.-

 

;-‘11 um

m: ‘

0:th d -:

  
   

rum in

   

:1;
. -. .n
- h

          

 

« 9 ,.‘
“m“
J“

 
 
 
  
   

 
  

 

.1

   

LIIEBRAR

WWW\IIHHWINHINIHNHHIHI “Ill

 

 

 

 

 

 

This is to certify that the
dissertation entitled
IDENTIFICATION AND CHARACTERIZATION OF CARBOHYDRATE
BINDING PROTEIN 35 GENE STRUCTURE

presented by

SHIZHE JIA

has been accepted towards fulfillment
of the requirements for

DOCTOR OF PHILOSOPHY degreein BIOCHEMISTRY

 

 

Engi LIB-t3

Major professor

~‘1/‘3 [7/

Date

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

___

 

LIBRARY
lMlchlgan State
: University

L.__

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE
l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r -"————r‘=“

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
cmmtlnnt

 

IDENTIFICATION AND CHARACTERIZATION OP CARBOHYDRATE

BINDING PROTEIN 35 GENE STRUCTURE

BY

SHIZHE JIA

A DISSERATATION
SUBMITTED TO
NICRIGAN STATE UNIVERSITY
IN PARTIAL FULFILLMENT OP REQUIREMENTS

FOR TEE DEGREE OP

DOCTOR OF PHILOSOPHY

DEPARTMENT OF BIOCRENISTRY

1990

éjj- a!) w

ABSTRACT

Inna-arcane»: nun marrow or cmoamm
BINDING PROTEIN 35 GENE STRUCTURE
3]!
Shirhe Jia

Carbohydrate Binding Protein 35 (C8235) is a galactose-specific
lectin identified in the cytoplasm.and nucleus of mouse 3T3 fibro-
blasts, as well as a number of other tissues and cell types.
Affinity purified antibodies directed against C8235 were used to
screen a lambda gtll expression library derived from.mRNA of mouse
3T3 fibroblasts. One positive clone containing cDNA for C8235 was
characterized by expression of a fusion protein containing beta-
galactosidase and C8235 sequences. Limited proteolysis of bacterial
lysates containing the fusion protein. followed by $88 polyacry-
lamide gel electrophoresis and immmoblotting with anti-C8235,
yielded.a peptide mapping pattern comparable to that obtained.frem
parallel treatment of authentic C8235. Such a limited proteolysis
followed by affinity chromatography on a column of Sepharose deriva-
tized with galactose also yielded a 30-kDa polypeptide that exhi-
bited carbohydrate-binding activity. This polypeptide can be insane-
blotted with anti-C8235. but not with antibodies directed against
beta-galactosidase. These results indicate that the positive clone
is be an authentic 68235 cDNA clone.

The complete nucleotide sequence of the cDNA clone for C8235 has
been determined. The deduced amino acid sequence showed that the

protein consists of two domains: (a) an amino terminal portion that
contains an internal sequence homology featured by 8 repeats of a
9 amino acid motif which is rich in proline and glycine residues and
which shows structure similarity to certain regions of proteins of
the heterogeneous nuclear ribonucleoprotein complex (hnRNP) ; and (b)
a carboxyl terminal portion that is homologous to beta—D-galactoside

specific lectins isolated from a number of animal tissues.

A cloned genomic DNA segment for C8235 gene was isolated . Nucleo-
tide sequence analysis of the cloned gene, as well as the results
of genomic Southern blot hybridization revealed that the gene is
unique and spans 9 kilobases of genomic DNA. The mRNA for the lectin
is encoded by five exons separated by four introns. Intron I must
necessarily be in the 5' transcribed but untranslated region of the
primary transcript. Examination of the nucleotide sequence 5' to the
transcription initiation site revealed characteristic TATA and CCAAT
boxes. A putative serum—responsive element has also been identified
about 200 nucleotides upstream from the site for initiation of
transcription. This may serve as a binding site for a class of
transcription factors responsible for the serum-stimulated

expression of the C8235 gene.

DEDICATED TO MY PARENTS

iv

ACKNOWLEDGEMENTS
I would like to express my sincere appreciation to Dr. John Wang
for all his encouragement and support. He has provided me with
opportunities such that I can become more mature during my graduate

career.

I would like to thank my committee, Drs. Lee McIntosh, Arnold
Revzin, Leonard Robbins and John Wilson for their assistance during

my education.

I thank Drs. Richard Anderson, Melvin Schindler, John Ho, Ioannis
Moutsatsos for their advice and discussion. I am grateful to my
laboratory mates for their cooperation and criticisms: Patty Voss,
Kristen Yang, Sung-Yuan Wang, Neera Agrwal, Kim Hamann, Jamie Laing

and Dr. Liz Cowles.

TABLE OF CONTENTS

PAGE
LIST OF TABLES .......................................... i
LIST OF FIGURES .. ........................................ ii
LIST OF ABBREVIATIONS .................................... 1v
CHAPTER I LITERATURE REVIEW .............................. l
I INTRODUCTION ......................................... 1
II ANIMAL LECTINS: STRUCTURAL MOTIFS ................... l
A: CLASSIFICATION ................................... l
B: C-TYPE LECTIN .................................... 3
C: S-TYPE LECTIN ...... . ......................... . 7
III CARBOHYDRATE BINDING PROTEIN 35 ..................... 10
A: ISOLATION AND CHARACTERIZATION ................... 10
B: DISTRIBUTION ........... ' .......................... 10
C: NUCLEAR LOCALIZATION OF CBP35 ...... - .............. 12
D: REGULATION OF EXPRESSION ... ...................... 14

IV IDENTITY BETWEEN CBP35 AND PROTEINS STUDIED UNDER
OTHER NAMES ........................................ . 15
A: L—34 ............................................. 16
B: Mac-2 ........................ . ................... 16
C: LAMININ BINDING PROTEIN (LBP35) ............. ..... 18
D: IgE BINDING PROTEIN (eBP) ........................ 18
V STRUCTURE OF 14 KD S-TYPE LECTINS ................... 19
VI GOAL OF THIS THESIS ................................. 20
REFERENCES .......................................... 20

CHAPTER II CARBOHYDRATE BINDING PROTEIN 35: MOLECULAR CLONING

AND EXPRESSION OF A RECOMBINANT POLYPEPTIDE WITH

vi

LECTIN ACTIVITY IN ESCHERICHIA COLI ...........
SUMMARY .................................................
INTRODUCTION .... ...... . .................................
MATERIALS AND METHODS ..................... ..............

(a) AFFINITY PURIFICATION OF ANTIBODIES AGAINST CBP35
(b) SCREENING OF THE LAMBDA gtll EXPRESSION LIBRARY .....
(c) ISOLATION OF RECOMBINANT POLYPEPTIDE WITH LECTIN
ACTIVITY FROM FUSION PROTEIN OF CLONE 1 ....... .....
(d) NORTHERN BLOTTING ANALYSIS .........................
RESULTS AND DISCUSSION ..................................
(a) IDENTIFICATION OF A cDNA CLONE FOR CBP[35 .. ........
(b) EXPRESSION OF A FUSION PROTEIN ENCODED EY LAMBDA gtll
CLONE 1 DNA ... ...... ................ ..... .... ......
(c) PEPTIDE MAPS DERIVED FROM FUSION PROTEIN AND CBP35
(d) ISOLATION OF RECOMBINANT CBP35 FROM FUSION PROTEIN
(e) NORTHERN ELOT ANALYSIS OF mRNA ENCODING CBP35 ......
ACKNOWLEDGMENTS ........................................ .

REFERENCES ..............................................

CHAPTER III CARBOHYDRATE BINDING PROTEIN 35: COMPLEMENTARY

DNA SEQUENCE REVEALS HOMOLOGY WITH PROTEINS OF

A: DNA CLONING .........................................
B: DNA SEQUENCING ......................................
RESULTS AND DISCUSSION ...... . ............. . ..............

REFERENCES . ............. , ................................
CHAPTER IV NUCLEOTIDE SEQUENCE OF THE MURINE GENE FOR

vii

24
25
26
27
27
28

29
30

32
32

32
35
38
45
49
50

54
55
56
57
58
58
58
59

74

FOOTNOTES ...............................................
SUMMARY .................................................
INTRODUCTION ............................................
EXPERIMENTAL PROCEDURES .................................
A: SCREENING OF THE GENOMIC LIBRARIES ..................
B: CLONING AND SEQUENCING ..............................

C: DIRECTIONAL DELETION OF A 2.7 KB FRAGMENT FOR DNA
SEQUENCING ..........................................
D: PRIMER EXTENSION ASSAY ..............................
E: SOUTHERN ANALYSIS OF GENOMIC DNA ....................
RESULTS ....... p ..........................................
ISOLATION AND CHARACTERIZATION OF GENOMIC CLONES
SEQUENCE ANALYSIS OF THE GENOMIC CLONES .............

STRUCTURAL FEATURES OF THE CBP35 GENE ...............

UOU’W

COMPARISONS OF THE SEQUENCES OF CBP35 WITH L—34,
Mac-2, AND eBP ......................................
DISCUSSION ..............................................

REFERENCES ..............................................

viii

76
77
78
79
81
81
81

84
85
86
87
87
88
94

 

LIST OF TABLES

TABLE
CHAPTER I
l. EXAMPLES OF MEMBRANE BOUND AND SOLUBLE LECTINS .....
2. COMPARISON OF C—TYPE AND S-TYPE ANIMAL LECTINS .....
CHAPTER IV
1. DIFFERENCES IN THE AMINO ACID SEQUENCES OF CBP35

VERSUS L-34 ........................... . ............

