked disaccharides vs. monosaccharide-GA standards.

Synthesized Standards Disaccharides α-D-Glcp-GA β-D-Glcp-GA α-D-Galp-GA α-D-Manp-GA α-D-Glcp-(1-3)-Glc 0.9977 0.8845 0.8823 0.7968 β-D-Glcp-(1-3)-Glc 0.8572 0.9840 0.7608 0.8568 α-D-Glcp-(1-3)-Fru 0.9930 0.8899 0.9027 0.8045 α-D-Galp-(1-3)-Gal 0.9178 0.7530 0.9838 0.6942 α-D-Manp-(1-3)-Man 0.8461 0.8702 0.7527 0.9891

2.4 Conclusions

Collisional activation of deprotonated 1-3 linked hexose-containing disaccharides

(m/z 341) generated a low-abundance m/z 221 product ion. By 18O-labeling the reducing sugar carbonyl oxygen of these disaccharides, at least two structural isomers of the m/z

221 ion with the main portion derived from either side of the glycosidic linkage could be mass-discriminated (m/z 221 vs. m/z 223), which enabled the isomers to be isolated and independently studied. The m/z 221 isomer containing the intact non-reducing sugar attached in glycosidic linkage to a glycolaldehyde aglycon was found to be analytically useful, since CID of this species provided the structural information that identified the stereochemistry and anomeric configuration of the non-reducing sugar. No structural information could be obtained from m/z 223 isomer(s) to determine the stereochemistry of the non-reducing sugar or its anomeric configuration. The formation of the diagnostic m/z 221 isomer was found to be affected by CID conditions and was favored under higher energy beam-type CID. It was demonstrated that under optimized CID conditions, this 48 structural isomer could be generated predominantly from five different 1-3 linked disaccharides without requiring 18O-labeling of the reducing sugar. Identification of the nonreducing sugar and the anomeric configuration therefore were achieved at a high confidence level by statistically comparing the CID data of m/z 221 ions generated from the disaccharide samples to those of the synthetic standards via spectral similarity scores.

This study also demonstrated that beam-type CID was a more desirable activation method as compared to ion trap CID to characterize disaccharides using the methodology based on the CID patterns of m/z 221 ions. This method now enables the anomeric configuration and stereochemistry of the m/z 221 ions derived from 2-, 3-, 4-, or 6-linked disaccharides to be assigned in the negative ion mode. This capability was afforded due to the specific arrangement of the triple quadrupole-linear ion trap instrument. It combined (1) selection of the precursor (m/z 341) in the first quadrupole with (2) higher energy dissociation in the second quadrupole collision cell followed by (3) buildup of the desired low abundance m/z 221 product ion in the linear trap, all three of which were necessary to obtain these structural details for 3-linked disaccharides.

2.5 References

(1) Fang, T. T.; Bendiak, B., J. Am. Chem. Soc. 2007, 129, 9721-9736.

(2) Fang, T. T.; Zirrolli, J.; Bendiak, B., Carbohydr. Res. 2007, 342, 217-235.

(3) Brown, D. J.; Stefan, S. E.; Berden, G.; Steill, J. D.; Oomens, J.; Eyler, J. R.; Bendiak, B., Carbohydr. Res. 2011, 346, 2469-2481.

(4) Dallinga, J. W.; Heerma, W., Biol. Mass Spectrom. 1991, 20, 215-231.

(5) Carroll, J. A.; Ngoka, L.; Beggs, C. G.; Lebrilla, C. B., Anal. Chem. 1993, 65, 1582- 1587.

(6) Garozzo, D.; Giuffrida, M.; Impallomeni, G.; Ballistreri, A.; Montaudo, G., Anal. Chem. 1990, 62, 279-286.

(7) Bendiak, B.; Fang, T. T., Carbohydr. Res. 2010, 345, 2390-2400.

(8) Wells, J. M.; McLuckey, S. A., Collision-Induced Dissociation (CID) of Peptides and Proteins. In Biol. Mass Spectrom., Volume 402 ed.; Burlingame, A. L., Ed. Academic Press: 2005; pp 148-185.

(9) Stein, S.; Scott, D., J. Am. Soc. Mass Spectrom. 1994, 5, 859-866.

(10) Frewen, B. E.; Merrihew, G. E.; Wu, C. C.; Noble, W. S.; MacCoss, M. J., Anal. Chem 2006, 78, 5678-5684.

(11) Lam, H.; Deutsch, E. W.; Eddes, J. S.; Eng, J. K.; King, N.; Stein, S. E.; Aebersold, R., Proteomics 2007, 7, 655-667.

(12) Lam, H.; Deutsch, E. W.; Eddes, J. S.; Eng, J. K.; Stein, S. E.; Aebersold, R., Nat. Methods 2008, 5, 873-875.

(13) Kameyama, A.; Kikuchi, N.; Nakaya, S.; Ito, H.; Sato, T.; Shikanai, T.; Takahashi, Y.; Takahashi, K.; Narimatsu, H., Anal. Chem. 2005, 77, 4719-4725.

(14) Zhang, Z., Anal. Chem. 2004, 76, 3908-3922. 50

CHAPTER 3 OBTAINING STEREO-STRUCTURAL INFORMATION WITH SINGLE-SUGAR RESOLUTION FROM LINEAR OLIGOSACCHARIDES

3.1 Introduction

MS3 CID of the m/z 221 diagnostic ions was demonstrated to assign the stereochemistry and anomeric configuration of the non-reducing sugar unit of D- aldohexose (D-aldoHex) containing disaccharides for 1-2, -4, and -6 linked disaccharides1 and 1-3 linked disaccharides2 in Chapter 2. This method was based on the unique fragmentation patterns of diagnostic product ions formed by MS2 CID of deprotonated disaccharide ions, having a non-reducing sugar glycosidically linked to a glycolaldehyde (D-aldoHex-GA, m/z 221, C8H13O7; and D-aldoHexNAc-GA, m/z 262,

3 C10H17NO7 from acetylamine modified disaccharides). By comparing the MS CID spectrum of thus obtained diagnostic ions from disaccharides to MS2 CID spectrum of the synthetic standards using spectral matching algorithm, highly confident structural assignment was achieved.2 Note that the diagnostic ion contains a substructure of a disaccharide with significantly reduced number of stereo-centers, and thus the number of standards for the diagnostic ion could be largely reduced (e.g. 16 for D-aldoHex-GA).

This is a significant reduction as compared to the fact that there are theoretically more than 104 possible structural isomers for disaccharides (only considering D-hexoses).3

Twenty different sugar-GA standards have been synthesized so far (8 D-hexoses, D- 51

GlcNAc, and D-GalNAc, each with 2 anomeric configurations).1 These standards form abundant diagnostic ions as the deprontated molecular ions ([M-H]-1) upon ESI in the negative ion mode. MS2 CID of the synthetic standards provides a straightforward way of generating standard spectra library on a given instrument. This approach has been successfully applied for the determination of non-reducing sugars within 13 disaccharides with a combination of 5 sugar units (D-Glc, D-Gal, D-Man, D-GlcpNAc, and D-

GalpNAc), 2 anomeric configurations, and 4 different linkage positions (1-2, -3, -4, and -

6).1,2

Based on the establishment of m/z 221 diagnostic ions for non-reducing sugar identity and anomeric configuration from disaccharides, we attempted to develop an MSn approach which can pinpoint individual sugar identity and anomeric configuration from an oligosaccharide. The key characteristics of this MSn (n = 4, 5) approach involves optimizing cleavages at glycosidic linkages during early stages of MSn to form overlapping disaccharide substructures, then enhancing cross-ring cleavages for the formation of the m/z 221 diagnostic ions, and finally obtaining the fragmentation pattern of m/z 221 ions which can lead to the assignment of sugar unit identities and anomeric configurations. A modified hybrid triple quadrupole/linear ion trap mass spectrometer was employed in this study to enable high efficiency bidirectional ion transfer between quadrupole arrays4 so that two types of CID, i.e., beam-type or on-resonance ion trap CID can be executed at any stage of MSn up to 5 stages (MS5). This capability was shown to be critical for optimizing the formation of the ladder of product ions involved in the MSn analysis. This MSn approach using highly modified instrument successfully lead us to garner detailed stereo-structural information of individual sugar units, i.e. sugar identity 52 and anomeric configuration, as well as their locations within two pentasaccharide isomers: [α-D-Glcp-(1-4)]4-D-Glc and [β-D-Glcp-(1-4)]4-D-Glc.

3.2 Experimental

3.2.1 Materials

[α-D-Glcp-(1-4)]4-D-Glc, [β-D-Glcp-(1-4)]4-D-Glc, 4-aminobenzoic acid (ABA), chloroform, sodium cyanoborohydride, acetic acid, and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich, Inc. (St. Louis, MO, USA). Sugar-GA standards were synthesized by Fischer glycosidation to generate allyl glycosides. They were converted to the sugar-GAs by ozonolysis.1

3.2.2 Reducing End Modification (Reductive Amination)

Reducing end modification of oligosaccharides with ABA was conducted via reductive amination.5 Briefly, oligosaccharides (0.5 mg) were dissolved in 10 μL of a solution containing 0.15 M ABA in acetic acid: water (3:17, v/v), and 10 μL of 1.0 M sodium cyanoborohydride dissolved in DMSO was added to achieve reductive amination.

After the vial was heated at 37 °C for 16 h, the solvent was removed under vacuum

(Speed-Vac, LABCONCO, Kansas City, MO) and the residue was dissolved in 300 μL water followed by an addition of 300 μL chloroform. The aqueous phase containing the derivatized oligosaccharide was collected after vortex mixing. This step was repeated with another 300 μL of water for extraction and the aqueous phases were combined and 53 vacuum dried. The dried sample was dissolved in 100 μL DMSO and 900 μL methanol.

This solution was further diluted 100-fold in methanol prior to MS analysis.

3.2.3 Mass Spectrometry

A modified hybrid triple quadrupole/linear ion trap mass spectrometer (QTRAP

4000, AB SCIEX, Concord, Ontario, Canada) was used for performing experiments with high efficiency bidirectional ion transfer between Q1, Q2, and Q3 quadrupole arrays.4

Ion selection in Q1 utilized RF/DC isolation, while ion isolation in Q2 and Q3 was typically performed by broadband waveform at q = 0.45 for the center of the frequency notch. Isolation of the m/z 221 diagnostic ion was performed with unit resolution before the last stage of CID, while the isolation windows for other precursors were varied from unit to 3 Th to achieve optimized sensitivity. Ion trap CID was performed in Q2 or Q3 through on-resonance excitation by applying a dipolar ac signal. Beam-type CID was performed with energetic ion transfer either from the Q1ÆQ2 or from Q3ÆQ2.

Derivatized samples were diluted to 0.01 mg/mL in methanol with 1% ammonium hydroxide for negative mode ESI, which was conducted at flow rates of 1-2 μL/min.

3.2.4 MS2 CID Database of 16 D-AldoHex-GA Synthetic Standards

MS2 ion trap CID of 16 deprotonated D-aldoHex-GA standards (m/z 221) was collected in the Q3 linear ion trap on a QTRAP 4000 instrument. To avoid space charge effects, the parent ion intensity before CID was kept around 1 x 106 counts per second

(cps). The remaining parent ion intensity after CID was kept around 18 ± 5% relative to 54 its base product ion (100%). Seven spectra from individual standards were collected over a one year period. Standard deviations of peak heights were calculated for 21 major peaks listed below, which were commonly observed from the standards. The averaged data

(from seven spectra) for the D-aldoHex-GA standards are shown in Figure 3.1, with the standard deviations for each peak indicated. It is worth noting that depending on the instrument and type of dissociation6 the specific product ion abundances for a single compound can vary, but are reproducible within measured error bars for a specific instrument, when energy input is defined by controlling the ratio of the precursor/base product ion. 55

α-D-Glcp-GA, m/z 221 β-D-Glcp-GA, m/z 221 131 131 100 100 101 161 % 99 % 87 113 129 129 159 161 221 87 221 203 101113 159 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Galp-GA, m/z 221 β-D-Galp-GA, m/z 221 101 131 100 100

% 161 % 99 129131 203 221 87 221 87 113 159 129 203 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Manp-GA, m/z 221 β-D-Manp-GA, m/z 221 159 101 100 100 131 161 % % 221 221 87101 129 203 87 99113129131 161 203 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Altp-GA, m/z 221 β-D-Altp-GA, m/z 221 101 100 131 100

% 161 % 99 87 221 113129 131 159 221 101 129 159 203 8597 161 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Gulp-GA, m/z 221 β-D-Gulp-GA, m/z 221 161 131 100 101 100 113 97 203 % 221 % 203 89 159 221 85 131 143 177 189 87 97101129 161 189 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Idop-GA, m/z 221 β-D-Idop-GA, m/z 221 131 101 131 100 100 203 113 87 % 161 % 97 221 85 221 8797101113 141159 203 129 141 161 189 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Allp-GA, m/z 221 β-D-Allp-GA, m/z 221 101 159 100 100 113 161 % % 8589 221 129 221 99 125131 161 203 85 99113 131 203 0 0 60 100 140m/z 180 220 60 100 140m/z 180 220 α-D-Talp-GA, m/z 221 β-D-Talp-GA, m/z 221 97 189 100 100 131 113 101 131 % 129 141 201203 % 113 87 159161 221 87101 221 85 125 189 85 161 203 060 100 140 180 220 060 100 140 180 220 m/z m/z

Figure 3.1 Averaged MS2 trap CID data of 14 deprotonated D-aldoHex-GA standards (m/z 221). The 21 major peaks were labelled with standard deviation based on 7 averaged spectra obtained over a 1 y period. 56

3.2.5 Spectral Matching by Similarity Score

Spectral similarity scores were calculated between the CID spectrum of m/z 221 ions obtained from MSn (n =4 or 5) of oligosaccharides and each of the averaged spectra in Figure 3.1 (CID spectra of m/z 221 ions of the D-aldoHex-GA standards). The following equation was used for similarity score:7

భȀమ כ൯ σ ூ כσ൫௞ூ ூ Spectral similarity score ೘ ೘ , and k ೘ ൌ ೖሺ಺భ శ಺మ ሻ ൌ σ ೘ ೘ σ ூ೘ మ

* where Im and Im are the normalized intensities (to the highest fragment peak) of an ion at m/z = m for the two spectra. 21 major peaks (m/z 85, 87, 89, 97, 99, 101, 111, 113, 125,

129, 131, 141, 143, 159, 161, 175, 177, 189, 201, 203, and 221) were chosen for this calculation. Note that the spectral similarity score always has a value between 0 and 1. If two spectra are exactly the same, then the spectral similarity score is equal to 1. In general, a large similarity score indicates close similarity between the two spectra and a large degree of structural similarity.

3.3 Results and Discussion

3.3.1 MSn Approach

The general strategy of the MSn approach is shown in Scheme 3.1. It involves optimizing cleavages at glycosidic linkages during early stages of MSn, then enhancing cross-ring cleavages only with disaccharides isolated as an ordered set of overlapping substructures (illustrated with a tetrasaccharide in Scheme 3.1, where sugar unit number was assigned from reducing end (i.e., unit 1Æ4). Key steps include: 1) dissociation of a precursor derivatized at the reducing end (M, Scheme 3.1) to a ladder of smaller oligosaccharides, each successively one sugar unit shorter, containing the reducing end

8 tag (Ym ions); 2) generation of disaccharide fragments (C2 ions) from each Ym ion, derived from the opposite end; 3) dissociation of the disaccharides to their m/z 221 diagnostic product ions (the sugar-GA); 4) fragmentation of these diagnostic ions; 5) spectral matching of the diagnostic ion CID data to that of a database of synthetic standards for assigning stereochemistry.

Scheme 3.1 The MSn (n = 4 or 5) approach for stereo-structural analysis of oligosaccharides. “M” stands for reducing end modification. Sugar unit number is assigned from the reducing end.

The above described MSn approach has the advantage of knowing the exact origin of each fragment with respect to its initial position within the oligomer. Also, the dissociation patterns of disaccharides in negative-ion mode provide information that can assign the linkage between them.9 Given the need to preserve the structural integrity of intact sugars and their linkage, disaccharides are key substructures to isolate during earlier stages of MSn. The reducing end modification (M) introduces a mass distinction between Y and C ions and also bears a negative charge that enables selection of the

2 ladder of singly charged Ym ions after MS . Yet, through charge-transfer, it permits 59

disaccharide (C2) ions to be isolated via a neutral loss of the derivatized end of the molecule (MS3). Note that this approach does not define the structure of the reducing end unit.

The success of this method highly depends on producing reasonable yields of the

n ladder of ions employed in MS , i.e., Ym, C2, and m/z 221 ions, from a given oligosaccharide. Formation of Ym and C2 ions requires glycosidic bond cleavages. These bonds can be cleaved preferentially using low-energy CID under most conditions.10 Our previous studies showed that the higher-energy beam-type CID condition was critical to produce m/z 221 diagnostic ions with both high structural purity and reasonable yield for many 3-linked disaccharides.2 On the other hand, ion trap CID of the m/z 221 ions was necessary to obtain reproducibly distinct fingerprint patterns for structural identification.

In order to have the capacity of optimizing the product ions involved in the tree of MSn analysis, it is desirable to perform experiments on an instrument platform which provides both beam-type and ion trap CID at any MS/MS stage.

3.3.2 A Hybrid Triple Quadrupole/Linear Ion Trap for MSn

We choose to work on a hybrid triple quardrupole/linear ion trap instrument since this type of instrument already has the capability of both beam-type and ion trap CID and therefore, efforts of instrument modification can be significantly reduced based on this platform. A schematic representation of the instrument is shown in Figure 3.2.

Conventionally, beam-type CID is performed by accelerating ions from Q0 to Q2 (Figure

3.2a), while the precursor ions are mass isolated in Q1 (RF/DC) during fly. In this mode, beam-type CID can only be performed in the MS2 stage. However, if the flow of ions 60 can be reversed, the mass analysis quadrupole arrays (Q1 and Q3) and reaction sections

(Q0 and Q2) can be re-accessed and the tandem-in-space experiments such as beam-type

CID can be achieved as shown in Figure 3.2b and 3.2c. Indeed, Thompson et al.11 and

Xia et al.4 have shown that with the use of high efficiency bidirectional ion transfer between quadrupole arrays, beam-type CID) can be performed in MSn (n>2) without significant hardware modification of a Q-q-TOF instrument. Given the above considerations, we employed a research grade triple quadrupole/linear ion trap mass spectrometer to implement the proposed MSn approach. In order to facilitate high efficiency bi-directional ion transfer, an axial electric field (LINAC)12,13 was superposed to the center axis of Q2 and Q3 quadrupole arrays. The direction and amplitude of the electric field were controlled from instrument software and were optimized during experiments. Results collected from this instrument showed ~80% transfer efficiency after three cycles of transfers (Q2ÆQ3, Q3ÆQ2, and Q2ÆQ3). Compared to beam-type

CID, it is relatively simple to perform ion trap CID at a given stage of MS/MS. The only requirement is to transfer ions to the quadrupole array which is capable of ion trap CID.

On this modified instrument, on-resonance ion trap CID can be either performed in Q2 or

Q3 quadrupole arrays via dipolar ac excitation. Ion trap CID in Q2 improved CID efficiency due to the higher pressure in Q2 (6 to 8 mTorr) as compared to that in Q3 and

Q1 (3 x 10-5 Torr), as previously described.14,15 With all above capabilities (i.e., reversing ion beams and ion trap CID in Q2 and Q3), MSn (n = 4 or 5) employing either of the two CID methods is possible to conduct at any stage of MS/MS.

Auxiliary AC

SK IQ1ST1 IQ2 IQ3ST3 EX -’ve ESI Q0Q1 Q2 Q3 10 mTorr 3e-5 Torr 6-8 mTorr 3e-5 Torr

Beam a) CID

Beam Beam c) CID CID b)

Figure 3.2 A schematic of a modified triple quadrupole/linear ion trap mass spectrometer (QTRAP 4000). Beam-type CID can be performed through energetic axial ion transfer between quadrupole arrays (Q0 to Q3): a) conventional method: Q1 to Q2 with mass isolation in Q1; b) reversed ion transfer from Q3 to Q2 with mass isolation in Q3; and c) reversed ion transfer from Q1 to Q0 with mass isolation in Q1.

3.3.3 Effect of CID Conditions on the Formation of Key Product Ions

A pentasaccharide modified with 4-aminobenzoic acid ([α-D-Glcp-(1-4)]4-D-Glc-

ABA, structure shown in the inset of Figure 3.3) was used as a model compound for method and instrumental parameter optimization. Based on the MSn method described in

Scheme 3.1, an MS5 CID with following sequence is necessary to obtain sugar unit 3

2- structural information: [474 ([M-2H] )Æ624 (Ym)Æ341 (C2)Æ221 (diagnostic ion)Æfragments]. Figure 3.3a shows the ion transfer scheme for conducting MS3 CID using the modified triple quadrupole/linear ion trap. In short, the doubly charged parent anions ([M-2H]2-, m/z 474) formed by ESI in negative ion mode, were mass selected in the Q1 quadrupole array (RF/DC mode) and accelerated to the Q2 collision cell to effect beam-type CID. Product ions were collected in Q3, followed by a broadband isolation of

Y3 (m/z 624) ions. Either beam-type or ion trap CID of m/z 624 ions could be performed 62 for MS3 by reversing the ion beam. Beam-type CID involved accelerating ions from Q3 to Q2 in the opposite direction of the traditional beam-type CID. For ion trap CID, m/z

624 ions were transferred from Q3 to Q2 with minimum kinetic energies, followed by on- resonance dipolar ac excitation in Q2.

3 Note that the goal of the MS is to maximize the formation of C2 (m/z 341) from

Y3 (m/z 624), which requires a glycosidic bond cleavage between the first and the second sugar units (refer to Figure 3.3 for the oligosaccharide nomenclature). A range of collision energies (CEs) (10 – 60 V) were tested using beam-type CID from Q3 and Q2, but m/z 341 ions were barely detectable under those conditions. At higher CEs, small fragments in the range of m/z 200-300 dominated due to sequential fragmentation of Y3 ions (Figure 3.3b, CE = 42 V). The ion trap CID spectrum of Y3 ions (Figure 3.3c, q =

0.45, 1.4 V, and 300 ms activation) consisted of fewer fragments, most of which derived from glycosidic bond cleavages including C2 ions (m/z 341) as well as m/z 462, 444, 300,

282. Ion trap CID consistently yielded more C2 ions from the set of singly charged Ym ions as compared to beam-type CID. This observation is consistent with glycosidic bonds being preferentially cleaved under slow heating conditions.16

Figure 3.3d shows the ion transfer scheme for conducting either beam-type CID or ion trap CID in the 4th stage of MS/MS. The goal of MS4 is to maximize the formation

1 2 of m/z 221 diagnostic ions from C2 (m/z 341) ions. The MS and MS steps was the same as Figure 3.3a and ion trap CID was employed in Q2 for the MS3 step. After that, the

3 MS product ions were sent to Q3 and C2 (m/z 341) ions were isolated in Q3 via broadband. Using reversed ion transfer (Q3ÆQ2), C2 ions were subjected to either beam-type or ion trap CID in Q2 for MS4. Our previous studies on disaccharides showed 63 that diagnostic product ions (m/z 221) were formed by cross-ring cleavages and usually required higher activation energies, especially for 3- and 4-linked structures. Similarly,

C2 ions isolated from oligosaccharides required higher-energy CID to generate m/z 221 ions. As shown in Figure 3.3e, the m/z 221 diagnostic ion is the second most abundant using beam-type CID (CE = 22 V), yet it is not detectable using ion trap CID (Figure

3.3f, q = 0.45, 1.4 V, and 300 ms activation). The data in Figure 3.3 demonstrate that the capability of employing different types of CID is extremely important to optimize the yields of key product ions, which is also critical to the success of the planned MSn experiments.

