Structural characterization of carbohydrate attached to the glycoprotein cellulase enzymes of Trichoderma reesei QM 9414


Material Information

Structural characterization of carbohydrate attached to the glycoprotein cellulase enzymes of Trichoderma reesei QM 9414
Physical Description:
xlii, 176 leaves : ill. ; 29 cm.
Du Meé, Charles Philip Roger de Chasteigner, 1955-
Publication Date:


Subjects / Keywords:
Mitosporic Fungi   ( mesh )
Glycoproteins   ( mesh )
Carbohydrate Sequence   ( mesh )
Biochemistry and Molecular Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- Biochemistry and Molecular Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1984.
Bibliography: leaves 166-174.
Statement of Responsibility:
by Charles Philip Roger de Chasteigner du Meé.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000424817
oclc - 12385308
notis - ACH3292
sobekcm - AA00004890_00001
System ID:

Full Text

OF Trichoderma reesei

QM 9414





I wish to dedicate this work first to my wife, Linda, who stood

by my side and who provided encouragement, love and emotional support

throughout and also to my two sons, Colin and lan, who were both born

during this time.

I would also like to dedicate this work to two people who

instilled in me the drive to succeed and to enjoy science and without

whom I may not have followed the course I did. Firstly, Dr. Bob Coley,

who taught me the fun of organic chemistry and secondly Dr. Bob

Dekker, who introduced me to biochemical research and helped develop

my confidence and ambition.


I would like to thank my advisor, Dr. Ross D. Brown, Jr., for

his support and enthusiasm for my doctoral work and also for his

advice and direction during my graduate career.

I would like to sincerely thank several colleagues who were

invaluable during various phases of the project: Ms. Cindy Jackson and

Dr. John Gander for their contributions and performance of NMR

experiments, Mr. Mike Trehy and Mr. Jim Templeton for performing the

gas chromatography/mass spectrometry analyses, Ms. Virginia Wiley for

operating the amino acid analyzer and Dr. Mikelina Gritzali for the

preparation of the antisera. I would also like to thank Mr. William

Chirico for his innumerable helpful suggestions and hints over the

past five years and for his being a good friend and confidant.

Last, but not least, I would like to thank my advisory committee

for their helpful suggestions and recommendations and also Dr. Peter

McGuire in whose laboratory I was able to learn the art of in vitro

translation systems.



ACKNOWLEDGEMENTS ............................................. iii

LIST OF TABLES ................................... ......... .. .. vi

LIST OF FIGURES ............................................ viii

ABBREVIATIONS .......................................... .... xi

ABSTRACT .................................................... xii

INTRODUCTION ................................................. 1

Glycoproteins ......................................... 1

Fungal Glycoenzymes .................................... 4

Cellulases ............................................. 6

Structure:Function Relationships for Fungal Glycoenzymes 14

Assessment .................................. .......... 15

EXPERIMENTAL PROCEDURES ...................................... 17

Materials ............................................ 17

Methods ........................................... 20

RESULTS AND DISCUSSION ....................................... 56

Cellobiohydrolase I(D) ................................. 56

Cellobiohydrolase II ................................. 88

Endoglucanases ......................................... 118

SUMMARY .................................................... 144

APPENDICES ................................................ 152

A. Electron impact mass spectra of peracetylated
alditols ........................................... 152

B. Electron impact mass spectra of peracetylated
aldononitriles ..................................... 154

C. Electron impact mass spectra of partially methylated
alditol acetates .................................. 156

D. Chemical ionization mass spectra of peracetylated
alditols .......................................... 159

E. Chemical ionization mass spectra of peracetylated
aldononitriles .................................... 161

F. Chemical ionization mass spectra of partially
methylated alditol acetates ........................ 163

REFERENCES .......................................... ............. 166

BIOGRAPHICAL SKETCH .......................................... 175


Table No. Page

I. Carbohydrate content of Trichoderma
cellulase enzymes ................................ 8

II. Neutral carbohydrate composition of
cellobiohydrolase I(D) ........................... 60

III. Oligosaccharides released by preparative
reductive B-elimination of
cellobiohydrolase I(D) ........................... 64

IV. Composition of oligosaccharides released from
cellobiohydrolase I(D) by reductive
B-elimination .................................... 65

V. Selected 13C-NMR signals due
carbohydrate covalently attached to
cellobiohydrolase I(D) and due to purified
oligosaccharides released from this enzyme by
reductive B-elimination .......................... 81

VI. Proton-NMR signal assignments of anomeric protons
of purified oligosaccharides released from
cellobiohydrolase I(D) reductive B-elimination ... 86

VII. Neutral carbohydrate composition of
cellobiohydrolase II ............................. 92

VIII. Oligosaccharides released by preparative
reductive B-elimination of
cellobiohydrolase II ............................ 96

IX. Composition of oligosaccharides released from
cellobiohydrolase II by reductive
B-elimination .................................... 97

X. Selected 1C-NMR signals due
cellobiohydrolase II and due to the purified
trisaccharide released from this enzyme by
reductive B-elimination ......................... 110

XI. Proton-NMR signal assignments of anomeric protons
of purified oligosaccharides released from
cellobiohydrolase II by reductive B-elimination .. 116

Table No. Page

XII. Selected 13C-NMR signals due to
carbohydrate covalently attached to the
endoglucanases of T. reesei QM 9414 .............. 138


Figure No. Page

1. Isocratic elution pattern from DEAE-Sephadex column
chromatography of the extracellular protein
preparation from T. reesei QM 9414 arown on
cellulose ......................................... 22

2. Isocratic elution pattern from SP-Sephadex column
chromatography of cellobiohydrolase II and the
endoqlucanases which had been eluted from a
DEAE-Sephadex column ............................... 25

3. Gas chromatoaraphic separation of peracetylated
alditol and aldononitrile derivatives of neutral
monosaccharides released from cellobiohydrolase I(D)
after reductive B-elimination and subsequent acid
hydrolysis ........................................ 32

4. Efficacy of sugar release from cellobiohydrolase I(D)
by reductive B-elimination .......................... 35

5. Quantitation of aldoses as the peracetylated
aldononitriles ...................................... 38

6. Quantitation of alditols as the peracetylated
alditols .......................................... 40

7. Molecular ion scanning of peracetylated alditol and
aldononitrile derivatives of monosaccharides released
from cellobiohydrolase I(D) following reductive
B-elimination and subsequent acid hydrolysis ........ 50

8. Polyacrylamide disc gel electroohoresis of crude
extracellular protein prepared from T. reesei
QM 9414 and highly purified cellobiohydrolase I(D) .. 58

9. Separation on a Biogel P-2 column of oliqosaccharides
released from cellobiohydrolase I(D) by reductive
B-elimination ....................................... 62

10. Methylation analysis of cellobiohydrolase I(D) and
the oligosaccharides released from cellobiohydrolase
I(D) by reductive B-elimination ..................... 68

Figure No. Page

11. Comparison of electron impact spectra obtained from
partially methylated alditol acetates of the
tetrasaccharide from cellobiohydrolase I(D) with
standards ........................................... 71

12. HPLC separation of the products of sequential
glycosidase digestion of oligosaccharides released
from cellobiohydrolase I(D) by reductive
B-elimination ......................................... 73

13. HPLC separation of the products of acetolysis of
oligosaccharides released from cellobiohydrolase I(D)
by reductive B-elimination .......................... 77

14. Proton decoupled 13C-NMR spectra of
cellobiohydrolase I(D) and the oliqosaccharides
released from cellobiohydrolase I(D) by reductive
B-elimination ......................... ............. 80

15. Proton-NMR spectra at 300 MHz of oligosaccharides
released from cellobiohydrolase I(D) by reductive
B-elimination ..................................... 85

16. Polyacrylamide disc gel electroohoresis of crude
extracellular protein prepared from T. reesei
QM 9414 highly purified cellobiohydrolase II ........ 90

17. Separation on a Biooel P-2 column of oligosaccharides
released from cellobiohydrolase II by reductive
B-elimination ...................................... 95

18. Methylation analyses of cellobiohydrolase II and the
oligosaccharides released from cellobiohydrolase II
by reductive B-elimination ......................... 100

19. HPLC separation of the products of sequential
glycosidase digestion of oligosaccharides released
from cellobiohydrolase II by reductive
B-elimination ....................... .... ..... ...... 103

20. HPLC separation of the products of acetolysis of the
trisaccharide released from cellobiohydrolase II by
reductive B-elimination ........................... 107
21. Proton decoupled 1C-NMR spectra of
cellobiohydrolase II and the trisaccharide released
from cellobiohydrolase II by reductive
B-elimination ....................................... 109

22. Proton-NMR spectra at 300 MHz of oligosaccharides
released from cellobiohydrolase II by reductive
B-elimination ................... ........... .... .. 115

Figure No. Page

23. Polyacrylamide disc gel electrophoresis of successive
protein fractions eluted isocratically from an
SP-Sephadex column during chromatography of
T. reesei endoglucanases .......................... 120

24. Elution pattern from Sephacryl S-200 column
chromatograohy of endoglucanase E ................... 123

25. Polyacrylamide disc gel electrophoresis of successive
protein fractions eluted from a Sephacryl S-200
column during chromatography of endoglucanase E ..... 125

26. Elution pattern from SP-Sephadex column
chromatography of endoolucanase B and endoglucanase D
using a pH gradient ............................... 128

27. Polyacrylamide disc ael electrophoresis of successive
protein fractions eluted from an SP-Sephadex column
during chromatoaraphy of endoglucanase B and
endoalucanase D using a pH gradient ................. 130

28. Elution pattern from Sephadex G-75 column
chromatography of endoglucanase B ................... 132

29. Polyacrylamide disc gel electrophoresis of successive
protein fractions eluted from a Sephadex G-75 column
during chromatography of endoglucanase B ............ 134

30. Polyacrylamide disc gel electroohoresis of crude
extracellular protein prepared from T. reesei
QM 9414 and purified endoglucanases ................. 136

31. Proton decoupled 13C-NMR spectra of the
endoolucanases purified from T. reesei QM 9414
grown on cellulose .................................. 140

32. Proposed structures of the oligosaccharides
covalently attached to the glycoprotein cellulases
from T. reesei QM 9414 ............................ 146


M,M 2M G,
M2 2

- Cellobiohydrolase I(D) from T. reesei QM 9414
- Cellobiohydrolase II from T. reesei QM 9414
- Gas-liquid chromatography
- Gas chromatography/mass spectrometry
- Gas chromatograph/mass spectrometer/data system
- Electron impact
- Chemical ionization
- Molecular ion
- Nuclear magnetic resonance
- High pressure (performance) liquid chromatography
- Peracetylated alditol
- Peracetylated aldononitrile
- Diethylaminoethyl
- Sulfopropyl
- Serine
- Threonine
- AsDaragine
- a-aminobutyric acid
- Xylose
- Mannose
- Glucose
- N-acetylglucosamine

- Oligosaccharides released from the cellobiohydrolases
(where M = Man and G = Glc)
- Dimethylsulfoxide
- (Trimethylsilane)-l-propane sulfonate

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

OF Trichoderma reesei QM 9414



August, 1984

Chairman: Dr. Ross D. Brown, Jr.
Major Department: Biochemistry and Molecular Biolooy

Cellobiohydrolases I(D) and II were purified by ion exchange

chromatography from an extracellular culture filtrate of Trichoderma

reesei QM 9414. Neutral sugar composition of each was determined by

gas-liquid chromatographic analysis of the peracetylated alditol and

aldononitrile acetate derivatives of sugars released by either

reductive B-elimination or acid hydrolysis. Mannose and glucose were

found to be the only neutral sugars and were covalently attached to

the proteins through alkali-labile mannosyl residues. Analysis of the

alkaline borohydride-released carbohydrate by high pressure liquid

chromatography (HPLC) demonstrated that each molecule of

cellobiohydrolase I(D) contained 5.9% carbohydrate comprising 0.7

tetra-, 4.2 tri-, 1.1 di- and 1.2 monosaccharides and that

cellobiohydrolase II contained 18.9% carbohydrate comprising 14.9

tri-, 1.6 di- and 9.0 monosaccharides. The purified oligosaccharides

were shown by methylation analysis to contain (1-2) and (1-6)

glycosidic linkages and the position of 6-substituted residues was

confirmed by acetolysis. The sequence and anomeric nature of the sugar

residues in each oligosaccharide was determined by sequential

glycosidase digestion and all the residues were found to be a-linked.

Proton decoupled 13C-NMR analysis suggested that for

cellobiohydrolase I(D), each oligosaccharide was attached to a

threonyl residue on the polypeptide and for cellobiohydrolase II, each

oligosaccharide was attached to threonyl and seryl residues on the

DolypeDtide. These data were supported by amino acid analysis for

a-aminobutyric acid after B-elimination and palladium chloride

treatment of each protein. Coupled 13C-NMR analysis confirmed that

the anomeric carbon of each sugar residue was in the a-configuration

and 1H-NMR identified the presence of mannosyl and a-alucosyl

residues in the purified oliqosaccharides. The structures of the

oligosaccharides attached to the enzymes were determined to be

mannosyla(1-2)glucosyla(1-6)mannosyla(l-2)mannose, qlucosyla(1-6)

mannosyla(l-2)mannose, mannosyla(l-2)mannose and mannnose. This is the

first report of a unique glucosyla(l-6)mannose linkage. Thus, work

with the cellobiohydrolases and the endoglucanases indicates that the

predominant cellulase enzymes secreted by T. reesei QM 9414 are each

glycosylated with similar oligosaccharides.