PAGE

 

LIST OF FIGURES

FIGURE
CHAPTER I
1. SUMMARY OF STRUCTURAL FEATURES OF C—TYPE ANIMAL
LECTINS ................ . ...........................
2. SUMMARY OF STRUCTURAL FEATURES OF S-TYPE ANIMAL
LECTINS ............................................
CHAPTER II '
1. ISOLATION OF CLONE 1 AND CDNA INSERT OF CLONE 1
FOR CBP35 CDNA ............. , ........................
2. DETECTION OF EXPRESSION OF FUSION PROTEIN BY
IMMUNOBLOTTING ANALYSIS ............................
3. COMPARISON OF TEE PEPTIDE MAPS DERIVED FROM THE FUSION
PROTEIN AND CEP35 ....... ...........................
4. IMMUNOBLOT ANALYSIS OF RECOMBINANT POLYPEPTIDE WITH
LECTIN ACTIVITY USING AFFINITY PURIFIED ANTI-CBP35

5. NORTHERN BLOT ANALYSIS OF THE POLYA+ FRACTION OF

CHAPTER III

1. RESTRICTION MAP OF THE CDNA INSERT OF CBP35 CLONE 1

STRATEGY FOR SEQUENCING .. ..........................

2. NUCLEOTIDE AND DEDUCED AMINO ACID SEQUENCE FOR CBP35

3. COMPARISON OF THE AMINO ACID SEQUENCE OF CBP35
WITH THE AMINO ACID SEQUENCES OF SIX OTHER BETA-D-

GALACTOSIDE BINDING PROTEINS .......................

PAGE

34

37

40

43

47

61

63

66

ALIGNMENT OF THE AMINO ACID RESIDUES SHOWING INTERNAL

SEQUENCE HOMOLOGY .................................. 69
6. COMPARISON OF THE DEDUCED AMINO ACID SEQUENCE OF CBP35

WITH THE AMINO ACID SEQUENCES OF hnRNP PROTEINS .... 71

CHAPTER Iv

1. RESTRICTION MAP AND SEQUENCING STRATEGY FOR MURINE

CBP35 GENE ...... .................................. 83
2. SOUTHERN BLOT HYBRIDIZATION OF MURINE GENOMIC DNA

WITH cDNA FOR CBP35 ............................... 90
3. NUCLEOTIDE SEQUENCE OF THE MURINE CBP35 GENE ...... 93
4. IDENTIFICATION OF THE TRANSCRIPTION INITIATION SITE OF

THE MURINE CBP35 GENE BY PRIMER EXTENSION ASSAY ... 96
5. COMPARISON OF THE NUCLEOTIDE SEQUENCE OF THE CODING

REGION OF THE CBP35 GENE AND CORRESPONDING SEQUENCE

REPORTED FOR L-34 .................................. 100
6. COMPARISON OF THE NUCLEOTIDE SEQUENCE OF THE 5'

UNTRANSLATED REGION OF L-34, Mac-2, eBP AND CBP35 .. 114

xi

LIST OF ABBREVIATIONS

ASGP: asialoglycoprotein

beta-galase: beta—galactosidase

bp: base pair(s)

CBP: carbohydrate binding protein

CLL-I: Chicken lactose lectin I

CLL—II: Chicken lactose lectin II

CRD: carbohydrate recognition domain

eBP: IgE binding pro:ein

EDTA: (ethylenedinitrilo)-tetraacetic acid
Gal: galactose

hnRNP: heterogenous nuclear ribonucleOprotein Complex
8R2: horseradish peroxidase

IPTG: isopropyl~beta-D-thio—galactopyranoside
IVS: Intervening sequence

kb: kilobase(s)

kD: kilodalton(s)

LBPBS: laminin binding protein 35

L—34: mouse lectin 34

M-G—P: Mannose 6-phosphate

Mac-2: mouse macrophage surface antigen
PAGE: polyacrylamide gel electrophoresis
PMSF: phenylmethyl sulfonyl fluoride

SDS: sodium dodecyl sulfate

SRE: serum response element

TE: 125mM Tris-HCl, lmM EDTA, pH 6.8

Tris: tris (hydroxymethyl) aminoethane

xii

CHAPTER I

LITERATURE REVIEW

I INTRODUCTION

Carbohydrate binding proteins bind to saccharide-containing regions
of glycoproteins and glycolipids. Lectins are carbohydrate binding
proteins that were originally identified. in grunn: extracts as
agglutinins of erythrocytes (1). In the more recent literature,
lectin and carbohydrate binding protein have become interchangeable
terms, and these will be used similarly in this thesis. Although
lectins were originally identified in plants, they have also been
found in many other organisms including bacteria, slime molds and
vertebrates (2). In this thesis, the gene structure of an animal
lectin, designated Carbohydrate Binding Protein 35 (CEP35), will be
described. The literature review will therefore focus on certain

structural themes of animal lectins.

II ANIMAL LECTINS : STRUCTURAL MDTIFS

A: CLASSIFICATION

There are two main Classification systems for animal lectins. The
first system classifies the proteins according to solubility (Table

I): a) integral membrane lectins which require detergents for

their solubilization; and b) lectins soluble in aqueous buffer (17).

2
Key examples of the membrane bound lectin group include the

asialoglycoprotein (ASGP) receptors (3, 4) that may function to
clear serum glycoproteins and the mannose 6—phosphate (M6P)
receptors that facilitate the transport of lysosomal enzymes from
the golgi to their target organelles (5, 6). The soluble lectins in
turn fall into two major families, one with molecular weights of
14,000—16,000 (14 kD —16 kD) and a second with molecular weights of

29,000 -35,000 (29 kD-35 kD).

Another classification of animal lectins is based on the dependence
of the carbohydrate binding activity on cations (C—type lectins )
or on thiols (S—type lectins) (18). The carbohydrate binding
activity of C—type lectins is Ca2+—dependent, while the S—type
lectins are thiol—dependent for the activity. Table II summarizes
the distinguishing features of the two types of lectins (18). From
numerous structural studies, it is now clear that both C-type and
S-type lectins consist of two different domains. The first domain,
if one exists, is termed the special effector domain; this confers
unique properties on the particular lectin. The second domain is the

carbohydrate recognition domain (CRD).

B: C-TYPE LECTINS

Figure l is a summary of the structural features of C—type lectins
(18) . An important feature of the Structure is that all these lectins
have a CRD of 130 amino acid residues, in which 18 amino acid re—
sidues are conserved (20). Of particular interest are the four Cys
residues that are involved in disulfide bonds. It should be noted
that the carbohydrate binding specificities of different C-type

lectins differ: some are specific for galactose (e.g. ASGP receptor)

Table I Examples of Membrane Bound and Soluble Lectins

 

CLASSIFICATION LECTIN SUBUNIT MW LIGAND SOURCE REFERENCE

 

 

 

 

 

 

 

Membrane ASGP 26 K GlcNAc Chicken 3
bound receptor 52 K D-Gal Rabbit 4
M-6-P 46 K _ Man-6-P Bovine 5
receptor 225K Man-6-P Human 6
CLL—I/CLL-II 14—16K D-Gal Chicken 7-9
Soluble
L-34 /CBP35
Mac~2/LBP35 29-35 K D—Gal Mouse 14—16
eBP 31K D—Gal Rat 36

 

Table II Comparison of C—type and S—type Animal Lectins

 

 

PrOperty C-type lectins S-type lectins
Caz+ requirement Yes No
State of cysteines Disulfides Free thiols
Solubility Variable Aqueous solutions
Location Extracellular Intracellular and

extracellular

Carbohydrate
specificity Various Mostly B—galactosides

 

Figure 1.

Summary of the structural features of C-type animal

lectins. The invariant residues found in the common

carbohydrate-recognition domain of the C-type lectins
are shown, flanked by schematic diagrams of the
special effector domains (if any) found in individual

members of the family . EGF, epidermal growth factor;

GAG, glycosaminoglycan; (adapted from ref. 18)

a mrzmaa

 

       
 

 

 

    
   

     
       

   

mpquawm
3:..-qu $8
.. o ... 0 2350865
5.38%:
mZEIOO
52-30 flb\lx\o€ mzaxo So
a. i. E l E
L F . -....- figuroooflomm
. a ma
...TV a: 1
/ . _
// 0..»0 2.83 2.5% «mm
/
/ 2.83 5..
/
/
// mxfiwmlk Fzfioqumam $420132
C 255.:
/ o ozaza- mmozzdz
$5323
quomrmwizqu «638% we... wtoorazz _
2.83 0:3er 286:;
2228
till... 29:58? - r «058%

 

mkdmotlommdo z_w.»O¢QOU>JOOJSm<
“mark: 0 .

 

7
while others are specific for mannose. Thus, the conservation of
amino acids in the CRD is for carbohydrate—binding, not necessarily

for a specific saccharide.

A second feature of the structure is that different lectins in the
C-type category have distinct effector domains. For example, the
ASGP receptor described earlier is a membrane bound lectin; it
therefore has a membrane anchor as its effector domain (Fig. 1).
Similarly, the cartilage proteoglycan core protein has a CRD that
is linked.to an effector domain to which many glycosaminoglycans are
attached. Finally, certain C-type lectins, such as the fly lectin
and the sea urchin lectin, contain only the CRD; they lack the

effector domain.

Thus, although the exact functions of all of the C—type lectins are
still not entirely understood, the structure information of this
class of lectins is relatively clear. In fact, some of the structural
information on the effector domain provides clear cut notions of
their overall function. This theme is also shared by the domain

delineation in the S-type lectins.
C: S-TYPE LECTINS

In S—type lectins, there are two groups of proteins: 14 —16 kD and
29-35 kD lectins. The 14—16 kD lectin group exhibits only one domain,
the CRD. In addition to the CRD, the 29-35 kD lectins group exhibits
a second domain. These features are schematically illustrated in

Figure 2.