Unit # 5 4 [α-D-Glcp-(1-4)]4-D-Glc-ABA 3 2 1 Precursor ion Ion for further CID

Y3, 624 C2, 341 3 a) MS : 100 300 [474Æ624Æfragments] b) Beam CID MS3 222 282 % Q1 Q2 Q3 264 341 624 246 204 365 444462 0 200 300 400 500 600 700 Iso 474 Beam CID Iso 624 m/z 100 624 c)Beamc))BBeam CCICIDD c) Trap CID MS3 d)Trapd)Trap CICCIDD MSAEM % ion transfer 462 282300 444 341 179 263 606 0 200 300 400 500 600 700 4 m/z d) MS : 161 [474Æ624Æ341Æfragments] 100 e) Beam CID MS4

Q1 Q2 Q3 % 221 179 143 131 263 341 0 Iso 474 100 200 300 400 Beam CID Iso 624 m/z 161 Trap 100 f) 4 CID Iso 341 Trap CID MS

e)Beame))BBeam CIDCCID % f)Trapf)T p 179 221 323 CIDCIC D 341 MSAEM 143 263 281 ion transfer 0 100 200 300 400 m/z

3 4 Figure 3.3 MS and MS experiments of a pentasaccharide, [α-D-Glcp-(1-4)]4-D-Glc- ABA. (a) Ion transfer scheme for conducting either beam-type CID or ion trap CID in rd 3 the 3 stage MS/MS for the formation of C2 ions (m/z 341) via MS [474(2- )Æ624Æfragments]). (b) MS3 beam-type CID from Q3 to Q2, CE = 42 V; (c) MS3 ion trap CID in Q2, q = 0.45, 1.4 V, and 300 ms of activation. (d) Ion transfer scheme for conducting beam-type or ion trap CID in the 4th stage MS/MS for the formation of m/z 221 diagnostic ions via MS4 [474(2-)Æ624Æ341Æfragments] (e) MS4 beam-type CID from Q3 to Q2, CE = 22 V; (f) MS4 ion trap CID in Q2, q = 0.45, 1.4 V, and 300 ms of activation. The structure of [α-D-Glcp-(1-4)]4-D-Glc-ABA, is shown in the inset.

3.3.4 Determination of Individual Sugar Identity and Anomeric Configuration from Pentasaccharides

The feasibility of determining individual sugar unit identity and anomeric configuration within an oligosaccharide was tested with two model pentasaccharides: [α-

D-Glcp-(1-4)]4-D-Glc-ABA and [β-D-Glcp-(1-4)]4-D-Glc-ABA, isomers having D- and

β-anomeric configurations, respectively. To determine the structural information for sugar units from the non-reducing end (i.e. unit 5Æ2), the following ions need to be

2 produced at MS : m/z 341 (C2), 786 (Y4), 624 (Y3), and 462 (Y2), respectively. Both singly (1-) and doubly (2-) deprotonated parent ions were formed from ESI in the negative mode. Beam-type CID (CE = 21 – 24 V) of 2- charge state parents yielded higher intensities for most product ions thus beam-type CID was chosen for MS2 prior to higher stages of MSn. Using sugar unit 3, for example, the set of MS5 experiments required to obtain its stereo-structural information were: [474(2-

)Æ624Æ341Æ221Æfragments]. Based on previous study, ion trap CID was used for

MS3 of m/z 624 ions and beam-type CID was used for MS4 CID of m/z 341. Ion transfer scheme specific to this MS5 experiment is shown in Figure 3.4a and CID conditions of each step were summarized in Table 3.1. MS5 CID of m/z 221 was obtained for both pentasaccharide isomers differing only in their anomeric configurations. For [α-D-Glcp-

(1-4)]4-D-Glc-ABA, the m/z 221 fragment derived from sugar unit 3 (Figure 3.4b) yielded a fragmentation pattern characteristic of α-D-Glcp-GA (Figure 3.4d, error bars from 7 experiments acquired over a 1 yr period). Spectral matching to the D-aldoHex-GA standard database (16 structures: 8 sugar identities × 2 anomeric configurations) was based on similarity scores calculated as described in the experimental section. α-D-Glc 66 was identified as the top match for sugar unit 3 with a high similarity score (0.9771) as compared to the second match (β-D-Ido, 0.9171, Table 3.1). Figure 3.4c shows the MS5

n CID spectrum of [β-D-Glcp-(1-4)]4-D-Glc-ABA acquired from the same MS sequence.

This spectrum is markedly different from the corresponding α-anomer (Figure 3.4b) and almost identical to the ion trap CID data of synthetic β-D-Glcp-GA (Figure 3.4e). Here,

β-D-Glc was the top match (similarity score: 0.9654) in comparing to the database of standards. The alternate anomers matched poorly (scores of 0.75-0.86, Table 3.1). The data in Figure 3.4 demonstrate that even for a subtle structural difference, i.e. anomeric configuration, sugar unit 3 could be confidently determined.

5 a) MS : [α-D-Glcp-(1-4)]4-D-Glc-ABA [E-D-Glcp-(1-4)]4-D-Glc-ABA [474Æ624Æ341Æ221Æfrag] 131 131 100 100 (b) (c) 5 MS5 Q1 Q2 Q3 101 MS 161 % % 161 87 113 159 221 87 159 221 0 0 Iso 474 Beam CID Iso 624 100 150 200 100 150 200 m/z m/z Trap 131 2 CID Iso 341 131 2 MS 100 (d) MS 100 (e) 161 Beam CID 101 % 99 113 % Iso 221 87 129 129 159161 221 87 221 203 101113 159 Trap 0 0 CID MSAE 100 150 200 100 150 200 m/z m/z

5 Figure 3.4 MS experiments of pentasaccharides, [α-D-Glcp-(1-4)]4-D-Glc-ABA and [β- 5 D-Glcp-(1-4)]4-D-Glc-ABA: (a) Ion transfer scheme for conducting MS CID for both 5 pentasaccharides. MS ion trap CID of (b) [α-D-Glcp-(1-4)]4-D-Glc-ABA and (c) [β-D- 2 Glcp-(1-4)]4-D-Glc-ABA. MS ion trap CID of synthetic standards: (d) α-D-Glcp-GA and (e) β-D-Glcp-GA.

Determining sugar units 5 and 4 involved MS4 [474(2-)Æ341Æ221Æfragments] and MS5 [474(2-)Æ786Æ341Æ221Æ fragments] experiments. The data are shown in

Figures 3.5 and 3.6 for [α-D-Glcp-(1-4)]4-D-Glc-ABA and [β-D-Glcp-(1-4)]4-D-Glc-

ABA, respectively. Conditions for each MSn stage are summarized in Table 3.2. Sugar 2 could not be identified since m/z 221 ions were not formed by CID of the Y2 ion. Table

3.1 summarizes the top 3 similarity scores of diagnostic ions derived from the two pentasaccharides against the D-aldoHex-GA standard database. A full list is reported in

Table 3.3. The highest scores were assigned to the correct identity and anomeric configuration for each sugar unit. The stereo-structures of 3 individual sugar units were successfully determined for each pentasaccharide based on MSn (n = 4 or 5) experiments and spectral matching to standards.

Unit# 5 4 3 2 1 [α-D-Glcp-(1-4)]4-D-Glc-ABA

Precursor ion Ion for further CID

545 161 100 100 100 131 a) b) c) MS2 MS3 MS4 179 % 159 [M-2H]2- 101 221 587 341 113 161 474 221 99 263 338 129 203 768 281 87 141 341 444 503 624 786 0 0 0 300 400 500 600 700 800 900 150 200 250 300 350 100 150 200 m/z m/z m/z

383 221 131 100 100 100 d) MS3 e) MS4 f) MS5

101 % 425 786 161 113 341 444 99 129 159 221 282 143 179 87 161 606 768 263 141 203 0 0 0 200 300 400 500 600 700 800 150 200 250 300 350 100 150 200 m/z m/z m/z

624 161 131 100 g)100 h)100 i) MS3 MS4 MS5 101

I% 221 462 113 161 282 300 444 179 87 159 221 143 341 179 263 606 131 263 341 0 0 0 200 300 400 500 600 700 100 200 300 100 150 200 m/z m/z m/z

Figure 3.5 MSn (n = 4 or 5) for the determination of stereo-structural information of sugar units 3-5 from [α-D-Glcp-(1-4)]4-D-Glc-ABA. Spectra (a) to (c) show the sequential MS2 to MS4 data based on [474(2-)Æ341Æ221Æfragments] to determine sugar unit 5 structural information. Spectra (d) to (f) show the sequential MS3 to MS5 data based on [474(2-)Æ786Æ341Æ221Æfragments] to determine sugar unit 4. Spectra (g) to (i) show the sequential MS3 to MS5 data based on [474(2- )Æ624Æ341Æ221Æfragments] to determine sugar unit 3. The experimental conditions for each step can be found in Table 3.2.

Unit# 5 4 3 2 1 [E-D-Glcp-(1-4)]4-D-Glc-ABA

Precursor ion Ion for further CID

282 161 131 100 [M-2H]2- 100 100 a)474 MS2 b)MS3 c) MS4 161 % 221 587 179 222 341 503 143 221 393 545 786 624 341 87 101 0 0 0 200 500 800 100 200 300 400 100 150 200 m/z m/z m/z 161 100 393 100 100 131 d)282 MS3 e)MS4 f) MS5 161 221 % 179 221 143 341 624 263 341 87 101 0 0 0 200 400600 800 100 200 300 400 100 150 200 m/z m/z m/z 100 462 100 161 100 g)MS3 h)MS4 i) 131 MS5 300 444 % 282 161 263 179 143 341 624 179 221 222 606 87 113 263 73 159 221 0 0 0 200 300 400 500 600 100 200 300 400 100 150 200 m/z m/z m/z

Figure 3.6 MSn (n = 4 or 5) for the determination of stereo-structural information of sugar units 3-5 from [β-D-Glcp-(1-4)]4-D-Glc-ABA. Spectra (a) to (c) show the sequential MS2 to MS4 data based on [474(2-)Æ341Æ221Æfragments] to determine sugar unit 5 structural information. Spectra (d) to (f) show the sequential MS3 to MS5 data based on [474(2-)Æ393(2-)Æ341Æ221Æfragments] to determine sugar unit 4. Spectra (g) to (i) show the sequential MS3 to MS5 data based on [474(2- )Æ624Æ341Æ221Æfragments] to determine sugar unit 3. The experimental conditions for each step can be found in Table 3.2.

Table 3.1 The top 3 ranked D-aldoHex candidates based on spectral similarity scores (indicated in parenthesis).

Spectral Similarity Score sample Unit # 1st 2nd 3rd α-D-Glc β-D-Ido β-D-Alt 3 (0.9771) (0.9171) (0.9084) [α-D-Glcp-(1-4)] - α-D-Glc β-D-Alt β-D-Ido 4 4 D-Glc-ABA (0.9939) (0.9082) (0.9070) α-D-Glc β-D-Alt β-D-Ido 5 (0.9862) (0.8913) (0.8902) β-D-Glc α-D-Ido α-D-Alt 3 (0.9654) (0.9407) (0.9253) [β-D-Glcp-(1-4)] - β-D-Glc α-D-Ido α-D-Alt 4 4 D-Glc-ABA (0.9802) (0.9649) (0.9498) β-D-Glc α-D-Ido α-D-Alt 5 (0.9748) (0.9639) (0.9499)

Like any MSn experiment, each stage of MS/MS enhances the selectivity of the experiment at the cost of reduced ion signal. The signal loss arises from partitioning of parents among fragments and ion loss during isolation/dissociation. Six orders of magnitude reduction of signal was observed in MS5 experiments. This led to long signal averaging periods to achieve good signal/noise (S/N) ratios. For instance, Figure 3.4b accumulated 31160 scans (~11 hr) to reach a S/N of 50, using 6-7 μg of sample. The number of scans, however, can be lowered using the spectral matching approach. For example, fewer scans (3422, ~1.2 hr) than the spectrum shown in Figure 3.4b with a lower S/N (7) still provides the first-ranked spectral similarity score (0.9475) against α-

D-Glcp-GA (second match: β-D-Idop-GA, 0.8928). Note that these spectra could not be obtained at all using only one type of dissociation methods, for example; there was simply not enough m/z 221 product ion to afford a spectrum without beam-type CID.

Table 3.2 Experimental conditions for conducting MSn experiments.

Sample Unit # MS2 MS3 MS4 MS5

474(2-)Æ[S2a] 474(2-)Æ624Æ[S2g] 474(2-)Æ624Æ341Æ[S2h] 474(2-)Æ624Æ341Æ221Æ[S2i, 4a] 3 Q1ÆQ2 Beam (24 V) Q2 Trap (q = 0.45, 1.4 V, 300 ms) Q3ÆQ2 Beam (24 V) Q3 Trap (q = 0.23, 15 mV, 100 ms)

[α-D-Glcp-(1-4)] - 474(2-)Æ 474(2-)Æ786Æ[S2d] 474(2-)Æ786Æ341Æ[S2e] 474(2-)Æ786Æ341Æ221Æ[S2f] 4 4 D-Glc-ABA Q1ÆQ2 Beam (23 V) Q2 Trap (q = 0.45, 1.8 V, 100 ms) Q3ÆQ2 Beam (24 V) Q3 Trap (q = 0.23, 18 mV, 100 ms)

474(2-)Æ 474(2-)Æ341Æ[S2b] 474(2-)Æ341Æ221Æ[S2c] 5 ------Q1ÆQ2 Beam (23 V) Q3ÆQ2 Beam (23 V) Q3 Trap (q = 0.23, 18 mV, 100 ms)

474(2-)Æ 474(2-)Æ624Æ[S3g] 474(2-)Æ624Æ341Æ[S3h] 474(2-)Æ624Æ341Æ221Æ[S3i, 4b] 3 Q1ÆQ2 Beam (23 V) Q2 Trap (q = 0.45, 1.4 V, 300 ms) Q3ÆQ2 Beam (22 V) Q3 Trap (q = 0.23, 14 mV, 100 ms)

[β-D-Glcp-(1-4)] - 474(2-)Æ 474(2-)Æ393(2-)Æ[S3d] 474(2-)Æ393(2-)Æ341Æ[S3e] 474(2-)Æ393(2-)Æ341Æ221Æ[S3f] 4 4 D-Glc-ABA Q1ÆQ2 Beam (21 V) Q3ÆQ2 Beam (15 V) Q3ÆQ2 Beam (16 V) Q3 Trap (q = 0.23, 12 mV, 100 ms)

474(2-)Æ[S3a] 474(2-)Æ341Æ[S3b] 474(2-)Æ341Æ221Æ[S3c] 5 ------Q1ÆQ2 Beam (22 V) Q3ÆQ2 Beam (23 V) Q3 Trap (q = 0.23, 15 mV, 100 ms)

Superscripts indicate the figures, above, from which corresponding MS data was derived.

71 2

Table 3.3 Spectral similarity scores for each unit in oligosaccharides against 16 D-aldoHex-GA standards.

Unit Aldohexose-GA Synthetic Standards Sample # α-D-All β-D-All α-D-Alt β-D-Alt α-D-Gal β-D-Gal α-D-Glc β-D-Glc α-D-Glu β-D-Glu α-D-Ido β-D-Ido α-D-Manβ-D-Man α-D-Tal β-D-Tal

3 0.8094 0.7673 0.8656 0.9084 0.8994 0.7857 0.9771 0.8556 0.8543 0.8572 0.8599 0.9171 0.8140 0.8631 0.9008 0.7838 [α-D-Glcp-(1-4)] - 4 4 0.7940 0.7544 0.8863 0.9082 0.8844 0.8193 0.9939 0.8629 0.8006 0.8651 0.8723 0.9070 0.8137 0.8521 0.8567 0.7411 D-Glc-ABA 5 0.7701 0.7685 0.8862 0.8913 0.8517 0.8312 0.9862 0.8597 0.7785 0.8836 0.8867 0.8902 0.8361 0.8207 0.8406 0.7183

3 0.4633 0.7389 0.9253 0.5929 0.6621 0.8145 0.7962 0.9654 0.6712 0.8273 0.9407 0.6991 0.8466 0.5312 0.7047 0.5996 [β-D-Glcp-(1-4)] - 4 4 0.4708 0.7286 0.9498 0.6032 0.6647 0.8636 0.7910 0.9802 0.6544 0.8391 0.9649 0.6710 0.8520 0.5627 0.6690 0.6040 D-Glc-ABA 5 0.4099 0.7075 0.9499 0.5427 0.6094 0.8721 0.7559 0.9748 0.6092 0.8527 0.9639 0.6476 0.8507 0.5132 0.6443 0.5728

The numbers in red indicate the highest scores (first match) while the numbers in black bold indicate the second highest scores (second match). 72 73

3.4 Conclusions

An MSn approach to pinpoint the stereo-structures (sugar identity, anomeric configuration, and location) of individual sugar units within linear oligosaccharides was described. The approach involved first optimizing the isolation of disaccharide units as an ordered set of overlapping substructures via glycosidic bond cleavages. Subsequently, cross-ring cleavages were optimized for individual disaccharides to yield key diagnostic product ions (m/z 221). High confidence stereo-structural determination was achieved by matching MSn CID of the diagnostic ions to synthetic standards via a spectral similarity scores. By using this approach, structural information (identity and amoneric configuration) of 3 individual sugar units (out of 5) was successfully determined for two isomeric pentasaccharides. To our knowledge, this is the first report showing that specific sugar identities, their anomeric configurations, and their positions within oligosaccharides can be determined by MS. The above described approach of higher stages of MSn could be used as a complementary means to NMR for oligosaccharide structural analysis while offering enhanced sensitivity. A larger pool of oligosaccharides will be tested for the applicability in future studies. In the case of branched oligosaccharides (N-glycans), exoglycosidases will be used to convert the branched structure into linear structures before subjected mass spectrometric analysis. We will also explore different types of reducing end modifications which could enhance fragmentation at selected bonds within the oligosaccharides. Several instrumentation methods are being evaluated for enriching minor fragment ions at any MSn stage for improved sensitivity.

3.5 References

(1) Fang, T. T.; Bendiak, B., J. Am. Chem. Soc. 2007, 129, 9721-9736.

(2) Konda, C.; Bendiak, B.; Xia, Y., J. Am. Soc. Mass Spectrom. 2012, 23, 347-358.

(3) Laine, R. A., Glycobiology 1994, 4, 759-767.

(4) Xia, Y.; Thomson, B. A.; McLuckey, S. A., Anal. Chem. 2007, 79, 8199-8206.

(5) Harvey, D. J., J. Am. Soc. Mass Spectrom. 2000, 11, 900-915.

(6) Bendiak, B.; Fang, T. T., Carbohydr. Res. 2010, 345, 2390-2400.

(7) Zhang, Z., Anal. Chem. 2004, 76, 3908-3922.

(8) Domon, B.; Costello, C. E., Glycoconjugate J. 1988, 5, 397-409.

(9) Dallinga, J. W.; Heerma, W., Biol. Mass Spectrom. 1991, 20, 215-231.

(10) Zaia, J., Mass Spectrom. Rev. 2004, 23, 161-227.

(11) Thomson, B. A. Apparatus and method for MSnth in a tandem mass spectrometer system. U.S. Patent 7,145,133 B2, December 5, 2006.

(12) Thomson, B. A.; Jolliffe, C. L. Spectrometer with axial field. U.S. Patent 5,847,386, December 8, 1998.

(13) Loboda, A.; Krutchinsky, A.; Loboda, O.; McNabb, J.; Spicer, V.; Ens, W.; Standing, K., Eur. J. Mass Spectrom. 2000, 6, 531-536.

(14) Collings, B. A., J. Am. Soc. Mass Spectrom. 2007, 18, 1459-1466.

(15) Morris, M.; Pierre, T.; B., R. K., J. Am. Soc. Mass Spectrom. 1994, 5, 1042-1063.

(16) Carroll, J. A.; Willard, D.; Lebrilla, C. B., Anal. Chim. Acta 1995, 307, 431-447. 75

CHAPTER 4 Z1 IONS AS DIAGNOSTIC IONS FOR LINKAGE DETERMINATION

4.1 Introduction

In terms of linkage determination, disaccharides have been extensively studied as model systems in the positive ion mode as metal ion adducts1-6 and in the negative ion mode as deprotonated molecular ions7-11 or anion adducts of the molecules.12 Based on the patterns of A-type product ions, such as m/z 221, 251, 263, and 281, formed by cross- ring cleavages within the reducing sugar via various dissociation methods, the linkage positions can be determined.1-15 Furthermore, the relative intensities of A ions as well as the cleavage products from either side of the glycosidic oxygen provided additional evidence for linkage assignment in the negative ion mode.11 Linkage determination for oligosaccharides takes the same approach by detecting a certain combination of A ions.

Data analysis, however, can be much more challenging for oligosaccharides due to the co-existence of isomeric and isobaric peaks from MS/MS, potentially derived from either end of the molecule. Permethylation16 or reducing-end modification17 have therefore been utilized to simplify the situation by providing mass discrimination of the product ions.

Currently, MS2 employing collision-induced dissociation (CID) is still the most widely used strategy for linkage determination. However, depending on the nature of analytes

(trisaccharides or larger) and CID conditions, key diagnostic ions (i.e. A type ions) may be missing, leading to either miss-assigned or un-assigned linkage positions.5,8,9 In order 76 to maximize the structural information inherent in fragmentation patterns using MS2 for oligosaccharide analysis, different activation/dissociation methods have been investigated, including infrared multiphoton dissociation (IRMPD),18 electron capture dissociation (ECD),19,20 electron-induced dissociation (EID),21 electron detachment dissociation (EDD),22 and electron excited dissociation (EED).23 The application of the above methods has been limited to research groups owing these instrumental capabilities.

Overall, MS2, the simplest MS/MS, has advantages such as good sensitivity, high throughput, and simplicity in performance and instrumentation. However, structural information to firmly establish anomeric configurations and stereochemistries of individual monosaccharides and some of their linkages within oligosaccharides often cannot be obtained using MS2 alone.