Ghose [1] has estimated the annual worldwide production of bio-

degradable substances through photosynthesis at approximately 1.8 x

1012 tons, of which 40% is cellulose. As much as 25% of this [2]

could be made available for conversion processes, a significant amount

of which occurs as agricultural and municipal wastes. Therefore, in

the future, cellulose must be regarded as an important potential

source of fuel, food and chemical feedstocks.

Enzymatic saccharification of cellulose has proved to be both

specific and efficient, and among the organisms tested, Trichoderma

species showed the highest levels of extracellular activity [3]. Most

fungal cellulases have been shown to contain covalently attached carbo-

hydrate, although the complete structure of the oligosaccharides

attached to the polypeptide of a cellulase enzyme has never been



Proteins which have covalently associated carbohydrate are

termed glycoproteins [4]. These are a diverse group of macromolecules

found throughout nature in plants, animals and microorganisms.

Glycoproteins are implicated in a wide variety of functions including

cell adhesion, molecular recognition, structural support, lubrication,


hormone control, blood clotting and enzyme catalysis. With the

increased sophistication of modern analytical methods, the structure

of the oligosaccharides attached to many glycoproteins have been

determined and are well documented [4-8]. The oligosaccharide side

chains of glycoproteins fall into two general classes, (i) those which

are attached via an N-glycosidic linkage from an N-acetylglucos-

aminyl residue to the amide nitrogen of an asparaginyl residue on the

protein and (ii) those which are 0-glycosidically linked from a

neutral sugar or an N-acetylgalactosaminyl residue to either a seryl,

threonyl, hydroxyprolyl or hydroxylysyl residue on the protein.

The N-linked oligosaccharides are similar in that they all

contain a pentasaccharide core structure, Mana(1-3)[Man a(1-6)]Man-

B(1-4)G1cNAcB(1-4)GlcNAcB-(Asn), attached to the protein. This

class of oligosaccharides comprises three general groups, (i) those

that have only mannosyl residues in addition to the core are termed

"simple", (ii) those that have one or more of a variety of sugars,

galactose, N-acetylqalactosamine, sialic acid and/or fucose in

addition to the core, are termed "complex" and (iii) those that have a

mixture of "simple" and "complex" are termed "hybrid". The N-linked

class of glycoproteins has a specific amino acid sequence requirement

for addition of the carbohydrate to the protein. The sequence Asn-X-

Thr(Ser) must be present, where X can be almost any amino acid except

proline [5].

Oligosaccharides of this class are transferred from a lipid-

linked precursor en bloc to the protein. This precursor is a large

branched structure containing mannose and glucose. Following transfer,

the glucosyl residues and most of the mannosyl residues are removed

and the core is subsequently enlarged to the simple, complex or hybrid

forms. The glycoproteins containing these types of oligosaccharides

are represented diversely in nature. They are cell surface [9,10],

secretary [11,12], plasma [13,14] and hormone polypeptides [15,16] to

mention a few.

The 0-linked oligosaccharides are not known to have such an

amino acid sequence requirement, although it has been suggested that a

conformational prerequisite, i.e. a B-turn, is necessary. Sugars are

thought to be added directly to the protein one by one rather than by

transfer of an oligosaccharide en bloc. This class of oliqosaccha-

rides also comprises three general groups, (i) those that have

N-acetylgalactosamine attached to the hydroxyl oroup of serine or

threonine, (ii) those that have neutral sugars attached to the

hydroxyl group of hydroxylysine or hydroxyproline and (iii) those that

have neutral sugars attached to the hydroxyl group of serine or

threonine. The oligosaccharides of the first group are generally

short, one to six residues in length, although some are much larger.

They may also contain other sugars such as qalactose, fucose,

N-acetylglucosamine and/or sialic acid on the chains. This group is

exemplified by the mucins, but also includes antifreeze glycoproteins

[17], chorionic gonadotropin [18], blood group substances [19] and

cartilage keratan sulfate [20]. The second class is characterized by

several structural proteins, the collagens [21,22] and basement

membrane proteins [23,24] in animals, which have galactose 0-linked

to hydroxylysine and a group of plant glycoproteins called extensins,

which have arabinose and galactose attached to hydroxyproline. For the

third group, the neutral sugar can be xylose in the case of

chondroitin sulfate and dermatan sulfate [25], galactose in the case

of cuticle collagen [26,27], fucose in the case of several mammalian

cell line glycoproteins [28,29] and mannose in the case of fungal

[30-38] and yeast [39,40] glycoproteins.

Fungal Glycoenzymes

With the exception of cuticle collagen from Nereis virens,

which was found by Spiro and Bhoyroo [26] to contain a unique acidic

disaccharide of 6-O-a-D-glucuronosyl-O-D-mannose, all of the qlyco-

oroteins currently known to contain a mannosyl-0-Thr(Ser) linkage

are of fungal origin. Of these, the majority are fungal carbohydrases

which are secreted from the respective organisms.

The qlucoamylases ((1,4)(1,6)-cL-D-glucanglucohydrolases) from a

number of Aspergillus species have been analyzed for carbohydrate

content and composition [31-35]. Pazur et al. [31] removed the neutral

carbohydrate of glucoamylase I from A. niger with mild alkaline

treatment. The oligosaccharide chains released ranged in size from one

to five sugars, with 20 monosaccharides, all being mannose; 11

disaccharides, all mannobiose; and the larger chains which were

combinations of glucose, galactose and mannose. Methylation and

G.C./M.S. analysis determined the disaccharides to be (1-2) linked and

the larger species to be branched and containing (1-3) and (1-6)

glycosidic bonds. The structures of these larger species, however,

were never definitively described. Manjunath and Rao [33,34] prepared

a series of glycopeptides from glucoamylase II of A. niger following

extensive pronase digestion. Monosaccharides were subsequently

released from the glycopeptides by sequential glycosidase digestion

and the products separated and identified by paper chromatography. A

series of trisaccharides were described with the general formula,

X-Man-Man-, where X was either glucose, galactose, mannose, N-acetyl-

glucosamine or xylose. The residues were a-linked in each case. The

procedure used, however, did not exclude the possibility of branched

structures, as substitution of individual sugars was not determined by

classical methylation analysis. Carbohydrate and amino acid analysis

following reductive B-elimination identified mannose as the only

sugar attached to either threonine or serine on the polypeptide.

Comparison of glucoamylases I and II from A. niqer 1,2 and 3, Asper-

qillus foetidus and Aspergillus candidus showed that most of the

enzymes contained mannose, glucose, galactose, xylose and glucosamine

[35]. The Aspergillus glucoamylases are heavily glycosylated and

contain between 48 and 151 monosaccharides covalently attached to each

protein molecule [35,36]

Rosenthal and Nordin [37] have characterized a mycodextranase

(endo-1,4-a-D-glucanase) from Penicillium melinii. Glycosidase

digestions, Smith degradation and methylation analysis indicate that

the oligosaccharides are present as about 25 mannose, Glca(1-2)Man and

Mana(1-2)Glca(1-2)Man side chains.

Raizada and Schutzbach [38] characterized a Man

a(1-2)Mana(1-2)Mannitol oligosaccaharide from a major cell envelope

glycoprotein of the fungus imperfectus, Cryptococcus laurentii. The

trisaccharide was removed from the protein by reductive B-elimination

using H-NaBH4 and the radiolabelled product isolated by gel

filtration. Characterization was performed by glycosidase digestion

and comparison of the products with standards by paper chromatography.

While not enzymatic in nature, Saccharomyces cerevisiae cell

wall mannan was shown by Nakajima and Ballou [39] to contain a number

of similar 0-linked mannooligosaccharides. These were released from

the protein by mild alkaline treatment and individually purified by

gel filtration. Glycosidase digestion and gas chromatography were used

to identify a series of related oligosaccharides, the largest being

Man a(1-3)Man a(1-2)Mana( 1-2)Man.


Most cellulases are of fungal origin and are glycoprotein in

nature, but relatively little is known about their structure. Unlike

their a-glucanase counterparts such as the glucoamylases and the

dextranases, the cellulases are thought to exist as a system of


The cellulase system of Trichoderma reesei QM 9414 comprises

three types of enzymes which act synergistically and can effect the

complete conversion of both amorphous and crystalline cellulose to

glucose [41,42]. The system consists of several endo-1,4-f3-Dglucana-

ses (EC, two exo-1,4-B-D-cellobiohydrolases (EC

and one or more B-D-glucosidases (EC[43,44]. Endoglucan-

ases release soluble oligosaccharides and free chain ends from

cellulose, cellobiohydrolases release cellobiose from free chain ends

and B-glucosidases release glucose from soluble cellooligo-


It has been shown that the biosynthesis of endoglucanases and

cellobiohydrolases can be induced by addition of sophorose (0-B-D-

glucopyranosyl-1,2-0-glucopyranose) to resting cells of this

organism [43,45-47]. Gritzali and Brown [43] have shown that cellobio-

hydrolase I, form D (CBH I(D)) from cellulose-grown and sophorose-

induced Trichoderma cultures have the same electrophoretic mobility,

amino acid composition and total carbohydrate content suggesting that

they are the same protein. CBH I(D) is so named as it has been shown

to have the same amino acid composition as three other forms of

cellobiohydrolase I, A, B and C from Trichoderma viride [43,48]. Gum

and Brown [48] proposed that these forms only differ structurally in

the amount of covalently attached carbohydrate. Cellobiohydrolase II

(CBH II) from cellulose-grown cells also exhibited the same

electrophoretic pattern as its induced counterpart although no

comparative compositional studies on the induced enzyme were

performed. CBH I(D) and CBH II have been shown by several criteria to

be separate proteins. Apart from electrophoretic mobility, the two

iiave distinct amino acid compositions [42,49]. Fagerstam and

Pettersson [42] have also determined that the sequence of the

N-terminal 20 amino acids for each polypeptide was different, although

both N-terminii were blocked by a pyroglutamic acid residue.

Most of the cellulolytic enzymes isolated from Trichoderma are

glycoprotein in nature [30,43,48-56], although some have no covalently

bound carbohydrate [57,58]. The carbohydrate content of each type of

cellulase from this genus varies considerably and is summarized in

Table I. Gum and Brown [30] have presented the only attempt so far at

structural characterization of the carbohydrate of a cellulolytic




0= 0

"- c L

', I ( -)0

L t-
C) =1-

0 O-


"O 0
0 Z


0 0

0 0



0 S
0 (/Ir
U Zv

0 0 -4 C\ c O co
In in In LO qa z

OO 0

x (D m m x C m c D '- CD C U( D CD U U C U
* 3 ^ (" 3 3 3 0 ~ 5t '

CV CM l3- 00 K3T

Om -i

"-a wI-

o 0 *r- *> *- *- *
Vl l~ -o -l > la

'1' *f r1 *
M I- IM I- I l- l i- l > 1t 1 Il






a U

CO C) C) C)
C4 c,

0 0 0
0 0 0
0 0 0
C) r) C

m Imm CD
CD C) C0
mZ m m

<- -> C-5

0 0
0 0
0 C*"

ID *1-


Q0 0
en enz

Co C)

LO -n (r) Ln Mn 0 -O C'J -n

* .. *-* %*. %* % %.. *.. ..
0 0 00 00 0 e r
Z Z I I( nD xCD m

* ** -
zzzzooCJ <


'n in
r-iq -*

-i-i i-i

Cy C7 CY
*r- 'r- *r- F- lZ 0 0
O1C7CT a) w a) C ) ) 0
*,- *r- *r- *r- *r- *f *r- *,- *r- *C A- 0
CCCC S- S- S.- S-S- S S- 0)
o o o C -- .- a-
-R i > I >- > >> >I >

i- ii'i-V i- i- i- i-ii' ; I' 1-1 I-

-i Z1m mO II'JOr

m e r '3- -1 I-i

000000000 0

1z 1 C I I i I I I CD






a) 0
o U
c C

a) a
> 4-
M- a)

c s-

o a t
EE -

a) C

C 0) 0

S-- E 4-
a; a 4-
0 S- &- 0 e
' -Ca) a)

M41 a n3 C
i- 4.-

-T Va CO
0 4-4 4-
o '-

C L- S- &-
., 4 =- )

aj u 3
0> a)

c > I 4-

a) 0) 0) a) cx

S a* u a

o 1- s a) a)

a CC3 S.. L

CA r -o10 0

E 1o <. So

N), EE

L1 *r Iz x

S_ S S 4 1
0 0-0 S-

glycoenzyme. All the neutral carbohydrate was found to be released

from cellobiohydrolase C from Trichoderma reesei QM 9414, with

alkaline borohydride (reductive B-elimination), whereas the amino

sugars remained attached to the polypeptide. It was calculated that

the neutral carbohydrate was attached to serinyl and threoninyl

residues at an average of 16.7 sites per enzyme molecule. A range of

small oligosaccharides were released, with the mono- and

trisaccharides being the most abundant. Periodate oxidation and

G.C./M.S. of partially methylated residues determined that (i) the

oligosaccharides were unbranched, (ii) most of the linkages between

the neutral sugars in the chains are (1-6) and (iii) mannose was the

sugar linking the oligosaccharides to the hydroxy amino acids.