Like the C-type lectins, the CRD of S—type lectin exhibits

Figure 2.

Summary of the structural features of S-type animal
lectins. Conserved residues found in all of the

family members so far sequenced are shown. In

addition , the extra domain found in CBP35 is shown

schematically.

 

N mismam

 

.....ollamzamnoia V
pp»... z..zm.£o..m.mmla
w
t....32-u_l<-e.....o¢..xuz

wz_<s_oo
10E can.

 

4

2.4200

 

wk<mo>zomm<o
math ..m

zo_._._zooomm lli..|.'

Sea 9. 3-3

265 9. m. - r

10
conservation of certain residues, implying that these residues are
necessary for carbohydrate binding. The 39 conserved amino acid
pattern contains several charged residues. The CRD of S—type lectins
is completely different frOm the CRD of C-type lectins. Unlike the
C-type CRDs, the Cys residues are not conserved in the S-type CRDs.
As indicated earlier, different CRDs of C-type lectins bind to
different saccharides (Fig. 1). In contrast, all of the CRDs of 8—
type lectins thus far identified are specific for galactose/lactose-

containing glycoconjugates.

At present, there has been no definitive identification of the
function of any of the S-type lectins. The structural results,
however, do provide ideas concerning the function of these proteins
that can be tested. This will be the subject of more detailed

discussion in the context of the structure of CBP35.

III CARBOHYDRATE BINDING PROTEIN 35

A: ISOLATION AND CHARACTERIZATION

CBP35 was initially isolated from the extracts of cultured 3T3
fibroblasts by fractionation on Sepharose columns covalently
derivatized with asialofetuin (16). Three CBPs, with molecular
weights of 35 kD, 16 kD and 13.5 kD, respectively, were purified from
the columns. CBP35 wasfurther purified from the other two CBPs by
Sephadex G-150 chromatography. When purified CBP35 was subjected to
sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis
(PAGE) analysis, only one band was observed at 35 kD. CBP35 exhibits
the following features: (a) it has agglutination activity when

assayed with rabbit erythrocytes; (b) it specifically recognizes

11
galactose-containing glycoconjugates; and (c) it requires b-
mercaptoethanol to retain carbohydrate binding activity, which is

a characteristic of the S-type lectins.

B: DISTRIBUTION

Purified CBP35 was injected into rabbits to produce anti-CBP35
antibodies. The antibodies directed against CBP35 did not cross—react
with the 13.5 kD and 16 kD lectins from.3T3 cells. When the extracts
of 3T3 cells were subjected to SDS PAGE and immunoblotting analysis,
only one band in the 35 kD position was observed (23). Using the
antibodies directed against CBP35, the distribution of CBP35 was

studied both in various tissues and in 3T3 cells (21, 23).

CBP35 was found in human, chicken and mouse fibroblasts and in a
macrophage—like cell line, P388D1. In a study of the distribution
of CBP35 in various tissues, both embryonic and adult tissues were
searched for cross-reactivity with anti-CBP35. A cross-reactive
polypeptide of about 35 kD was found in liver, lung, thymus, skin
and muscle of embryonic tissues and.in the lung, artery, thymus, and

spleen of the adult mouse.

In 3T3 cells, the majority (>95%) of the CBP35 is found inside the
cell. Although CBP35 could be detected at the cell surface ( e. g.
by anti-CBP35 antibodies), the interpretation of such results has
not been clear cut. Much of these extracellular CBP35 could be due
to cell lysis and leaking of intracellular lectins, which become
bound to the cell surface. Therefore, most of the attention.has been

focused on intracellular CBP35.

12
The results from immunofluorescence studies with anti-CBP35
antibodies on formaldehyde-fixed and detergent permeabilized 3T3
cells have been particularly interesting. First, the lectin was
predominantly localized in the nucleus of the proliferating 3T3
cells, while in quiescent 3T3 cultures the majority of the CBP35 was
found in the cytoplasm. Second, CBP35 had a distinct punctate
staining in the 3T3 cell nucleus (21). This indicated that CBP35

may be associated with some subnuclear structure.

C: NUCLEAR LOCALIZATION OF CBP35

Immunochemical studies have shown that CBP35 may not be present in
a free form in 3T3 cells but rather may be associated with the
heterogeneous nuclear ribonucleoprotein complex (hnRNP) in the
nucleus (27) and ribonucleOprotein complexes (RNPs) in cytoplasm

(Laing, J., and Wang, J., L. unpublished data).

When unfixed and detergent-permeabilized 3T3 cells were digested with
ribonuclease A (RNase A) and deoxyribonuclease I (DNase I), only
RNase A could release CBP35 from the nucleus. This indicated that
the nuclear CBP35 may be associated with the ribonucleoprotein
fraction in the nuclei. When subnuclear fractions were isolated by
sucrose or cesium sulfate gradient centrifugation, CBP35 was found
in the same fractions as hnRNP. Moreover, when nucleoplasm was
subjected to affinity chromatography on a galactose—Sepharose column,
the bound fraction eluted with lactose contained CBP35 as well as
polypeptides corresponding to those identified in hnRNP ( 27 ). This
was interpreted to indicate that CBP35 in the complex bound to the

column which in turn co-purified the hnRNP components as a complex.

13

Since CBP35 was identified in the hnRNP complex, it is helpful to
summarize some literature<m1 hnRNP, particularly in the structural
study area. In eukaryotic cells, nascent RNA polymerase II
transcripts are rapidly assembled.into ribonucleoprotein structures
called heterogeneous nuclear ribonucleoprotein (hnRNP) complexes
(41). The mammalian cell has a 408 hnRNP complex and is most likely
a key structure in hnRNA processing : splicing, packaging and
transporting out of the nucleus (42). The complex contains at least
20 polypeptides. The protein moiety of hnRNP particle has been
characterized and shown to contain six major polypeptides, called

"core proteins": Al/A2, Bl/B2, C1/C2 (41).

The cDNA sequences of an hnRNP protein in Artemia salina (shrimp)
(GRP~33) (44), the rat A1 protein (45) and human hnRNP proteins A1,
A2, Bl, C1 and C2 (46), have been determined. The hnRNP protein
sequences deduced. from. the cDNA sequences have the following
features: (a) the distribution of Gly and Pro residues in the poly-
peptide chains is nonuniform: for both GRP-33 and rat hnRNP A1 pro-
tein, approximately 77% of the Gly residues are located in the car-
boxyl-terminal portion of the protein while in the human C1 hnRNP
protein, all the Pro residues are found in the amino-terminal half
of the polypeptide chain (47); (b) there are domains containing con-
sensus sequences characteristic of the RNA binding protein family
(47); (c) some hnRNP proteins are almost identical to each other.
For example, A2 and B1 cDNA.are identical except for a 36 nucleotide
in frame insert in B1, while C2 contains an extra 39 in frame nucle0*
tides compared to C1 cDNA. The last feature may indicate that some
of the diversity apparent among the hnRNP proteins is due to alter-

native hnRNP pre-mRNA splicing (46, 48 ).

14

D: REGULATION OF EXPRESSION

From quantitative immunoblotting, it was documented that the level
of CBP35 was 3.5—fold higher in sparse proliferating 3T3 cells than
in quiescent confluent monolayers of the same cells. When quiescent
3T3 cells were stimulated by the addition of serum, the level of
CBP35 increased during the early part of G1, well before the onset
of the S-phase of the cell cycle (21). Consistent with.these results
at the protein level, analysis of the expression of the CBP35 gene
showed that there is elevation of both transcription and accumulation

of mRNA for CBP35 in proliferating cells.

Using Northern blotting and nuclear transcription run-off assays,
it was demonstrated that: (a) the mRNA level was increased as early
as 30 minutes, and reached.a peak at about 3 hours, after the addi—
tion of serum.to stimulate the cells; (b) the rate of transcription
of the CBP35 gene was elevated within 3 hours and reached its maximal
level at about 10 hours after the serum stimulation; and (c) the
elevation of the transcription rate.of the CBP35 gene is not affeced
by the presence of 10 ug/ml cycloheximide, suggesting that the
increased transcription is a primary event of serum stimulation (24) .
Thus , the regulation of the expression of the CBP35 gene appears
to be similar to that of other growth regulated genes, including c—
fos, c-myc and actin genes (25,26). One might predict that there
may be some structural similarity in the regulatory regions of CBP35

and those genes.

c-Fos and actin genes are serum stimulated genes. One mechanism for
regulation of the expression of these genes involves the serum

response element (SRE), which is present in the 5' flanking regions

15
of both genes (39, 40). Using the serum.stimu1ation assay to test
the effects of different deletions of 5' flanking regions of the
human c-fos gene, Treisman defined the SRE to be between nucleotides
—332 to -276 relative to the mRNA cap site (39). A similar SRE was
also found in the actin promoter region, -94 to -75 nucleotides from
the transcriptional start site (40). The SREs identified show the
same consensus sequence: C- C— A- A/T- A— T— A/T- A/T— G- G. Both

SRE sequences appear to bind the same serum response factors (40).

IV IDENTITY BETWEEN CBP35 AND PROTEINS STUDIED UNDER OTHER NAMES

From independent structural determinations, it is clear now that a
protein correSponding to CBP35 has been identified and purified in
several different laboratories. In many of these cases, the basis
for the interest in the protein and for the purification protocol
were totally independent of carbohydratewbinding activity. Thus,
CBP35 exists in the literature under several names. These include:
(a) lectin 34 (L—34) from a mouse fibrosarcoma (29); (b) mouse
macrophage cell surface protein, Mac—2 (15); (c) a.non~integrin type
laminin-binding protein of mouse macrophage (LBP35) (49); and (d)

an IgE-binding protein (eBP) from rat basophilic leukemia cells (36) .