In this chapter, we describe an MSn approach to extract linkage information from linear oligosaccharides together with their locations within the molecules with greater confidence than is currently possible. We discovered that CID of Z1 ions derived from the reducing sugar of deprotonated disaccharides showed distinct fragmentation patterns virtually solely based on their linkage types and not significantly affected by the sugar unit identities or their anomeric configurations. An MSn CID (n = 3-5) strategy was developed that permitted linkages to be determined more confidently within a linear oligosaccharide by decomposing the oligomer in the gas-phase into a set of overlapping disaccharide structures with known origins within the oligosaccharide. Z1 ions were then optimized for each disaccharide subunit, and repeatable CID spectra were obtained for Z1 ions derived from independent disaccharides. Spectral similarity scores were employed for linkage determination via spectral matching of the thus obtained Z1 CID data to the 77 standard disaccharide database. At a stage of proof-of-principle, this approach was successfully demonstrated with two trisaccharides and a pentasaccharide.

4.2 Experimental

4.2.1 Materials

A list of 17 disaccharides and 3 oligosaccharides is shown in Table 4.1. All samples were purchased from commercial sources (indicated by the superscripts in Table

4.1) and used without further purification. 4-Aminobenzoic acid (ABA), ethyl 4-

18 aminobenzoate (ABEE), H2 O (97 atom %), chloroform, sodium cyanoborohydride, acetic acid, and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich, Inc.

(St. Louis, MO, USA). Detailed procedures of 18O-labeling24 and reductive amination25,26 of reducing saccharides with ABA and ABEE were described in the previous chapters

(18O-labeling in chapter 2 and reductive amination in chapter 3). α1-2,3 Mannosidase with G6 reaction buffer and BSA was purcharsed from New England Biolabs, Inc.

(Ipswich, MA, USA). Digestion condition was followed the procedure from manufacture.

Table 4.1 List of disaccharides and oligosaccharides being studied.

Analytes were purchased from: a Carbosynth, Ltd. (Berkshire, UK) and b Sigma-Aldrich, Inc. (St. Louis, MO, USA).

4.2.2 Reducing End Modification (Amination without Reduction)

Similar to the reductive amination except for using sodium cyanoborohydride was used. Briefly, disaccharides or oligosaccharides (0.5 mg) were dissolved in 10 μL of a solution containing 0.15 M ABA, ABEE, or APA in acetic acid: DMSO (3:7, v/v), and the solution was incubated at 60 °C for 10-12 h. The solvent was removed under vacuum

(Speed-Vac, LABCONCO, Kansas City, MO) and the residue was dissolved in 300 μL water followed by an addition of 300 μL chloroform. The aqueous phase containing the derivatized sugar was collected after vortex mixing. This step was repeated with another

300 μL of water for extraction and the aqueous phases were combined and vacuum dried. 79

The dried sample was dissolved in 100 μL DMSO and 900 μL methanol. This solution was further diluted 100-fold in methanol prior to MS analysis.

4.2.3 Mass Spectrometry

All samples were analyzed in the negative ion mode on a hybrid triple quadrupole/linear ion trap mass spectrometer (QTRAP4000 Applied Biosystems,

Toronto, Canada) equipped with a home-built nanoelectrospray ionization (nanoESI) source. A schematic diagram of the instrument is shown in Scheme 2.2. Two types of low energy collisional activation were accessible on this instrument, i.e. beam-type CID and ion trap CID (via on-resonance single frequency excitation). In beam-type CID, the precursor ions were isolated in Q1, accelerated to the Q2 collision cell for collisional activation, and all products were analyzed in the Q3 linear ion trap. The collision energy

(CE, in the lab frame) was defined by the difference of the dc rod offsets between Q0 and

Q2. For ion trap CID, the precursor ions were isolated in the Q3 linear ion trap via

RF/DC mode and a dipolar excitation was applied at a Mathieu q value of 0.23. MS2 CID experiments were carried out by either beam-type or ion trap CID. Higher stages of CID, i.e., MS3 to MS5, were carried out by ion trap CID in Q3. MSn (n≥3) experiments were performed through access to the original method table. Analyst 1.5 software was used for instrument control, data acquisition, and processing. The typical parameters of the mass spectrometer used in this study are listed in Supporting Information. The standard

18 spectra of Z1 ions (either m/z 163 from O-labeled glucose homodimers or m/z 161 from intact glucose homodimers) were generated based on the average of 6 spectra over a 6- month period. Standard deviations of peak heights were calculated for major 14 80 fragments including m/z 73, 83, 89, 97, 101, 103, 113, 115, 131, 133, 135, 143, and 145 for CID spectra of m/z 163 and m/z 73, 83, 87, 89, 97, 101, 103, 113, 115, 125, 131, 133,

143, and 161 for CID spectra of m/z 161. These peaks were also used to calculate spectral similarity scores.

4.2.4 HPLC

ABEE modified branched mannose 5 and exoglycosidase digested mannose 3 were separated using Agilent 1200 series HPLC system (Agilent Technologies, Santa

Clara, CA, USA). Separation was carried out on a HILIC (PolyLC, Inc., Columbia, MD,

USA) at a flow rate of 1 mL/min with a linear gradient of 20-35% solvent A in 30 min.

Solvent A was a mixture of 5mM ammonium acetate in water and solvent B contained 5 mM ammonium acetate in 100% acetonitrile. The eluent was detected at a wavelength of

305 nm.

4.3 Results and Discussion

3 4.3.1 Disaccharide Linkage Analysis Based on MS CID of Z1 Ions

Disaccharides, the smallest substructures of oligosaccharides containing linkage information between two complete sugar molecules were used as model systems for evaluating mass spectra of product ions that may yield detailed information about the linkage positions. Initially, we used 18O-labeling at the carbonyl oxygen of 15 disaccharides to introduce a mass discrimination between X, Y, and Z vs. A, B and C 81 product ions,27 recognizing that some of the ions might be isomeric and need to be resolved as isotopomers. MS2 beam-type CID spectra of m/z 343 precursor ions of glucose homodimers (D-anomeric configurations) with four different linkage positions are shown in Figure 4.1. In these spectra, m/z 223, 253, 265, and 283 ions have incorporated 18O (compared to the non-labeled data), suggesting that these fragments contained a labeled component derived from the reducing sugar unit. Depending on the disaccharide, some of the other product ions were observed as pairs varying in mass by 2 m/z units, the higher m/z isomer not having been observed in unlabeled disaccharides. Of pertinence to this paper, peaks at m/z 161 and 163 were observed, indicative of two isomers which would have been indistinguishable without 18O-labelling. A zoomed-in region showing m/z 161 and 163 product ions is presented in insets for each linkage position. Ions at m/z 163 should be Z1 ions, formed by cleavage at the reducing side of the glycosidic oxygen. This ion species is unlikely to be formed as a sequential water loss

2 18 from Y1 ions (m/z 181), which we have not observed for MS CID of O-labeled disaccharides. The assignment is also supported by the fact that MS3 CID of m/z 163 product ions showed totally different fragmentation patterns as compared to those

18 generated from H2O loss of O-labeled monosaccharides (data not shown). Furthermore,

MS3 CID of m/z 161 product ions (Figure 4.2) from the 18O-labeled disaccharides yielded mass spectra that were very different from those derived from the m/z 163 ions, suggesting that they have different origins and structures. These above results indicated that from disaccharides, m/z 163 product ions were generated by dissociation pathways unique to the substitution position of the reducing sugar. 82

(a) α-D-Glcp-(1-2)-D-Glc-18O 265 343 100

179 50 343 161 163 223 325

Rel. Int. (%) Int. Rel. 113 119 101 143161 0 60 100 140 180 220 260 300 340 m/z (b) α-D-Glcp-(1-3)-D-Glc-18O 115 343 100

163 161 163 343 50 161 113 179

Rel. Int. (%) Int. Rel. 283 145 227 89 143 281 253 325 0 60 100 140 180 220 260 300 340 m/z (c) α-D-Glcp-(1-4)-D-Glc-18O 343 163 100

50 161 161 163 145 343

Rel. Int. (%) Int. Rel. 143 179 281 103 263 283 0 60 100 140 180 220 260 300 340 m/z (d) α-D-Glcp-(1-6)-D-Glc-18O 343 100 179

343

161 163 50 119 163 221 113 161 325 Rel. Int. (%) Int. Rel. 323 101 143 251 281 0 60 100 140 180 220 260 300 340 m/z

Figure 4.1 MS2 beam-type CID of 18O-labeled α-D-glucose homodimers, m/z 343 ([M- H]-), with different linkage positions: (a) 1-2, CE: 8 V; (b)1-3, CE: 12 V; (c) 1-4, CE: 10 V; and (d) 1-6, CE: 10 V. Insets in each of the spectra show the expanded region covering m/z 161 and 163. 83

(a) α-D-Glcp-(1-2)-D-Glc-18O (b) α-D-Galp-(1-3)-D-Gal-18O 113 343 113 343 100 100 161 161 143 131 50 50 143 161 Rel. Int. (%) Int. Rel. 161 (%) Int. Rel. 131 83 87 101 125 0 60 80 100 120 140 160 0 m/z 60 80 100m/z 120 140 160 (c) α-D-Glcp-(1-3)-D-Glc-18O (d) α-D-Manp-(1-4)-D-Man-18O 113 343 113 343 100 100 161 161 131 131 143 50 143 50 83 101 101 161 Rel. Int. (%) Int. Rel.

Rel. Int. (%) Int. Rel. 73 87 125 161 8387 0 0 60 80 100m/z 120 140 160 60 80 100m/z 120 140 160 (e) α-D-Glcp-(1-4)-D-Glc-18O (f) β-D-Galp-(1-4)-D-Man-18O 343 113 343 100 113 100 161 161 131 131 50 143 50 143 161 125 161 (%) Int. Rel. 83 101 Rel. Int. (%) Int. Rel. 8387 101 0 0 60 80 100 120 140 160 60 80 100m/z 120 140 160 m/z (g) α-D-Glcp-(1-6)-D-Glc-18O (h) α-D-Galp-(1-6)-D-Glc-18O 113 343 113 343 100 100 161 161 143 131 131 143 50 50 161 161 Rel. Int. (%) Int. Rel. Rel. Int. (%) Int. Rel. 8387 101 125 101 0 60 80 100 120 140 160 0 m/z 60 80 100m/z 120 140 160

Figure 4.2 MS3 ion trap CID of m/z 161 product ions derived from 18O-labeled disaccharides (D-anomeric configuration) with different sugar identities and linkage positions: (a) α-D-Glcp-(1-2)-D-Glc, (b) α-D-Galp-(1-3)-D-Gal, (c) α-D-Glcp-(1-3)-D- Glc, (d) α-D-Manp-(1-4)-D-Man, (e) α-D-Glcp-(1-4)-D-Glc, (f) β-D-Galp-(1-4)-D-Man, (g) α-D-Glcp-(1-6)-D-Glc, and (h) α-D-Galp-(1-6)-D-Glc.

3 18 MS ion trap CID spectra of m/z 163 ions (Z1 ions) were collected for 15 O- labele disaccharides. Figure 4.3 shows the data sets obtained from glucose homodimers with 4 different linkage positions. Note that each panel is an average of 6 spectra (3 spectra each from α- and β-anomeric configurations) collected over 6 months. The error bars on the major peaks represent the standard deviation of the six spectra. In the standard spectra, CE was controlled to reduce the parent ion intensity to 18% of the base peak, which was found to provide good reproducibility of the fragmentation patterns. Clearly,

CID of m/z 163 produces distinct fragmentation patterns characteristic of the linkage positions. There are 13 common peaks (>3% relative intensity) observed from these spectra, however with quite different relative intensities for different linkage positions.

These include ions at m/z 73, 83, 89, 97, 101, 103, 113, 115, 131, 133, 135, 143, and 145.

For the 1-2 linkage (Figure 4.3a), the base fragment peak is at m/z 103 and other signature peaks include m/z 101, 115, 133, 135, and 145. We did notice that the α- and β- anomers showed a small difference in the CID fingerprints, which was reflected by the larger standard deviations of peak intensities at m/z 115 and 145 (22% and 9%), respectively. These two peaks are of higher relative intensities in the β-anomer as compared to the α-anomer. The comparisons of MS3 CID of m/z 163 from the α- and β- anomers can be found in Figure 4.4. For the 1-3 linkage (Figure 4.3b), the base peak shifts to m/z 115 and peaks at m/z 115 and 145 are relatively abundant. All other fragment ions are present with relatively low intensities (<5%). The signatures of the 1-4 linkage

(Figure 4.3c) include dominant fragments at m/z 83 and 103 (base peak), while other major fragments (m/z 73, 89, 97, 115, 133, and 145) have similar intensities around 20-

30%. The 1-6 linkage (Figure 4.3d) produces a relatively simple fragmentation spectrum 85 with m/z 101 as the base peak, while other fragments (m/z 103, 131, and 135) show relative intensities lower than 10%. Note that these fragmentation patterns are very distinctive according to the linkage positions and repeatable as reflected by the relatively small standard deviations of all major peaks. These characteristics of Z1 fragmentation make it suitable as a diagnostic ion for linkage determination. With 18O-labeling on the reducing sugar carbonyl group, dissociation profiles of the m/z 163 ion enable the linkage position to be determined with high confidence.

Another interesting aspect that we discovered from CID of Z1 ions from 15 disaccharides was that neither the stereochemistry (identity) of the sugar units nor the anomeric configuration (except for the 1-2 linkage discussed above) had any noticeable effects on the fragmentation patterns. CID of Z1 ions from 7 disaccharides other than glucose homodimers were compiled and shown in Figure 4.5. For example, the five different 1-4 linked disaccharides (Figure 4.5a to e) showed almost identical fragmentation patterns to that of the 1-4 linked glucose homodimer (Figure 4.3c). This was consistently observed for other disaccharides containing 1-3 and 1-6 linkages (for instance, compare Figure 4.5f to Figure 4.3b and Figure 4.5g to Figure 4.3d). Although this phenomenon precludes obtaining stereo-structural information from the m/z 163 (Z1) product ion (i.e., identity and anomeric configuration) for the reducing sugar unit, it is advantageous for linkage determination due to the simplification of producing data from standards and in making comparisons to unknowns for spectral analysis.

(a) 1-2 linkage 103 100 343

115 133 163 101 135145 163 Rel. Int., %

0 60 80 100 120 140 160 180 m/z (b) 1-3 linkage 115 100 343

163 145 163 Rel. Int., %

0 60 80 100 120 140 160 180 m/z (c) 1-4 linkage 103 100 83 343

163 73 97 115 133 145 163 Rel. Int., % 89 0 60 80 100 120 140 160 180 m/z (d) 1-6 linkage 101 100 343

163

163 Rel. Int., % 131 103 135 0 60 80 100 120 140 160 180 m/z

3 18 Figure 4.3 Averaged MS ion trap CID data of Z1 ions (m/z 163) derived from O- labeled glucose homodimers with different linkage positions: (a) 1-2, (b) 1-3, (c) 1-4, and (d) 1-6. The error bars represent standard deviations of peak intensities based on 6 averaged spectra (3 spectra each from D- and E-anomers) obtained over a 6-month period. 87

α-D-Glcp-(1-2)-D-Glc-18O -H -

103 100 (a) 343 163 133 50 101 135 163 Rel. Int. (%) Int. Rel. 113 143 0 60 80 100 120 140 160 m/z β-D-Glcp-(1-2)-D-Glc-18O -H -

103 100 343 (b) 133 163 115 50 101 135 Rel. Int. (%) Int. Rel. 113 145 163 131 143 0 60 80 100 120 140 160 m/z

Figure 4.4 MS3 CID spectra of m/z 163 product ions from (a) α-D-Glcp-(1-2)-D-Glc-18O and (b) β-D-Glcp-(1-2)-D-Glc-18O. Additional peaks (m/z 115 and 145) were found in the β configuration.

(a) α-D-Manp-(1-4)-D-Man-18O (b) β-D-Galp-(1-4)-D-Glc-18O 83 103 100 103 343 100 343

163 163 83

50 50 73 Rel. Int. (%) Int. Rel. 73 (%) Int. Rel. 133 115 133 163 97 115 145 163 89 97 145 89

0 60 80 100 120 140 160 0 60 80 100 120 140 160 m/z m/z (c) α-D-Galp-(1-4)-D-Gal-18O (d) β-D-Galp-(1-4)-D-Man-18O 83 103 103 100 343 100 83 343 163 163

50 50 73 73 Rel. Int. (%) Int. Rel. (%) Rel. Int. 133 97 115 133 145 163 115 145 163 89 89 97

0 60 80 100 120 140 160 0 60 80 100 120 140 160 m/z m/z (e) α-D-Manp-(1-4)-D-Glc-18O 103 100 83 343

163

50 73 Rel. Int. (%) Int. Rel. 97 115 133 145 163 89

0 60 80 100 120 140 160 m/z 18 (f) α-D-Galp-(1-3)-D-Gal-18O (g) α-D-Galp-(1-6)-D-Glc- O 115 101 100 343 100 343

163 163

50 50 145 Rel. Int. (%) Int. Rel. Rel. Int. (%) Int. Rel. 83 103 163 113 163 103 131135 0 0 60 80 100 120 140 160 60 80 100 120 140 160 m/z m/z

Figure 4.5 MS3 ion trap CID of m/z 163 product ions from (a) α-D-Manp-(1-4)-D-Man- 18O, (b) β-D-Galp-(1-4)-D-Glc-18O, (c) α-D-Galp-(1-4)-D-Gal-18O, (d) β-D-Galp-(1-4)- D-Man-18O, (e) α-D-Manp-(1-4)-D-Glc-18O, (f) α-D-Galp-(1-3)-D-Gal-18O, and (g) α-D- Galp-(1-6)-D-Glc-18O. 89

4.3.2 The Effect of CID Conditions on the Formation of Z1 Ions

The data in Figure 4.3 demonstrate that CID of Z1 ions (m/z 163) can be used for linkage determination. Note that the formation of Z1 ions was typically accompanied by its structural isomer, ions at m/z 161 (insets of Figure 4.1). Collisional activation of the m/z 161 ions from 18O-labeled disaccharides did not provide any distinguishable fragmentation patterns according to their sugar unit identities, anomeric configurations, or linkage types within the disaccharides. Although reducing-end labeling with 18O can be used to clearly distinguish Z1 ions vs. m/z 161 ions, it was also highly desirable to optimize the formation of Z1 ions while minimizing the abundance of their structural isomer(s) (m/z 161) when studying native disaccharides or disaccharide units formed by gas-phase dissociation of oligosaccharides. We investigated the effect of CID (beam-type

18 and ion trap) conditions on the formation of Z1 ions vs. their structural isomers using O- labeled glucose homodimers (1-2, 1-3, 1-4, and 1-6 linkages and both α and β anomers).

The results are summarized in Figure 4.6 showing one anomer type as an example for each linkage because the same trend of fragmentation behavior was observed for both anomeric configurations. These spectra were collected using a wide isolation window

(around 5 Da) to observe both m/z 161 and 163 ions. Data from three different CEs (low, medium, and high) under beam-type CID (left column) and ion trap CID (right column) are compared side-by-side.

The best condition for forming Z1 ions from 1-2 linked disaccharides was the lower energy ion trap CID (Figure 4.6a, 5 mVpp), from which m/z 163 accounted for 80% of the summed intensities of m/z 161 and 163. Higher CE ion trap CID (7.5 and 10 mVpp) significantly increased the relative intensity of m/z 161. Note that when beam-type CID 90

was employed, m/z 161 was the dominant peak, while Z1 ions (m/z 163) were not detected above noise level. On the other hand, 1-3 and 1-4 linked disaccharides showed dominant formation of Z1 ions relative to m/z 161 under both beam-type and ion trap CID conditions (Figure 4.6b and c, respectively). For 1-6 linked disaccharides, the best conditions for the formation of Z1 ions were found using lower CEs for both beam-type and ion trap CID as shown in Figure 4.6d. The data in Figure 4.6 clearly demonstrate that the formation of m/z 161 and 163 is strongly affected by collisional activation conditions and the best common condition to obtain relatively pure Z1 ion for all the linkage types was to use ion trap CID with relatively low energy (around 5 mVpp). It should be noted that even using these optimized conditions, the purity of Z1 ions from 1-2 linked disaccharides was about 80%. After determining the CID conditions for preferential formation of Z1 ions, the spectra in Figure 4.3 were re-collected for native unlabeled glucose homodimers with relatively low energy ion trap CID. The data are shown in

Figure 4.7. Characteristic and highly reproducible fragmentation patterns were obtained,

18 which were almost identical to the CID of Z1 ions derived from O-labeled samples. The data indicated that under the optimized CID conditions, the existence of small impurities within the Z1 ions did not significantly affect their distinct fragmentation patterns. These m/z 161 CID spectra were later used as standard spectra for Z1 ion fragmentation derived from native disaccharides or disaccharide substructures of oligosaccharides.

3 We also performed MS CID of Z1 ions from two fructose containing disaccharides (α-D-Glcp-(1-3)-D-Fru and β-D-Galp-(1-4)-D-Fru) with and without 18O- labeling. Fructose has a ketone group at C2 position and 18O is incorporated in hydroxyl group at C2 position as compared to the C1 position for aldohexoses. Collisional 91

18 activation of Z1 ions from the two O-labeled fructose containing disaccharides showed characteristic fragmentation patterns of 1-3 and 1-4 linkages (Figure 4.8a and c), respectively. The fragment fingerprint patterns are identical to the data obtained from the

1-3 and 1-4 linked glucose homodimers (Figure 4.3b and c). The 1-4 linked fructose containing disaccharide data (Figure 4.8c), however, showed 2 m/z shifts for fragments such as m/z 75, 87, 99, 117, and 135, as compared to the aldohexose containing disaccharide data (Figure 4.3c). The mass shifts are likely contributed by the difference of 18O-labeling position between an aldohexose and a ketohexose. Using the optimized

3 conditions for Z1 ion formation (ion trap CID, 5 mVpp), MS CID of Z1 ions of the two unlabeled fructose containing disaccharides also showed characteristic fragmentation patterns of 1-3 and 1-4 linkages (Figure 4.8b and d). These results confirmed that MS3

CID of Z1 ions can be applied to disaccharide units containing fructose for determination of linkage position.

Beam-type CID Ion trap CID

(a) 1-2 linkage 5 5 10 7.5 15 10 161 163 161 163 m/z m/z

(b) 1-3 linkage 5 5 10 7.5 15 10 161 163 161 163 m/z m/z

(d) 1-6 linkage 5 5 10 20 15 30 161 163 161 163 m/z m/z

Figure 4.6 The effect of CID conditions on the formation of Z1 (m/z 163) vs. its structural isomers (m/z 161). Data were obtained from 18O-labeled glucose homodimers: (a) α-D- Glcp-(1-2)-D-Glc, (b) β-D-Glcp-(1-3)-D-Glc, (c) β-D-Glcp-(1-4)-D-Glc, and (d) α-D- Glcp-(1-6)-D-Glc. 93

(a) 1-2 linkage 101 100 113 341

161 131 143 161 Rel. Int., % 133 0 60 80 100 120 140 160 180 m/z (b) 1-3 linkage 113 100 341

161 143 161 Rel. Int., %

0 60 80 100 120 140 160 180 m/z (c) 1-4 linkage 83 101 100 341

161

73 115 97 133143 161 Rel. Int., % 87 0 60 80 100 120 140 160 180 m/z (d) 1-6 linkage 101 100 341

161

Rel. Int., % 131 0 60 80 100 120 140 160 180 m/z

3 Figure 4.7 Averaged MS ion trap CID data of Z1 product ions (m/z 161) derived from native unlabeled glucose homodimers with different linkage positions: (a) 1-2, (b) 1-3, (c) 1-4, and (d) 1-6. The error bars show standard deviation of peak intensities based on 6 averaged spectra (3 spectra each from D- and E-anomers) obtained over a 6-month period. 2 Ion trap CID with 5 mVpp was used for MS CID of m/z 341 step for (a) through (d).