Glycosidase experiments suggested that a-mannose was present at the

non-reducing ends of some chains and that the disaccharides were


More recently, Fagerstam et al. [60] have sequenced most of CBH

I(D) and found that all of the 0-linked carbohydrate was attached

within a short region, 20 amino acids in length, located 32 residues

from the C-terminus of the polypeptide. Due to the presence of the

carbohydrate, this short region was never sequenced, but was found to

contain seven threonyl and three seryl residues. Shoemaker et al. [61]

have now sequenced the gene encoding CBH I obtained from

Trichoderma reesei strain L27. The nucleic acid sequence was shown

to contain two introns and, by comparison to the previously available

amino acid sequence, the region of carbohydrate attachment was shown

to contain eight threonyl and three seryl residues.

The B-glucosidases produced by Trichoderma have been found to

contain little or no carbohydrate. Berghem and Pettersson [62]

reported no carbohydrate attached to an extracellular B-glucosidase

from T. reesei using an orcinol-sulfuric acid method. Chirico [44]

found that a B-glucosidase from the same organism would not bind to

Concanavalin A-Sepharose, indicating the lack of available a-glucosyl

or a-mannosyl residues. However, four residues of neutral sugar and

two residues of N-acetylqlucosamine were estimated per enzyme

molecule based on phenol-sulfuric acid and amino sugar analysis,

respectively. The identity of the neutral sugar was not determined.

This was quite different from the enzyme isolated by Emert [63] from

T. viride, which was found to contain 10.3% carbohydrate. Gong et

al. [56] also determined that a cellobiase (.-glucosidase) activity

from another T. viride preparation did not bind to Concanavalin

A-Sepharose. Wood and McCrae [64] purified two B-glucosidases, BG1

and BG2, from Trichoderma koningii. Neither enzyme bound to Concan-

avalin A-Sepharose although BG2 was shown to contain 2% neutral carbo-

hydrate by a phenol-sulfuric acid determination. By the same method,

no carbohydrate was found associated with BG1.

Glycosylated cellulases are also produced by organisms other

than Trichoderma, although no detailed carbohydrate analyses have

been performed on them. Eriksson and Pettersson [65] have purified

five endoglucanases from the white-rot fungus, Sporotrichum pulveru-

lentum, of which four are glycosylated. The four glycoenzymes, T1,

T2b, T3a and T3b were found to contain 10.5%, 7.8%, 4.7% and
2.2% carbohydrate, respectively, as determined by gas-liquid chromato-

graphy as the alditol acetates. T1 was found to have 19 mannosyl and

2 glucosyl residues and T2b was found to have 5 mannosyl, 7

galactosyl, 1 glucosyl and 1 arabinosyl residue per enzyme molecule.

Kanda et al., using the phenol-sulfuric acid method, showed that both

an endocellulase [66] and an exocellulase [67] from the fungus

Irpex lacteus were glycosylated, containing 12.2% and 2.4% carbo-

hydrate, respectively. Beguin and Eisen [68] purified three endocellu-

lases, I, CA and CB, from Cellulomonas. While enzymes CA and CB were

found to stain with periodic acid-fuchsin and also bound to Concana-

valin A-Sepharose, typical of glycoproteins, enzyme I did neither and

was proposed to be unglycosylated. The amount and nature of the carbo-

hydrate were not determined. Tong et al. [69] determined that three

cellulases, I, II and III, and a 1-glucosidase from the thermophilic

fungus Thermoascus aurantiacus, were all glycosylated based on the

results of anthrone-sulfuric acid analysis. They were found to contain

33%, 5.5%, 2.6% and 1.8% carbohydrate, respectively. Ait et al. [70]

found that electrophoretic gels of a cellulase from Clostridium

thermocellum stained with periodic acid-fuchsin suggesting that it

also is glycosylated.

Wood and McCrae [71] separated four cellobiohydrolase forms, A,

B, C and D, from Fusarium solani, by isoelectric focussing. Phenol-

sulfuric acid analysis indicated that they contained 21%, 10%, 12% and

1% carbohydrate, respectively. These same workers later compared the

carbohydrate composition of this cellobiohydrolase activity with that

of cellobiohydrolases from Penicillium funiculosum and T.

koningii [72]. Of the carbohydrate attached to F. solani cello-

biohydrolase (all four species), 83% was mannose, 7% was xylose, 4%

was glucose and 1% was galactose. The carbohydrate component of the

cellobiohydrolase activity of T. koningii (two species) was found

to be composed of 73% mannose, 27% glucose and a trace amount of

xylose. The structure of the carbohydrate of any of these enzymes has

not been reported.

Whereas the B-glucosidases of Trichoderma are either not

qlycosylated or are so at a low level [44,56,62-64], enzymes from

other species are reported to be glycosylated to a greater extent.

Shewale and Sadana [73] purified four B-alucosidases from the fungus

Sclerotium rolfsii and suggested that they were glycoproteins due

to their affinity for Concanavalin A-Sepharose and to a positive

reaction to periodic acid-fuchsin staining. Rudick and Elbein [74]

determined that a B-glucosidase from Aspergillus fumiqatus

contained glucosamine and mannose based on the results of paper and

gas-liquid chromatography. Approximately 15 moles of mannose and 2

moles of glucosamine were found per mole protein. That either sugar

could only be released by strong alkaline treatment with 1N NaOH at

1000C for 6 hours suggested that the carbohydrate was associated

with the protein via a glucosaminyl-peptide linkage. Hirayama et al.

[75] purified a glycoprotein B-glucosidase from the phytopathogenic

fungus, Pyricularia oryzae. Although the enzyme contained 101

mannosyl and 13 glucosyl residues per protein molecule, as determined

by gas-liquid chromatography of the trifluoroacetyl derivatives, only

5% of the carbohydrate could be selectively removed using an


Structure:Function Relationships for Fungal Glycoenzymes

While little is known about the structure of the carbohydrate

attached to fungal cellulases, even less is known about its function.

Hayashida and Yoshioka [76] were able to partially remove the carbo-

hydrate attached to an exocellulase (Avicelase) and an endocellulase

(CMCase) from the thermophilic fungus Humicola insolens YH-8 by

either chemical or enzymatic treatment. Treatment of either with a

mixed-glycosidase preparation was shown to release about 65% of the

carbohydrate. The removal of the carbohydrate by this method did not

appear to affect the ability of either enzyme to hydrolyze crystalline

cellulose (Avicel), although the thermal and oH stability of both were

decreased. Treatment with periodate oxidation and Smith degradation

was shown to release about 90% of the carbohydrate from each enzyme.

After this chemical degradation, activity of each on Avicel was

decreased about 30% and pH and thermal stability of each was also

further reduced. There was no apparent effect of covalently bound

carbohydrate on substrate specificity since cellobiose was the

predominant product of enzymatic hydrolysis of cellulose by either the

native, 65%-carbohydrate depleted or 90%-carbohydrate depleted exo- or

endocellulase. However, in neither treatment was all the carbohydrate

removed and in the case of the chemical treatment, the effect on the

polypeptide was not determined. These results suggest that the

carbohydrate is important primarily for stability of the enzymes.

Extensive periodate oxidation of glucoamylase I from Asper-

gillus niger was shown to cause a marked loss in the stability of

the enzyme [77]. Glucoamylase I, normally heavily glycosylated at

about 45 sites on the polypeptide, was demonstrated to precipitate out

of solution when two-thirds of the carbohydrate was removed by this

method. Thus, for this enzyme, solubility and perhaps transfer to the

aqueous phase of the culture medium may be affected significantly by

the glycosylation.

To understand the importance of carbohydrate covalently-bound to

glycoproteins it is necessary to know the role of the protein to which

it is attached. Whereas the role of some glycoproteins is not known,

glycoenzymes offer a dynamic function for which one can study binding

and catalysis. Thus, definitive information can be achieved either by

preventing glycosylation or by removing the carbohydrate from the

post-translationally modified polypeptide. In spite of our knowledge

concerning the enzymic properties of fungal glycoenzymes, and

cellulases in particular, there is little known regarding the three-

dimensional structure of these enzymes. If the carbohydrate can be

removed from these glycoenzymes, the polyoeptide may be more amenable

to crystallization and X-ray diffraction studies.


Physico-chemical studies have revealed the molecular weight,

isoelectric point, amino acid composition and more recently, the amino

acid sequence of CBH I(D) and the genetic sequence encoding the

information for synthesis of CBH I(D) messenger RNA. Peptide isolation

has shown that the 0-glycosylated residues are confined to a short

sequence on the protein. With the availability of sophisticated

instrumentation such as high resolution nuclear magnetic resonance

(NMR) and gas chromatography/mass spectrometry (GC/MS), the stucture

of oligosaccharides attached to glycoproteins can be accurately

elucidated. The role the carbohydrate plays in cellulase function

cannot be fully understood unless the structure is known; this

information will enable further exploration of the molecular basis for

cellulase biosynthesis, secretion, stability, activity and microhetero-

geneity. Successful analyses, of course, depend on pure preparations

of proteins and oligosaccharides to ensure a definitive structural




Enzymes--Crude extracellular preparation from

Trichoderma reesei QM 9414, grown on Avicel (microcrystalline

cellulose), was a gift from Gulf Oil Chemicals Company, Merriam, KS.

B-Glucosidase from Trichoderma reesei QM 9414; prepared by

William Chirico.

a-Glucosidase maltasee) from Yeast (Lot #202900); Calbiochem,

Los Angeles, CA.

a-Glucosidase (Type I) from Yeast (Lot #398-5370), a-glucosidase

(Type VI) from Yeast (Lot #21F-8106), pullulanase from

Enterobacter aerogenes (Lot #99C-02731), amyloglucosidase from

Aspergillus niaer (Lot #72F-0560); Sigma Chemical Company, St.

Louis, MO.

a-Mannosidase from Jack Bean (Lot #7064101); Boehringer Mannheim

Corporation, New York, NY.

Chromatographic Supplies--DEAE-Sephadex A-50, SP-Sephadex

C-50; Pharmacia Fine Chemicals, Piscataway, NJ.

Dowex 50W (200-400 mesh; H+ form), Biogel P-2 (-400 mesh);

BioRad Laboratories, Richmond, CA.

Amberlite MB-3 (mixed bed resin, Amberlite IR-120 and Amberlite

IRA-410, fully regenerated); Mallinckrodt, Inc., Paris, KY.

OV-225 (cyanopropylmethyl phenyl-methyl silicone), Chromosorb

G-HP (80-100 mesh); Varian Associates, Sunnyvale, CA.

Neopentyl glycol succinate; Alltech Associates, Deerfield, IL.

Gas Chrom Q (100-120 mesh); Applied Science, State College, PA.

Whatman Partisil PXS 10/25 PAC column (bonded cyano-amino type,

polar phase); Whatman Chemical Separations Division, Clifton, NJ.

Substrates--CM-Cellulose 7HP; Hercules Powder Company,

Wilmington, DE.

Avicel PH 101 (microcrystalline cellulose); American Viscose

Division, FMC Corporation, Newark, DE.

Walseth cellulose (phosphoric acid-swollen cellulose); prepared

from Avicel PH 101 by the method of Wood [78].

D_-Nitrophenyl-a-D-mannopyranoside (Lot #701993); Calbiochem,

Los Angeles, CA.

p-Nitrophenyl-a-D-glucopyranoside (Lot #09053-3); Pierce

Chemical Company, Rockford, IL.

e-Nitrophenyl -B-D-mannopyranoside (Lot #62C-1270),

p-nitrophenyl-B-D-glucopyranoside (Lot #88C-5039); Sigma Chemical

Company, St. Louis, MO.

Carbohydrate Standards--Dextrose; National Bureau of

Standards, Washington, DC.

D-Xylose, D-xylitol, D-galactose, gentiobiose, Yeast mannan;

Sigma Chemical Company, St. Louis, MO.

D-Mannose; Calbiochem, Los Angeles, CA.

Cellobiose; Eastman Kodak Company, Rochester, NY.

Kojibiose, nigerose; gifts from Dr. Seiya Chiba, Department of

Agricultural Chemistry, Hokkaido University, Sapporo, Japan.

Isomaltose; Applied Science Laboratories, Inc., State College,


Dextran T-10; Pharmacia Fine Chemicals, Piscataway, NJ.

Chemicals--Acetic Anhydride; Mallinckrodt, Inc., Paris, KY.

Acetonitrile (HPLC grade), dimethyl sulfoxide (kept dry over 3A

molecular seives, methyl iodide, palladium (II) Chloride,

hydroxylamine hydrochloride; Fisher Scientific Company, Fair Lawn, NJ.

Acrylamide (>99.9%); BioRad Laboratories, Richmond, CA.

Ampholytes (Ampholine,pH 3.5-9.5); LKB Produkter AB, Bromma,


Basic Fuchsin (91% dye); Eastman Kodak Company, Rochester, NY.

Bromophenol Blue; Canalco, Rockville, MD.

Coomas s i e Bril l iant Blue R250, Column Coat,

N,N-methylene-bis-acrylamide, ammonium persulfate; Miles Laboratories,

Inc., Elkhart, IN.