Once the amino acid sequence of each of these proteins became
available, from direct sequencing at the protein level or deduced
from the nucleotide sequence, searches in the sequence data bank
showed them to be almost identical or homologous (e.g. rat versus
mouse) to CBP35. This in turn led to tests of carbohydrate—binding
activity and the results indicated that these proteins also exhibited

galactose-specific binding activity. Thus, it was found that LBP35

16
bound to laminin and eBP bound to IgE via carbohydrate-specific
recognition and that each of these interactions can be blocked by

galactose/lactose.

A: L-34

L-34 was originally isolated from a mouse fibrosarcoma cell line by
galactose-specific affinity chromatography. The lectin has a single
polypeptide chain with a molecular weight of 34, 000. It can recognize

galactose—containing glycoconjugates (14).

A monoclonal antibody directed against L-34 was produced . L-34 was
localized at the fibrosarcoma cell surface by the antibody (50) . It
was found that the monoclonal antibody was capable of inhibiting
colony formation of some tumor cells under both anchorage-independent
and anchorage—dependent growth conditions, suggesting that L-34 might

be involved in regulation of tumor cell growth (50).

A cDNA clone from a mouse fibrosarcoma library was identified by
the anti-L-34 antibodies and the DNA sequence of the clone was
determined“ The clone consisted of 144 5'-f1anking , 792 coding, and
124 3' —-f1anking nucleotides (29) . A comparison of CBP35 and L-34 cDNA
coding regions shows that the CBP35 and L-34 sequences are
identical. However, the 5’ -untranslated regions of the CBP35 and L-34

clones were different; this will be discussed in Chapter IV.

B: Mac-2

Mac-2 is a 32 kD murine macrophage cell surface antigen. It was

originally identified by Ho and Springer (54) . Although found largely

17
in the cytosol, the Mac-2 antigen also appears in the extracellular
mediun: and at the cell surface. The expression of Mac-2 is
heterogeneous among different macrophage populations and in different

macrophage cell lines ( 51).

Several cDNA clones of Mac-2 were identified from a library derived
from mouse macrophage mRNA (15). One clone was expressed.to produce
Mac-2 antigen. It was found that this recombinant protein was able
to bind to an asialofetuin-sepharose column, suggesting that Mac-2
has the same galactose-specific binding activity as CBP35. The cDNA
sequences of Mac—2 clones were determined. For the coding region,
the Mac-2 sequences are identical to CBP35. It is interesting that
different cDNA.clones have the same coding sequences; however, there
are some difference in the 5’—f1anking regions (15). In order to
prove that the difference is not a cloning artifact, polymerase chain
reaction (PCR) was used to synthesize :mouse macrophage cDNA. A 20-
nucleotide oligomer primer designed to hybridize to the mRNA about
130 nucleotides downstream of the ATG starting codon was used to
extend the cDNA towards the S'end, and the reverse direction was
achieved with a 28-nucleotide primer which was complementary to the
poly-G tail added to the 5' end of the mRNA. The cDNAs derived from
the experiment confirmed that there were two species of mRNA in
mouse macrophages. One PCR-derived clone is highly homologous with
the original L—34 clone in the 5'-flanking region. The second cDNA
is identical to the first Mac-2 cDNA except for a 27-bp inframe
insertion. One cDNA sequence may encode for a putative signal peptide
for entry into the endoplasmic recticulm, ultimately transporting
Mac-2 to the cell surface while the other mRNA, without the signal

peptide, may be intracellular Mac~2.

 

18

C: Laminin binding protein (LBP35)

1A laminin-affinity column was used.to fractionate extracts of mouse
macrophages. One protein with molecular weight of 35,000 was bound
to the column and eluted only in high salt. This was designated as
a laminin binding protein (LBP35). When the macrophage was surface
labeled , the 35 kD LBP was the predominant surface protein that

bound to the laminin affinity column (49).

Two peptide fragments of LBP35 were subjected to amino acid seq-
uence analysis. The sequences obtained from the two fragments were
100% identical with CBP35 (49): Ile Val Leu Asp Phe Xxx Xxx Gly
Asn Asp Val Ala Phe Xxx Asn Pro in LBP35 and Ile Val Leu Asp Phe Arg
Arg Gly Asn Asp Val Ala Phe His Asn Pro in CBP35; Ile Gln Val Leu
Val Glu Ala Asp Xxx Phe Lys Val in LBP35 and Ile Gln Val Leu Val Glu
ALa Asp His Phe Lys Val in CBP35. On this basis, the binding of LBP35
to laminin was tested for carbohydrate specificity. It was found that

galactose and lactose inhibited the binding of LBP35 to laminin.
D: IgE binding protein

IgE binding protein (eBP) was isolated from rat basophilic leukemia
cells (RBL) on the basis of its binding to an IgE—containing affini—
ty column (52). The protein has a molecular weight of 31,000.

By immunoblotting of subcellular fractions of rat basophilic
leukemia cells with anti-eBP, the majority of the antigenic
determinant was found within the cell, both in the cytoplasm.and in

the nucleus (53).

The eBP cDNA structure was determined and the amino acid sequence

19
of eBP showed 85% identity with CBP35, including the CRD. eBP can
specifically bind to a galactose affinity column; conversely, CBP35
also can bind to an IgE affinity column (37). The binding of CBP35
and eBP to IgE can be inhibited by galactose. Antibodies directed
against CBP35 and eBP were cross reactive with both proteins. Thus,

eBP may be the rat homolog of CBP35.
‘V STRUCTURE OF 14 kD S-TYPE LECTINS

As stated earlier, the second subfamily of the S-type lectins
consists of 14—16 kD soluble lectins (Fig. 2). A number of cDNA
clones from different species, encoding all or parts of 14 kD
lectins, have been sequenced. These include: chicken skin lectin
(28), mouse fibrosarcoma lectin L—14 (55), bovine heart lectin (30),
electric eel lectin (31) and human lung lectin (32). Some of the
above lectins have been also sequenced by amino acid sequence analy-

sis of peptide fragments (32-33).

A comparison of the sequences of 14-16 kD lectins shows a strong
homology with each other (35) . There are two regions of the sequences
that are highly conserved in all proteins; featured by consensus seq—
uences "His-Phe-Asn-Pro—Arg-Phe" and "Trp—Gly-Xxx—Glu-Xxx-Arg-Glu"
(Fig. 2). The second.consensus sequence, containing a tryptophan and
two glutamic acid residues, corresponds to the proposed galactoside-

binding site of the eel lectin (31).

The 14 kD chicken skin lectin genomic clone was isolated.and.the DNA
sequence was determined (38). The lectin gene. Spanning 4 kb, con—
sists of four exons and three introns. A TATA box, located at ~24

to ~31 nucleotides, and a possible CCAAT box, located at -80 to —86

20
nucleotides 5' to the transcription starting site, were identified.
The polyadenylation signal sequence (AATAAA) was identified in the
3' flanking region. It is interesting that the first exon of the gene
is unusually small, only nine nucleotides in length. There is no con—

sensus sequence for SRE in the 5’ flanking region.
VI GOAL OF THIS THESIS

In this thesis work, one cDNA clone and two genomic DNA clones for
CBP35 have been isolated, identified and characterized. The goal is
to provide the structural basis for: a) the carbohydrate binding
activity of the lectin; b) the nuclear localization property; c) the
serum stimulated regulation of expression; and d) the dual localiza-

tion of CBP35 intracellularly and external to the cell.

REFERENCES

l. Barondes, S.H. (1981) Annu. Rev. Biochem. 50, 207—231.

2. Goldstein, I. J., and Hayes, C. E. (1978) Adv. Carb. Chem.
Biochem. 35, 127-340.

3. Tanabe, T., Pricer, W. E., and Ashwell, G. (1978) J. Bio. Chem.
254, 1038—1043.

4. Kawasaki, T., and Ashwell, G. (1977) J. Biol. Chem. 251, 1296-
1302.

5. Hoflack, B., and Kornfeld, S. (1985) J. Biol. Chem. 260, 12008—
12014.

6. Sahagian, G. G., Distler, J., and Jourdian, G. W. (1981) Proc.

Natl. Acad. Sci. U. S. A. 78, 4289-4293.

21

7. Beyer,_E.C., and Barondes, S. H. (1982) J. Cell Biol. 92, 27-37.
8. Barondes, S. K., and Haywood~Reid, P. (1981) J. Cell Biol. 91,
568-572.

9. Beyer, B., and Barondes, S. H. (1982) J. Cell Biol. 92, 28-33.
10. Oda, Y., and Kasai, K. (1983) Biochem. Biophys. Acta. 761, 237—
245.

11. Childs, R. A., and Feizi, T. (1980) Cell Biol. Int. Rep. 4, 755-
761.

12. Teichberg, V. I., Silman, I., Beitsch, D. D., and Rescheff, G.
(1975) Proc. Natl. Acad. Sci. U.S.A. 72, 1383-1387.

13. Powell, J. T. (1980) Biochem. J. 187, 123—129.

14. Raz, A., Meromsky, L., and Lotan, R. (1986) Cancer Res. 46, 3667-
3672.

15. Cherayil, B. J., Weiner, S. J., and Piliai, S. (1989) J. Exp.
Med. 170, 1959-1972.

16. Roff, C. F., and Wang, J. L. (1983) J. Biochem. Chem. 258, 10657-
10663.

17. Barondes, S. H. (1984) Science 233, 1259-1264.

18. Drickamer, K. (1988) J. Biol. Chem. 263, 9557-9560.

19. Briles, E. B., Gregory, W., Flettcher, P., and Kornfeld, S.
(1979) J. Cell Biol. 81, 528-537.

20. Drickamer, K. (1987) Kidney Int. 32, $167—$180.

21. Moutsatsos, I, K., Wade, M., Schindler, M., and Wang, J. L.
(1987) Proc. Natl. Acad. Sci. U.S.A. 84, 6452—6456.

22. Raz, A., and.Lotan, R. (1987) Cancer Metastasis Rev. 6, 433-452.
23. Crittenden, S. L., Roff, A. F., and Wang, J. L. (1984) Mol. Cell.
Biol. 4, 1252-1259.