(a) α-D-Glcp-(1-3)-D-Fru-18O (c) β-D-Galp-(1-4)-D-Fru-18O 115 103 100 343 100 83 343

163 163

50 145 50 75 Rel. Int. (%) Rel. Int. 163 (%) Rel. Int. 99 117 135 163 87 145

0 60 80 100 120 140 160 0 60 80 100 120 140 160 m/z m/z (b) α-D-Glcp-(1-3)-D-Fru (d) β-D-Galp-(1-4)-D-Fru 113 101 100 100 341 83 341

161 161 143 50 50 73

Rel. Int. (%) Int. Rel. (%) Int. Rel. 133 161 97 115 143 161 87 0 60 80 100 120 140 160 0 60 80 100 120 140 160 m/z m/z

3 18 Figure 4.8 MS ion trap CID of the Z1 ion from O-labeled fructose containing 3 disaccharides and MS ion trap CID of the Z1 ion generated by ion trap CID with low energy from native unlabeled fructose containing disaccharides, (a) α-D-Glcp-(1-3)-D- Fru-18O, (b) α-D-Glcp-(1-3)-D-Fru, (c) β-D-Galp-(1-4)-D-Fru-18O, and (d) β-D-Galp-(1- 4)-D-Fru.

n 4.3.3 Linkage Determination for Oligosaccharides via MS CID of Z1 Ions

3 Given that linkage positions can be determined from MS CID of Z1 ions from deprotonated disaccharides, we further extended this method to linear oligosaccharides.

The approach involved optimizing glycosidic bond cleavages during early stages of MSn so that an ordered set of overlapping disaccharide substructures could be formed, generating Z1 ions from each disaccharide unit, and finally obtaining the fragmentation fingerprints of Z1 ions for linkage determination. The concept is illustrated with a pentasaccharide in Scheme 4.1, with the numbering of the sugar units starting from the 95 reducing-end (sugar 1). The key steps include: 1) dissociation of a precursor derivatized at the reducing-end (M, Scheme 1) to a ladder of smaller oligosaccharides, each successively one sugar unit shorter, containing the reducing-end tag (Ym ions); 2) generation of disaccharide fragments (C2 ions) from each of the Ym ions; 3) dissociation of the disaccharide units to form Z1 ions; 4) CID of Z1 ions and 5) spectral matching of the Z1 CID data to the standard database (derived from disaccharides, Figure 4.3 or

Figure 4.7) for linkage determination. Linkage positions between the first and the second sugar units could be directly obtained from 18O-labled oligosaccharides via MS3 CID of m/z 163. The above approach has the advantage of knowing the exact origin of each fragment with respect to its initial position within the oligomer. The reducing-end modification M introduces a mass distinction between Y and C ions. In addition, the group used to derivatize the reducing- end bears a negative charge and enabled the

2 selection of the ladder of charged Ym ions after MS . Yet, through charge-transfer, it permitted disaccharide (C2) ions to be formed and isolated. For example, to obtain linkage information between sugar units 3 and 4, the following MS5 is needed: [[M-H]-

ÆY4ÆC2ÆZ1Æfragments] (shown in Scheme 4.1). 96

5 O 4 O 3 O 2 OM1 Y4 HO 4 O 3 O 2 OM1

C2 Gas-Phase Dissociation HO 4 O 3 OH

3 OH MS5 CID Linkage Position Unique Dissociation Pattern Disaccharide Standard Spectral Matching Database

Scheme 4.1 The MSn approach for linkage determination of an oligosaccharide. “M” stands for reducing-end modification.

Two trisaccharides and one pentasaccharide were used to test the MSn approach for extracting linkage information using this approach. The data in Figure 4.9 show a series of experiments (MS2 to MS4) using 18O-labeled α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-

D-Glc (Mw: 506 Da). In order to determine the linkage positions of linkages 1 and 2

(counting from the reducing- end), the following experiments were needed: MS3 CID

[505Æ163Æ fragments] and MS4 CID [505Æ341Æ161Æfragments]. MS2 CID (CE =

15 V, beam-type) of the deprotonated molecular ions ([M-H]-) is shown in Figure 4.9a, in

3 which the formation of Z1 ions (m/z 163) and C2 ions (m/z 341) can be clearly seen. MS

CID of Z1 ions (ion trap CID, 27 mVpp, Figure 4.9c) showed an almost identical spectrum to the 1-4 linkage standard spectrum (Figure 4.3c). Therefore, linkage 1 can be confidently identified as a 1-4 linkage. Note that the linkage could be assigned directly from the m/z 163 ion derived from the labeled trisaccharide without requiring prior 97 dissociation and isolation of the m/z 343 reducing-end disaccharide. This is particularly worth emphasizing as the m/z 505 precursor ion gave rise to negligible quantities of the m/z 343 product ion (Figure 4.9a), thus information could not be obtained for this linkage either from the direct dissociation pattern of its reducing disaccharide (m/z 343) or from the 163 product ion that may have been derived from it. Another important point is that

2 since m/z 161 ions are also formed in MS CID (due to sequential fragmentation of C1 ions), it is necessary to use m/z 163 ions to achieve correct linkage information for the

3 reducing end. Low energy MS CID (ion trap, 5 mVpp) of C2 ions (m/z 341, Figure 4.9a) produced a reasonable intensity of m/z 161 ions (Figure 4.9b). Note that the best CID conditions characterized from disaccharide studies (ion trap CID, 5 mVpp, 200 ms) were

4 used to favor the formation of Z1 ions relative to their structural isomers. Indeed, MS

CID (ion trap, 35 mVpp) of m/z 161 showed a fragmentation pattern characteristic of the

1-6 linkage position (compare Figure 4.9d to the standard spectrum in Figure 4.7d). For the other trisaccharide sample, 18O-labeled α-D-Galp-(1-3)-β-D-Galp-(1-4)-D-Gal, the same set of MS3 and MS4 experiments were required to obtain information for linkages 1 and 2 : ([505Æ163Æ fragments] and [505Æ341Æ161Æfragments]). Abundant Z1 and

2 C2 ions were formed under MS beam-type CID (CE = 15 V) as shown in Figure 4.10a.

3 MS CID (ion trap, 27 mVpp) of the Z1 ion clearly showed the distinct fragmentation

3 pattern of a 1-4 linkage position (Figure 4.9e). MS CID (ion trap, 5 mVpp) of C2 ions

4 (m/z 341) (Figure 4.10b) produced abundant Z1 ions (m/z 161). MS CID (ion trap, 35 mVpp) of the Z1 ions (m/z 161) optimized in abundance is shown in Figure 4.9f. The fragmentation pattern matched well to the 1-3 linkage standard spectrum (Figure 4.7b) and allowed confident assignment of this linkage position. 98

α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-18O, m/z 505

341 100 C2 505 505 (a) Z1 C1 179 383 50 161 163 143 113 221 323 Rel. Int. (%) Int. Rel. 101 161 281 425 0 100 200 300m/z 400 500 100 179 505 (b) 341 323 281 50 161 221 Rel. Int. (%) Int. Rel. 143 341 251 311 0 60 140 220 300 m/z 505 100 103 101 (c) 505 (d) 83 341 163 161 50

Rel. Int. (%) Rel. Int. 145 73 115 133 163 161 8997 131 0 60 80 100 120 140 160 60 80 100 120 140 160 m/z m/z α-D-Galp-(1-3)-β-D-Galp-(1-4)-D-Gal-18O, m/z 505 113 505 100 83 505 (e) 103 (f) 341 163 161 50

73 115 133 Rel. Int. (%) Int. Rel. 145 143 161 8997 163 0 60 80 100 120 140 160 60 80 100 120 140 160 m/z m/z The parent ion The product ion for CID in the next stage

Figure 4.9 MSn CID spectra for the determination of linkage types within trisaccharides. α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-18O: (a) MS2 beam-type CID: [505Æfragments], (b) MS3 ion trap CID: [505Æ341Æfragments], (c) MS3 ion trap CID: [505Æ163Æfragments], (d) MS4 ion trap CID: [505Æ341Æ161Æfragments]. α-D- Galp-(1-3)-β-D-Galp-(1-4)-D-Gal-18O: (e) MS3 ion trap CID: [505Æ163Æ fragments] and (f) MS4 ion trap CID: [505Æ341Æ161Æfragments]. 99

α-D-Galp-(1-3)-β-D-Galp-(1-4)-D-Gal-18O, m/z 505

100 161 505 Z1 (a) 163 C2 505 341

C1 50 113 179 443 Rel. Int. (%) Int. Rel. 143 323 342 425 0 100 200 300 400 500

161 341 505 100 (b) 179 341

50 113

Rel. Int. (%) Int. Rel. 143 281 323 0 60 140 220 300 m/z The parent ion The product ion for CID in the next stage

Figure 4.10 MS2 and MS3 CID spectra from α-D-Galp-(1-3)-β-D-Galp-(1-4)-D-Gal-18O. (a) MS2 beam-type CID: [505Æfragments], and (b) MS3 ion trap CID: [829Æ341Æfragments].

In order to determine the individual linkages within the glucose homopentamer,

n [β-D-Glcp-(1-4)]4-D-Glc using the MS approach, formation of the following ions at the

2 18 MS stage was a prerequisite from the O-labeled sample: m/z 163 (Z1), 341 (C2), 505

2 (Y3), and 667 (Y4). MS CID of the deprotonated molecular ion (m/z 829, Figure 4.11a)

produced major peaks at m/z 341 (C2), 503 (C3), 665 (C4), and also a small amount of Z1

ions (m/z 163). However, no Y ions were detected above the noise level, which made it

impossible to conduct MSn to determine linkages 2 and 3. Reducing-end derivatization

with different types of functional groups has been shown to alter the fragmentation

patterns of oligosaccharides and generates X, Y, and Z product ions of

oligosaccharides.17 We tried reductive amination with ABA and ABEE at the reducing 100 sugar to form the reduced form of the corresponding Schiff base. Only data from ABEE are discussed here since higher intensity of the desired ions were observed. The full series

2 5 of MS to MS spectra of ABEE modified [β-D-Glcp-(1-4)]4-D-Glc is shown in Figure

4.12. MS2 beam-type CID of the doubly deprotonated molecular ion, m/z 487 (Figure

2- 4.12a), produced a variety of C and Y ions: m/z 341 (C2), 406 (Y4 ), 503 (C3), 652 (Y3), and 814 (Y4). The formation of reasonable intensities of Y3 and Y4 allowed further stages of MS/MS. Finally, the following MSn CID sequences were employed for the determination of linkages 1-4 individually: linkage 1, MS3 CID [829Æ163Æfragments] from the 18O-labeled sample; linkage 2, MS5 CID [487(2-

)Æ652Æ341Æ161Æfragments] from the ABEE labeled sample; linkage 3, MS5 CID

[487(2-)Æ406(2-)Æ341Æ161Æfragments] from the ABEE labeled sample; and linkage

4, MS4 CID: [829Æ341Æ161Æfragments] from 18O-labeled sample. These four spectra all showed characteristic fragmentation patterns of the 1-4 linkage position, allowing the linkage position to be confidently assigned as 1-4. The data for individual stages of

MS/MS within each sequence of MSn CID can be found in Figure 4.11 (for the 18O- labeled sample), and 4.12 (for the ABEE labeled sample).

101

18 C [β-D-Glcp-(1-4)]4-D-Glc- O, m/z 829 C 4 100 3 665 503 829 -H - (a) C C , m/z 341 Z , m/z 163 2 2 1 341

50 Z1 161 163 Rel. Int. (%) Int. Rel. 485 323 587 749 829 161 281 425 0 200400 600 800 m/z 83 103 829 161 829 83 829 100 100 100 101 (b) 163 (c) 341 (d) 341

161 50 50 341 50 115 117 73 Rel. Int. (%) Int. Rel. 89 73 75 97 135 145 163 179 115 133 161 281 8797 125 143 0 0 0 60 80 100 120 140 160 60 140 220 300 380 60 80 100 120 140 160 m/z m/z m/z The parent ion The product ion for CID in the next stage

2 4 18 Figure 4.11 MS and MS CID spectra from [β-D-Glcp-(1-4)]4-D-Glc- O. Experimental step for linkage 1 analysis is [829Æ163Æfragments], (a) MS2 beam-type CID: [829Æfragments], (b) MS3 ion trap CID: [829Æ163Æfragments]. Experimental step for linkage 2 analysis is [829Æ341Æ161Æfragments], (c) MS3 ion trap CID: [829Æ341Æfragments], and (d) MS4 ion trap CID: [829Æ341Æ161Æfragments].

102

487 [β-D-Glcp-(1-4)]4-D-Glc-ABEE, m/z 487(2-) 100 310 487(2-) 2- (a) -2H

50 C Y 2- Y3 Y4 2 4 652 Rel. Int. (%) Int. Rel. 341 C3 814 Y3, m/z 652 406 503 545 Y4, m/z 406(2-) 0 200 400 600 800 m/z 652 100 100 161 100 83 487(2-) (b) 487(2-) (c) 487(2-) (d) 101 652 652 652

341 341 50 50 341 50 161 73

Rel. Int. (%) Int. Rel. 328 179 115 133 161 310 341 472 490 606 87 97 113 125 143 0 0 0 200 400 600 100 180 260 340 60 80 100 120 140 160 m/z m/z m/z 161 100 652 100 100 83 (e) 487(2-) (f) 487(2-) (g) 487(2-)

406(2-) 406(2-) 101 406(2-) 310 341 50 50 341 341 50 161 161

Rel. Int. (%) Int. Rel. 407 161 250 73 179 115 133 143 341 281 87 97 125 0 0 0 200 400 600 100 180 260 340 60 80 100 120 140 160 m/z m/z m/z The parent ion The product ion for CID in the next stage

2 5 Figure 4.12 MS to MS CID spectra from ABEE modified [β-D-Glcp-(1-4)]4-D-Glc. Experimental step for linkage 2 analysis is [487(2-)Æ652Æ341Æ161Æfragments], (a) MS2 beam-type CID: [487(2-)Æfragments], (b) MS3 ion trap CID: [487(2- )Æ652Æfragments], (c) MS4 ion trap CID: [487(2-)Æ652Æ341Æfragments], and (d) MS5 ion trap CID: [487(2-)Æ652Æ341Æ161Æfragments]. Experimental step for linkage 3 analysis is [487(2-)Æ406(2-)Æ341Æ161Æfragments], (e) MS3 ion trap CID: [487(2-)Æ406(2-)Æfragments], (f) MS4 ion trap CID: [487(2-)Æ406(2- )Æ341Æfragments], and (g) MS5 ion trap CID: [487(2-)Æ406(2- )Æ341Æ161Æfragments].

As mentioned above, dissociation patterns of disaccharide ions (m/z 341) in the negative ion mode have been used to assign linkages using sector instruments, triple quadrupoles, Fourier transform ion cyclotrons and ion traps.7-11,13,24 Disaccharides having

2- or 6-linkages can be readily assigned in ion traps directly by MS2 as they yield relatively abundant product ions of m/z 221 and 263, or m/z 221, 251 and 281, respectively. For example, MS3 CID of the disaccharide substructure (α-D-Glcp-(1-6)-D- 103

Glc, m/z 341) from α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-18O (Figure 4.9b) produced abundant ions of m/z 221, 251, and 281, characteristic of the 1-6 linkage. However, especially for 3-linkages and to some extent for 4 linkages, depending on the specific sugars involved, the A ions derived by dissociation processes occurring within the reducing sugar can be relatively low in abundance (frequently less than 5% of the base product peak) with very abundant product ions observed at m/z 161 and 179 due to cleavage on either side of the glycosidic oxygen. In these cases, further isolation/dissociation of the m/z 161 ions are highly preferred specifically because, under defined dissociation conditions, they are nearly entirely comprised of the diagnostic Z1 isomers derived from the reducing sugar (Fig. 4.6b and c). For example, MS3 CID of the disaccharide substructure (α-D-Galp-(1-3)-D-Gal, m/z 341) derived from α-D-Galp-(1-3)-

β-D-Galp-(1-4)-D-Gal-18O (Figure 4.10b) only revealed one observable cross-ring fragment at m/z 281 (signature fragments of the 1-3 linkage: m/z 251 and 281 but no

263), making it difficult to assign the linkage position. Thus for 3- and 4-linked disaccharides, the CID data of their Z1 (m/z 161) ions are particularly useful. Another key advantage is that the reducing sugar of an oligosaccharide can be selectively 18O- labeled, whereby the Z1 product ion obtained directly from the reducing sugar (m/z 163) can be used to determine the reducing-end linkage, without requiring prior isolation of a reducing disaccharide fragment. In some cases this is very valuable because the reducing disaccharide may only be produced as a trace product ion or not, apparently, at all.

104

4.3.4 Linkage Determination for a Branched Oligosaccharide

In earlier discussions, we have shown that the Z1 ion approach can be applied to linear oligosaccharides. However, in a real situation, branched oligosaccharides are often encountered, such as N-linked glycans. We explored a way to convert a branched oligosaccharide into linear structure using exoglycosidases and applied Z1 ion approach for the smaller linear structures. Branched mannose 5 (Man5, having linkages of α1-3 and α1-6, structure shown in Scheme 4.2) is one of the common core structure in N- glycans and was chosen as a model to test this method. First step is the chromophore attachment to the reducing end to enable detection during later separation processes.

Schiff base formation with ABEE (without reduction) was employed for the initial tests.

Modified Man5 was purified by HPLC to eliminate substantial amount of impurities (i.e. smaller oligosaccharides and isomers from the supply of commercial sample). ABEE modified Man5 eluted around 10 min as shown in the chromatogram, Figure 4.13a.

There are two branching points within Man5 and they need to be cleaved to produce a linear structure. An exoglycosidase, α1-2,3 Mannosidase, which cleaves off α1-2 and α1-

3 bonds was used to convert Man5 into a linear structure as shown in step 3. After purification (step 4), a linear trisaccharide, α-D-Manp-(1-6)-α-D-Manp-(1-6)-D-Man-M, was obtained. This ABEE modified Man3 came out around 7 min in Figure 4.13b. The two HPLC separations (steps 2 and 4) are very important to eliminate possible impurities and extract only the digested products originally from Man5. Step 5 is the delabeling of chromophore and step 6 is 18O-labeling. 105

Step 1 Reducing end labeling (M)

Step 2 HPLC separation

Step 3 α1-2,3 Mannosidase digestion Step 4 HPLC separation M Step 5 Delabeling (M)

Step 6 18O-labeling 18O

Step 7 MS analysis

Scheme 4.2 Sample preparation steps for branched oligosaccharide before MS analysis.

106

A set of MS3 and MS4 experiments was required to obtain information for linkages 1 and 2 within 18O-labeled α-D-Manp-(1-6)-α-D-Manp-(1-6)-D-Man:

([505Æ163Æfragments] and [505Æ341Æ161Æfragments]). Z1 and C2 ions were formed under MS2 beam-type CID (CE = 20 V) as shown in Figure 4.14a. MS3 CID (ion trap, 10 mVpp) of the Z1 ion clearly showed the distinct fragmentation pattern of a 1-6

3 linkage position (Figure 4.14c). MS CID (ion trap, 5 mVpp) of C2 ions (m/z 341) (Figure

4 4.14b) produced Z1 ions (m/z 161). MS CID (ion trap, 20 mVpp) of the Z1 ions (m/z 161) is shown in Figure 4.14d. The fragmentation pattern of Figure 4.14c and d matched well to the 1-6 linkage standard spectra (Figure 4.3d and 4.7d, respectively) and allowed confident assignment of this linkage position.

The total amount of Man5 used for this experiment was only 1.2 nmol which is compatible to permethylation method. Permethylation has the advantage of obtaining the linkage information from the whole structure, it might experience the miss-assignment of linkage positions due to ion suppression. In that case, Z1 method with a combination of exoglycosidase can be used as a complimentary method to obtain confident linkage information.

107

(a) Step 2 HPLC sepration

2000

1600 Man5-M

1200

mAU 800

400

0 036912 Retention time (min) (b) Step 4 HPLC sepration 140 120 100 80

mAU 60 Man3-M 40 20 0 036912 Retention time (min)

Figure 4.13 HPLC chromatograms from (a) step 2 and (b) step 4 separations.

108

α-D-Manp-(1-6)-α-D-Manp-(1-6)-D-Man-18O, m/z 505

100 383 505 (a)

Z1 50 161 163 C2 341 Rel. Int. (%) Int. Rel. 443 505 179 281 413 0 100 200 300 400 500 m/z 505 100 341 (b) 341 281

50 179 Rel. Int. (%) Int. Rel. 161 221 251 323 0 60 140 220 300 m/z 101 505 505 100 101 (c) (d) 163 341 161 50 161

Rel. Int. (%) Int. Rel. 103 163 131 133135 0 60 80 100 120 140 160 60 80 100 120 140 160 m/z m/z

Figure 4.14 MSn CID spectra for the determination of linkage types within trisaccharides digested from Man5, α-D-Manp-(1-6)-α-D-Manp-(1-6)-D-Man-18O: (a) MS2 beam-type CID: [505Æfragments], (b) MS3 ion trap CID: [505Æ341Æfragments], (c) MS3 ion trap CID: [505Æ163Æfragments], (d) MS4 ion trap CID: [505Æ341Æ161Æfragments].

109

4.3.5 Algorithm-Assisted Linkage Determination

For the possibility of automation of the linkage analysis in future developments, it is desirable to use an algorithm for spectral-matching. Spectral similarity scores can be

n calculated between MS CID spectra of Z1 ions from oligosaccharides following selected dissociation pathways and the averaged CID spectra of Z1 ions from the disaccharide standards originating from 2-, 3-, 4-, or 6-linkages. The same equation introduced in

Chapter 2 (Eq. 2.1) was used.

For disaccharide standards, two sets of standard spectra were prepared. One was based on the averaged CID of m/z 163 spectra as shown in Figure 4.3 and the other was generated based on the averaged CID of m/z 161 spectra formed under ion trap CID with low energy from native glucose-dimers shown in Figure 4.7. The spectral similarity score for CID of m/z 163 from unknowns was calculated against CID of m/z 163 from standards while the spectral similarity score for CID of m/z 161 from unknowns was calculated against CID of m/z 161 from standards. Table 4.2 summarizes the similarity

n scores of MS CID spectra of Z1 ions for each linkage derived from the three oligosaccharides studied. The highest spectral similarity scores always corresponded to the correct linkage types with the values very close to unity. Note that the similarity scores for incorrect linkage assignments were typically smaller than 0.8.

110

n Table 4.2 Spectral similarity scores of MS CID of Z1 ion spectra derived from oligosaccharides vs. disaccharide standards.