Deuterium oxide (99.96 atom % D); Aldrich Chemical Company,

Inc., Milwaukee, WI.

Disodium succinate hexahydrate (A grade); Calbiochem, Los

Angeles, CA.

Periodic acid; G. Frederick Smith Chemical Company, Columbus,

Pyridine (silylation grade), N,N,N ,N -tetramethylethylene

diamine (TEMED); Pierce Chemical Company, Rockford, IL.

Riboflavin, L-a-amino-n-butyric acid, Freunds Complete Adjuvent;

Sigma Chemical Company, St. Louis, MO.

Sodium borohydride, sodium hydride (50% in oil dispersion); Alfa

Products, Danvers, MA.

All other chemicals were reagent grade.


Purification of Cellobiohydrolase I(D)--Cellobiohydrolase I(D)

(CBH I(D)) was purified from an extracellular culture filtrate of

Trichoderma by a modification of the method described by Gritzali

[49]. A crude precipitate produced by adding ammonium sulfate (65%

saturation) to a solution of the extracellular material from

Trichoderma reesei QM 9414 grown on Avicel was used as the source

of the enzyme. This precipitate was dissolved in distilled water,

dialyzed and lyophillized. The resulting pale yellow powder (5.0-7.5

g) was dissolved in 150 ml of 50 mM sodium succinate, 3 mM sodium

azide, pH 6.0. The mixture was then filtered slowly through glass

fibre paper (Whatman, grade 934 AH) and the filtrate applied to a

DEAE-Sephadex A-50 column (14 x 22 cm) equilibrated with the same

buffer. Cellobiohydrolase II, aryl-B-D-glucosidase and

endo-1,4-B-D-glucanase activity were eluted isocratically at a flow

rate of 400 ml/h (Fig. 1). CBH I(D) was obtained in high purity when

the pH of the elution buffer was lowered in a stepwise fashion to 3.6.

The previous method [49] involved elution of CBH I(D) in the presence



- ct
C 4- C

,-- n
L .0 -

*I- 0


C) m )

U C71-- a
.: ft0 -P
5- E t

U 0 CI-
a ) Wa)

0- S v


U C3 0-

oC C a-
ucn r:-





- (

. CU
O o
U, *4-



o r-
o o
0 0


0 5

(0-) Iuu08Z
; q

IV 30NV88OS88V

0 0 0 0 0
0 o 0 to
(0-C) [uw/u!W/dsv] AIlAllV 3SVNVOn1DOOaN3


of 0.5M NaCl, and while this was adequate for that cellulase

preparation, with this batch under the same conditions, several

impurities eluted with the enzyme. In the absence of the salt, CBH

I(D) eluted at pH 3.6 with approximately one bed volume of buffer and

was separated from minor contaminants.

Purification of Cellobiohydrolase II--Cellobiohydrolase II

(CBH II) was also purified by the method of Gritzali [49]. The

proteins eluting isocratically at pH 6.0 from the DEAE-Sephadex column

used to separate CBH I(D) (fractions 45-105) were pooled. This

material, containing CBH II, endoglucanase and B-glucosidase

activity, was dialyzed against 2mM sodium succinate containing 3mM

sodium azide at pH 4.5, concentrated by ultrafiltration to 30 ml and

applied to an SP-Sephadex C-50 column (4.4 x 55 cm). All endoglucanase

activity eluted isocratically (Fig. 2) and CBH II was obtained in high

purity when eluted in a batchwise manner with an elutant of 8mM sodium

succinate containing 3mM sodium azide at pH 4.5 (flow rate = 70 ml/h).

The previous method [49] eluted these enzymes in the same buffer at pH

5.0, but elution at the lower pH was conducted in this case to better

separate the endoglucanases for further purification. B-Glucosidase

activity was subsequently eluted from the column with 50mM sodium

succinate, 3mM sodium azide, pH 6.0.

Dialysis and Ultrafiltration--Protein concentration and

dialysis were performed on either an Amicon Model 2000 High

Performance Ultrafiltration Cell (volumes up to 2 1) or an Amicon

Model 202 Ultrafilration Cell (volumes up to 200 ml)(Amicon Scientific



o -
0 4-1
)( 0

') 5-


to (UC
C 4- .C

o .

C a)
S- a0 M

'u 0) rC C

. = 0 U C

C4 4-) = )
u o o Vo
S- E~( 0)1

U o~ r0 C
. -0 <
O = 0)

o -0 C0 L
U (L) CL






0 4

S- 4-u



or i-


.- S-

4- ) :



o o 0 0 0 0
ui C c 60



E C~



01 0


v 0 co CE co0
04 04 0- 0 00
c( 4 c'J 1N
(0-a) (jWIU!WrdS1jJ A.IAJOVFVNV~fl1)0GN3

Systems, Lexington, MA). Pressure was supplied with nitrogen gas at 40

psi and 30 psi, respectively. A Diaflo PM-10 Ultrafilter, with a

molecular exclusion limit of 10,000 daltons, was used with each cell.

Protein Determination--Protein concentrations were estimated

during purification by spectrophotometric determination at 280 nm,

using an approximate extinction coefficient (E1% at 280 nm) of 10.0.

For all quantitative analyses, protein concentration was determined

using experimentally derived extinction coefficients. Samples of

protein were dialyzed extensively against water, lyophillized and then

stored over phosphorous pentoxide for at least 10 days. Protein was

weighed and then made up into solutions of known concentrations of

between 0.3-0.8 mg/ml. The absorbance of each solution was determined

both at 260 nm and 280 nm on a Beckman DU-8 UV-Visible

Spectrophotometer (Beckman Instruments, Inc., Palo Alto, CA).

Enzyme Assays--Endo-1 ,4-B-D-glucanase activity was measured

by following the decrease in viscosity of a carboxymethylcellulose

solution as described by Shoemaker and Brown [79]. Specific activity

was expressed as the change in specific fluidity/min/mg protein.

Aryl-glycosidase activity was followed by monitoring the release of

i-nitrophenol from e-nitrophenyl -glycosides. A solution of 10 mM
substrate (0.5 ml) and an appropriate buffer (2.0 ml) were

preincubated in small test tubes for 10 min at 400C. For

a-mannosidase, digestions were performed with 50mM sodium citrate

containing 3mM sodium azide at pH 4.5 and for a-glucosidase reactions

were carried out in solutions of 65mM potassium dihydrogen phosphate

with 3mM sodium azide at pH 6.0. Aliquots of enzyme solution (2-25

pl) were then added, mixed and incubated at 400C for 20 min. The

tubes were placed in a boiling water bath for 10 min, and when cool, 1

ml sodium borate (0.5 M, pH 9.8) was added. The absorbance at 400 nm

was measured, compared to a p_-nitrophenol standard curve and

specific activity expressed as moles p-nitrophenol released/min/mg

protein. The solution was adjusted with the sodium borate buffer to pH

9.1, a value greater than one pH unit above the pKa of

p-nitrophenol, assuring complete dissociation.

Polyacrylamide Disc Gel Electrophoresis--Disc gel

electrophoresis was performed as a routine analysis of column

fractions during purification and was used as a criterion of

homogeneity. Electrophoresis was conducted using the discontinuous

buffer system No. 1 described by Maurer [80]. The pH of the stacking

gel and separating gels were 8.3 and 9.5, respectively, and the

polyacrylamide concentration was 7.5 percent (W/V). Power was supplied

at 2 mA/tube until the bromophenol blue tracking dye had entered the

separating gel, at which point it was increased to 3 mA/tube until the

dye band was within 0.5 cm of the bottom of the tube. Separations were

performed in a Canalco electrophoresis chamber (Miles Laboratories,

Inc., Elkhart, IN) with a Hoefer PS 1200 DC power supply (Hoefer

Scientific Instruments, San Fransisco, CA).

After electrophoresis, protein bands were fixed by immersion in

12 percent (W/V) trichloroacetic acid for 30 min after which the gels

were washed with water (3 x 5 min). Protein was stained with 0.1%

Coomassie Brilliant Blue R250 in methanol-water-acetic acid (45:45:10)

for 1 h and subsequently destined with 7% acetic acid at 400C with

frequent changes of destaining solution. Carbohydrates in the gels

were stained by the periodic acid-Schiff (PAS) method described by

Lang [81]. After fixing and washing, the gels were immersed in fresh

periodic acid (0.5% w/v) for 1 h in the dark. The acid was removed by

successive washes with 7% acetic acid (3 x 10 min). Gels were then

stained for 1 h with 1% Basic Fuchsin in 0.15N HC1 containing 1.9%

sodium metabisulfite. Excess dye was removed by destaining with 0.1%

sodium metabisulfite at 400C, with repeated changes of destaining


Isoelectric Focusing--Isoelectric points for homogeneous

proteins were determined using a BioRad Model 1415 Horizontal

Electrophoresis Cell (BioRad Laboratories, Richmond, CA).

Determinations were performed with cold water (40C) passing through

the gel bed using a Haake Model FE Constant Temperature Circulator

(Haake Instruments, Inc., Saddle Brook, N.J.). Polyacrylamide gel

slabs (100 x 125 x 0.8 mm) were used as described by BioRad [82],

without initial prefocusing and with the modification that the

ampholytes were replaced with 0.5 ml Ampholine, pH 3.5-9.5 (LKB

Produkter AB, Bromma, Sweden), as these produced more linear pH

gradients. Electrofocusing was performed for 2.5 h at 4.5 amps

constant current. Best results were obtained using lOg samples of

protein applied to small strips of glass fiber paper (0.3 x 0.5 cm)

placed at the center of each lane on the gel slab. The strips were

removed 30 min after the current was turned on. After electrofocusing,

sections were cut (0.5 x 0.5 cm) down each side of the gel and placed

in tubes containing 1 ml 0.1M KC1 (made with degassed, deionized

water). The pH of each section was determined after 1 hour at room

temperature and the gradient calculated. Fixing, staining and

destaining were performed as described by Winter et al. [83]. Proteins

were fixed (17.3 g sulfosalicylic acid and 57.5 g trichloroacetic acid

in 500 ml water) for 1 h at room temperature, stained with Coomassie

Brilliant Blue R250 (0.46 g in 400 ml destaining solution) for 1 h and

then destined (water-ethanol-acetic acid, 67:25:8) until the

background was clear.

Carbohydrate Composition--Total neutral carbohydrate was

determined by the phenol-sulfuric acid method of Dubois et al. [84].

Carbohydrate components of the enzymes were determined after either

hydrolysis or B-elimination and individual monosaccharides were

identified after separation by gas-liquid chromatography. Aldoses were

determined as the peracetylated aldononitriles (PAANs) and alditols

were determined as the peracetylated alditols (PAAs).

For phenol-sulfuric acid analysis, 0.1-1.0 mg of glycoprotein

(containing 2-15 pg neutral carbohydrate) was dissolved in 0.2 ml

water in a small test tube. Phenol (0.2 ml,3%) was added followed by

1.0 ml concentrated H2SO4 with vigorous mixing using a vortex

mixer. The samples were cooled to 400C and the absorbance determined

at 400 nm, using mannose as the standard, as this is the principal

carbohydrate constituent of these glycoproteins.

Lyophillized glycoproteins (1-2 mg) were hydrolyzed in 1 N HC1

(2 ml) for 2 h at 1000C in sealed 5 ml Reacti-Vials (Pierce Chemical

Company, Rockford, IL). When cool, xylose and/or xylitol were added

as internal standards and the mixtures immediately deionized usino

small columns (1.4 x 6 cm) of Amberlite MB-3 mixed bed resin. The

eluants were taken to dryness in a rotary evaporator, transferred to 5

ml Reacti-Vials and lyophillized. Monosaccharides were then either

converted to the peracetylated aldononitriles (PAAN) by a modification

of the method of Varma et al [85] or subsequently reduced and

converted to the peracetylated alditol by the method of Sawardeker et

al. [86]. It was essential that these derivatives be separable and a

profile of that separation by gas-liquid chromatography is shown (see

Fig. 3).

The PAAN preparative procedure involved adding 3-5 mg

hydroxylamine hydrochloride and 0.25 ml pyridine to the lyophillized

monosaccharide mixture (0.01-10mg), sealing the Reacti-Vials with

screw top teflon caps and heating at 900C for 45 min. The resultant

oximes were then peracetylated by addition of 0.25 ml acetic anhydride

and heating at 900C for 45 min. The products were taken to dryness

under nitrogen at 600C and then evaporated with toluene (3 x 1 ml) to

remove residual acetic anhydride using an N-Evap Model 106 Analytical

Evaporator (Organomation Associates, Inc., Shrewsbury, MA). The

products were subsequently partitioned between 3M HC1 (1 ml) and

CHCl3(1 ml), the aqueous layer removed and the organic layer washed

with water (2 x 1 ml). Washing involved vigorously mixing the two

phases in the Reacti-Vial and removing the aqueous layer with a

pipette after partitioning was complete. The CHC13 layers were then

washed with 0.5M sodium bicarbonate (1 x 1 ml) and then again with

water (1 x 1 ml). A spatula tip of MgSO4 powder was then added to

each to remove any excess water. The organic layers were then filtered


Gas chromatographic separation of peracetylated
alditol and aldononitrile derivatives of neutral
monosaccharides released from cellobiohydrolase I(0)
after reductive B-elimination and subsequent acid

Peaks identified are peracetylated 1. xylononitrile; 2.
xylitol; 3. mannononitrile; 4. glucononitrile; 5. mannitol.
Conditions of separation are described in "Experimental
Procedures". The peaks eluting before the xylononitrile
acetate peak were always present in PAAN preparations but
not in PAA preparations and are thought to be byproducts of
the reaction [88].