24. Agrwal, N., Wang, J. L., and Voss, P. G. (1989) J. Biol. Chem.

264, 17236-17242.

22

25. Lau, L. F., and Nathans, D. (1985) EMBO J. 4, 3145-3151.

26. Mohun, T., Garrett, N., and Trieisman, R. (1987) EMBO J. 6, 667—
673.

27. Laing, J. G., and Wang, J. L. (1988) Biochem. 27, 5329-5334.
28. Ohyama, Y., Hirabayashi, J., Oda, Y., Ohno, S., Kawasak, H.,
Suzuki, K., and Kasai, K. (1986) Biophys. Biochem. Res. Commun. 134,
51—56.

29. Raz, A., Pazerini, G., and Carmi, P. (1989) Cancer Res. 49, 3489-
3493.

30. Southan, C., Aitken, A., Childs, R. A., Abbott, W. M, and FeiZi,
T. FEBS Lett. 214, 301-304.

31. Paroutoud, P., Levi, G., Teichberg, V. I., and Strosberg, A. D.
(1987) Proc. Natl. Acad. Sci. U. S. A. 84, 6345-6348.

32. Gitt, M. A., and Barondes, S. H. (1986) Proc. Natl. Acad. Sci.
U. S. A. 83, 7603—7607.

33. Hirabayashi, J., Kawasaki, H., Suzuki, K., and Kasai, K. (1987)
J. Biochem. 101, 775-787.

34. Abbott, W. M., and Feizi, T. (1989) Biochem. J. 259, 291-294.
35. Jia. S., and Wang, J. L. (1988) J. Biol. Chem. 263, 6009-6011.
36. Albrandt, K., Orida, N. K., and Liu, F.-T. (1987) Proc. Natl.
Acad. Sci. U. S. A. 84, 6858-6863.

37. Laing, J., Robertson, M. W., Gritzmacher, C. A., Wang, J. L.,
and Liu, F.~T. (19890 J. Biol. Chem. 264, 1907-1910.

38. Ohyama, Y., and Kasai, K. (1988) J. Biochem. 104, 173-177.

39. Treisman, R. C. (1986) Cell 46, 567-574.

40. Gustafson, T. A., Miwa, T., Boxer, L. M., and Kedes, L. (1988)
Mol. Cell. Biol. 8, 4110-4119.

41. Dreyfuss, G. (1986) Ann. Rev. Cell Biol. 2, 459-498.

42 Choi, Y. D., Grabowski, P. J., Sharp, P. A., And Dreyfuss, G.

23
(1986) Science 231, 1534-1539.
44. Cruz-Alvarez, M., and Pellicer, A. (1987) J. Biol. Chem. 262,
1377-1380.
45. Cobianchi, F., Sengupta, D. W., Zmudzka, B. 2., and Wilson, 8.
H. (1986) J. Biol. Chem. 261, 3536-3543.
46. Burd, C. G., Swanson, M. S., Gorlach, M., and Dreyfuss, G. (1989)
Proc. Natl. Aced. Sci. U. S. A. 86, 9788—9792.
47. Swanson, S., Nakagawa, T. Y., LeVan, K., and Dreyfuss, G. (1987)
Mol. Cell. Biol. 7, 1731-1739.
48. Biamonti, G., Buvoli, M., Bussi, M. T., Morandi, C., Cobianchi,
P., and Riva, S. (1989) J. Mol. Biol. 207, 491—503..
49. Woo, H.-J., Shaw, L. M., Messies, J. M., and Mercurio, A. M.
(1990) J. Biol. Chem. 265, 7097-7099.
50. Lotan, R., Lotan, D., and Raz, A. (1985) Cancer Res. 45, 4349—
4353.
51. Nibbering, P. B., Leijh, P. C. J., and Furth, R. van. (1987) Cell
Immunol. 105, 374-380.
52. Liu, F.-T., and Orida, N. (1984) J. Biol. Chem. 259, 10649-10652.
53. Gritzmacher, C. A., Robertson, M. W., and Liu, F.-T. (1988) J.
Immunol. 141, 2801—2806.
54. Ho, M. K., and Springer, T. A. (1982) J. Immunol. 128, 1221-1230.
55. Raz, A., Carmi, P., and Pazerini, G. (1988) Cancer Res. 48, 645—

649.

24

CHAPTER II

Carbohydrate binding protein 35: molecular cloning and
expression of a recombinant polypeptide with lectin

activity in Escherichia coli

(recombinant DNA; cDNA clone; lambda gt vector; fusion
protein; immunoblot screening; peptide mapping; Northern
blotting)

Shizhe Jia, Robert P. Mee, Gerald Morford, Neera Agrwal,

Patricia G. Voss, Ioannis K. Moutsatsos and John L. Wang

Department of Biochemistry
Michigan State University

East Lansing, MI 48824

(This chapter has been published in Gene 60, 197—204 (1987) ; It has

been permitted for reprint from publisher)

25

SUMMARY

Affinity purified antibodies directed against carbohydrate binding
protein 35 (CBP35) , a galactose-specific lectin, were used to screen
a lambda gtll expression library derived from mRNA of 3T3 fibro-
blasts. This screening yielded several putative clones containing
cDNA for CBP35, one of which was characterized in terms of its
expression.of a fusion protein containing b~galactosidase and.CBP35
sequences. Limited proteolysis of lysates containing the fusion
protein, followed by SDS—polyacrylamide gel electrophoresis and
immunoblotting with anti-CBP35, yielded a peptide mapping pattern
comparable to that obtained from parallel treatment of authentic
CBP35. Such a limited proteolysis followed by affinity chroma—
tography on a column of Sepharose derivatized with galactose also
yielded a 30-kDa polypeptide that exhibited carbohydrate-binding
activity. This polypeptide can be immunoblotted with anti-CBP35, but
not with antibodies directed against b—galacto-sidase. These
results indicate that we have identified a cDNA clone for CBP35 that
yields a recombinant polypeptide with lectin activity produced in
E. coli. Using this cDNA clone as a probe, Northern blot analysis
showed an increased.expression of the CBP35 gene when quiescent. 3T3

cells were activated by the addition of serum growth factors.

26

INTRODUCT ION

Carbohydrate binding protein 35 (CBP35; 35-kDa) is a lectin that
binds specifically to galactoseandgalactose-N-acetylglucosamine
containing glycoconjugates (Roff and Wang, 1983; Roff et al., 1983) .
Using a highly specific antibody directed against CBP35, immuno—
blotting and immunofluorescence studies have shown that the lectin
can be found in the nucleus of a cell (Moutsatsos et al., 1986).
When serum-starved, quiescent cultures of 3T3 cells were activated
by the addition of serum, there was an elevation of the expression
of CBP35 (detected by immunoblotting) and an apparent translocation
into the nucleus (assayed by immunofluorescence) prior to the onset
of DNA synthesis in these synchronized cells (Moutsatsos et al.,
1987). Because of these interesting distributional properties of
CBP35, we have undertaken molecular cloning studies to obtain a
reliable and abundant source of material amenable for our purifica-
tion scheme, to develop a probe for studying the expression of the
CBP35 gene, and ultimately, to obtain information on the primary
structure of the protein. In this communication, we report the
identification of a cDNA clone for CBP35 from a lambda gtll expre-
ssion library. This clone expresses a fusion protein which, upon
proteolysis and affinity fractionation, yields a polypeptide that

exhibits carbohydrate-binding activity.

27
MATERIALS AND METHODS
(a) Affinity purification of antibodies against CBP35

CBP35 was purified from mouse lung by affinity chromatography on
Sepharose columns derivatized with N-e—aminocaproyl-D-galactosamine
(Allen and Neuberger, 1975). The generation and characterization
of the speci—ficity of rabbit antiserumudirected against CBP35 have
been reported previously (Roff and Wang, 1983; Crittenden et al.,
1984: Moutsatsos et al., 1986). Antibodies directed against CBP35
were affinity purified by specific adsorption to CBP35 on nitro-
cellulose membranes following theSmith and Fisher procedure (1984) .
CBP35 ( 20 ug) was subjected to SDS-PAGE (Laemmli, 1970). After
electrophoresis, the protein was transferred to a sheet of nitro—
cellulose membrane (200 mA, 3 h, 25°C) (Towbin et al., 1979).
Strips of nitrocellulose membrane at the left and right edges of
the sheet were excised and immunoblotted with rabbit antiserum
directed against CBP35 (1:250 dilution, 3 h, 259C), followed by
HRP-conjugated goat anti—rabbit immunoglobulin (Bio—Rad Lab—
oratories, 1:2000 dilution, 1 h, 25°C) . The immunoreactive material
was revealed via HRP activity with 4vchloro-1-naphthol and hydrogen
peroxide as substrates. Using the positions of migration of CBP35
on these edge strips as guides, a horizontal strip of nitro-
cellulose membrane, corresponding to the: CBP35 band and.designated
as the CBP35 strip, was excised from the preparative nitrocellulose
sheet. This CBP35 strip served as the affinity adsorbent for the
purification of anti-CBP35 antibodies. Rabbit antiserum directed
against CBP35 was diluted (1:150) in 20 ml Tris buffered saline (20