Linkage Disaccharide Standards (from Oligosaccharides reducing 1-2 1-3 1-4 1-6 -end)

α-D-Glcp-(1-6)-α- 1 0.7583 0.7053 0.9917 0.4566 D-Glcp-(1-4)-D-Glc 2 0.8668 0.5212 0.7971 0.9944 α-D-Galp-(1-3)-β- 1 0.7603 0.7045 0.9912 0.4180 D-Galp-(1-4)-D- Gal 2 0.7223 0.9937 0.4805 0.4848 1 0.7722 0.6979 0.9702 0.5169 2 0.7504 0.4878 0.9917 0.7907 [β-D-Glcp-(1-4)]4- D-Glc 3 0.7619 0.5230 0.9875 0.7670 4 0.7511 0.5253 0.9938 0.7762 α-D-Manp-(1-6)-α- 1 0.7807 0.5702 0.6247 0.9561 D-Manp-(1-6)-D- Man 2 0.8907 0.5850 0.7598 0.9770

111

4.4 Conclusions

A new approach for linkage determination of oligosaccharides has been evaluated

n 3 18 based on MS (n = 3-5) CID of Z1 ions. MS CID of O-labeled disaccharides

(deprotonated ions) enabled the discovery of Z1 ions as “diagnostic ions” for linkage determination. The fragmentation pattern of Z1 ions was only sensitive to their linkage positions and not to their sugar identities and anomeric configurations. This unique property allowed standard CID spectra of Z1 ions to be generated from a small set of disaccharides (possibly from 4 disaccharides) that were representative of many other possible isomeric structures. The formation of Z1 ions could be optimized using ion trap

CID at lower activation energies vs. their structural isomers. This enabled their analysis to be performed on native disaccharides or disaccharide subunits formed by MS/MS of larger oligosaccharides. This information can be used in conjunction with that from the m/z 341 disaccharide ion dissociation patterns to provide highly confident assignments of all linkages. It is worthy of note that for 3- and 4-linked disaccharides, information from the Z1 ions of unlabeled disaccharides is especially valuable for their linkage assignments as they highly predominate over other isomers under low or high energy dissociation conditions. With 2- and 6-linked disaccharides, lower energy ion-trap conditions are preferable to yield a preponderant Z1 ion. While a small amount of an isomeric species cannot be eliminated, fragmentation patterns of the ions isolated under optimal conditions still enabled the linkage to be assigned. Yet the 2- and 6-linkages yield distinct fragmentation patterns directly from dissociation of their disaccharide precursors with abundant and characteristic sets of product ions,5-9,28 so for a completely unknown, unlabeled disaccharide, this information would be highly useful in conjunction with 112

information from Z1 ion dissociation should there be any suspicion as to linkage assignment. As we have not analyzed all stereochemical variants of Z1 product ions arising from all linkages, it seems reasonable to surmise that higher statistical variance of the linkage-characteristic dissociation patterns shown in Figures 4.3 and 4.7 may be encountered, although observations so far indicate that linkage isomers dissociate with

n dramatic differences. MS CID of Z1 ions was applied to two trisaccharides and one pentasaccharide to assess whether their linkage positions could be determined. By

n comparing the MS CID spectra of Z1 ions derived from different locations within oligosaccharides with the standard CID spectra of Z1 ions from disaccharides, confident assignments for individual linkages together with their locations within the oligomers could be achieved for all three oligosaccharides studied as model compounds. This method was further applied to a branched oligosaccharide, Man5. An exoglycosidase (α1-

2,3 Mannosidase) converted Man5 into a linear trisaccharide having the structure of α-D-

Manp-(1-6)-α-D-Manp-(1-6)-D-Man and two linkage information were successfully obtained. The process of comparison of the Z1 ion spectra to standards was greatly enhanced by a spectral similarity score algorithm, which provided numeric values in the range of 0-1 with the highest scores indicative of the most likely assignment for the linkage position. Although the examples were demonstrated on a hybrid triple quadrupole/linear ion trap mass spectrometer, the MSn CID approach should in principle be possible to implement on any standalone ion trap instrument, depending on the relative abundances of isolated product ions derived from different oligosaccharide structures. 113

4.5 References

(1) Domon, B.; Müller, D. R.; Richter, W. J., Org. Mass Spectrom. 1989, 24, 357-359.

(2) Spengler, B.; Dolce, J. W.; Cotter, R. J., Anal. Chem. 1990, 62, 1731-1737.

(3) Zhou, Z.; Ogden, S.; Leary, J. A., J. Org. Chem. 1990, 55, 5444-5446.

(4) Hofmeister, G. E.; Zhou, Z.; Leary, J. A., J. Am. Chem. Soc. 1991, 113, 5964-5970.

(5) Dongre, A. R.; Wysocki, V. H., Org. Mass Spectrom. 1994, 29, 700-702.

(6) Ashline, D.; Singh, S.; Hanneman, A.; Reinhold, V., Anal. Chem. 2005, 77, 6250- 6262.

(7) Ballistreri, A.; Montaudo, G.; Garozzo, D.; Giuffrida, M.; Impallomeni, G.; Daolio, S., Rapid Commun. Mass Spectrom. 1989, 3, 302-304.

(8) Garozzo, D.; Giuffrida, M.; Impallomeni, G.; Ballistreri, A.; Montaudo, G., Anal. Chem. 1990, 62, 279-286.

(9) Dallinga, J. W.; Heerma, W., Biol. Mass Spectrom. 1991, 20, 215-231.

(10) Carroll, J. A.; Willard, D.; Lebrilla, C. B., Anal. Chim. Acta 1995, 307, 431-447.

(11) Mulroney, B.; Traeger, J. C.; Stone, B. A., J. Mass Spectrom. 1995, 30, 1277-1283.

(12) Guan, B.; Cole, R. B., J. Am. Soc. Mass Spectrom. 2008, 19, 1119-1131.

(13) Fang, T. T.; Zirrolli, J.; Bendiak, B., Carbohydr. Res. 2007, 342, 217-235.

(14) Guan, B.; Cole, R. B., Rapid Commun. Mass Spectrom. 2007, 21, 3165-3168.

(15) Carroll, J. A.; Ngoka, L.; Beggs, C. G.; Lebrilla, C. B., Anal. Chem. 1993, 65, 1582- 1587.

(16) Reinhold, V. N.; Reinhold, B. B.; Costello, C. E., Anal. Chem. 1995, 67, 1772-1784.

(17) Chen, S.-T.; Her, G.-R., J. Am. Soc. Mass Spectrom. 2012, 1-11.

(18) Xie, Y.; Lebrilla, C. B., Anal. Chem. 2003, 75, 1590-1598.

(19) Adamson, J. T.; Håkansson, K., Anal. Chem. 2007, 79, 2901-2910.

(20) Zhao, C.; Xie, B.; Chan, S.-Y.; Costello, C.; O’Connor, P., J. Am. Soc. Mass Spectrom. 2008, 19, 138-150. 114

(21) Wolff, J.; Laremore, T.; Aslam, H.; Linhardt, R.; Amster, I. J., J. Am. Soc. Mass Spectrom. 2008, 19, 1449-1458.

(22) Wolff, J.; Amster, I. J.; Chi, L.; Linhardt, R., J. Am. Soc. Mass Spectrom. 2007, 18, 234-244.

(23) Yu, X.; Huang, Y.; Lin, C.; Costello, C. E., Anal. Chem. 2012. 84, 7487-7494.

(24) Fang, T. T.; Bendiak, B., J. Am. Chem. Soc. 2007, 129, 9721-9736.

(25) Harvey, D. J., J. Am. Soc. Mass Spectrom. 2000, 11, 900-915.

(26) Chiesa, C.; Horváth, C., 1993, 645, 337-352.

(27) Konda, C.; Bendiak, B.; Xia, Y., J. Am. Soc. Mass Spectrom. 2012, 23, 347-358.

(28) Sheeley, D. M.; Reinhold, V. N., Anal. Chem. 1998, 70, 3053-3059.

115

CHAPTER 5 THE EFFECTS OF REDUCING END MODIFICATION ON GAS- PHASE CEHMISTRY OF SMALL OLIGOSACCHARIDES

5.1 Introduction

Derivatization of sugars such as permethylation1,2 of hydroxyl groups and reducing end modification3,4 is widely applied strategy in structural analysis of carbohydrates prior to mass spectrometric analysis to achieve improved sensitivity. These derivatizations largely affect gas-phase dissociation chemistry of glycan ions.5,6 One popular approach for reducing end derivatization is amination (formation of Schiff base) between an aromatic amine (e.g., 2-aminobenzamide,7 4-aminobenzoic acid ethyl ester8) and a reducing sugar due to the high reaction efficiency and simple one-pot procedure.

Reductive amination is the reduced form of amination and the reducing sugar stays in open-ring form.9,10 Her et al. previously reported that not only the type of derivatives but also the structure of reducing sugar ring (open or closed) affected the fragmentation patterns. They reported that closed-ring structure enhanced cross-ring cleavages as compared to open-ring form.11

MSn (n>2) approach has been demonstrated on a modified triple- quadrupole/linear ion trap instrument for stereo-structure characterization of individual sugar units within small linear oligosaccharides in Chapter 3. This approach (as shown in

Scheme 3.1) involves optimizing the formation of an ordered set of overlapping 116 disaccharide. Subsequently, the diagnostic product ion (m/z 221) having the structure of non-reducing sugar glycosidically linked to a glycolaldehyde formed by cross-ring cleavages were obtained. The success of this MSn approach largely depends on the formation of the correct ladder of precursor ions within the tree of analysis, including Ym,

C2, and m/z 221 ions from C2. Stereochemistry and anomeric configuration of 3 sugar units (3, 4, and 5 counting from the reducing-end sugar) in ABA modified (by reductive amination) pentasaccharide anomeric isomers were successfully obtained. However, sugar unit 2 information, which was theoretically possible to obtain by this approach, was unavailable due to the absence of diagnostic m/z 221 ions from CID of Y2 ions.

Herein, a couple of N-derivatives (ABA, ABEE) were applied to model disaccharides (α- and β-D-Glcp-(1-4)-D-Glc) by amination and reductive amination and their fragmentation behavior, especially for the formation of m/z 221 ion, was studied.

Reducing-end derivatized disaccharides are considered to have the same structure with the Y2 ions from oligosaccharides and can be used to simulate the condition without prior

MS/MS steps. Amination or reductive amination of N-derivatization showed a huge difference on their fragmentation. Reductive N-derivatization produced no m/z 221 ions while N-derivatization without reduction produced m/z 221 ions. This phenomenon is consistent with that m/z 221 ion formation (from cross-ring cleavages) is favored by having the reducing sugar in a closed-ring structure. N-derivatization without reduction was then applied to trisaccharides. However CID of the deprotonated ion did not produce Y2 ions. Since N-derivatization could not satisfy both criteria, we tested a new derivatization: O-derivatization. O-derivatization gives the closed-ring structure of the reducing-end sugar unit and the formation of m/z 221 ions was expected. To our 117 knowledge, there is no systematic studies on the fragmentation behavior caused by different types of derivatives and also this is the first time to introduce O-derivatized sugars for mass spectrometric analysis. Among five derivatives (HBA, HBEE, HBSA,

1,2HNA, and 3,2HNA), HBA and HBSA showed a formation of m/z 221 ions and were applied to trisaccharides to confirm whether they form Y2 ions or not. HBSA showed the formation of Y2 ions seemed successful, however, this derivative failed when linkage isomers of α- and β-D-Glcp-(1-4)-D-Glc were used.

5.2 Experimental

5.2.1 Materials

Eight types of reducing end modifications (shown in the table 1 with estimated pKa values) were performed on a series of disaccharides and trisaccharides to study their effect on the gas-phase fragmentation of oligosaccharide anions. β-D-Glcp-(1-4)-D-Glc,

α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc, [α-D-Glcp-(1-4)]2-D-Glc, and [β-D-Glcp-(1-4)]4-

18 D-Glc, H2 O, chloroform, sodium cyanoborohydride, acetic acid, dimethyl sulfoxide

(DMSO), acetic anhydride, dichloromethane (DCM), boron trifluoride diethyl etherate

(BF3·Et2O), sodium methoxide were purchased from Sigma-Aldrich, Inc. (St. Louis, MO)

Octaacetyl α-D-Glcp-(1-4)-β-D-Glc from Carbosynth, Ltd. (Berkshire, UK). Detailed procedures of 18O-labeling,12 reductive amination (R)3,13, and amination without reduction (NR) of reducing saccharides were described in the previous chapters (18O- labeling in chapter 2 and reductive amination in chapter 3, and amination without reduction in chapter 4). 118

Table 5.1 List of reducing end derivatives

Core R R (pKa of R ) Chemical Name (Abbreviation) Structure 1 2 2

COOH (4.87e) 4-aminobenzoic acid (ABA) NH2 COOCH2CH3 (-) Ethyl-4-aminobenzoate (ABEE)

COOH (4.57) 4-hydroxybenzoic acid (HBA)

OH COOCH2CH3 (-) Ethyl 4-hydroxybenzoate (HBEE)

SO3H(2.54) 4-hydroxybenzenesulfonic acid (HBSA)

OH COOH (3.28e) 1-hydroxy-2-napthoic acid (1,2HNA)

OH COOH (3.02e) 3-hydroxy-2-napthoic acid (3,2HNA) . *Superscript e beside pKa value indicates estimation.

5.2.2 O-Derivatization at Reducing Sugars

Overall reaction steps are illustrated in Scheme 3.1 and (1) through (3) in

Scheme3.1 correspond to the each following section.

5.2.2.1 Acetylation of hydroxyl groups on sugars

Disaccharides or oligosaccharides (30 nmol) were dissolved in 60 to 100 μl of acetic anhydride with 0.5 to 1 mg of iodine in an eppendorf tube and the reaction was conducted with a stirring magnet for 3 h. The progress of the reaction was monitored by

MS for each step. After the completion of acetylation of hydroxyl groups on sugars,

DCM and aqueous sodium thiosulphate solution were added to the reaction mixture and mixed thoroughly. The colorless organic layer was transferred to another eppendorf tube and was washed with aqueous sodium carbonate solution to neutrality. The organic layer was again transferred and dried completely by Speed-Vac. 119

5.2.2.2 O-derivatization to the acetylated sugars14

1 mg of acetylated sugar, 1.5 mg of derivatives (HBA, HBEE, HBSA, 1,2HNA, or 3,2HNA), 1 μL of BF3·Et2O were dissolved in 500 μL of DCM under nitrogen. The reaction mixture was placed in ultrasonic bath (B3500A-MT, VWR, Radnor, PA) for 1 to

2 h. After the completion of O-derivatization, aqueous sodium carbonate solution was added and mixed thoroughly. The organic layer was transferred and dried completely by

Speed-Vac.

5.2.2.3 Deacetylation of O-derivatized Acetylated Sugars

1 mg of O-derivatized acetylated sugars and 1mg of sodium methoxide were dissolved in 500 μL of methanol and incubated at 60 oC for 10-24 h.

5.2.2.4 Desalting of O-derivatized Sugars by HPLC

The reaction mixture of deacetylated O-derivatized sugars was desalted using

Agilent 1200 series HPLC system (Agilent Technologies, Santa Clara, CA). Desalting was carried out on a HILIC (PolyLC Inc., Columbia, MD) column at a flow rate of 0.4 mL/min with a linear gradient of 20-35% solvent A in 30 min. Solvent A was water and solvent B was acetonitrile. The eluent was detected at a different wavelength suitable for individual chromophores (HBA: 310 nm, HBEE: 252 nm, HBSA: 271 nm, 1,2HNA: 352 nm, 3,2HNA: 352 nm). 120

Scheme 5.1 Reaction steps for O-derivatization at reducing sugars

5.2.3 Mass Spectrometry

All samples were analyzed in the negative ion mode using a 4000Qtrap mass spectrometer (Applied Biosystems/Sciex, Toronto, Canada) equipped with a home-built nanoelectrospray (nanoESI) source. Two types of low energy collisional activation methods were accessible on this instrument, i.e. beam-type CID and ion trap CID.

Analyst 1.5 software was used for instrument control, data acquisition, and processing.

The typical parameters of the mass spectrometer used in this study were set as follows: spray voltage, -1.1 to -1.5 kV; curtain gas, 10; declustering potential, 50 V; beam-type

CID collision energy (CE), 5 to 30 V; ion trap CID activation energy (AF2), 5 to 60

(arbitrary units); scan rate, 1000 m/z/s; pressure in Q2, 5.0 x 10-3 Torr, and in Q3, 2.5 x

10-5 Torr. 121

5.3 Results and Discussion

In order to investigate the major fragments formed from a native disaccharide,

MS2 beam-type and ion trap CID spectra of 18O-labeled β-D-Glcp-(1-4)-D-Glc were collected and shown in Figure 5.1a. Product ions observed were m/z 161 (B1, -182 Da),

163 (Z1, -180 Da), 179 (C1, -164 Da), 221 (A2, -122 Da), 263 (A2, -78 Da), 283 (A2, -60

Da) and 325 (-18 Da). Product ions at m/z 161, 163, and 179 are formed from glycosidic bond cleavages and m/z 221, 263, and 283 are formed from cross-ring cleavages. The m/z 221 ion is the diagnostic ion for stereo-structure analysis and can be observed from any disaccharides (although abundance may vary), however, Y1 ion (m/z 181) was never observed in the product ions from CID of native disaccharide anions.

5.3.1 Gas-Phase Fragmentation Studies of N-Derivatized Disaccharides

MS2 beam-type and ion trap CID spectra of ABA derivatized disaccharide by amination (β-D-Glcp-(1-4)-D-Glc-NR-ABA, MW: 461 Da) are shown in Figure 5.1b.

The major fragments included peaks at m/z 161 (B1), 178 (X0), 179 (C1), 192 (X0), 220

(X0), 221 (A2), 234 (X0), 262 (X0), 263 (A0), and 280 (Z1). The relatively intensity of m/z

161, 179, 221, 263 ions were similar to the β-D-Glcp-(1-4)-D-Glc-18O (Figure 5.1a), however, a variety of X0 ions (cross-ring cleavages consisting of the reducing end) were produced abundantly which suggested that the charge stayed on reducing end due to the presence of carboxylic acid group in the derivative. As previously reported, beam-type

CID produced more fragments from cross-ring cleavages (A and X ions, especially m/z

221 ion) which required higher activation energies, while glycosidic bond cleavage (Z1 122 ion formation) was enhanced by on-resonance ion trap CID (slow heating of the ion which favors the lowest energy fragmentation channels).15

MS2 beam-type and ion trap CID spectra of ABA derivatized disaccharide by reductive amination (β-D-Glcp-(1-4)-D-Glc-R-ABA, MW: 463) are shown in Figure

5.1c. These spectra showed major peaks at m/z 161 (B1), 179 (C1), 204 (X0), 222 (X0),

234 (X0), 246 (X0), 264 (X0), 282 (Z1), 300 (Y1), and 402 (X1). Unlike non-reduced derivatization (Figure 5.1b), Y1 ion was formed, however, there were no A ions such as m/z 221, 263, and 281. Similar phenomena were observed from the pare of β-D-Glcp-(1-

4)-D-Glc-NR-ABEE (MW: 489) vs. β-D-Glcp-(1-4)-D-Glc-R-ABEE (MW: 491) as shown in Figure 5.1d and e, respectively. The A type ions were only observed from β-D-

Glcp-(1-4)-D-Glc-NR-ABEE while Y1 ion was only observed from β-D-Glcp-(1-4)-D-

Glc-R-ABEE. Comparing ABA and ABEE derivatives (e.g., Figure 5.1b and d, A, B, and C type ions consisting of the non-reducing end were always observed in higher intensities in ABEE than ABA which is consistent with the lower acidity of ABEE derivative than ABA. 123

18 (a) O 179 163

100 163 Z 283 Beam 343 type 161 221 B C 50 161 179 A A A Rel. Int. (%) Int. Rel. 263 283 343 145 221 0 100 200 300 400 500 100 163 Z Ion trap

50 C

Rel. Int. (%) Rel. Int. A 343 179 A A 221263 283 325 0 100 200 300 400 500 179 282 (b) NR-ABA 179 280 (c) R-ABA

220X 221 161 100 A 280 Z 161 100 282 Z 300 Beam type Beam type 462 221 A 460 263 X X C 234 50 X 50 X 178179 B Y1 262 222 246

Rel. Int. (%) Int. Rel. 460 161 B X C204 X 300 462 161 192 234 179 264 0 0 100 200 300 400 500 100 200 300 400 500 280 Z 282 Z 100 100 Ion trap Ion trap X A X X 222 234 50 263 50 C 220 X A 246 179 281 B 204 Y Rel. Int. (%) Rel. Int. 262 460 C 1 462 B 161 X 161192 442 179 264 300 402 0 0 100 200 300 400 500 100 200 300 400 500 179 308 179 310 (d) NR-ABEE 281 (e) R-ABEE

221 Z 161 328 161 310 100 248 X 100 Z 488 Beam type Beam type 328 Y1 C 308 X 490 A 164 179 232X 50 263 X 50 B 290 161 C 250 B A 179 Rel. Int. (%) Int. Rel. 488 490 161 221 A 283 281 0 0 100 200 300 400 500 100 200 300 400 500 100 248X 100 310 Z Ion trap Ion trap A A 221 263 Z 164 50 C 308 50 B C 179 X A 161 179 Y1 X Rel. Int. (%) Int. Rel. B 220 281X 488 X X 430 490 161 290 250 328 400 0 0 100 200 300 400 500 100 200 300 400 500 m/z m/z

Figure 5.1 MS2 beam-type vs. ion trap CIDs from (a) β-D-Glcp-(1-4)-D-Glc-18O, (b) β- D-Glcp-(1-4)-D-Glc-NR-ABA, (c) β-D-Glcp-(1-4)-D-Glc-R-ABA, (d) β-D-Glcp-(1-4)- D-Glc-NR-ABEE, and (e) β-D-Glcp-(1-4)-D-Glc-R-ABEE. 124

5.3.2 Gas-Phase Fragmentation Studies of O-Derivatized Disaccharides

MS2 beam-type and ion trap CID spectra of β-D-Glcp-(1-4)-D-Glc-HBEE (MW:

489), β-D-Glcp-(1-4)-D-Glc-HBA (MW: 461), and β-D-Glcp-(1-4)-D-Glc-HBSA (MW:

497) are shown in Figure 5.2a, b, and c, respectively. β-D-Glcp-(1-4)-D-Glc-HBEE showed major peaks at m/z 165 (Y0), 323 (B2), and 411 (X1) and small amount of m/z 221 ion was formed by beam-type CID (Figure 5.2a). α-D-Glcp-(1-4)-D-Glc-HBA (Figure

5.2b) showed major peaks at m/z 137 (Y0), 263 (A2), and 323 (B2). α-D-Glcp-(1-4)-D-

Glc-HBSA (Figure 5.2c) showed major peaks at m/z 173 (Y0), and 335 (Y1) while small amount of m/z 221 (A2) ion was formed by both beam-type and ion trap CID. HBEE and

HBA derivatives showed abundant glycosidic bond cleavages between a reducing end sugar and a derivative, forming Y0 and B2 ions depending on the charge locations. The ratio of these two ions changes depending on the type of CID used. Bean-type CID produced Y0 ions as the base peak while ion trap CID produced B2 ions as the base peak.