5 10 15 20 25


through glass wool, concentrated to about 20-50 pl and 1 p1 aliquots

were analyzed by gas-liquid chromatography.

In the procedure for peracetylated alditols, monosaccharides

were reduced overnight with aqueous sodium borohydride (10 mg in 1 ml

water) at 40C. The reactions were stopped by dropwise addition of 2N

acetic acid until the effervescence stopped, using 3 drops of

n-octanol to suppress foaming. The mixtures were taken to dryness

under a stream of dry nitrogen at 600C and the residues evaporated

with methanol (3 x 1 ml) to remove excess borate as the volatile

methyl ester. The mixtures were then deionized by passage through

small Amberlite MB-3 columns and the eluants were taken to dryness on

a rotary evaporator. The samples were transferred to 5 ml

Reacti-Vials and lyophillized. Pyridine (0.25 ml) and acetic anhydride

(0.25 ml) were added and the vials sealed and heated at 90C for 45

min. The mixtures were taken to dryness under a stream of dry nitrogen

at 60C and then evaporated with toluene (3 x 1 ml). The residues

were dissolved in 20-50 pl CHC13 and 1 pl aliquots analyzed by

gas-liquid chromatography.

0-linked oligosaccharides were released by reductive

B-elimination of the glycooroteins, by a modification of the method

of Nakajima and Ballou [87]. To 1 or 2 milligrams of the glycoproteins

in a 5 ml Reacti-Vial were added 2 ml of 0.1N NaOH and 0.3M NaBH4.

The solution was incubated at 400C for 48 hours. During a 72 h

control experiment, no further sugars were released from the protein

after 3 hours of reductive B-elimination (Fig. 4). After

neutralization by dropwise addition of 2 N acetic acid, the samples

were evaporated with methanol to remove borate esters and then


Efficacy of sugar release from cellobiohydrolase I(D)
by reductive B-elimination.

Oligosaccharides released were subsequently hydrolyzed to
constituent sugars. The resultant monosaccharides were then
converted to the peracetylated alditols and aldononitriles
and analyzed by gas-liquid chromatography on 1% OV-225 as
described under "Experimental Procedures". A, Mannose;O,
Mannitol;O, Glucose.




2 4.0


(3 2.0



20 40 60

deionized by passage over Amberlite MB-3. The eluate was dried under

reduced pressure and hydrolyzed as previously described. The sugars

released were converted to either a mixture of PAANs and PAAs by the

PAAN method described previously or were further reduced with NaBH4

and all converted to the PAAs. Formation of mixed derivatives is

feasible as alditols, unlike aldoses, do not form the corresponding

nitrile in the presence of pyridine and hydroxylamine hydrochloride;

but do undergo peracetylation upon addition of acetic anhydride. This

derivatization method following B-elimination provided information

regarding, not only the total neutral carbohydrate composition of the

0-linked oligosaccharides, but also identification of the type of

sugar attached to the protein (leading to the formation of the

corresponding PAA) and an estimation of the number of sites at which

carbohydrate had been attached to the polypeptide.

The area of peaks resulting from gas-liquid chromatography (see

below) of the sugar derivatives were subjected to integration and were

compared to internal standards. Positive identification of derivatives

was achieved by gas chromatography/mass spectrometry (see below).

Quantitative determination of the peracetylated aldononitriles of

mannose and glucose were derived from the linear relationship of their

GLC peak areas with respect to that of the xylose internal standard as

were those of the peracetylated alditols of mannitol and glucitol with

respect to the peracetylated xylitol internal standard (Figs. 5 and


While alditol acetate formation proceeds essentially to

completion, Furneaux [88] has shown that formation of some

peracetylated aldononitriles occurs with the formation of several


Quantitation of aldoses as the peracetylated

Aldoses (mannose (0), glucose (0) and xylose (internal
standard)) were converted to the PAANs and the derivatives
were separated by gas-liquid chromatography on 1% OV-225 at
1900C as described under "Experimental Procedures". The
ratio of peak areas of sample/standard was plotted against
the molar ratio of the sample/standard. Correction factors
necessary for calculation of unknowns were determined to be:
Man/Xyl, 0.85; Glc/Xyl, 0.94.




z 1.0

0.5 1.5 2.5




Quantitation of alditols as the peracetylated

Alditols (mannitol(D), glucitol (U) and xylitol (internal
standard)) were converted to the PAAs and the derivatives
were separated by gas-liquid chromatography on 1% OV-225 at
190C as described under "Experimental Procedures". The
ratio of peak areas of sample/standard was plotted against
the molar ratio of the sample/standard. Correction factors
necessary for the calculation of unknowns were determined to
be: Mannitol/Xylitol, 0.87; Glucitol/Xylitol, 0.94.



1 .0

0.5 1.5 2.5



furanose and pyranose byproducts. The PAANs are more labile than the

PAAs; the former partially decompose in the injector of the gas-liquid

chromatograph Observations in this laboratory support these

conclusions as we have found that the yield of PAANs is lower than

that of the corresponding peracetylated alditols from derivatization

of equal amounts of the respective aldoses and alditols. Several small

unidentified peaks also eluted after these derivatives and by elution

at higher temperatures on 1% OV-225 as was seen by Furneaux. This

discrepancy was accounted for, however, by the use of internal

standards in each sample, since it was found that the aldoses behaved

reproducibly as a group, as do the alditols (Fig. 5 and 6).

Preparation of reduced oligosaccharides--01igosaccharides

0-linked to the protein were released by treating 50-200 ma of each

glycoorotein with 0.1N NaOH and 0.3M NaBH4 at 40C for 48 hours

[87]. The reactions were performed in sealed culture tubes with teflon

screw caps in a volume of 15 ml. Reaction products were neutralized

with 2 N acetic acid using 5 drops of n-octanol to control foaming and

taken to dryness on a rotary evaporator at 50C. The residues were

evaporated with methanol several times to remove excess borate and

then deionized through a small column (50 ml) of Amberlite MB-3. The

eluant was taken to dryness, dissolved in 400 pl of water and

injected onto a Biogel P-2 (-400 mesh) column (0.25 in x 8 ft)

attached to a Waters high performance liquid chromatograph (HPLC;see

below). Oligosaccharides were eluted with water and the eluant stream

monitored with a differential refractometer. Oligosaccharides were

Bradbury, A.G.W. (1984) Personal Communication.

purified either by rechromatography of pooled fractions under the same

conditions, or by passage through a Partisil PXS 10/25 PAC polar

reverse-phase column, also by HPLC. Composition of each

oligosaccharide and identification of the sugar attaching the

oligosaccharide to the protein was determined following mild acid

hydrolysis of each and conversion of the product to a mixture of the

alditol and aldononitrile acetates by the PAAN method previously

described. Derivatives were then analyzed by gas-liquid

chromatography and gas chromatography/mass spectrometry.

Preparation of Methylsulfinyl Carbanion--The methylsulfinyl

carbanion suspension for methylation analysis was prepared by an

adaptation of the methods of Lindberg [89] and Spiro [90]. Sodium

hydride was obtained in a 50% mineral oil dispersion as it reacts

explosively with water; all glassware was acid cleaned, baked and then

kept over phosphorous pentoxide until use. Under a dry nitrogen

stream, 4.75 g of sodium hydride was weighed into a 250 ml erlenmeyer

flask, equipped with a two-hole rubber stopper (with a CaC12 trap

and a tube for nitrogen) and a small magnetic stirrer bar. To this was

added 100 ml DMSO (dried over 3A molecular seive) to achieve a final

concentration of 2M sodium hydride, and the mixture stirred under a

nitrogen stream for one hour. Aliquots (25 ml) of the suspension were

poured immediately into 30 ml serum bottles, sealed with rubber caps

and placed on ice. The suspensions were still generating hydrogen gas;

so the caps were punctured with two syringe needles and the bottles

flushed with nitrogen until the mixture froze solid. The

methylsulfinyl carbanion suspension was stored under nitrogen (with

Drierite) at -200C. Under these conditions the preparation was good

for at least four years. When ready for use, the caps were punctured

again with two syringe needles and flushed with nitrogen in a

sonicator bath for one hour. Aliquots of the suspension were removed

with an 18 gauge needle (smaller needles clogged up with the thick

slurry) and the mixture frozen on ice again under a nitrogen stream.

Methylation Analysis--Glycoproteins and oligosaccharides

released by reductive B-elimination were methylated by the method of

Hakomori [91] and samples were subsequently hydrolyzed, reduced and

acetylated by a method adapted from those of Bjbrndal et al. [92] and

Gum [93]. Samples oligosaccharidess, 0.5-1.5 mg; glycoproteins, 10

mg) were first dried over phosphorous pentoxide overnight in 5 ml

Reacti-Vials, dissolved in 1 ml dry DMSO and closed with

teflon/silicone seals. The seals were punctured with two syringe

needles, 1.0 ml of 2M methylsulfonylmethylsodium was injected and the

needles removed after flushing the vials with nitrogen. The vials were

agitated in an sonicator bath for one hour. After another 2 hours, 1

ml methyl iodide was injected dropwise and the vials sonicated for 30

min. The products were partitioned into CHC13 (5 ml) and water (7.5

ml) and the water layers washed again with CHC13 (2 x 3 ml). The

organic layers were pooled and repeatedly washed with water (3 x 5

ml). Excess water was removed from the organic layers by addition of

a spatula tip of MgSO4 and the supernatant filtered through glass

wool into 5 ml Reacti-Vials and evaporated to dryness under nitrogen.

Methylated oligosaccharides were then hydrolyzed with 88% formic acid

(2 ml) for 2 hours at 1000C. The products were taken to dryness

under nitrogen and the hydrolysis completed with 0.25 M H2SO4 (1

ml) at 1000C for 12 hours. Samples were transferred to small test

tubes, 0.54 g BaCO3 (10% excess) was added to neutralize the

hydrolysates and the supernatants removed after centrifugation (5000

rpm, 10 min). The pellets were extracted washed with 1 ml water,

centrifuged again and the supernates pooled. The partially methylated

monosaccharides, then in 3 ml of water, were then reduced by addition

of 20 mg NaBH4 at room temperature for 2 hours. The supernatants

were then taken to dryness in 5 ml Reacti-Vials and evaporated with

methanol (3 x 2 ml). After drying overnight with P205, the

partially methylated alditols residues were acetylated with pyridine

(0.25 ml) and acetic anhydride (0.25 ml) at 900C for 1 hour. The

reaction mixtures were taken to dryness under nitrogen at 600C,

evaporated with toluene (3 x 1 ml) and dissolved in 20-50 pl CHC13.

Products were analyzed by gas-liquid chromatography on 3% neopentyl

glycol succinate, and positively identified by gas chromatography/mass

spectrometry (see below).

Tetramethyl and trimethyl alditol acetate standards for GLC and

GC/MS comparison were prepared by methylation of either (i) 5 mg of

yeast mannan for the mannose series or (ii) 5 mg each of the glucose

disaccharides, kojibiose (al-2), nigerose (al-3), cellobiose (81-4)

and gentiobiose (B1-6). Pentamethyl alditol acetate standards were

prepared after reduction of 5 mg of each of the glucose disaccharides

described with 20 ma sodium borohydride in 2 ml of water at room

temperature overnight. The reactions were stopped with 2 N acetic

acid, in the presence of n-octanol, and each sample taken to dryness

under reduced pressure. Excess borate was removed by evaporation as

the volatile methyl ester by addition of aliquots of methanol (3 x 10

ml) and the samples subsequently transferred to 5 ml Reacti-Vials and

lyophillized in preparation for methylation.

Acetolysis--Reduced tri- and tetrasaccharides were subjected

to acetolysis by the method of Tai et al. [94] as modified by Li et

al. [95]. Oligosaccharides (100 pg) were dissolved in 0.5 ml pyridine

and 0.5 ml acetic anhydride in 5 ml Reacti-Vials and left at room

temperature for 70-74 h. The vials were then heated at 800C for 4 h.

Samples were taken to dryness under nitrogen and evaporated with

toluene (3 x 1 ml). Acetylated oligosaccharides were then acetolyzed

with 1.0 ml acetic anhydride-acetic acid-H2SO4 (10:10:1) at 40C

for 16 h. The products were then partitioned into CHC13 (1 ml) and

water (1 ml). The aqueous layer was removed and the organic layer

washed repeatedly with water to remove the color (3 x 3 ml). The

aqueous washes were then extracted with CHC13 (3 x 1 ml) and all the

CHC13 fractions pooled and dried. The residue was then deacetylated

with 1.0 ml 0.1% sodium methanolate for 30 min at room temperature.

The reaction was neutralized with 2.0 ml ethyl acetate and the

products evaporated to dryness under nitrogen. The residue was then

dissolved in 10-30 pl of water and the products separated by polar

reverse-phase HPLC on a Whatman Partisil PXS 10/25 PAC column. The

compositions of purified oligosaccharides were then determined by

gas-liquid chromatography as the peracetylated alditol and/or

aldononitriles (see below).