mM Tris, 0.5 M NaCl, pH 7.5)‘ containing 0.5% Tween 20 and was

28

incubated with the CBP35 strip overnight at 4°C. The supernatant
was saved and reused (see below). The CBP35 strip was washed for
5 min in Tris buffered saline. Antibodies bound to the CBP35 on
this strip were eluted by the addition of 500 ul of 5 mM glycine,
0.5 M NaCl, 100 ug/ml bovine serum.albumin, 0.5% Tween 20, pH 2.3.
This solution was transferred to a tube containing 25 ul of 1 M
NaZHPO4 to neutralize the {TL The elution was repeated with an
additional aliquot of 500 ul of elution solu-tion; this second
aliquot was neutralized and combined with the first elution. This
sample was designated as affinity purified anti-CBP35. The CBP35
strip can also be reused (see below).The. supernatant solution
containing the diluted antiserum (1:150) was "boosted" by the
addition of 50 ul of fresh antiserum. This was then reincubated
with the CBP35 strip overnight at 4°C and the whole procedure
repeated“ We have found that the CBP35 strip can be used for 10-12
adsorption-elution cycles. This conclusion is based on comparing
the titers of the affinity purified anti—CBP35 from the various
cycles by immunoblotting individual nitrocellulose strips containing

a known amount of CBP35.
(b) Screening of the lambda gtll expression library

The lambda gtll cDNA expression library (Clone Tech Laboratories)
was prepared using mRNA from mouse 3T3 fibroblasts; it contains
about 7.3 x 105 independent clones, with an average insert size of
0.94 kb. E. coli (Y1090) were infected with the lambda phage and
grown on agar plates (2 x 105 plaques-lplate) (Young and Davis,
1983). Six positive plaque colonies were identifed and rescreened
by immunoblotting with affinity-purified anti-CBP35. The clones

were isolated and amplified twice. The recombinant lambda gtll DNA

29

was isolated from the positive clones by the method described in

Maniatis et al., 1982.

(0) Isolation of recombinant polypeptide with lectin activity from

fusion protein of clone 1

The procedures for detecting expression of a fusion protein followed
those described previously (Huynh.et al., 1984). The fusion protein
was detected by immunoblotting using either: (a) affinity-purified
rabbit anti-CBP35 plus HRP-conjugated goat anti—rabbit immunoglo—
bulin, or (b) mouse anti-E. coli b-galase (Accurate Chemical
Scientific Corp) plus HRP-conjugated goat anti-mouse immuno-
globulin. Iysates containing fusion protein were digested with
protease S. aureus V-8 (Miles Scientific) (Cleveland et al., 1977).
In parallel, authentic CBP35 (Crittenden et a1. , 1984) was digested.
The samples were then loaded on SDS—PAGE. The peptide maps were

revealed by immunoblotting with affinity-purified anti-CBP35.

E. coli lysogen cells were suspended in TE buffer (125 mM Tris—HCl,
1 mM EDTA, pH 6.8) containing 2 mM b-mercaptoethanol and were lysed
by freezing and thawing. The lysate was treated with deoxyribo-
nuclease I (Sigma; 0.2 mg/ml) and was then centrifuged at 13,000 x
g for 2 minutes at 4°C. The pellet fraction was incubated at 37°C
for 30 minutes with: (a) TE buffer; (b) TE containing 0.5 M lac—
tose; (c) TE containing 1% Triton X-100; (d) TE containing ribo-
nuclease A (Sigma; 10 ug/ml); (e) TE containing 5 M urea; (f) TE
containing 8 M urea; and (g) TE containing 0.1% SDS. The mixtures
were then centrifuged at 14,000 x g for 2 minutes and the super-

natant and pellet fractions were analyzed by SDS-PAGE and immuno-

30

blotting with anti-CBP35.

To test material derived from the fusion protein for carbohydrate-
binding activity, the lysates of E. coli lysogen were extracted for
30 minutes at 37°C with TE buffer containing 0.1% SDS and 2 mM
b—mercaptoethanol. After centrifugation at 13,000 x g for 10
minutes, the pellets were reextracteds Protease V—8 was added.to the
combined supernatants (total protein concentration 20 mg/ml) to a
concentration of 0.1 mg/ml. The mixture was incubated at 37°C for
30 minutes, after which the solution was made 1 mM in PMSF and 1%
in Triton X—100. This material was dialyzed at 4°C for overnight
against buffer A (75 mM Tris-H01, 50 mM CaCl2, 2 mM b-mercapto-
ethanol, 1 mM PMSF, 1 mM NaNg, pB 7.2) containing 1% Triton X-100.
After dialysis, the mixture was incubated overnight at 4°C with 2
ml of Sepharose derivatized with N-e-aminocaproyl-D-galactosamine
(Allen and Neuberger, 1975). The beads were then packed into a
column (1.4 x 2.8 cm), washed with 15 ml buffer A containing 1%
Triton X-100, and 30 ml of buffer A without Triton X-100 (until the
absorbance (280 nm) was less than 0.003). The column was eluted
with 15 ml of buffer A containing 0.2 M lactose and the effluent

material was analyzed by SDS-PAGE and immunoblotting.

(d) Northern blotting analysis

The cDNA insert of clone 1 was derivatized for detection using the
BluGENE system (Bethesda Research Laboratories) by nick translation
with biotin—7-dATP. Colorimetric detection of the hybridized probe
was accomplished by the addition of strep-tavidin-alkaline
phosphatase conjugate and incubation in the substrates nitroblue

tetrazolium and 5-bromo-4-chloro-3-indolyl-phosphate (Amorese,

31
1986). Serumrstarved, quiescent 3T3 cells were synchronously
reactivated by the addition of serum (Moutsatsos et al., 1987); the
poly A+ fraction of mRNA derived from quiescent 3T3 cells and cells
at various times post serum addition were isolated, electrophoresed,
transferred onto nitrocellulose, and probed with the clone 1 cDNA
following previously described procedures (Stuart et al., 1985).
A cDNA clone for mouse b-2-microglobulin (Stuart et al., 1985) was
similarly derivatized for detection by the BluGBNE system and this
probe was used as an internal control of the amount of RNA electro—

phoresed in each lane of the gel.

32

RESULTS AND DISCUSSION
(a) Identification of a cDNA clone for CBP35

Using affinity purified anti-CBP35, we have screened a lambda gtll
expression library of 3T3 cell cDNA” Sinpositive clones were iden—
tified and rescreened. The clones were then isolated and amplified
twice. The purity of one of the clones, designated as clone 1, was
confirmed upon secondary rescreening, in vfldrxxall.plaques yielded
positive reactions (Fig. 1a). The use of normal rabbit immunoglo-
bulin failed to yield any positive reaction. In addition, parallel
screening of E. coli infected with gtll phage (no recombinants) did
not show positive reaction with anti-CBP35. These results suggest
that lambda gtll clone 1 contained cDNA insert that coded for seq—
uences specifically immunoreactive with affinity purified anti—

CBP35.

DNA from clone 1 was digested with the restriction endonuclease
EcoRl to excise the cDNA insert. The digest was analyzed by agarose
gel electrophoresis. Based on this analysis, the size of the cDNA

insert was estimated to be approximately 900 base pairs (Fig. 1b).

(b) Expression of a fusion protein encoded by lambda gtll clone 1

DNA

The 3T3 cell cDNA is inserted near the 3'-end of the b-galase gene

in the lambda gt ll cloning vector (Young and Davis, 1983). The

33

Figure 1: (a) Specific binding' of affinity purified rabbit

anti-CBP35 to proteins of lambda gtll clone 1. This clone, positive

upon initial screening of a gtll expression library of mouse 3T3

fibroblast CDNA, was purified and amplified twice. E. coli (Y1090)

were infected with the lambda gtll clone 1 phage and grown on agar:

a nitrocellulose filter, saturated with IPTG, was overlaid on the

37° C. It was then

'-CBP35 (1:50 dilution, 3

h 250 C) followed by HRP-conjugated goat anti-rabbit immunoglobulin.

Lane 1: DNA digest from gtll clone 1: Lane 2: DNA (HindIII
digest) and phiX174 DNA (HaeIII digest) DNA size Harkers. The
position of migration of the cDNA insert is indicated by an

arrowhead and is estimated to be 940 base pairs.

 

34

Figure 1

 

35

fusion protein encoded by the recombinant lambda gtll DNA would
therefore be composed of b—galase interrupted by the 3T3 cell pro—
tein 18 amino acids from the carboxyl terminus. The expression. of
this fusion protein was examined by infecting E. coli (Y1089) with
lambda gtll clone 1 phage. Lysogens containing lambda gtll clone
1 DNA were selected on the basis of their growth at 32°C and non-
growth at 42°C. After induction with IPTG, the cells were harve-
sted, lysed, and the lysate was analyzed by SDS—PAGE. The lysate
of lysogen l, immunoblotted with anti-b-galase, yielded two major
two bands (153 and 150 kD) (Fig. 2a, lane 2). For comparison, E.
coli b-galase was also subjected to SDS-PAGE and immunoblotting with
anti—b-galase antibodies. This yielded a prominent band at a posi-
tion corresponding to 116 kD, the subunit molecular weight of b—
galase (Fig. 2a, lane 1). These results suggest that in the lysogen
1 lysate, the bands at 153 kDa and 150 kDa represent fusion protein
components. This conclusion is supported by immunoblotting of the

same bands in lysogen 1 lysate with anti-CBP35 (Fig. 2b, lane 4).