Based on these results, the majority of the original deprotonation sites can be considered as located at the derivatized function groups and beam-type CID (fast reaction, < 1 ms) produces abundant Y0 ions. The deprotonation sites are moved down to the sugar side while ions were staying in the trap by ion trap CID (slow reaction, > 100 ms). HBSA showed a little different effect on the formation of Y0 and B0 ions. Beam-type CID produced 100% of Y0 ion (base peak) and about 40% of Y1 ion while no B2 ion was formed. Ion trap CID reduced the formation of Y0 ion and increased the formation of Y1 ion, however, again, no B2 ion was generated. This is probably due to the higher acidity of HBSA (pKa = 2.54) as compared to HBEE and HBA (pKa = 4.57) and the negative charge tends to be localized on sulfonic acid group. 125

MS2 beam-type and ion trap CID spectra of β-D-Glcp-(1-4)-D-Glc-1,2HNA

(MW: 511) and β-D-Glcp-(1-4)-D-Glc-3,2HNA (MW: 511) are shown in Figure 5.2d and e, respectively. Both medication groups showed very similar fragments, having a base peak at m/z 229 (X0), and fragments at m/z 187 (Y0), 313 (X0), and 331 (Z1). All of the fragmentation appeared were X, Y, Z type of ions and no A, B, C ions were formed. This is also due to the higher acidity of HNA (pKa=3.02 – 3.28) as compared to HBEE and

HBA (pKa = 4.57). 126

179 323 331 187 (a) HBEE (d) 1,2HNA 165 Y 221 100 0 161 165 100 229 X Beam 489 Beam type 511 type

50 B2 50 B Y0 161 323 Rel. Int. (%) Int. Rel. A A X 489 187 X Z X 511 178 221245 411 313 331 391 0 0 100 200 300 400 500 100 200 300 400 500 B2 100 323 100 229 X Ion trap Ion trap X 50 50 185 361 375 443 X Y0

Rel. Int. (%) Int. Rel. Y 411 489 187 X 511 0 467 165 244 287 435453 157 313 347 493 0 0 100 200 300 400 500 100 200 300331 400 500 (b) HBA 179 323 (e) 3,2HNA 229 X 187 100 137 Y0 100 157 511 Beam 461 Beam 137 type 161 type X 493 50 B 50 295 Z 2 185 Y0 323 331 Rel. Int. (%) Int. Rel. A A 461 143 187 X X 479 511 245 263 167 271 313 473 0 0 100 200 300 400 500 100 200 300 400 500 B 100 323 2 100 229 X Ion trap Ion trap Z X 331 493 50 50 295 Y 157 0 Y 483

Rel. Int. (%) Int. Rel. Y 0 A A 461 187 X X 1 511 137 245263 185 271 313 349 0 0 100 200 300 400 500 100 200 300 400 500 317 m/z (c) HBSA 281 100 173 Y Beam 0 221 173 497 type 335 Y1 50 172 241 335

Rel. Int. (%) Int. Rel. X 255 Z 497 215 283 317 0 100 200 300 400 500 100 335 Y Ion trap 1 Y0 173 50 A 221 Rel. Int. (%) Int. Rel. X A X 497 215 281 417 437 0 100 200 300 400 500 m/z

Figure 5.2 MS2 beam-type vs. ion trap CID from (a) β-D-Glcp-(1-4)-D-Glc-HBEE, (b) β-D-Glcp-(1-4)-D-Glc-HBA, (c) β-D-Glcp-(1-4)-D-Glc-HBSA, (d) β-D-Glcp-(1-4)-D- Glc-1,2HNA, and (e) β-D-Glcp-(1-4)-D-Glc-3,2HNA. 127

5.3.3 Comparisons between N-Derivatized (by Amination) and O-Derivatized Trisaccharides

Non-reduced form of ABA and ABEE, HBA and HBSA derivatized disaccharides produced m/z 221 ions (diagnostic ions for stereo-chemical information) and these derivatives are thus applied to trisaccharides to test whether Y2 ions can be formed or not.

The formation of Y2 ions is the key to obtain sugar 2 information from above listed derivatized oligosaccharides because the MSn CID steps needed to follow for sugar 2 are:

- 2 [M-H] ÆY2Æ221Æfragments). MS beam-type and ion trap CID spectra of α-D-Glcp-

(1-6)-α-D-Glcp-(1-4)-Glc-NR-ABA and α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-NR-

ABEE are shown Figure 5.3a and b, respectively. α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-

Glc-NR-ABA showed major peaks at m/z 161 (B1), 179 (C1), 192 (X0), 220 (X0), 221

(A2), 262 (X0), 280 (Z1), 323 (A2), 341 (C2), 383 (A3), 425 (A3), and 562 (X2). α-D-Glcp-

(1-6)-α-D-Glcp-(1-4)-D-Glc-NR-ABEE showed very similar fragmentation patterns, having different masses for X, Y, Z ions due to the different masses in derivatives.

However, no Y2 ions were produced from CID of α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-

Glc-ABA and α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-ABEE and these derivatizations were not applicable to our MSn (n>2) approach.

O-derivatized trisaccharides were then examined to see the formation of Y2 ions

2 2 from MS CID. MS beam-type and ion trap CID spectra of [α-D-Glcp-(1-4)]2-D-Glc-

HBEE and [α-D-Glcp-(1-4)]2-D-Glc-HBSA are shown Figure 5.4b and c, respectively.

Spectra from [α-D-Glcp-(1-4)]2-D-Glc-HBEE showed a base peak at m/z 485 (B3) and small amount of product ions at m/z 165(Y0), 323 (B2), 383 (A3), 425 (A3), and 573(X2).

As was previously observed in α-D-Glcp-(1-4)-D-Glc-HBEE, glycosidic bond cleavages 128 between a reducing end sugar and a derivative was the main fragmentation site. α-D-

Glcp-(1-4)-D-Glc-HBEE showed different base peaks (Y0 for beam-type and B2 for ion trap), there was no significant difference between beam-type and ion trap CIDs for [α-D-

Glcp-(1-4)]2-D-Glc-HBEE. Spectra from [α-D-Glcp-(1-4)]2-D-Glc-HBSA showed major peaks at m/z 173 (Y0), 335, (Y1), 497 (Y2) with small amount of product ions at m/z 215

(X0), 221 (A2), 317 (Z2), 383 (A3), 479 (Z2), and 581 (A3). HBSA derivatized trisaccharide produced a variety of product ions, most importantly having Y2 ion and m/z

221 ions.

(a) NR-ABA 161 179 221323 622 341 Beam 280 383A2 100 type B 192 X0 A2 2 C1 220 321323 622 A2 50 179 C2 B 221 341 1 Z Rel. Int. (%) Int. Rel. 161 1 A 280 2 X2 425 562 0 100 200 300 400 500 600 700 m/z (b) NR-ABEE 161 179 221323

341 650 X0 308 Beam 248 B2 100 type 323 A2 A Z1 2 650 (%) 221 308 383 50 220 C B1 2

Rel. Int. Rel. B 161 C1 2 341 A2 179 323 425 0 100 200 300 400 500 600 700 m/z Figure 5.3 MS2 beam-type CID from (a) α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-NR- ABA, and (b) α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-D-Glc-NR-ABEE. 129

161 161 179 179 18 341 341 (a) O 163 (c) HBA 281 137 323 323 161B 383A 100 1 100 3 623 341C2 505 A Beam Z1 Beam type 3 A A3 425 type 163 3 383 425 Y0 C C1 A 381 50 1 50 137 2 407 B 179 A A 179 A 311 3 X A A 2 3 B 2 B2 485 2 Rel. Int. (%) Int. Rel. 2 2 281 B 443 505 1 221 545 623 263 2 161 A2/Z2 323 221 323 281 587 0 0 100 200 300 400 500 100 200 300 400 500 600

341C 485 B3 100 2 100 Ion trap Ion trap

(%) Z1 C 2 407 A3 X2 50 B1 163 50 A2 305B 341 605 A 2 A3 425 545 161 3 A2 A 221 323 Y2 Rel. Int. Rel. C1 A2 3443 505 281 383 623 425 179 281 487 461501 587 0 0 100 200 300 400 500 100 200 300 400 500 600 (b) HBEE (d) HBSA 479 485 165 317 497 173 323 485B 335 100 3 651 100 173Y0 659 Beam Beam Y2 type type Y1 50 50 335 497 A A 3 X0 2 Z A Z X 659 Rel. Int. (%) Int. Rel. Y B A X 651 1 X 2 0 2 383 3 2 3 1 2 165 323 425 512 573 623 215 221 317 383419 479 581 0 0 100 200 300 400 500 600 100 200 300 400 500 600

485 B 497 Y2 100 3 100 Ion trap Ion trap

50 50 Y1 X A3 A3 2 651 X A 335 A 659 Rel. Int. (%) Int. Rel. 0 2 3 383 425 512 573 623 215 281 383 0 0 100 200 300 400 500 600 100 200 300 400 500 600 m/z m/z

2 18 Figure 5.4 MS beam-type vs. ion trap CID from (a) [α-D-Glcp-(1-4)]2-D-Glc- O, (b) [α-D-Glcp-(1-4)]2-D-Glc-HEE, (c) [α-D-Glcp-(1-4)]2-D-Glc-HBA, and (d) [α-D-Glcp-(1- 4)]2-D-Glc-HBSA. 130

The relative intensity (normalized to the base peak) of the specific product ions formed by different types of derivatizations was summarized in Table 5.2. Specific product ions of our interest were m/z 221 ions from disaccharide-derivatives and Y2 ions from trisaccharide-derivatives. Both criteria were only satisfied by HBSA. HBSA was then derivatized to disaccharide linkage isomers (e.g., α-D-Glcp-(1-2)-D-Glc, α-D-Glcp-

(1-3)-D-Glc, β-D-Glcp-(1-3)-D-Glc, α-D-Glcp-(1-6)-D-Glc, and β-D-Glcp-(1-6)-D-Glc) to test the applicability of this derivatives to a variety of disaccharides. Unfortunately, m/z 221 ion could not be obtained from linkage isomers listed above.

Table 5.2 Summary of the normalized product ion intensity of characteristic ions by a variety of derivatizations

Relative intensity of product ions (%) Disaccharide Trisaccharide Type of derivatives m/z 221 ion Y2 ion Native 5 0 NR-ABA 15 0 R-ABA 0 - NR-ABEE 14 0 R-ABEE 0 - HBA 0 4 HBEE 3 0 HBSA 2 100 1,2HNA 0 - 3,2HNA 0 131

5.4 Conclusions

The fragmentation behavior of oligosaccharides can be affected by the chemical nature of reducing-end derivatizations. HBSA was the only derivative which produced both Y2 and m/z 221 ions from a derivatized model trisaccharide, [α-D-Glcp-(1-4)]2-D-

Glc-HBSA. However, this derivative only worked for 1-4 linked sugars and could not be applied for other linkage isomers.

We continue to explore different derivatives, especially having the pKa value of between 3.5 and 4.5 would be a good choice since HBSA (pKa = 2.54) was too acidic and produced mainly Y ions while HBA (pKa = 4.57) was not acidic enough to produce

Y ions. 132

5.5 References

(1) Ciucanu, I.; Kerek, F., Carbohydr. Res. 1984, 131, 209-217.

(2) Mechref, Y.; Kang, P.; Novotny, M. V., Solid-Phase Permethylation for Glycomic Analysis. In Glycomics, Packer, N.; Karlsson, N., Eds. Humana Press: 2009; Vol. 534, pp 53-64.

(3) Harvey, D. J., J. Am. Soc. Mass Spectrom. 2000, 11, 900-915.

(4) Harvey, D. J., J. Chromatogr. B 2011, 879, 1196-1225.

(5) Cancilla, M. T.; Penn, S. G.; Carroll, J. A.; Lebrilla, C. B., J. Am. Chem. Soc. 1996, 118, 6736-6745.

(6) Orlando, R.; Allen Bush, C.; Fenselau, C., Biol. Mass Spectrom. 1990, 19, 747-754.

(7) Ruhaak, L. R.; Huhn, C. H.; Deelder, A. M.; Wuhrer, M., Anal. Chem. 2008, 80, 6119-6126.

(8) Cheng, H.; Pai, P.; Her, G.-R., 2007, 18, 248-259.

(9) Her, G. R.; Santikarn, S.; Reinhold, V. N.; Williams, J. C., 1987, 6, 129-139.

(10) Nishikaze, T.; Kaneshiro, K.; Kawabata, S.; Tanaka, K., Anal. Chem. 2012, 84, 9453-9461.

(11) Chen, S.-T.; Her, G.-R., J. Am. Soc. Mass Spectrom. 2012, 23, 1408-1418.

(12) Fang, T. T.; Bendiak, B., J. Am. Chem. Soc. 2007, 129, 9721-9736.

(13) Chiesa, C.; Horváth, C., 1993, 645, 337-352.

(14) Smits, E.; Engberts, J. B. F. N.; Kellogg, R. M.; van Doren, H. A., J. Chem. Soc. Perkin Trans. 1 1996, 0, 2873-2877.

(15) Zaia, J., Mass Spectrom. Rev. 2004, 23, 161-227. 9,7$ 133

VITA

Chiharu Konda was born on December 22, 1979, in Nagano, Japan. She is the daughter of Yoshie and Hiromich Konda. While enrolling at Meiji Gakuin

Higashimurayama High School (Tokyo, Japan), she transferred to Notre Dame College

School (Ontario, Canada) and graduated from both high schools in 1998. Chiharu started working as a sales engineer at one of the largest cell phone companies in Japan, NTT

Docomo in 1999. After working for 6 years, she pursued her dream to study abroad for higher education and enrolled at Valdosta State University (Georgia, USA) in 2005. She worked in Professor De La Garza’s lab from 2006-2008, focusing on the fabrication of

TiO2 film by electro-deposition. In August of 2009, Chiharu graduated with a Bachelor of

Science in Chemistry with Magna Cum Laude and began her graduate career in the

Chemistry Department at Purdue Univeristy. She joined Professor Yu Xia’s lab in the fall of 2009, where her research focused on structural analysis of carbohydrates using mass spectrometry. She defended her Ph. D. thesis in October 2013. Following graduate school, she will be working at the new division of NTT Docomo where developing the synergetic technology involving analytical science and cell phone network. 38%/,&$7,216 134

PUBLICATIONS

1. Du, Y. M.; Konda, C.; Xia, Y.; Ouyang, Z. “Statistical Analysis Model for Classifying Stereo-Structures of Oligosaccharides Using Tandem Mass Spectrometry”, In Preparation.

2. Konda, C.; Bendiak, B.; Londry, F. A.; Xia, Y. “Stereochemistry and Anomeric Configuration with Single-Sugar Resolution for Oligosaccharides via MSn (n>2)”, In Preparation.

3. Konda, C.; Bendiak, B.; Xia, Y. “Linkage Determination of Linear Oligosaccharides n by MS (n>2) Collision-Induced Dissociation (CID) of Z1 Ions in Negative Ion Mode Tandem Mass Spectrometry”, J. Am. Soc. Mass Spectrom., accepted.

4. Konda, C.; Bendiak, B.; Xia, Y. “Differentiation of the Stereochemistry and Anomeric Configuration for 1-3 linked Disaccharides via Tandem Mass Spectrometry and 18O- Labeling”. J. Am. Soc. Mass Spectrom., 2012, 23, 347-358. 135

B American Society for Mass Spectrometry, 2011 J. Am. Soc. Mass Spectrom. (2012) 23:347Y358 DOI: 10.1007/s13361-011-0287-5

RESEARCH ARTICLE Differentiation of the Stereochemistry and Anomeric Configuration for 1-3 Linked Disaccharides Via Tandem Mass Spectrometry and 18O-labeling

Chiharu Konda,1 Brad Bendiak,2 Yu Xia1 1Department of Chemistry, Purdue University, West Lafayette, IN 47907-1393, USA 2Department of Cell and Developmental Biology, and Program in Structural Biology and Biophysics, University of Colorado Denver, Anschutz Medical Campus, Aurora, CO, USA

Abstract Collision-induced dissociation (CID) of deprotonated hexose-containing disaccharides (m/z 341) with 1–2, 1–4, and 1–6 linkages yields product ions at m/z 221, which have been identified as glycosyl-glycolaldehyde anions. From disaccharides with these linkages, CID of m/z 221 ions produces distinct fragmentation patterns that enable the stereochemistries and anomeric configurations of the non-reducing sugar units to be determined. However, only trace quantities of m/z 221 ions can be generated for 1–3 linkages in Paul or linear ion traps, preventing further CID analysis. Here we demonstrate that high intensities of m/z 221 ions can be built up in the linear ion trap (Q3) from beam-type CID of a series of 1–3 linked disaccharides conducted on a triple quadrupole/linear ion trap mass spectrometer. 18O-labeling at the carbonyl position of the reducing sugar allowed mass-discrimination of the “sidedness” of dissociation events to either side of the glycosidic linkage. Under relatively low energy beam-type CID and ion trap CID, an m/z 223 product ion containing 18O predominated. It was a structural isomer that fragmented quite differently than the glycosyl- glycolaldehydes and did not provide structural information about the non-reducing sugar. Under higher collision energy beam-type CID conditions, the formation of m/z 221 ions, which have the glycosyl-glycolaldehyde structures, were favored. Characteristic fragmentation patterns were observed for each m/z 221 ion from higher energy beam-type CID of 1–3 linked disaccharides and the stereochemistry of the non-reducing sugar, together with the anomeric configuration, were successfully identified both with and without 18O-labeling of the reducing sugar carbonyl group.

Key words: Oligosaccharides, Stereochemistry, Anomeric configuration, Collision-induced dissociation, Tandem mass spectrometry

Introduction structural components for plant cell walls. By conjugating with proteins and lipids, carbohydrates are widely involved arbohydrates play important roles in biological systems, in cell–cell interactions, cell signaling, and self and non-self Csuch as providing energy to cells and functioning as recognition events [1, 2]. Carbohydrates can form almost unlimited variations in their structures due to their structural Electronic supplementary material The online version of this article complexity. In order to elucidate the structure of an (doi:10.1007/s13361-011-0287-5) contains supplementary material, which is available to authorized users. oligosaccharide, it is necessary to characterize the stereochemistry of each monosaccharide unit, the anomeric Correspondence to: Yu Xia; e-mail: [email protected] configuration of the glycosidic bonds, linkage positions, Received: 23 August 2011 Revised: 20 October 2011 Accepted: 21 October 2011 Publishedo nline: 18 November 2011 136

348 C. Konda et al.: Structure Determination for Disaccharides

and the sequence of the individual monosaccharides in enabled linkage positions to be determined, both in the the oligomer. When enough sample is available, higher- negative and positive ion modes, and linkage sites can be dimensional NMR is a powerful tool to obtain detailed established for neutral disaccharides as positive ion adducts structural information [3–5]. of lithium or sodium ions or as protonated adducts [22–27]. Mass spectrometry is a widely applied method in Determination of anomeric configuration has been demon- structural analysis of carbohydrates, due to its capability strated for underivatized disaccharides in the negative ion of providing detailed molecular information and high mode [15, 17, 18], for derivatized 1-4- and 1-6-linked sensitivity [6]. The molecular weight information of disaccharides [28], and in the positive ion mode with alkali carbohydrates is readily obtained from soft ionization metals [22–27] and lead cationization [20]. Prior knowledge methods such as electrospray ionization (ESI) [7, 8]and such as linkage position, ring form, and stereochemistry [15, matrix-assisted laser desorption/ionization (MALDI) [9– 17, 18, 20, 22–28], however, was typically required to 11]. Tandem mass spectrometry based on collision- assign anomeric configuration, which was difficult to apply induced dissociation (CID) is heavily relied on to obtain to larger oligosaccharide systems [18]. CID of metal structural information for carbohydrates. Glycosidic bond cationized monosaccharides derivatized as a Schiff base cleavages and/or cross-ring cleavages are typically ob- with diethylenetriamine [29] and CID of metal cationized N- served from collisional activation. Following the nomen- acetylhexosamine diastereomers [30] have been shown to clature proposed by Domon and Costello , A, B, and C produce distinct fragmentation patterns according to the ions are fragments containing the non-reducing terminus stereochemistry. The assignment of stereochemistry has also while X, Y, and Z ions include the reducing end [12]. been obtained by matching the CID spectrum of acetylated Four types of fragments, B, C, Y, and Z ions, are monosaccharides in oligosaccharides with reference spectra formed from cleavages on either side of the glycosidic [31, 32]. Using additional statistical analysis, anomeric oxygen. B and Y ions are cleaved at the non-reducing configuration has been identified for metal cationized side of a glycosidic oxygen and C and Z ions are glucopyranosyl-glucose disaccharides (1–2, –3, –4, and –6 cleaved at the reducing side of a glycosidic oxygen. linkages) [26]. Also, the linkage positions and stereo- Based on the specific mass differences of fragments chemistries of non-reducing units can be discriminated for resulting from glycosidic bond cleavages, sequence glucose, galactose, and mannose containing disaccharides information for both linear and branching oligosacchar- having 1–2, –3, and –4 linkages [27]. ides as well as glycotypes such as the complex, hybrid, Recently, a tandem mass spectrometry approach has been or high-mannose can be determined for methylated developed to differentiate the stereochemistry and anomeric oligosaccharides [8, 13]. configuration for the non-reducing unit of hexose-containing A and X ions result from cross-ring cleavages and they disaccharides having any of the 16 possible stereochemical are typically more informative for the structural analysis of variants [33, 34]. In this method, diagnostic ions at m/z 221 carbohydrates. It has been shown that the relative abundan- were formed from CID of deprotonated disaccharide ions ces of these ions can be correlated to the linkage position, (m/z 341). It was established that the m/z 221 ions consisted and in some cases, the anomeric configuration and stereo- of the intact non-reducing sugar glycosidically linked to chemistry of each monosaccharide. For example, CID glycolaldehyde, as indicated in Scheme 1 (where GA spectra of deprotonated di- and oligosaccharide alkoxy abbreviates glycolaldehyde). Note that an open-chain form anions in the negative ion mode showed distinguishable for the reducing sugar is indicated in Scheme 1 and also for fragmentation patterns for each linkage position, which was other disaccharides discussed later. This is based on the successfully applied to di-, tri-, and hexasaccharides [14– observation of absorbance in the carbonyl stretch region in 17]. Negative ion adducts [18, 19] and positive adducts [20, variable wavelength infrared radiation photo-dissociation of 21] of deprotonated di- and oligosaccharides have also deprotonated monosaccharide anions in the gas phase [35].