Glycosidase digestion--Oligosaccharides released from the

glycoproteins by reductive B-elimination were subjected to sequential

glycosidase digestion which provided information as to the anomeric

nature of each sugar residue and also the sequence of sugars in each

oligosaccharide. Samples of each oligosaccharide (100 pg) were placed

in 250 pl Microfuge tubes (Beckman Instruments, Inc., Palo Alto, CA)

and lyophillized. Incubation with yeast a-glucosidase maltasee) was

performed with one unit of enzyme in 0.05 ml of 0.05 M potassium

phosphate, pH 6.8, for 12 hours at 400C with 10 pl of toluene added

to ensure sterility. One unit of enzyme was then added every 12 hours

for the next 36 hours until four units of enzyme had been added.

Incubation with jack bean a-mannosidase was performed with one unit of

enzyme in 0.05 ml of 0.05 M sodium citrate, pH 4.5 for 48 hours at

400C with 10 pl toluene added. Products were separated directly by

HPLC on a Partsil PXS 10/25 PAC polar reverse-phase column, equipped

with an Bondapak AX/Corasil anion exchange precolumn. Oligosaccharides

were subsequently incubated with a different glycosidase enzyme and

the resulting monosaccharides were analyzed as the alditol and/or

aldononitrile acetates by gas-liquid chromatography.

Gas-Liquid Chromatography--Monosaccharide components of

oligosaccharides and glycoproteins were identified and quantified

after separation of acetylated derivatives by gas-liquid

chromatography and as such it was necessary to achieve a separation of

alditol and aldononitrile acetates of component sugars of the

glycoproteins (see Fig. 3). Although separation of PAANs [85] and PAAs

[86] have been reported, these methods were extensively modified to

permit simultaneous separation of the two classes of sugar

derivatives. All gas-liquid chromatography was performed on a Varian

Model 2700 gas chromatograph (Varian Associates, Sunnyvale, CA)

equipped with a flame ionization detector and attached to an Autolab

System 1 computing integrator (Spectra-Physics, Santa Clara, CA).

Alditol and aldononitrile acetates were separated on a 1% OV-225 on

Chromosorb G H/P (80-100 mesh) glass column (2 mm X 8 ft.),

isothermally at 1800C. Partially methylated alditol acetates were

separated on a 3% neopentyl glycol succinate on Gas Chrom Q (100-120

mesh) glass column (2 mm x 20 ft.), with isothermal elution at 1900C.

The injector and detector temperatures were 2300C and 2600C,

respectively, for both columns. Trimethyl alditol acetates were found

to be less well resolved on packed columns of 1% OV-225, whether

separated isothermally or under a variety of temperature programmed


Gas Chromatography/Mass Spectrometry--While separations of

the derivatized sugars can be performed by gas-liquid chromatography,

definitive identification was achieved by this technique in

conjunction with mass spectrometry. Two types of fragmentation were

used in this work, electron impact (EI) and chemical ionization (CI),

and each yields different information regarding the molecules being

identified. Mass spectrometry is not normally used for quantitation of

samples as a relatively low fraction of the sample is ionized (about

1% for El) and so all quantitation of derivatives was performed on the

gas chromatograph. El fragmentation is severe as the sample is exposed

directly to an electron beam in the ionizing chamber; hence El spectra

represent an array of structural fragments (fingerprints) of the

molecules. It should be mentioned that peracetylated alditol (Appendix

A) and peracetylated aldononitrile (Appendix B) derivatives of mannose

and glucose fragment identically by EI; hence the need for prior

separation by gas-liquid chromatography. This is more important when

comparing the partially methylated alditol acetates, where

coordinately methylated mannose and glucose derivatives fragment

identically. Differentially methylated residues, while having many

similar fragments, give rise to unique spectra (see Appendix C).

CI fragmentation is less severe as the ionizing chamber is

filled with an excess of reagent gas, i.e. isobutane, and thus it is

the gas molecules that are ionized directly. These ionized gas

molecules then collide with sample molecules causing a milder

fragmentation than that which occurs in El, as less energy is

transferred. CI spectra thus provide very accurate molecular weight

information by a monitoring of the molecular ion (M ), generated by

loss of an electron from the molecule during collision. As

galactononitrile acetate and mannitol acetate were found to coelute on

a packed 1% OV-225 column, mass fragment scanning was useful in ruling

out the presence of galactose in glycoprotein samples following

reductive B-elimination and hydrolysis. This is shown in Fig. 7,

where the mass ion for galactononitrile acetate is 356 and none was

shown to be present under the mannitol acetate peak (M+=375). CI

spectra for peracetylated alditol (Appendix D) and peracetylated

aldononitrile (Appendix E) standards are shown. By providing primarily

molecular weight information, CI spectra cannot distinguish between

the various trimethyl alditol acetates prepared as standards but of


Molecular ion scanning of peracetylated alditol and
aldononitrile derivatives of monosaccharides released
from cellobiohydrolase I(D) following reductive
B-elimination and subsequent acid hydrolysis.

Molecular ions were selected out of the profile following
chemical ionization gas chromatography/mass spectrometry.
Scanned were the M ions for: pentitol acetates (303),
hexitol acetates (375), pentononitrile acetates (256) and
hexononitrile acetates (328). The bottom panel represents
the reconstituted ion current (RIC), or the total material
in the sample. Elution times are shown for peracetylated 1.
xylononitrile; 2. xylitol; 3. mannononitrile; 4.
glucononitrile; 5. mannitol.




Z 100.0


- 20.5


6:40 13:20 20:00 26:40




256 0

328 0

course can distinguish between the tri- and tetramethyl alditol

acetates (see Appendix F).

Gas chromatography/mass spectrometry was performed either on a

Finnegan 4021 GC/MS/DS or on a Finneqan TSQ GC/MS/MS/DS instrument.

All chromatographic conditions were the same as those for gas-liquid

chromatography except the 3% neopentyl glycol succinate column was

only 8 ft long. Electron impact spectra were recorded at an ionizing

potential of 70 eV with an ion source temperature of 3000C. Chemical

ionization spectra were recorded at an ionizing potential of 55eV and

a filament current of 0.3A, using isobutane as the reagent gas at 5 x

10-5 torr.

High Performance Liquid Chromatography--HPLC separations were

performed with a Waters Model 6000 pump, Model U6K injector and Model

R401 differential refractometer (Waters Associates, Inc., Milford,

MA). Analytical separations were performed by an adaptation of the

method of Gum and Brown [96], on a Partisil PXS 10/25 PAC polar

reverse-phase column (Whatman Inc., Clifton, NJ) equipped with a

Bondapak AX/Corasil anion exchange precolumn (Waters Associates).

Elution was performed at a flow rate of 1.5 ml/min at room

temperature, with either 75% or 77% acetonitrile in water and at a

pressure of 1400 psi. Preparative separations were performed on a

water-jacketed Biogel P-2 (-400 mesh) gel filtration column (0.25 in x

8 ft), with no precolumn, using a Haake Constant Temperature

Circulator. Oligosaccharides were eluted with water at 600C at a flow

rate of 0.7 ml/min and at a pressure of 700 psi.

NMR Analyses--NMR analyses were performed on both

glycoproteins and on oligosaccharides isolated from them after

preparative reductive B-elimination and the spectra generated

provided information about the entire molecule being analyzed. For

H-NMR, analysis of the glycoprotein is not practical due to the

excessive number of signals generated. For the oligosaccharides, the

H-NMR provided information as to the number of anomeric residues

present and also, in the case of glucose, the nature of the anomeric

linkage. For mannose, the splitting of the anomeric proton for both a-

and B-anomers is less than 2 Hz, which is below the resolution of the

instrument at 300 MHz.

Characterization of oligosaccharides attached to the

glycoproteins was possible by proton coupled and proton decoupled
C-NMR as these signals (between 60-110 ppm) were far removed from

those of amino acid residues. The decoupled spectra generated signals

for each carbon nucleus in the sample, although due to the closeness

of many of the signals, it was not possible to assign all of them. The

signals generated by the anomeric carbons, located between 98-106 ppm,

were distinct and were far downfield from the other carbohydrate

signals. These can be assigned, on the basis of literature values and

by analysis of the compounds by chemical methods. The number of

hydroxymethyl carbons can also be determined as they generate signals

at the upfield end of the carbohydrate region between 62 and 66 ppm.

But like the 1H-NMR, decoupled 13C-NMR spectra are a

characteristic fingerprint of the molecule, even though it is not

possible to assign all the signals. Coupled 1C-NMR, like H-NMR,

provided information as to the anomeric nature of the sugar residues

in the oligosaccharides. The coupling constants generated from the

anomeric carbons were characteristic of either a-linkages (170 Hz) or

B-linkages (160 Hz), although the resolution of signals in coupled

spectra is much poorer than those acquired in the decoupled mode.

CBH I(D) was prepared by dialyzing against water, lyophillizing

and dissolving the protein into 0D0 (98.5%). CBH II was less soluble

and counterions were needed, so the protein was dialyzed against 50 mM

NaCl and made to at least 60% D20. The endoglucanases were prepared

in the same way, except that 20 mM NaCl was used. The reduced

oligosaccharides were lyophillized and picked up in 100% D20 several

times for all the NMR analyses.

1C-NMR spectra were obtained on a Nicolet NMR Spectrometer

with a 70.5 kG field and a broadband tunable probe operating in the

Fourier Transform mode at 75.46 MHz. D20 was used as solvent in all

samples and the field was locked on the deuterium signal. Spectra were

collected with broad band proton decoupling either on or off, as

indicated in the particular experiment. Pulse widths and individual

recycle times used were 29.00 psec and 3.00 msec, respectively, and

in all experiments, the post acquisition delay time was greater than

5T1. The spectra were filtered to improve the signal to noise ratio

and this treatment broadened the lines by 1 Hz. Chemical shifts are

reported in parts per million (ppm) from internal

(trimethylsilane)-l-propane sulfonate (TSP). Under these conditions,

the anomeric carbon of B-D-glucopyranose resonates at 98.76 ppm.

H-NMR spectra were obtained on the same instrument with a

fixed frequency (300 MHz) probe and spectra were acquired at both

250C and 700C to eliminate the possibility of signals "hidden" under

the large HOD signal (4.77 ppm at 250C).

Amino Acid Analysis--One mi lligram samples of the

glycoproteins were B-eliminated with 0.1M NaOH and 0.3M NaBH4 for

48 hours at 400C. The resultant unsaturated amino acids were

subsequently reduced with palladium chloride and polypeptides

hydrolyzed to the free amino acids as described by Downs et al. [97].

Reactions were performed in sealed screw-cap culture tubes in a volume

of 2.0 ml. After alkaline treatment, 5 drops of n-octanol and a small

stirrer bar were added. One milliliter 0.8M HC1 was added followed by

0.1 ml 0.08M PdCl2 and the solutions mixed. Two milliliters NaBH4

(0.66M in 0.1M NaOH) and 2.0 ml PdC12 (0.016M in O.8M HC1) were then

added simultaneously in a dropwise fashion, while mixing rapidly. The

total volume was now 7.1 ml and to this was added an equal volume of

concentrated HC1. The tubes were resealed with teflon liners and

heated at 1100C for 24 h. Each sample was cooled, transferred

quantitatively to a 100 ml round bottom flask and taken to dryness on

a rotary evaporator at 50C. Residual HC1 was removed by repeated

evaporation with water (3 x 10 ml). The final volume was carefully

dissolved in 1 ml lithium citrate buffer (0.2M, pH 2.2) and

transferred to a 5 ml volumetric flask. The round bottom flask was

then rinsed repeatedly (4 x 1 ml) and the washings added to the 5 ml

volumetric flask and mixed. Samples containing PdCl2 contained a

small amount of black particulate precipitate, which was allowed to

settle before analysis. Aliquots (0.1 ml) were injected onto a Beckman

Model 119CL amino acid analyzer (Beckman Instruments, Inc., Palo Alto,

CA) in the physiological column mode. a-Aminobutyric acid (product of

threonine B-elimination and reduction) was detected with ninhydrin

[98] and eluted between 76-77 minutes. The peak was integrated and

compared to an internal standard, a-aminoguanidinopropionic acid.

Controls were performed on (i) samples which had undergone PdC12

treatment without previous B-elimination, to determine if any

background a-aminobutyric acid was produced and (i) on samples which

had undergone neither PdCl2 treatment nor B-elimination to

determine any possible effects of these treatments on any amino acids

released by hydrolysis.

Antibody Preparation and Immunochemical Analysis--Antisera to

CBH I(0), CBH II and the endoglucanases were prepared in New Zealand

White rabbits by a method adapted from that of Hurn and Chantler [99].

Samples of protein (2.5-3.0 mg) which had been dialyzed to water and

lyophillized were dissolved in 1 ml water and emulsified with 3.5 ml

Freunds Complete Adjuvent. Each rabbit, 1-2 Kg in size, was injected

with a total of 2 ml of one of the proteins, 1 ml into each thigh

muscle. Booster injections were aiven three weeks later under the same

conditions. Animals were test bled periodically and each antisera

titer determined by immunodiffusion on agar against all the

cellobiohydrolases and endoglucanases to check for cross-reactivity.