Control experiments were carried out to ascertain the specificity
of the immunoblotting reactions. First, anti—CBP35 immunoblots only
CBP35 (Fig. 2b, lane 3) but not b-galase (Fig. 2b, lane 1). Second,
when lysates of lysogen 1 were prepared in the absence of IPTG,
anti—CBP35 failed to detect any polypeptide bands in the immuno-
blots (Fig. 2b, lane 2). All of these results strongly suggest that
the 153 kD and 150 kD bands (Fig. 2a, lane 2 and Fig. 2b, lane 4)
represent fusion protein components containing both b-galase and

CBP35 sequences recognized by the respective antibodies.

(0) Peptide maps derived from fusion proteins and CBP35

36

Figure 2: Detection of expression of fusion protein by

immunoblotting analysis. E. coli (Y1089) was infected with lambda

gtll clone 1 phage. Lysogens, selected on the basis of their growth

at 32°C and nongrowth at 42°C, were incubated at 32°C to an

absorbance value (595 nm) Of 0.5; The temperature was shifted to

45°C for 20 minutes, then IPTG was added to 10 mM. After further

incubation at 38° C for 1 hour, the cells were harvested in TE

buffer containing 0.1% SDS, lysed by freezing and thawing, and

subjected to SDS-PAGE. (a) Immunoblot analysis with mouse anti-E-

coli b-galase. Lane 1: (1) E.

coli b-galase; Lane 2: lysate 0f

lysogen 1 derived from lambda gtll clone 1 induced with IPTG. (b)
Immunoblot analysis with affinity purified rabbit anti-CBP35. Lane

1: E. coli b-galase; Lane 2: lysate of lysogen 1 derived from.lambda

gtll clone 1 without IPTG induction; Lane 3:

authentic CBP35; Lane
4: lysate of lysogen 1 derived from lambda gt11 clone 1 induced with

 

37

 

4200

< 116
< 66
< 45

 

Figure 2

38

Lysates of lysogen 1 were digested with V—8 protease at 37°C for 30
min and then subjected to SDS-PAGE and immunoblotting with
anti-CBP35. This yielded two predominant bands: 33 kD and 30 kD
(Fig. 3, lane 2). These bands correspond closely to products of V—8
digestion of authentic CBP35 subjected.to SDS-PAGE and.immunoblott-
ing in parallel (Fig. 3, lane 3). Neither of these two ibands could
be immunoblotted with antibodies directed against b-galase. More-
over, analysis of the V-8 digestion products of authentic b—galase
by SDS-PAGE and immunoblotting with anti-b—galase did not yield the
peptide fragments observed.in Figure 3. These results suggest that
the cDNA insert of the lambda gtll clone 1 contains the CBP35 coding

sequence.

This conclusion is supported.by results of our preliminary sequence
analysis of clone 1” Three portions of the nucleotide sequence
yielded the following deduced amino acid sequences: (i)Gly-Asn-Asp-—
Val-Ala-Phe—His-Phe-Asn-Pro-Argv (ii) Gln—Asp-Asn-Asn-Trp—Gly-Lys;
and (iii) Gln—Ser-Ala-Phe-Pro- Phe-Glu-Ser-Gly-Lys, corresponding
to three tryptic peptides. These sequences show extensive homology
to those of other b—galactoside binding lectins, including bovine
heart lectin (Southan. et al., 1987) and chicken skin lectin
(Hirabayashi et al., 1987). These results confirm that cDNA clone

1 codes for CBP35.
(d) Isolation of recombinant CBP35 from fusion protein

Several experiments were performed to test the possibility that V—8
digestion of the fusion protein.may generate a recombinant 3T3 cell
protein that retains the carbohydrate-binding capacity of CBP35.

To test for carbohydrate-binding activity; it was important to

39

Figure 3: Comparison of the peptide maps derived from the fusion

protein of clone 1 and CBP35. Lysates containing fusion protein

(protein concentration 20 mg/ml) were digested with v-8 protease

(0.1 mg/ml, 37°C, 30 minutes). Purified CBP35 was digested in

parallel. The samples were subjected to SDS-PAGE and immunoblott-

ing analysis with affinity purified anti—CBP35. Lane 1: lysate 0f

lysogen 1 derived from lambda gtll clone 1; Lane 2: lysate Of

lysogen 1 derived from lambda gtll clone 1 after digestion with V-

protease; Lane 3:

8

CBP35 digested with V—8 protease; and Lane 4:

authentic CBP35.

 

40

Figure 3

 

 

41

retain the protein in native form.or to be able to renature it from
buffers containing denaturants. Although the fusion protein was
soluble in lysates of lysogenl prepared in SDS-PAGE buffer, we wish-
ed to ascertain its solubility in aqueous buffers without de-
naturants. Lysates of lysogen 1 were prepared in various buffers
and the extraction mixtures were centrifuged and separated into the
supernatant and pellet fractions. These fractions were then analyz-
ed for fusion protein by SDS-PAGE and immunoblotting with anti-
CBP35. I“; found that the fusion potein could be solubilized
(detected in the supernatant fraction) in TE buffers containing 0.1%
SDS or'8 M urea. In contrast, TE buffers containing no denaturants,
or TE buffers containing non-ionic detergents (1% Triton X—100) or
even low concentrations of urea (5 M) failed to solubilize the fu-
sion proteins Finally, TE buffers containing 10 ug/ml ribonuclease

also did not solubilize the fusion protein.

Lysates of lysogen l were prepared in TE buffer containing 0.1% SDS
and 2 mM b—mercaptoethanol. The supernatant fraction was digested
with V-8 protease (0.5 h at 37°C), dialyzed versus Tris buffer con-
taining 1% Triton X—100 (but without SDS), and then fractionated
over an affinity column of Sepharose derivatized with N-e-amino—
caproyl—D-galactosamine. The bound material was eluted with 0.2
M lactose and analyzed by SDS-PAGE and immunoblotting with anti-
CBP35. V-8 digestion of the fusion protein yielded two polypeptides
(33 auxi 30 kDa) that reacted with anti-CBP35 (Fig. 4, lane 2).
After purification over the affinity column, however, only the 30
kDa polypeptide was detected by immunoblotting with anti-CBP35 (Fig.
4, lane 3). The position of migration of this polypeptide closely
matched that of one of the fragments derived from CBP35 after V-8

digestion (see Fig. 4, lane 3). When antibodies directed against

42

Figure 4: Immunoblot blot analysis of recombinant polypeptide with

lectin activity using affinity purified rabbit anti-CBP35. Lane 1:

authentic CBP35; Lane 2: lysate of lysogen 1 derived from lambda

gtll clone 1 after digestion with V—
37°C);

8 protease (0.1 mg/ml, 30 min,

Lane 3: material that was eluted with 0.2 M lactose when

lysate of lysogen 1 derived from lambda gtll clone 1 was digested

with V-8 protease and then purified by affinity chromatography on-

a column (1.4 X 2.8 cm) of Sepharose derivatized with

N-e—aminocaproyl-D—galactosamine.

 

 

43

 

 

Figure 4

 

 

44
b-galase were used to immunoblot this material, there was no reac-
tiond Thus, it appears that the 30 kDa.polypeptide contained little
or no b-galase sequences. These results strongly suggest that the
3T3 cell polypeptide contained within the fusion protein carried

carbohydrate-binding activity.

The fact that the 3T3 cell portion of the fusion protein, after V—8
digestion, can be isolated by galactose—Sepharose affinity chromato—
graphy suggests that we have quite fortuitously come upon a way to
purify the eukaryotic polypeptide synthesized in E. coli. Consis-
tent with many reports on the production of eukaryotic polypeptides
in E. coli (Marston, 1986), our recombinant polypeptide accumulates
intracellularly, but in an aggregated and insoluble form. This
fact determines, in part, the strategy for purification. On the one
hand, the aggregation of recombinant protein in a discrete form, as
dense inclusion bodies (Prouty et al., 1975), allows us to use low
speed centrifugation ( 13,000 x g) to separate it from most of the
bacterial cell debris. On the other hand, the recombinant fusion
protein has to be solubilized, from the aggregate by the use of
strong denaturants such as SDS or urea. Once the fusion protein was
solubilized, we then had to liberate the recombinant 3T3 cell
polypeptide from the fusion protein. In previous reports, various
molecular engineering strategies were devised to place a cleavage
site between the carboxyl terminus of the prokaryotic sequence and
the amino terminus of the eukaryotic coding sequence (Marston,
1986). In the case of the lambda gtll cloning vector, the juncture
of the b-galactosidase sequence and the cDNA insert is through an
EcoRI cleavage site, which contains a glutamic acid residue. V-8
protease cleaves on the carboxyl terminal side of glutamic acid

residues in polypeptides (Houmard and Drapeau, 1972) . In our fusion

45
protein, it appears that this site was indeed susceptible to
cleavage by the enzyme since antibodies against b-galase failed to
immunoblot the carbohydrate-binding 30 kD polypeptide. As a result,
lambda gtll clone 1 appears to be a good source of CBP35 with
lectin activity amenable for our purification scheme. The question
is then raised whether V-8 digestion and subsequent purification can
be applied to other eukaryotic proteins cloned in lambda gtll

expression vectors.

(e) Northern blot analysis of mRNA encoding CBP35

In previous studies, we had shown that when quiescent, serumrstarved
3T3 cells were stimulated by the addition of serum, there was an
increase in the level of CBP35 as quantitated by immunoblotting
analysis (Moutsatsos et al., 1987). This increase occurred within
8 hours after serum addition, well before the synchronized cells
entered the first period of DNA synthesis (between 16-24 hours after
serum.addition). The availability of a cDNA probe for CBP35 provid—
ed the opportunity to quantitate the expression of the CBP35 gene

at the level of mRNA.