-HH -

-H - CID CID

α-D-GlcDGlD Glcp-(1-4)-Glc (14)Gl(1 4) Glc, m/z/ 341α-D-Glc D GlGlcp-GA GAGA, m/z/ 221

Scheme 1. Formation of glycosyl-GA anions at m/z 221 from CID of deprotonated 1–4 linked disaccharides 137

C. Konda et al.: Structure Determination for Disaccharides 349

When m/z 221 ions were further dissociated by collisional Mass Spectrometry activation, disaccharides having different non-reducing sugar All samples were analyzed in the negative-ion mode on a units and anomeric configurations showed distinct fragmen- QTRAP 4000 mass spectrometer (Applied Biosystems/ tation patterns that matched synthetic glycosyl-GAs. This SCIEX, Toronto, Canada) equipped with a home-built method was shown to be useful for assigning the stereo- nanoelectrospray ionization (nanoESI) source. A schematic chemistry as well as the anomeric configuration of the presentation of the instrument ion optics is shown in the glycosidic bond for the non-reducing sugar in disaccharides Supporting Information, Figure S1. Two types of low energy having 1–2, 1–4, and 1–6 linkages [33, 34]. However, due to collisional activation methods were accessible on this the low abundance of m/z 221 ions produced from 1–3 instrument, i.e., beam-type CID and ion trap CID. In linked disaccharides, MS3 CID of m/z 221 ions could not be beam-type CID, the precursor ions (m/z 341 or 343) were performed, and it was unclear whether the fragmentation isolated in Q1, accelerated in the Q2 collision cell for patterns could be used for assigning either their stereochem- collisional activation, and all products were analyzed in the istry or anomeric configuration. Q3 linear ion trap. Collision energy (CE) was defined by the Herein, a series of 1–3 linked disaccharides were studied potential difference (absolute value) between Q0 and Q2. In on a triple quadrupole-linear ion trap mass spectrometer ion trap CID, the precursor ions were isolated in the Q3 (QTRAP 4000). MS3 CID data of m/z 221 ions from the 1–3 linear ion trap via the RF/DC mode and a dipolar excitation linked disaccharides were obtained for the first time. The was used for collisional activation. In order to perform ion formation of m/z 221 ions was examined using different trap CID at different Mathieu q-parameters, an AC (alter- collisional activation methods, i.e., beam-type CID and ion nating current) generated from an external waveform trap CID of the deprotonated disaccharides. 18O-labeling of generator (Agilent Technologies, Santa Clara, CA, USA) the reducing carbonyl oxygen in 1–3 linked disaccharides was used for resonance excitation. Frequency and the low was used to enable mass-discrimination of structural isomers mass cut-off were calculated by SxStability (Pan Galactic of the (usually) m/z 221 ions. By choosing the proper CID Scientific, Omemee, Ontario, Canada). MS3 CID experi- conditions, the diagnostic m/z 221 ions (the glycosyl-GAs) ments were carried out by first performing beam-type CID could be formed as the dominant isomer. Their CID of precursor ions in Q2. The fragment ions of interest were fragmentation patterns could be used to establish the isolated in Q3 and then subjected to ion trap CID. Analyst stereochemistry and anomeric configuration of the non- 1.5 software was used for instrument control, data acquisi- reducing sugar unit from 1–3 linked disaccharides. tion, and processing. The typical parameters of the mass spectrometer used in this study were set as follows: spray voltage, –1.1 to −1.5 kV; curtain gas, 10; declustering Experimental potential, 50 V; beam-type CID collision energy (CE), 5 to Materials 30 V; ion trap CID activation energy (AF2), 5 to 60 (arbitrary units); scan rate, 1000 m/z; pressure in Q2, 5.0× α – β – –3 –5 -D-Glcp-(1 2)-D-Glc (kojibiose), -D-Glcp-(1 2)-D-Glc 10 Torr, and in Q3, 2.5×10 Torr. Ion injection time was α – α (sophorose), -D-Glcp-(1 3)-D-Glc (nigerose), -D-Glcp- controlled to keep a similar parent ion intensity: typically 3× – 18 6 2 (1 3)-D-Fru (turanose), and H2 O were purchased from 10 counts per second (cps) for MS CID experiments and β – Sigma-Aldrich, Inc. (St. Louis, MO, USA); -D-Glcp-(1 3)- 1×106 cps for MS3 CID experiments. Activation time was α – α D-Glc (laminaribiose), -D-Manp-(1 3)-D-Man (3 -manno- kept constant at 200 ms for all ion trap CID experiments. α – α biose), and -D-Galp-(1 3)-D-Gal (3 -galactobiose) were Seven spectra were collected for CID of m/z 221 ions from α β purchased from Carbosynth, Ltd. (Berkshire, UK). - and - synthesized monosaccharide-GA standards (deprotonated monosaccharide-glycolaldehyde standards, glucopyranosyl- molecules) and disaccharides over a 1 y period. Standard glycolaldehydes (Glcp-GA), galactopyranosyl-glycolaldehydes deviations of peak heights were calculated for major frag- (Galp-GA), and mannopyranosyl-glycolaldehydes (Manp- ments such as m/z 87, 99, 101, 113, 129, 131, 159, 161, 203, GA) were synthesized as previously described [33]. Dis- and 221, which were observed from all the standards and accharides and synthetic standards were dissolved in meth- disaccharides studied here except β-D-Glcp-GA and β-D- anol to a final concentration of 0.01 mg/mL and NH4OH Glcp-(1–2)-D-Glc, which showed no peaks at m/z 99. was added to a final concentration of 1% immediately before use. Results and Discussion – 18O-Labeling of Reducing Disaccharides Ion trap CID of deprotonated 1 3 linked disaccharides (m/z 341) typically generates ions at m/z 221 in trace abundance 18 Disaccharides (1 mg) were dissolved in 100 μLofH2 O for on a Paul trap instrument, and isolation or further CID of m/z 3 to 10 d at room temperature. The solution was further 221 ions have not been achieved before [14, 15, 34, 36]. The diluted to 0.1 mg/mL with methanol before mass spectro- 4000QTRAP mass spectrometer used in this study has a metric analysis. unique triple quadrupole-linear ion trap configuration, 138

350 C. Konda et al.: Structure Determination for Disaccharides

offering high sensitivity due to the large capacity of the intensities of m/z 221 ions (indicated by an arrow in linear ion trap, and allowing either beam-type or ion trap Figure 1) were very low and their relative intensities were collisional activation. In beam-type CID, the precursor ions less than 1% (normalized to the base peak in the spectrum). were isolated in Q1 and accelerated in Q2 for collisional This phenomenon was generally observed for all 1–3 linked activation, while ion trap CID was conducted in Q3 with a disaccharides studied herein. The insets in Figure 1 demon- dipolar excitation for collisional activation. Since CID strate the isolated m/z 221 ions (with a 2 m/z isolation fragmentation patterns can be sensitive to the means of window) from each set of dissociation conditions. For beam- activation, the formation of m/z 221 ions from five 1–3 type CID, 1×106 cps of m/z 221 ions could be accumulated linked disaccharides was investigated via both beam-type with an injection time of 1 s, which was sufficient for and ion trap CID. Figure 1 compares the MS2 beam-type and performing the next stage of tandem mass spectrometry ion trap CID of deprotonated β-D-Glcp-(1–3)-D-Glc (m/z (MS3 in this case) with reasonable ion statistics and 341) using low energy CID conditions. A relatively low CE sensitivity. Far lower abundance of the m/z 221 ions (4.6× (6 V) was used for beam-type CID; in ion trap CID, the AF2 104 cps) could be isolated from ion trap CID of m/z 341, for an AC dipolar excitation was set to 25 (arbitrary units) even after doubling the injection time to 2 s. As a result, it for 200 ms. Under either activation condition, the absolute was not feasible to obtain MS3 CID for m/z 221 ions generated from m/z 341 precursor ions initially isolated 341 within the trap. In experiments described below, beam-type 100 221 ()(a) 1.05e6 CID was used to dissociate disaccharide precursor anions in the Q2 collision cell thereby generating m/z 221 product ions

ps in high enough abundance to acquire their spectra in the cp % linear trap reproducibly. y, y, % sity sity ns ns

te – ten CID of m/z 221 Ions Generated from 1 3 Linked In Int 222 Disaccharides e I

ive Previous studies have demonstrated that m/z 221 product at 161 179 ions formed from collisional activation of disaccharide Rel R anions typically consist of an intact non-reducing sugar with 113 143 a 2-carbon aglycone derived from the reducing sugar [34]. Three dominant fragment peaks are commonly observed 0 60 100 140 180 220 260 300 340 from CID of m/z 221: m/z 101, 131, and 161. The relative m/z intensities of these peaks, together with some other fragment ions, can be used to establish the fragmentation patterns and 100 179 to distinguish the stereochemistry and anomeric configura- (b) 161 4.6e446e4 221 tion of the non-reducing sugar. Given that the CID patterns of m/z 221 ions will be used for structural identification, s spectral reproducibility is an important issue. Similar to the cps % 341 y, c

y, % findings from a Paul trap instrument [33, 34], we noticed ty ity that the number of ions (m/z 221) in the linear ion trap and nsi nsi en en the energy input into an ion were among the most important nte nte In parameters affecting spectral reproducibility. To ensure e In

ve reasonable ion statistics and avoid adverse space charge

ativ effects, the intensity of the m/z 221 ions was kept at 1× 6 3 ela 113 10 cps before MS CID. Based on previous studies, the CID Re R 143 energies were tuned so that the ratio of remaining precursor ion to the most abundant product ion was kept around 18%± 0 3% [33]. Figure 2a, b, e, and f were the averaged spectra 60 100 140 180 220 260 300 340 from seven repetitions collected over a 1 y period, and they m/z were further used to make spectral comparisons in later discussion. Error bars in the spectra indicate the standard Figure 1. MS2 CID spectra in the negative ion mode deviation of the peak intensity for 10 major fragment ions, obtained from deprotonated β-D-Glcp-(1–3)-D-Glc (m/z 341) under low energy dissociation conditions: (a) beam-type CID which were frequently observed for all the disaccharides (CE=6 V), and (b) ion trap CID (AF2=25). Insets in (a) and (b) studied herein (m/z 87, 99, 101, 113, 129, 131, 159, 161, show the isolation of m/z 221 ions generated from beam-type 203, and 221). The standard deviations for these peaks were CID (injection time = 1 s) and ion trap CID (injection time = 2 s), less than 5% in most cases, indicating high reproducibility of respectively the spectra from day to day by controlling the ion counts in 139

C. Konda et al.: Structure Determination for Disaccharides 351

α-D-Glcp-(1-3)-Glc, m/z 341 α-D-Glcp-(1-2)-Glc, m/z 341 α-D-Glcp-GA, m/z 221 -H - -H - -H - m/z 221 m/z 221

131 161 131 100 (a) 100 (c) 100 (e)

101 203 101 99 99 113 131 113 113 87 221 87 129 159161 221 129 159161 221 Relative Intensity, % Relative Intensity, 203 % Relative Intensity, Relative Intensity, % Relative Intensity, 203 0 0 60 100 140 180 220 060 100 140 180 220 60 100 140 180 220 m/z m/z m/z

β-D-Glcp-(1-3)-Glc, m/z 341 β-D-Glcp-(1-2)-Glc, m/z 341 β -D-Glcp-GA, m/z 221 -H - -H - -H - m/z 221 m/z 221

131 131 100 161 100 (b) 100 (d) (f)

161 161 113 131 203

221 221 221 87 129

129 % Relative Intensity,

Relative Intensity, % Relative Intensity, 87 Relative Intensity, % Relative Intensity, 101 113 159 203 101113 159 0 0 0 60 100 140 180 220 60 100 140 180 220 60 100 140 180 220 m/z m/z m/z

Figure 2. MS2 ion trap CID spectra of m/z 221 ions derived from synthetic standards (a) α-D-Glcp-GA, AF2=25 and (b) β-D- Glcp-GA, AF2=18. MS3 CID spectra of m/z 221 ions derived from low energy beam-type CID of glucose-containing disaccharides: (c) α-D-Glcp-(1–3)-Glc, CE=6 V for MS2 andAF2=25forMS3, (d) β-D-Glcp-(1–3)-Glc, CE=6 V for MS2 and AF2=25 for MS3, (e) α-D-Glcp-(1–2)-Glc, CE=5 V for MS2, and AF2=27 for MS3,and(f) β-D-Glcp-(1–2)-Glc, CE= 5VforMS2, and AF2=25 for MS3. The error bars in the spectra show the standard deviation of the peak intensity based on seven spectra collected over a 1 y period the trap before CID and the energy input to the ions. Since either the stereochemistry or anomeric configurations of the the CID patterns upon dissociation of m/z 221 ions can differ ions. Note that beam-type CID was used to generate the m/z to some extent from instrument to instrument [37], CID 221 ions from disaccharides shown in Figure 2, a condition spectra of the synthetic monosaccharide-GA were collected differing from previous studies where ion trap CID had been as standards for comparisons. Figure 2a and b show the CID used [33]. This difference in activation could have contributed data of α- and β-D-Glcp-GA, respectively. The abundant to the formation of structural isomers observed for the m/z 221 peaks at m/z 101 and 131 in Figure 2a are a signature of a product ions. In order to test this hypothesis, m/z 221 ions non-reducing glucose with an α anomeric configuration. of 1–2 linked disaccharides, α-D-Glcp-(1–2)-D-Glc and β-D- Note that a distinct fragmentation pattern is observed for the Glcp-(1–2)-D-Glc, were formed using similar beam-type CID β configuration (Figure 2b), where m/z 131 and 161 ions are conditions and further subjected to MS3 CID (Figure 2e dominant. The same collisional activation conditions were and f). Except for a larger fluctuation in peak intensity applied to m/z 221 ions derived from α-D-Glcp-(1–3)-D-Glc for m/z 203, almost identical fragmentation patterns to and β-D-Glcp-(1–3)-D-Glc, anomeric isomers containing a the standards were observed (compare Figure 2a to e and non-reducing glucose. It is obvious that the spectra from the btof), strongly indicating that the expected D-Glcp-GA two anomeric isomers (Figure 2c and d) were drastically structures were formed. We further investigated a wide different from their corresponding D-Glcp-GA standards, variety of disaccharides and found that the CID patterns however, were similar to each other. This indicates that the of m/z 221 ions matched with their corresponding monosac- m/z 221 ions generated using low collision energies from m/z charide-GA standards with the exception of 1–3 linked 341 precursors have different structures from the D-Glcp-GA disaccharides when low collision energy beam-type CID standards, and that their CID patterns cannot be used to assign conditions were used to dissociate the disaccharides. 140

352 C. Konda et al.: Structure Determination for Disaccharides

18O-labeling at the carbonyl position of the reducing The inset in Figure 3a shows data collected using a wide sugar was used to mass-discriminate the “sidedness” of isolation window (6 m/z units) around m/z 221 after CID of dissociation events to either side of the glycosidic linkage m/z 343 precursor ions. A peak at m/z 223, due to and thus the origins of the m/z 221 and/or potential 223 incorporation of 18O, appeared with much higher abundance product ions. Figure 3 compares relatively low energy beam- than m/z 221 ions for both low-energy beam-type and ion type and ion trap CID of 18O-labeled deprotonated α-D- trap CID. Note that if the expected D-Glcp-GA structure Glcp-(1–3)-D-Glc, m/z 343. Similar fragments were ob- were formed, it should consist of the intact non-reducing served for both conditions; however, the ion abundance for sugar unit with a 2-carbon aglycon derived from the m/z 283 (loss of 60 Da, C2H4O2) was much higher in beam- reducing sugar (C-2 and C-3 or C-3 and C-4). In this case, type CID relative to ion trap CID. It is possible that this 18O should not be incorporated into the product ion and it fragmentation channel requires higher activation energy and should still appear at m/z 221. Therefore, the observation is promoted, even in lower-energy beam-type CID, since of abundant m/z 223 ions indicated that under relatively higher collision energies (several eV) may have been low energy CID conditions, most of the m/z 221 ions obtained as compared to ion trap CID (hundreds of meV). formed from α-D-Glcp-(1–3)-D-Glc do not have the D- Glcp-GA structure which is the structural isomer needed to distinguish the stereochemistry and anomeric configuration of the non-reducing sugar. 343 100 223 Ions at m/z 221 and m/z 223 were further subjected to ion (a) 2.9e5 trap CID. Figure 4 compares the CID spectra of the isolated m/z 221 ions and the m/z 223 ions generated from 18O- s ps % labeled α-D-Glcp-(1–3)-D-Glc and β-D-Glcp-(1–3)-D-Glc. cp

y, % The CID spectra of the m/z 221 ions (Figure 4a and c) from ity ty, ty, 221 the two anomeric isomers are distinct from each other and sit ns α β

ten almost identical to those of the corresponding and -D- ens te nt Glcp-GA standards (Figure 2a and b). The fragmentation In eI patterns of m/z 223 (Figure 4b and d), however, were similar ive 0

ati 283 to each other yet were very different from the synthetic

ela glucosyl-GA standards.

Re 163 179 The major fragments that resulted from CID of the m/z 223 ions included product ions at m/z 205, 163, 131, and 0 113. These ions are likely due to losses of water (−18 Da), a 60 100 140 180 220 260 300 340 2-carbon piece, C2H4O2 (−60 Da), a 3-carbon piece 18 18 m/z including O, C3H6O2+ O(−92 Da), and sequential or concerted losses of a three-carbon piece plus water 343 18 − 100 223 including O( 110 Da), respectively. Interestingly, neither (b) 404.0e4 4 the loss of water nor the loss of 60 Da significantly involves loss of the 18O oxygen. The m/z 223 ions are hypothesized s ps % to have a structure in which the reducing sugar is connected c y, % to a two-carbon piece from the non-reducing sugar as shown ty,

sity in the scheme above Figure 4b and d. Note that C-1 is no sit ns 221 longer chiral on the piece from the (former) non-reducing ten ens Int

nte sugar, which also explains the similarity in the CID data of In e I the m/z 223 ion derived from the two anomeric isomers ive 0 (Figure 4b and d). We also noticed some subtle differences at 179 between Figure 4b and d. For example, the relative Rel R 163 intensities of m/z 205 and 159 are higher (more than 10%) in Figure 4b than in d. These differences may be due to the 0 existence of a small fraction of structural isomers other than 100 140 180 220 260 300 340 60 that hypothesized for the m/z 223 ions. m/z Figure 3. MS2 spectra of 18O-labeled α-D-Glcp-(1–3)-D-Glc The Effect of CID Conditions on the Formation under (a) relatively low-energy beam-type CID (CE=6 V), and of m/z 221/223 Product Ions from 3-Linked (b) ion trap CID (AF2=25). Insets in (a) and (b) show isolation Disaccharides and their 18O-Labeled Isotopomers of m/z 221 and 223 ions generated from beam-type CID (injection time = 1 s) and ion trap CID (injection time = 2 s), As demonstrated in Figure 4, abundant structural isomers of respectively m/z 221 product ions, (m/z 223 ions from the 18O-labeled 141

C. Konda et al.: Structure Determination for Disaccharides 353 α-D-Glcp-(1-3)-Glc, m/z- 343 α-D-Glcp-(1-3)-Glc, m/z 343 -H -H - m/z 223 18 m/z 221 18 -H - -H -

163 100 131 100 (a) (b)

101 205 99 113 131 87 113 221 159 159 Relative Intensity, % Relative Intensity,

Relative Intensity, % Relative Intensity, 223 161 203 0 60 100 140 180 220 060 100 140 180 220 m/z m/z

β-D-Glcp-(1-3)-Glc, m/z 343 β-D-Glcp-(1-3)-Glc, m/z 343 18 -H - m/z 221 18 -H - m/z 223 -H - -H -

131 163 100 100 (c) (d)

161 113 131 205 221

Relative Intensity, % Relative Intensity, 87 Relative Intensity, % Relative Intensity, 159 223 101 159 203 0 60 100 140 180 220 0 60 100 140 180 220 m/z m/z

Figure 4. MS3 spectra of m/z 221 and 223 ions derived from 18O-labeled α-D-Glcp-(1–3)-D-Glc and β-D-Glcp-(1–3)-D-Glc. (a) CID of m/z 221 ions, CE=15 V (MS2), AF2=15 (MS3), (b) CID of m/z 223 ions, CE=5 V (MS2), AF2=14 (MS3) from 18O-labeled α- D-Glcp-(1–3)-D-Glc, and (c) CID of m/z 221 ions, CE=15 V (MS2), AF2=28 (MS3), (d) CID of m/z 223 ions, CE=5 V (MS2), AF2= 24 (MS3) from 18O-labeled β-D-Glcp-(1–3)-D-Glc disaccharides) were observed under relatively low-energy m/z 221 ions. When the CE was relatively low (CE=5 V), m/z dissociation conditions of 1–3 linked disaccharides, either 223 ions were predominantly formed, with four times higher using beam-type or ion trap CID. This prevents the intensity than that of m/z 221. At a higher CE (CE=10 V), m/z assignment of the stereochemistry or anomeric configuration 221 and 223 ions were seen at nearly equal intensities. Once the of the non-reducing sugar in a typical scenario where either a CE was increased to 15 V, m/z 221 ions became the dominant disaccharide is unlabeled or when it is isolated (unlabeled) peak, accounting for 80% of the total intensities from m/z 221 from a larger oligosaccharide structure. It would be highly and 223. Further increasing CE, however, resulted in a huge desirable to optimize CID conditions or, for that matter, to loss of ion abundance possibly due to competitive ion ejection find any dissociation conditions whereby the relatively pure, thus the ratio was not improved. structurally informative glycosyl-GA (m/z 221) ions could Parameters that might affect the formation of m/z 221 ions be formed predominantly. Figure 5 shows the effect of versus m/z 223 ions were also examined for ion trap CID. collision energies on the formation of m/z 221 and 223 ions When ion trap CID of m/z 343 was performed under the under beam-type and ion-trap CID, using 18O-labeled β-D- instrument default Mathieu q-parameter (q=0.235), m/z 223 Glcp-(1–3)-D-Glc as an example. The data were collected ions were formed exclusively independent of activation using a wide isolation window around m/z 221 to observe energies (data not shown). By changing the activation Mathieu both m/z 221 and 223 ions. q-parameter to a higher value, precursor ions are placed under a It is clear from Figure 5a–c that the collision energy in higher potential well depth, and higher activation energies can beam-type CID affects the absolute and relative intensities of be applied. An AC generated from an external waveform 142

354 C. Konda et al.: Structure Determination for Disaccharides

223 221 221 3.9e5 2.5e5 3.9e5 ()(a) (b) 223 (c) s s s ps ps cps c c y, c ty, ty, ty, sit sit nsi en ens ens 221 nte nte nte In In In 223

0 00 221 2233 221 223 221 223 m/z m/z m/z

223 223 223 91e49.1e4 10e51.0e5 221 (d) 20e52.0e5 (e) (f) ps ps ps cp cp cp y, y, y, sity sity 221 sity ns ns ns ten ten ten Int Int Int I I