Rabbits were then sacrificed after seven weeks, 70-100 ml of whole

blood collected from each and serum prepared after removal of whole

cells by centrifugation.


Cellobiohydrolase I(D)

Molecular Properties of CBH I(D)--CBH I(D) has been shown to

constitute almost 60% of the extracellular protein secreted by

Trichoderma reesei QM 9414 grown on cellulose. This protein has also

been found to have a molecular weight of 54,000 daltons, as determined

by sedimentation equilibrium and amino acid analysis [49]. CBH I(D)

was obtained in highly purified form here after elution from

DEAE-Sephadex at pH 3.6, as determined by native polyacrylamide qel

electrophoresis (Figs. 1 and 8). Horizontal isoelectric focusing of

this material gave rise to a single band which was isoelectric at pH

4.14 + 0.06. Antisera produced against CBH I(D) in rabbits generated

a single precipitin band after immunodiffusion against the homologous

protein, but had no cross-reactivity at all to CBH II or to any of the

endoglucanases tested. An experimentally determined extinction

coefficient of 13.80 + 0.20 (for a 1% solution at 280 nm) was used

in each calculation and pure CBH I(D) was found to have an

A280/A260 ratio of 1.82-1.84.

Neutral Carbohydrate Composition--The carbohydrate was removed

from CBH I(D) either by reductive 3-elimination to release 0-linked

oligosaccharides or by mild acid hydrolysis to release all sugars as


Polyacrylamide disc qel electrophoresis of crude
extracellular protein prepared from T. reesei
QM 9414 and highly purified cellobiohydrolase I(D).

Lanes 1 and 2 contain 100 pq each of crude extracellular
protein that was applied to the DEAE-Sephadex column. Lanes
3 and 4 represent 40 pq each of highly purified CBH I(D)
(from a pool of fractions 600-620) eluted from DEAE-Sephadex
at pH 3.6 (see Fig. 1). Lanes 1 and 3 were stained with
Coomassie Blue for protein and lanes 2 and 4 were stained
with the periodic acid-Schiff reagent for carbohydrate.

L.i 11


23 4

monosaccharides. The products were converted to the respective

peracetylated alditols and aldononitriles and analyzed by gas-liquid

chromatography (see Table II). The protein was found to contain about

13 mannosyl and 5 glucosyl residues which account for 6.0% of the 54

Kdalton molecular weight. This composition agrees well with previous

data (Table I,[493), even though protein was determined by the Lowry

colorimetric assay and not the extinction coefficient, as in the case

of the present work. This also agrees well with a value of 5.8 weight

percent carbohydrate obtained from phenol-sulfuric acid determination

using mannose as a standard.

Analysis of the products released by reductive B-elimination as

both the alditol acetates (derived from sugars directly attached to

the protein) and aldononitrile acetates provided information as to the

total carbohydrate composition, the number of 0-linked attachment

sites to the protein and the type of sugar attached (Table II). For

CBH I(D), mannose (identified as peracetylated mannitol) was the only

O--linked saccharide found to be attached to the protein and at an

average of 5.9 + 0.6 sites. The fact that the total number of

neutral sugars recovered from either reductive B-elimination or acid

hydrolysis were approximately the same suggests that all the neutral

carbohydrate was alkali-labile (i.e. 0-linked). This ruled out the

possibility of large N-linked structures, common in many

glycoproteins, as they are stable under the mild alkaline conditions


0-linked 01 igosaccharides--O-linked oligosaccharides

released from CBH I(D) by reductive B-elimination were separated by



CBH I(D) was treated with either mild acid or alkaline borohydride
and the products analyzed either as a mixture of PAANs and PAAs or, after
reduction with sodium borohydride, as the PAAs. For further details see
"Experimental Procedures." Values are expressed in moles/mole protein.

Method of Carbohydrate Release
Peracetylated Alditols
and Aldononitriles Acid Hydrolysis Alkaline Borohydride

Mannononitrile 15.2 + 1.0 6.2 + 0.6
Glucononitrile 5.9 + 0.5 5.2 + 0.8
Mannitol 5.9 + 0.6

Mannitol 12.8 + 0.5 13.8 + 0.5
Glucitol 5.3 + 0.5 4.9 + 0.3


Separation on a Biogel P-2 column of oligosaccharides
released from cellobiohydrolase I(D) by reductive

200 mq CBH I(D) was B-eliminated with 0.1M NaOH and 0.3 M
NaBH4 at 400C for 48 hours. The reaction was stopped
with 2N acetic acid and the mixture taken to dryness.
Following methanol evaporation, the reduced oligosaccharides
were deionized over Amberlite MB-3, lyophillized and then
applied to a Biogel P-2 column. For further details see
"Experimental Procedures". Peaks labelled M, M2, M,,G
and M G correspond to Mono-, di -, tri- a rd
tetrasaccharides, respectively.




w C

LL / l


80 100 120

gel filtration on Biogel P-2 (Fig. 9), purified by rechromatography on

the same column and weighed (Table III). Analysis of the eluate

permitted an estimate of an average of 4.2 chains of trisaccharide,

1.1-1.2 of the di- and monosaccharides and 0.7 chains of

tetrasaccharide per molecule of glycoprotein; calculations were made

on the assumption that each oligosaccharide chain has one attachment

site to the protein. These account for about 7.2 oligosaccharide

chains O-linked to CBH I(D) which is in fair agreement with an

estimate of 5.9 + 0.6 attachment sites obtained by gas

chromatography data (Table II). The 5.9% total neutral carbohydrate

found also agrees well with previous evidence from gas chromatography

(6%) and colorimetric determination (5.8%).

Analysis of the composition of each of the oligosaccharides by

gas-liquid chromatography (see Table IV) provided the first evidence

that these may be a related series. All contain mannose (as mannitol)

at the reducing terminus, and the tri- and tetrasaccharides were also

found to contain one residue of glucose. Since the peak shapes

generated from Biogel P-2 separation were symmetrical and the

stoichiometry of the gas chromatographic analysis of the di- and

trisaccharides yielded nearly integral values for each residue

indicating that each oligomer comprises one species, and is not,

therefore, a heterogeneous population. These results further

demonstrated the usefulness of the mixed alditol/aldononitrile acetate

technique in determining not only the linking sugars, but also the

composition of each oligosaccharide.



Oligosaccharides released by reductive B-elimination of CBH I(D) were
separated on Biogel P-2 (see Fig. 9), pooled and lyophillized. Samples
were then dried over Pe 0 and weighed. Calculations were made
assuming a molecular weight fr CBH I(D) of 54,000 daltons [49].

Weight Percent Weight Percent Moles per
Oligosaccharide of Carbohydrate of Glycoprotein Mole CBH I(D)

Tetra- 13.8 0.8 0.7

Tri- 66.9 4.0 4.2

Di- 12.1 0.7 1.1

Mono- 7.1 0.4 1.2

5.9% 7.2



Reduced oligosaccharides, purified on Biogel P-2, were hydrolyzed with
mild acid and the products analyzed by gas-liquid chromatography as a
mixture of the alditol and aldononitrile acetates. For further details,
see "Experimental Procedures".

Peracetylated Saccharide
Derivative Mono- Di- Tri- Tetra-

Mannitola 1.0 1.0 1.0 1.0

Mannononitrile 1.2 1.1 1.5

Glucononitrile 1.2 0.9

a mannitol hexaacetate has been normalized to 1.0 assuming one reducing
end per oligosaccharide chain prior to B-elimination.

Methylation Analysis--The types of linkages between the sugar

residues in the oligosaccharides attached to CBH I(D) were studied by

methylation analysis. The partially methylated alditol acetates

generated were subsequently separated and identified by gas

chromatography/mass spectrometry (see Fig. 10).

The intact glycoprotein yielded three methylated species

corresponding to the 2,3,4,6-tetramethyl, 3,4,6-trimethyl and

2,3,4-trimethyl alditol acetates which indicated the presence of

non-reducing terminal, 2-substituted and 6-substituted hexoses,

respectively (Fig. 10,Panel A). No dimethyl hexitol acetates were

observed indicating that all oligosaccharide chains are unbranched.

The ratio of the 3,4,6-/2,3,4-trimethyl species is 0.85, which agrees

well with the ratio expected (0.83) from previously determined amounts

of tri- and disaccharide attached to CBH I(D) (see Table III). The

disaccharide yielded two partially methylated alditol acetates

corresponding to the 1,3,4,5,6-pentamethyl and 2,3,4,6-tetramethyl

species indicating a 1-2 linked mannobiitol (Fig. 10,Panel B).

Analysis of the reduced trisaccharide also reveals three peaks

corresponding to the 1,3,4,5,6-pentamethyl, 2,3,4,6-tetramethyl and

2,3,4-trimethyl hexitol acetates (Fig. 10,Panel C); these indicated

the presence of 2-substituted (reducing end), non-reducing end and

6-substituted residues, respectively. The tetrasaccharide generated

four peaks corresponding to the 1,3,4,5,6-pentamethyl,

2,3,4,6-tetramethyl, 3,4,6- and 2,3,4-trimethyl hexitol acetates (Fig.

10,Panel D); these would be expected to arise from 2-substituted

alcohol (reducing-end), non-reducing end, 2-substituted and

6-substituted residues, respectively. The absence of the formation of


Methylation analyses of cellobiohydrolase I(D) and
the oligosaccharides released from cellobiohydrolase
I(D) by reductive B-elimination.

G.C./M.S. profiles of partially methylated alditol acetates
generated from (A) CBH I(D) (B) reduced disaccharide, (C)
reduced trisaccharide and (D) reduced tetrasaccharide.
Individual peaks were positively identified by their
electron-impact spectra compared with prepared standards for
further details, see "Experimental Procedures". Arrows
indicate elution of standards: 1. 2-0-acetyl-1,3,4,5,6-
penta-0-methyl glucitol; 2. 1,5-di-0-acetyl-2,3,4,6-
tetra---methyl mannitol; 3. 1,2,5-tri-0-acetyl-3,4,6-
tri-0-methyl mannitol; 4. 1,5,6-tri-0-acetyl-2,3,4-
tri-0-methyl mannitol.

10:00 13:20






(A) +




dimethyl species from any of the oligosaccharides suggested that none

of the chains are branched.

Electron impact spectra of all the partially methylated alditol

acetates were compared with those prepared standards (Appendix C) and

examples of those from the tetrasaccharide (Fig. 10,Panel 0) are shown

(Fig. 11). Glucitol and mannitol peracetates methylated in the same

positions coeluted on 1% OV-225 and also gave identical

electron-impact fragmentation patterns (data not shown). The

methylation profiles, in conjunction with the evidence from

glycosidase digestions (see below), suggest a series of related

oligosaccharides corresponding to Man(1-2)mannitol and

Glc(1-6)Man(1-2)mannitol for the di- and trisaccharides, respectively.

The data for the tetrasaccharide can be interpreted as either a

Man(1-2)Glc(1-6)Man(1-2)mannitol or a Man(1-6)Glc(1-2)Man(1-2)mannitol

oligosaccharide. Evidence from acetolysis (shown below) will support

the former structure.

Glycosidase Digestions--Oligosaccharides released from CBH

I(D) by reductive B-elimination were subjected to sequential

glycosidase digestion to determine the sequence of residues and the

anomeric nature of glycosidic bonds (see Fig. 12). Each of the

glycosidases were assayed for contaminating glycosidase activity using

the appropriate p-nitrophenyl glycosides. a-Glucosidase from yeast

was found to have minor levels of other glycosidase activities

(expressed as percent of a-glucosidase activity) as follows:

B-glucosidase (7.9 x 10-"5), a-mannosidase (5.6 x 10-%) and

B-mannosidase (4.7 x 10-5%). The a-mannosidase from jack bean also


Comparison of electron impact spectra obtained from
partially methylated alditol acetates of the
tetrasaccharide from cellobiohydrolase I(D) with

Panels (A), (B), (C) and (D) are spectra taken from the four
peaks in Figure 6, Panel (D). Panel (E) is a standard
pentamethyl alditol acetate from derivatization of
kojibiitol, and represents 2-O-acetyl-1,3,4,5,6-penta-O-
methyl glucitol. Panels (F), (G) and (H) are standards taken
from the partially methylated alditol acetates of yeast
mannan, and represent 1,5-di-0-acetyl-2,3,4,6-tetra-
0-methyl mannitol, 1,2,5-tri-0-acetyl-3,4,6-tri-
0-methyl mannitol and 1,5,6-tri-0-acetyl-2,3,4-
tri-0-methyl mannitol, respectively. The fragmentation
pattern of the pairs (A) and (E), (B) and (F), (C) and (G),
(D) and (H) were compared for positive identification.



1 U





I1I Ige

S(A) 7, (E)






TO II ] 14'

(D) (, 1 )

71 71 [ !
i I 1 0

1(1 1 s




HPLC separation of the products of sequential
glycosidase digestion of oligosaccharides released
from cellobiohydrolase I(D) by reductive

Shown are products of reduced oligosaccharides from (A) CBH
I(D) tetrasaccharide + a-mannosidase, (B) trisaccharide from
(A) + a-glucosidase, (C) disaccharide from (B) +
a-mannosidase, (D) CBH I(D) trisaccharide + a-glucosidase,
(E) disaccharide from (D) +a-mannosidase and (F) CBH I(D)
disaccharide + a-mannosidase. Elution times of standards are
shown: 1. mannose (glucose); 2. mannitol; 3. CBH I(D)
reduced disaccharide; 4. CBH I(D) reduced trisaccharide; 5.
CBH I(D) reduced tetrasaccharide. Separations were performed
as described in "Experimental Procedures".

12 3

5 (A)







0 10 20 30 40

1 1



was found to have minor contaminating activities of 8-glucosidase

(5.8 x 10-4%), B-mannosidase (2.2 x 10-3%) and a-glucosidase (2.8
x 10-4%). These levels of contaminating activities are so low that

they would not be expected to be a substantial influence, unless the

incubations were for a long period of time and/or with large amounts

of enzyme.

The products of digestion were separated by polar reverse-phase

HPLC. The monosaccharides released were converted to peracetylated

alditols and/or aldononitriles and were identified by gas-liquid

chromatography; subsequently the oligosaccharides were degraded with

an alternate glycosidase. All monosaccharides released with

a-glucosidase (Fig. 12,Panels B and 0) were identified as the

peracetylated glucononitrile. The trisaccharide(s) was not a aood

substrate for the yeast a-glucosidase and so more units of this enzyme

were necessary than in the case of the a-mannosidase. Consequently,

small amounts of mannose and mannitol were produced by the minor

a-mannosidase contamination. It should be mentioned that digestion of

the trisaccharide with B-glucosidase from Trichoderma did not cause

the release of glucose, indicating that the non-reducing terminal

glucose residue is not B-linked. All non-reducing terminal mannose

residues, however, were readily cleaved by a-mannosidase (Fig.

12,Panels A,C,E and F) and were identified as the mannononitrile

acetates. Recovery of the products after each digestion and the HPLC

chromatography step was about 85% in each case.

From the HPLC profiles, it was evident that all the glycosidic

linkages are in the a-configuration. The sequence of the tetra- (Fig.

12,Panels A, B and C), tri- (Fig. 12,Panels D and E) and disaccharides

(Fig. 12,Panel F) was proposed to be ManaGlcaManamannitol,

GlcaManamannitol and Manamannitol, respectively.

Acetolysis--The reduced tri- and tetrasaccharide from CBH I(D)

were each subjected to acetolysis in order to confirm the position of

(1-6) glycosidic linkages within the oligosaccharide chains (see

Experimental Procedures). After acetylation, oligosaccharides have

been shown to be differentially susceptible to hydrolytic cleavage

depending on the nature of the glycosidic linkages. The rate of

cleavage of acetylated a-linked mannooligosaccharides has been shown

to be (1-6)>>(1-3)>(1-2) [71], thus under controlled conditions, (1-6)

glycosidic bonds may be hydrolyzed preferentially.

The products of the reaction were separated by HPLC (see Fig.

13) and then further analyzed as the peracetylated alditols and/or

aldononitriles by gas-liquid chromatography. The trisaccharide was

cleaved into two products, a monosaccharide coeluting with a glucose

standard and reduced disaccharide coeluting with authentic reduced

disaccharide from CBH I(D) (Fig. 13,Panel A). The monosaccharide was

converted to the peracetylated glucononitrile and the disaccharide

into a 1:1 mixture of peracetylated mannitol and peracetylated

mannononitrile. This array of products from acetolysis is consistent

with a terminal glucose attached by a (1-6) glycosidic linkage to

mannobiitol. The tetrasaccharide also was cleaved into two products,

a reduced disaccharide which coeluted during HPLC analysis into

authentic reduced disaccharide from CBH I(D) and a (unreduced)

disaccharide (Fig. 13,Panel B). The reduced disaccharide was

converted by hydrolysis into a 1:1 mixture of peracetylated mannitol


HPLC separation of the products of acetolysis of
oligosaccharides released from cellobiohydrolase I(D)
by reductive B-elimination.

(A) Reduced trisaccharide and (B) reduced tetrasaccharide
were subjected to acetolysis as described under
"Experimental Procedures". 1 and 2 refer to the elution
times of glucose and reduced disaccharide from CBH I(D),
respectively. Separations were performed on a Whatman
Partisil PXS 10/25 PAC column eluted with 75% acetonitrile
at a flow rate of 1.5 ml/min.

I t (A)


10 20

and peracetylated mannononitrile and the disaccharide was converted to

a 1:1 mixture of peracetylated mannononitrile and peracetylated

glucononitrile. This acetolysis cleavage pattern is consistent with a

(mannose,glucose) disaccharide attached by a (1-6) glycosidic bond to


13C-NMR--Proton decoupled 13C-NMR at 75.5 MHz was

performed on CBH I(D) and the reduced tri- and disaccharides released

from CBH I(D) with alkaline borohydride (Fig. 14 and Table V). CBH

I(0) generated three signals in the anomeric region (C1) at 105.1,

102.5 and 100.8 ppm, which indicates the presence of three predominant

types of hexoses (Fig. 14,Panel A). It should be mentioned that at

this resolution, minor species were not distinguishable, so it is

assumed that these signals reflect the trisaccharide (major species)

attached to the protein. The signal at 100.8 ppm has been assigned to

the glucosyl residue (C-1") at the non-reducing end. The signal at

105.1 ppm was assigned to the central mannosyl residue (C-1') of the

trisaccharide and, although the signal is far down field from any

known standards, it is suggested that this is due to the unique nature

of the substitution of this residue. By comparison with published

values [100], the signal at 102.5 ppm was assigned to a mannosyl

residue (C-1) a-linked to a threonyl residue on the peptide. The

strength of this signal relative to those of the other anomeric

carbons suggest that most or all of the oligosaccharides are attached

to threonyl residues, rather than to seryl residues. Allerhand et al.

[100] have demonstrated that ManaThr and ManaSer standards generate

C1 signals 1 ppm apart (at 102.8 and 101.8 ppm, respectively), so


Proton decoupled 13C-NMR spectra of
cellobiohydrolase I(D) and the oligosaccharides
released from cellobiohydrolase I(D) by reductive

Spectra of (A) 400 mg CBH I(D), (B) 10.5 mg reduced
trisaccharide from B-elimination of CBH I(D) and (C) 2.0 mg
reduced disaccharide from B-elimination of CBH I(D), were
recorded at 75.46 MHz. Chemical shifts are reported in parts
per million from internal TSP. Spectra were the result of
(A) 35712 accumulations, (B) 4304 accumulations and (C)
16328 accumulations, respectively. For further details, see
"Experimental Procedures."

100 90 80 70

100 90 0o 70

100 90 P 0 70



Proton coupled and decoupled 13C-NMR spectra were performed at 75.5
MHz. 200 mg CBH I(D), 10.5 mg reduced trisaccharide and 2.0 mq reduced
disaccharide yere used. Chemical shifts are expressed in PPM and, in
parentheses, J values are expressed in Hz. For further details,
see "Experimental PFcedures".

Chemical Shifts Relative to TSP

Carbon CBH I(D) Reduced Reduced Residue
Atom Trisaccharide Disaccharide

Anomeric Carbons

C-1 102.5 (171.1) Man

C-1' 105.1 (173.6) 104.6 (169.2) 104.3 Man

C-1" 100.8 (171.5) 100.9 (170.5) Glc

Hydroxymethyl Carbons

C-1 66.0 66.0 Mannitol

C-6 63.9 63.9 64.0 Man

C-6' 63.8 Man

C-6" 63.4 63.4 Glc

attachment to more than one type of hydroxy amino acids would be

expected to result in separate signals, which is not the case here.

Protein folding is thought not to affect the chemical shift of

carbohydrate residues [101], as it does the chemical shift of amino

acid residues.

The above data is in agreement with sequence evidence from

Pettersson [60] which suggested that all of the 0-linked

carbohydrate is present in a 20 amino acid region located 33 residues

from the C-terminus of CBH I(D); the work of Shoemaker et al. [61],

who have sequenced the gene for the enzyme, suggests that the region

contains 8 threonyl and 3 seryl residues. This is also in agreement

with amino acid analysis of the polypeptide from which the

carbohydrate had been removed by B-elimination (see below).

Comparison of the CBH I(D) anomeric carbon signals with those of

the reduced trisaccharide shows the loss of the signal at 102.5 ppm

(Fig. 14,Panel B). This was expected as the linking sugar has now been

reduced and thus has no anomeric carbon atoms. This further confirms

that this signal is due to the mannosyl residue attached to the

protein. The signals at 105.1 ppm (C-i') has been shifted to 104.6

ppm, suggesting this mannosyl residue is close to the reducing end.

The signal at 100.9 ppm is unchanged, suggesting that this glucosyl

residue is furthest removed from the reducing end. This data is in

agreement with the chemical evidence of the sequence of sugars in the

trisaccharide (see above). Spectra of the reduced disaccharide reveal

a single signal at 104.3 ppm (C-l') due to the non-reducing terminal

mannosyl residue (Fig. 14,Panel C). Comparison of the signals in the

hydroxymethyl region (63-66 ppm), indicates that the carbohydrate of

intact CBH I(D) (Fig. 14,Panel A) generates two signals. Thus the

trisaccharide, as it is attached to the protein, has only two

hydroxymethyl groups and must contain one 6-substituted residue. The

reduced trisaccharide reveals three signals, and thus it must contain

one 6-substituted residue (Fig. 14,Panel B). The reduced disaccharide

also gives rise to three signals, which suggests that it cannot

contain a 6-substituted residue (Fig. 14,Panel C).

Coupled 1C-NMR spectra performed on both CBH I(D) and the

reduced trisaccharide, provide coupling constants for anomeric carbons

and yield information on the nature of those carbons. All

JCH-values (Table V) for anomeric carbons of both CBH I(D) and
the reduced trisaccharide were between 169.2-173.6 Hz indicative of

a-linked sugar residues (B-linkages generate J-values of about 160

Hz), in agreement with the results of the glycosidase digestion (see


H-NMR--The H-NMR signals of anomeric protons at 300 MHz

of the reduced oligosaccharides B-eliminated from CBH I(D) are shown

(Fig. 15). As with the 13C-NMR, the reducing end residues do not

contain anomeric carbons and do not generate signals in this region.

At this resolution, with a 300 MHz instrument, distinction between

signals separated by less than 2 Hz was not possible and thus it was

not possible to distinguish between a- and B-anomers of mannose by

these soectra. Signal assignments are shown in Table VI. The glucose

signals generated from the tri- and tetrasaccharides are split by 3.5

Hz and 3.4 Hz, respectively, due to the equatorial proton at C1 and

the axial proton at C2, which indicates that both residues are

U >
-o :3
C0 3
U S-


0 (

C) -

r 0

0 0

5.- .-
m- r

O 0

S4- 4-

0 0-

0- r- c
+' to

CL. o c
C 0) *r

- +'0 4-Q M C,-
0 u 40 =r-

s- (1) U) (I a
(0 -DI CL. o0 E
-C L L c S-.
U O L0
J 0 4- ( -
- rL. 0.

l) r> a 0 o U)-

U = I
3 u C
-0 0 n E m m

S0) O 4-*

S_ l -CC ..,- $.-
4r 0 --

"- O 0

L- S- C) 1) S-
<:5 U a: L

- O 0 3
to o a C 4 .
S_ 5. 0 +O *r- Li

s c o -( o 0
S 3 L ..t

0 0 0

CE U )-

MM a)U) -4 a)

-E.lCU *
) 3U LU0
smU s aD

a? On "3 aj r'
5- 0 .0 (/ 0 0 3















Oligosaccharide Anomeric Sugar Chemical JH-H Anomer
Proton Shift

Disaccharide H' Man 4.953 <2.0 a or B

Trisaccharide H' Man 4.995 <2.0 a or B
H" Glc 4.969/ 3.5 a

Tetrasaccharide H' Man 4.998 <2.0 a or 3
H" Glc 4.916/ 3.4 a
H'" Man 4.942 <2.0 a or B

present in the a-configuration. The addition of a terminal mannosyl

residue on the tetrasaccharide compared to the trisaccharide resulted

in an upfield shift of the anomeric glucose signal of 0.05 ppm. The

signals at 4.995 ppm and 4.998 ppm were assigned to the mannosyl

residues penultimate to the reducing end of the reduced tri- and

tetrasaccharides as these residues were in equivalent positions in the

respective molecules. Confirmation of these spectral assignments is

not possible at this time as these are unique oligosaccharides and

standards are not available.

Amino Acid Analysis--a-Amino-butyric acid (Abu) produced from

the B-elimination and reduction of CBH I(D) was measured to determine

the possible number of oligosaccharide chains attached to threonine on

the proteins. A control experiment without B-elimination but in the

presence of PdCl2 yielded a background of 4.9 + 0.2 moles of Abu

per mole CBH I(D). The origin of this background Abu is not known.

Following B-elimination, 11.8 + 0.6 moles of Abu per mole CBH I(D)

was found. Accounting for the control, there are 6.9 + 0.7 moles

Abu per mole CBH I(D) produced following B-elimination of 0-linked

oligosaccharides. This value of 6.9 + 0.7 oligosaccharide side

chains agrees with the value for side chains determined from analysis

of carbohydrate composition by GLC (5.9 + 0.6) and HPLC (7.2)

analyses. It also is in agreement with evidence from the 13C-NMR

and other sequence data [60,61] that most or all of the carbohydrate

is attached to threonine.