0 0 221 223 0 221 223 221 223 m/z m/z m/z

Figure 5. Isolation of m/z 221 and 223 ions derived from 18O-labeled β-D-Glcp-(1–3)-D-Glc under different collisional activation conditions. Beam-type CID: (a) CE=5 V, (b) CE=10 V, (c) CE=15 V. Ion trap CID at q=0.4, f=119.248 kHz, excitation time = 50 ms: (d) 100 mVp-p, (e) 250 mVp-p, (f) 400 mVp-p generator was used for resonance excitation at q=0.4. As relative intensity) under ion trap CID conditions, it is shown in Figure 5d to f, the ratio of m/z 221 to m/z 223 ions was reasonable to conclude that the formation of these ions from increased from almost zero to about 1 as the activation 1–2 linkages needs less energy than required for their amplitude was increased from 100 mVp-p to 400 mVp-p generation from 1–3 linkages. Therefore, the formation of (activation time: 50 ms for all cases). Further increasing the the glycosyl-glycolaldehyde product ions is a much lower activation amplitude resulted in a decrease in m/z 221 to 223 energy dissociation channel for 1–2 linked disaccharides but ratio as well as a huge ion loss. a fragmentation pathway for this isomeric species can only The data in Figure 5 suggest that m/z 221 ions, which be promoted for 1–3 linked disaccharides when the collision have the desired monosaccharide-GA structures, are gener- energy is higher. ated more favorably under relatively high collision energy Given the high pressure in the collision cell (~5 mTorr), conditions in both beam-type and ion trap CID. Compared multiple collisions happen in beam-type CID, and the first- with ion trap CID, beam-type CID provided more abundant generation product ions may also be subjected to collisional and higher relative intensities of the m/z 221 ions that were activation once they are formed within the collision cell wanted for discrimination of the stereochemistry and especially under higher CE conditions. In this sense, beam- anomeric configuration of the non-reducing sugar. Evidently, type CID is less selective than ion trap CID, where fragment a higher activation energy is needed for the formation of ions are not typically further activated. Indeed, MS3 CID these m/z 221 product ions, and the pathway to generate the studies in the ion trap showed that many fragment ions, glycosyl-GAs is favored when the internal energies of the including m/z 325, 323, 283, 281, 253, and 251 generated molecular ions increase. In beam-type CID, much higher m/z 221 ions, which might contribute to the observation of collision energies can be applied (typically more than 10 V) higher intensity m/z 221 ions under beam-type CID due to as compared to ion trap CID (hundreds of mV), which leads secondary dissociation. to a shift in the internal energy distribution of the molecular ions to the high energy direction [38]. It is interesting to Identification of the Non-Reducing Sugar point out that the glycosyl-glycolaldehyde product ions are and its Anomeric Configuration for 1–3 Linked virtually the only isomeric species generated under ion trap Disaccharides or low-energy beam-type CID of the 1–2 linked disaccharide anions [34]. Since much higher relative intensities of the Since relatively pure m/z 221 ions containing the intact non- glycosyl-glycolaldehyde product ions (10%–40%, normal- reducing sugars could be formed using beam-type CID with ized to the most abundant peak) can be formed from 1–2 high collision energies, it was possible to differentiate the linkages compared with that of 1–3 linkages (typically G1% stereochemistry and anomeric configuration of the non- 143

C. Konda et al.: Structure Determination for Disaccharides 355

α-D-Glcp-(1-3)-Glc, m/z 341 β-D-Glcp-(1-3)-Glc, m/z 341 α-D-Glcp-(1-3)-Fru, m/z 341 -H - -H - -H -

m/z 221 m/z 221 m/z 221

131 131 131 100 (a)100 (b) 100 (c)

161 101 99101 99 113 161 113 87 87 161 129 159 221 87 221 129 159 221

Relative Intensity, % Relative Intensity, 113 Relative Intensity, % Relative Intensity, Relative Intensity, % Relative Intensity, 203 203 101 129 159 203 0 0 0 60 100 140 180 220 60 100 140 180 220 60 100 140 180 220 m/z m/z m/z

α-D-Galp-(1-3)-Gal, m/z 341 α-D-Manp-(1-3)-Man, m/z 341 -H - -H -

m/z 221 m/z 221

101 159 100 (d) 100 (e) 131 161

161 131 99 221 101 129 221 113129 87 203 Relative Intensity, % Relative Intensity, Relative Intensity, % Relative Intensity, 87 159 203 99 113 0 0 60 100 140 180 220 60 100 140 180 220 m/z m/z

Figure 6. MS3 CID of m/z 221 ions generated via using high CE (CE=13 to 22 V) for the dissociation of deprotonated disaccharide ions (m/z 341). (a) α-D-Glcp-(1–3)-D-Glc, CE=15 V (MS2), AF2=26 (MS3), (b) β-D-Glcp-(1–3)-D-Glc, CE=13 V (MS2), AF2=30 (MS3), (c) α-D-Glcp-(1–3)-D-Fru, CE=18 V (MS2), AF2=25 (MS3), (d) α-D-Galp-(1–3)-D-Gal (3α-Gal-Gal), CE=22 V (MS2), AF2=35 (MS3), and (e) α-D-Manp-(1–3)-D-Man (3α-Man-Man), CE=20 V (MS2), AF2=36 (MS3). The error bars in the spectra show the standard deviation of peak intensities based on seven spectra collected over a one year period

reducing sugar in disaccharides without 18O-labeling. Figure 6 configuration can be clearly identified for Figure 6a (α-D- shows the MS3 CID spectra of m/z 221 ions generated by Glcp-(1–3)-D-Glc) and Figure 6c (α-D-Glcp-(1–3)-D-Fru), beam-type CID with relatively high CE (13 to 22 V) from five which is distinct from the β-anomeric isomer as shown in 1–3 linked disaccharides. Each spectrum was an average of Figure 6b (β-D-Glcp-(1–3)-D-Glc, compare all three to the seven spectra and the error bars indicate standard deviations of synthetic standards shown in Figure 2a and b). Figure 6d the peak intensities. The standard deviations were found to be shows the characteristic fragmentation profile for the m/z higher (0%–12%) for the disaccharide samples than those from 221 ion of disaccharides having galactose as the non- standards (0%–4%). This larger degree of spectral variation is reducing sugar and having an α-anomeric configuration likely contributed by the fluctuation in ion intensity of the low (compare to Figure S3a,CIDofm/z 221 from α-D-Galp-GA abundance m/z 221 isomers under slightly different instrument standard, Supporting Information). The MS3 CID of α-D- conditions, and these isomers fragment differently from the Manp-(1–3)-Man (Figure 6e) was similar to that of the α-D- diagnostic and more abundant m/z 221 ions that have the Manp-GA (Supporting Information, Figure S3b). It is also monosaccharide-GA structures. important to note that under low-energy dissociation condi- Note that α-D-Glcp-(1–3)-D-Glc, α-D-Glcp-(1–3)-D-Fru, tions, the spectra of the m/z 221 product ions derived from the and β-D-Glcp-(1–3)-Glc are disaccharide isomers containing disaccharides α-D-Glcp-(1–3)-D-Fru, α-D-Galp-(1–3)-D-Gal a glucose as the non-reducing sugar; however, each has and α-D-Manp-(1–3)-Man did not match those of their either a different anomeric configuration or reducing sugar. respective glycosyl-glycolaldehydes (Supporting Informa- The characteristic fragmentation profile for disaccharides tion, Figure S2b and Figure S3c and d). We conclude having glucose as the non-reducing sugar and an α-anomeric that this is due to the presence of alternate isomers, pos< 144

356 C. Konda et al.: Structure Determination for Disaccharides

Table 1. Spectral Similarity Scores for 1–3 Linked Disaccharides Versus Monosaccharide-GA Standards

Disaccharides Synthesized Standards

α-D-Glcp-GA β-D-Glcp-GA α-D-Galp-GA α-D-Manp-GA

α-D-Glcp-(1–3)-Glc 0.9977 0.8845 0.8823 0.7968 β-D-Glcp-(1–3)-Glc 0.8572 0.9840 0.7608 0.8568 α-D-Glcp-(1–3)-Fru 0.9930 0.8899 0.9027 0.8045 α-D-Galp-(1–3)-Gal 0.9178 0.7530 0.9838 0.6942 α-D-Manp-(1–3)-Man 0.8461 0.8702 0.7527 0.9891

sibly related in their origins to the hypothetical structures Conclusions shown in Figure 4b and d but having different reducing monosaccharides. Collisional activation of deprotonated 1–3 linked hexose- The methodology for assigning the stereochemistry and containing disaccharides (m/z 341) generated a low-abundance anomeric configuration for the non-reducing sugar unit m/z 221 product ion. By 18O-labeling the reducing sugar within a disaccharide is based on the comparison of the carbonyl oxygen of these disaccharides, at least two CID patterns of m/z 221 ions to those of the synthetic structural isomers of the m/z 221 ion with the main portion monosaccharide-GA standards [33]. A high similarity derived from either side of the glycosidic linkage could be between the compared spectra indicates a large likelihood mass-discriminated (m/z 221 vs. m/z 223), which enabled of them sharing the same structure. Spectral similarity the isomers to be isolated and independently studied. The scores, which have been widely used in mass spectral library m/z 221 isomer containing the intact non-reducing sugar search for both small molecules [39], peptides [40–42], and attached in glycosidic linkage to a glycolaldehyde aglycon oligosaccharides [43], were chosen to facilitate these was found to be analytically useful, since CID of this comparisons. The spectral similarity scores were calculated species provided the structural information that identified between each of the averaged spectra in Figure 6 and the the stereochemistry and anomeric configuration of the non- averaged spectra from the monosaccharide-GA standards reducing sugar. No structural information could be obtained based on the following equation [44], from m/z 223 isomer(s) to determine the stereochemistry of the non-reducing sugar or its anomeric configuration. The P 1=2 P kI 1 I 2 I 2 formation of the diagnostic m/z 221 isomer was found to Spectral similarity score ¼ P m m ; and k ¼ P m 1 þ 2 1 be affected by CID conditions and was favored under kIm Im I 2 m higher energy beam-type CID. It was demonstrated that under optimized CID conditions, this structural isomer 1 2 where Im and Im are the normalized intensities of an ion at could be generated predominantly from five different 1–3 m/z = m for the two spectra. Note that the spectral similarity linked disaccharides without requiring 18O-labeling of the score always has a value between 0 and 1. If two spectra are reducing sugar. Identification of the non-reducing sugar and exactly the same, the spectral similarity score becomes 1. In the anomeric configuration, therefore, were achieved at a general, a large similarity score indicates close similarity high confidence level by statistically comparing the CID between the two spectra and a large degree of structural data of m/z 221 ions generated from the disaccharide similarity. As shown in Table 1, the spectral similarity samples with those of the synthetic standards via spectral scores between a standard and a disaccharide having the similarity scores. This study also demonstrated that beam- same stereochemistry and anomeric configuration for the type CID was a more desirable activation method compared non-reducing side were the highest scores, ranging between with ion trap CID to characterize disaccharides using the 0.9838 and 0.9977. When a disaccharide’s stereochemistry methodology based on the CID patterns of m/z 221 ions. and anomeric configuration on the non-reducing side did not This method now enables the anomeric configuration and match with the standard, the spectral similarity score was stereochemistry of the m/z 221 ions derived from 1–2, –3, –4, significantly lower, between 0.6942 and 0.9178. Clearly, by or –6 linked disaccharides to be assigned in the negative ion comparing the spectral similarity scores, assigning the mode. This capability was afforded due to the specific stereochemistry and anomeric configuration for the non- arrangement of the triple quadrupole-linear ion trap instru- reducing side of the 1–3 linked disaccharides could be ment. It combined (1) selection of the precursor (m/z 341) in achieved with high confidence. Note that this was only the first quadrupole with (2) higher energy dissociation in the possible under high-energy beam-type CID conditions where second quadrupole collision cell followed by (3) buildup of the m/z 221 product anions containing the intact non- the desired low abundance m/z 221 product ion in the linear reducing sugars were optimally generated from precursor trap, all three of which were necessary to obtain these disaccharides. structural details for 1–3 linked disaccharides. 145

C. Konda et al.: Structure Determination for Disaccharides 357

21. Staempfli, A., Zhou, Z., Leary, J.A.: Gas-phase dissociation mecha- Acknowledgment nisms of dilithiated disaccharides: Tandem mass spectrometry and semiempirical calculations. J. Org. Chem. 57, 3590–3594 (1992) B.B. acknowledges support from the National Science 22. Laine, R.A., Pamidimukkala, K.M., French, A.D., Hall, R.W., Abbas, Foundation CHE-0137986. S.A., Jain, R.K., Matta, K.L.: Linkage position in oligosaccharides by fast atom bombardment ionization, collision-activated dissociation, tandem mass spectrometry and molecular modeling. L-Fucosylp- ↦ → β → β References ( 1 X)-D-N-acetyl-D-glucosaminylp-( 1 3)-D-galactosylp-( 1-O- methyl) where X=3, 4, or 6. J. Am. Chem. Soc. 110, 6931–6939 (1988) 1. Taylor, M.E., Drickamer, K.: Introduction to Glycobiology, 2nd edn. 23. Hofmeister, G.E., Zhou, Z., Leary, J.A.: Linkage position determination Oxford University Press, Oxford (2006) in lithium-cationized disaccharides: tandem mass spectrometry and 2. Varki, A., Cummings, R.D., Esko, J.D., Freeze, H.H., Stanley, P., semiempirical calculations. J. Am. Chem. Soc. 113, 5964–5970 (1991) Bertozzi, C.R., Hart, G.W., Etzler, M.E.: Essentials of Glycobiology. 24. Asam, M.R., Glish, G.L.: Tandem mass spectrometry of alkali Cold Spring Harbor Laboratory Press, Cold Spring Harbor (2009) cationized polysaccharides in a quadrupole ion trap. J. Am. Soc. Mass 3. Bush, C.A., Martin-Pastor, M., Imberty, A.: Structure and conformation Spectrom. 8, 987–995 (1997) of complex carbohydrates of glycoproteins, glycolipids, and bacterial 28 – 25. Polfer, N.C., Valle, J.J., Moore, D.T., Oomens, J., Eyler, J.R., Bendiak, polysaccharides. Annu. Rev. Biophys. Biomolec. Struct. , 269 294 B.: Differentiation of isomers by wavelength-tunable infrared multiple- (1999) photon dissociation-mass spectrometry: Application to glucose-contain- 4. Bendiak, B., Fang, T.T., Jones, D.N.: An effective strategy for structural ing disaccharides. Anal. Chem. 78, 670–679 (2005) elucidation of oligosaccharides through NMR spectroscopy combined 26. Simoes, J., Domingues, P., Reis, A., Nunes, F.M., Coimbra, M.A., with peracetylation using doubly 13C-labeled acetyl groups. Can. J. 80 – Domingues, M.R.: Identification of anomeric configuration of under- Chem. , 1032 1050 (2002) ivatized reducing glucopyranosyl-glucose disaccharides by tandem 5. Armstrong, G.S., Mandelshtam, V.A., Shaka, A.J., Bendiak, B.: Rapid mass spectrometry and multivariate analysis. Anal. Chem. 79, 5896– high-resolution four-dimensional NMR spectroscopy using the filter 5905 (2007) diagonalization method and its advantages for detailed structural 27. Zhang, H., Brokman, S.M., Fang, N., Pohl, N.L., Yeung, E.S.: Linkage elucidation of oligosaccharides. J. Magn. Reson. 173, 160–168 (2005) position and residue identification of disaccharides by tandem mass 6. Zaia, J.: Mass spectrometry of oligosaccharides. Mass Spectrom. Rev. spectrometry and linear discriminant analysis. Rapid Commun. Mass 23, 161–227 (2004) Spectrom. 22, 1579–1586 (2008) 7. Fenn, J., Mann, M., Meng, C., Wong, S., Whitehouse, C.: Electrospray ionization for mass spectrometry of large biomolecules. Science 246, 28. Mendonca, S., Cole, R., Zhu, J., Cai, Y., French, A., Johnson, G., 64–71 (1989) Laine, R.: Incremented alkyl derivatives enhance collision induced glycosidic bond cleavage in mass spectrometry of disaccharides. J. Am. 8. Reinhold, V.N., Reinhold, B.B., Costello, C.E.: Carbohydrate molecular Soc. Mass Spectrom. 14,63–78 (2003) weight profiling, sequence, linkage, and branching data: ES-MS and CID. Anal. Chem. 67, 1772–1784 (1995) 29. Gaucher, S.P., Leary, J.A.: Stereochemical differentiation of mannose, glucose, galactose, and talose using zinc(II) diethylenetriamine and ESI- 9. Karas, M., Bachmann, D., Bahr, U., Hillenkamp, F.: Matrix-assisted 70 – ultraviolet laser desorption of non-volatile compounds. Int. J. Mass ion trap mass spectrometry. Anal. Chem. , 3009 3014 (1998) Spectrom. 78,53–68 (1987) 30. Desaire, H., Leary, J.A.: Differentiation of diastereomeric N-acetylhexosamine monosaccharides using ion trap tandem mass spectrometry. 10. Tanaka, K., Waki, H., Ido, Y., Akita, S., Yoshida, Y., Yoshida, T., 71 – Matsuo, T.: Protein and polymer analyses up to m/z 100,000 by laser Anal. Chem. , 1997 2002 (1999) ionization time-of-flight mass spectrometry. Rapid Commun. Mass 31. Mueller, D.R., Domon, B.M., Blum, W., Raschdorf, F., Richter, W.J.: Spectrom. 2, 151–153 (1988) Direct stereochemical assignment of sugar subunits in naturally 11. Harvey, D.J.: Quantitative aspects of the matrix-assisted laser desorp- occurring glycosides by low energy collision induced dissociation. 15 – tion mass spectrometry of complex oligosaccharides. Rapid Commun. Application to papulacandin antibiotics. Biol. Mass Spectrom. , 441 Mass Spectrom. 7, 614–619 (1993) 446 (1988) 12. Domon, B., Costello, C.E.: A systematic nomenclature for carbohydrate 32. Domon, B., Muller, D.R., Richter, W.J.: Determination of interglyco- fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconj. sidic linkages in disaccharides by high performance tandem mass 100 – J. 5, 397–409 (1988) spectrometry. Int. J. Mass Spectrom. Ion Process. , 301 311 (1990) 13. Domon, B., Costello, C.E.: Structure elucidation of glycosphingolipids 33. Fang, T.T., Bendiak, B.: The stereochemical dependence of unim- and gangliosides using high-performance tandem mass spectrometry. olecular dissociation of monosaccharide-glycolaldehyde anions in the Biochemistry 27, 1534–1543 (1988) gas phase: A basis for assignment of the stereochemistry and anomeric 14. Garozzo, D., Giuffrida, M., Impallomeni, G., Ballistreri, A., Montaudo, configuration of monosaccharides in oligosaccharides by mass spec- G.: Determination of linkage position and identification of the reducing trometry via a key discriminatory product ion of disaccharide end in linear oligosaccharides by negative ion fast atom bombardment fragmentation, m/z 221. J. Am. Chem. Soc. 129, 9721–9736 (2007) mass spectrometry. Anal. Chem. 62, 279–286 (1990) 34. Fang, T.T., Zirrolli, J., Bendiak, B.: Differentiation of the anomeric 15. Dallinga, J.W., Heerma, W.: Reaction mechanism and fragment ion configuration and ring form of glucosyl-glycolaldehyde anions in the structure determination of deprotonated small oligosaccharides, studied gas phase by mass spectrometry: isomeric discrimination between m/z by negative ion fast atom bombardment (tandem) mass spectrometry. 221 anions derived from disaccharides and chemical synthesis of m/z Biol. Mass Spectrom. 20, 215–231 (1991) 221 standards. Carbohydr. Res. 342, 217–235 (2007) 16. Carroll, J.A., Ngoka, L., Beggs, C.G., Lebrilla, C.B.: Liquid secondary 35. Brown, D.J., Stefan, S.E., Berden, G., Steill, J.D., Oomens, J., Eyler, J. ion mass spectrometry/Fourier transform mass spectrometry of oligo- R., Bendiak, B.: Direct evidence for the ring opening of monosaccha- saccharide anions. Anal. Chem. 65, 1582–1587 (1993) ride anions in the gas phase: photodissociation of aldohexoses and 17. Mulroney, B., Traeger, J.C., Stone, B.A.: Determination of both linkage aldohexoses derived from disaccharides using variable-wavelength position and anomeric configuration in underivatized glucopyranosyl infrared irradiation in the carbonyl stretch region. Carbohydr. Res. disaccharides by electrospray mass spectrometry. J. Mass Spectrom. 30, 346, 2469–2481 (2011) 1277–1283 (1995) 36. Carroll, J.A., Ngoka, L., Beggs, C.G., Lebrilla, C.B.: Liquid secondary 18. Guan, B., Cole, R.B.: MALDI linear-field reflectron TOF post-source ion mass spectrometry/Fourier transform mass spectrometry of oligo- decay analysis of underivatized oligosaccharides: Determination of saccharide anions. Anal. Chem. 65, 1582–1587 (1993) glycosidic linkages and anomeric configurations using anion attach- 37. Bendiak, B., Fang, T.T.: Assignment of the stereochemistry and ment. J. Am. Soc. Mass Spectrom. 19, 1119–1131 (2008) anomeric configuration of structurally informative product ions derived 19. Jiang, Y., Cole, R.B.: Oligosaccharide analysis using anion attachment from disaccharides: infrared photodissociation of glycosyl-glycolalde- in negative mode electrospray mass spectrometry. J. Am. Soc. Mass hydes in the negative ion mode. Carbohydr. Res. 345, 2390–2400 Spectrom. 16,60–70 (2005) (2010) 20. Firdoussi, A.E., Lafitte, M., Tortajada, J., Kone, O., Salpin, J.-Y.: 38. Wells, J.M., McLuckey, S.A. Collision-Induced Dissociation (CID) of Characterization of the glycosidic linkage of underivatized disaccharides Peptides and Proteins. In: Biol. Mass Spectrom. Burlingame, A.L. (ed.) by interaction with Pb2+ ions. J. Mass Spectrom. 42, 999–1011 (2007) Academic Press, Amsterdam, pp. 148–185 (2005) 146

358 C. Konda et al.: Structure Determination for Disaccharides

39. Stein, S., Scott, D.: Optimization and testing of mass spectral library 42. Lam, H., Deutsch, E.W., Eddes, J.S., Eng, J.K., Stein, S.E., Aebersold, search algorithms for compound identification. J. Am. Soc. Mass R.: Building consensus spectral libraries for peptide identification in Spectrom. 5, 859–866 (1994) proteomics. Nat. Methods 5, 873–875 (2008) 40. Frewen, B.E., Merrihew, G.E., Wu, C.C., Noble, W.S., MacCoss, M.J.: 43. Kameyama, A., Kikuchi, N., Nakaya, S., Ito, H., Sato, T., Analysis of peptide MS/MS spectra from large-scale proteomics experi- Shikanai, T., Takahashi, Y., Takahashi, K., Narimatsu, H.: A ments using spectrum libraries. Anal. Chem. 78,5678–5684 (2006) strategy for identification of oligosaccharide structures using 41. Lam, H., Deutsch, E.W., Eddes, J.S., Eng, J.K., King, N., Stein, S.E., observational multistage mass spectral library. Anal. Chem. 77, Aebersold, R.: Development and validation of a spectral library 4719–4725 (2005) searching method for peptide identification from MS/MS. Proteomics 44. Zhang, Z.: Prediction of low-energy collision-induced dissociation 7, 655–667 (2007) spectra of peptides. Anal. Chem. 76, 3908–3922 (2004)