Northern blot analysis was carried out on the poly A+ fraction of
mRNA derived from quiescent 3T3 cells and from cells at 8, 16, and
24 hours after serum stimulation. The cDNA probe for CBP35 revealed
a mRNA species of about 1.3 kb that is absent or present at very low
levels in quiescent cells and increases as a function of time after
serum addition (Fig. 5A). The level of mRNA for CBP35 increased
within 8 hours after stimulation. The same blot was subjected to
a probe for b-2-microglobulin (Fig. SB), whose mRNA level does not

vary substantially during the cell cycle (Kelley et al., 1983).

46

Figure 5: Northern blot analysis of the poly A+ fraction of mRNA
derived from: (1) serum—starved, quiescent 3T3 fibroblasts; (2) 3
hours, (3) 16 hours, and (4) 24 hours after the addition of serum
(10%). (A) The blot was probed with the cDNA of clone 1. The
numbers on the right indicate the positions of migration of size
markers (in kb). (B) The same blot was subjected.to a second probe,

b-2—microglobulin, as an internal control of the amount of RNA

electrophoresed in each lane.

 

 

47

 

 

Figure 5

48
Lane 2 in Fig. 53 showed a less intense b-Z-microglobulin signal,

suggesting that a less amount of total RNA was loaded, probably due
to experimental errors. These results corroborate at the mRNA level
previous immunoblotting and immunofluorescence observations on the
proliferation dependent expression of CBP35 (Moutsatsos et al.,

1987).

The elevated level of the CBP35 mRNA after serum stimulation was
observed even in the presence of cycloheximide. This suggests that
the stimulation of expression of the CBP35 gene is a ”primary“ event
of growth factor addition, without the requirement for protein
synthesis. Moreover, the data suggest that the CBP35 mRNA was
"super-induced” (higher level of mRNA in the presence of
cycloheximide than in its absence) in a fashion similar to that
observed for the oncogenes c-myc and c-fos (Lau and Nathans, 1987) .
The possible explanation is that the mRNA for CBP35 is stabilized

by the presence of cycloheximide.

Finally, the fact that the size of the cDNA insert of clone 1 ( 1
kb) approximates the size of the mRNA detected on Northern blots
(1. 3 kb) suggests that the cDNA clone we have identified is nearly
full length. This is consistent with the conclusion that the cDNA
clone codes for sufficient structural information for carbohydrate

binding activity.

 

 

 

49

ACKNOWLEDGMENTS

We thank Drs. Sue Conrad and.Moriko Ito for their advice throughout
these studies and Mrs. Linda Lang for her help in the preparation
of the manuscript. This work was supported by grants GM—27203 and
GM—32310 from the National Institutes of Health. J.L. Wang was
supported by Faculty Research Award FRA-221 from the American Cancer

Society.

50

REFERENCES

Allen, A.K. and Neuberger, A.: A simple method for the preparation
of an affinity absorbent for soybean agglutinin using galactosamine
and CH-Sepharose. FEBS (Fed. Eur. Biochem. Soc.) Lett. 50 (1975)
362-364.

Amorese, D. :Nonradioactive detection of RNA using the BluGENE

system. Focus 8 (1986) 6—8.

Cleveland, D., Fischer, 8., Kirschner, M. and Laemmli, U. :Peptide
mapping by limited proteolysis in sodium dodecyl sulfate and
analysis by gel electrophoresis. J. Biol. Chem" 252 (1977).
1102-1106.

Crittenden, S.L., Roff, C.F. and Wang, J.L. :Carbohydrate-Binding
Protein 35: Identification of the galactose-specific lectin in

various tissue of mice. Mol. Cell. Biol. 4 (1984) 1252-1259.

Hirabayashi, J., Kawasaki, H., Suzuki, K. and Kawai, K. :Complete
amino acid sequence of 14 kDa b-galactoside-binding lectin of chick

embryo. J. Biochem. (Tokyo) 101 (1987) 775—787.

Houmard, J. and Drapeau, G. :Staphylococcal protease: a proteolytic
enzyme specific for glutamoyl bonds. Proc. Natl. Acad. Sci. U.S.A.

69 (1972) 3506-3509.

51
Huynh, T., Young, R. and Davis, R. :Constructing and screening cDNA
libraries 1J1 lambda gth and gtll. In Glover, D. (Ed.), DNA
cloning techniques: a practical approach. IRL Press, Oxford, 1984,

pp. 49—78.

Kelley, K., Cochran, B.H., Stiles, C.D. and.Leder, P.: Cell—specific
regulation of the c-myc gene by lymphocyte mitogens and

platelet-derived growth factor. Cell 35 (1983) 603-610.

Laemmli, U.K. :Cleavage of structural proteins during the assembly

of the head.of'bacteriophage»T4. iNature (Lond.) 227 (1970) 680-685.

Lau, K.F. and Nathans, D. :Expression of a set of growth related
immediate early genes in Balb/c 3T3 cells: coordinate regulation
with c-fos or oemyc. Proc. Natl. Acadn Sci. USA 84 (1987)

1182—1186.

Maniatis, T., Fristch, E.F. and Sambrook, J. :Molecular Cloning.

Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1982.

Marston, F. :The purification of eukaryotic polypeptides synthesized

in Escherichia coli. Biochem. J. 240 (1986) 1—12.

Moutsatsos, I.K., Davis, J.M. and Wang, J.L. :Endogenous lectins
from cultured cells: subcellular localization of Carbohydrate-
binding protein 35 in 3T3 fibroblasts. J. Cell Biol. 102 (1986)

477-483.

Moutsatsos, I.K., Wade M., Schindler, M. and.Wang, J.L. :Endogenous

lectins from cultured cells: nuclear localization of Carbohy-

52
drate-binding protein 35 in proliferating 3T3 fibroblasts. Proc.

Natl. Acad. Sci. U.S.A. 84 (1987) 6452-6456.

Prouty, W.F., Karnovsky, M.J. and Goldberg, H.L. :Degradation of
abnormal. proteins in Escherichia. coli” Formation. of jprotein
inclusions in cells exposed to amino acid analogs. J. Biol. Chem.

250 (1975) 1112—1122.

Roff, C.F., Rosevear, P.R., Wang, J.L. and Barker, R. : Iden—
tification of Carbohydrate—binding proteins from mouse and human

fibroblasts. Biochem. J. 211 (1983) 625—629.

Roff, C.F. and Wang, J.L. :Endogenous lectins from cultured cells.
Isolation and characterization of Carbohydrate—binding proteins from

3T3 fibroblasts. J. Biol. Chem. 258 (1983) 10657-10663.

Smith, D.E. and Fisher, P.A. :Identification, developmental
regulation, and response to heat shock of two antigenically related

forms of a major nuclear envelope protein in Drosophila embryos:

application of an improved method for affinity purification of
antibodies using polypeptides immobilized on nitrocellulose blots.

J. Cell Biol. 99 (1984) 20-28.

Southan, C., Aitken, A., Childs, R.A., Abbott, W.M. and Frizi, T.
: Amino acid sequence of {-galactoside-binding bovine heart lectin.
member of a novel class of vertebrate proteins. FEBS (Fed. Eur.

Biochem. Soc.) Lett. 214 (1987) 301-304.

53
Stuart, P., Ito, M., Stewart, C. and Conrad, S.E. :Induction of

cellular thymidine kinase occurs at the mRNA level. Mol. Cell Biol.

5 (1985) 1490-1497 .

Towbin, B., Staehelin, T. and Gordon, J. :Electrophoretic transfer
of proteins from polyacrylamide gels to nitrocellulose sheets:
procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76

(1979) 4350-4354 .

Young, R. and Davis, R. :Efficient isolation of genes by using

antibody probes. Proc. Natl. Acad. Sci. U.S.A. 80 (1983) 1194—1198.

54

CHAPTER III

CARBOHYDRATE BINDING PROTEIN 35: COMPLEMENTARY DNA
SEQUENCE REVEALS HOMOLOGY WITH PROTEINS OF THE

hnRNP*

Shizhe Jia and John L. Wang

Department of Biochemistry
Michigan State University

East Lansing, MI 48824

Running Title: Carbohydrate Binding Protein 35

(This chapter has been published, J. Biol. Chem. 263, 6009-6011

(1988); it has been permitted for reprint from publisher)

55
FOOTNOTES
* This work was supported by grant GM—27203 from the National
Institutes of Health. J.L.W. was supported by Faculty Research

Award FRA-221 from the American Cancer Society.

1 The abbreviations used are: CBP, carbohydrate binding protein;

hnRNP, heterogeneous nuclear ribonucleoprotein complex.

56

SUMMARY

The complete nucleotide sequence of a cDNA clone for Carbohydrate
Binding Protein 35, a galactose specific lectin identified in the
nucleus of mouse 3T3 fibroblasts, has been determined. The deduced
amino acid sequence suggests that the protein consists of two
domains: (a) an amino terminal portion that is homologous to
certain regions of proteins of the heterogeneous nuclear ribo-
nucleoprotein complex (hnRNP); and (b) a carboxyl terminal portion
that is homologous to b-D-galactoside specific lectins isolated from
, .zmber of animal tissues. This two domain motif is reminiscent

of several DNA- and RNA—binding proteins.

57

INTRODUCTION

Carbohydrate Binding Protein 35 (CBP35; Mr 35,000) is a galactose
specific lectin identified in the nuclei of cultured mouse 3T3
fibroblasts (1-3). The level of expression of this lectin and its
nuclear localization.are dependent on the proliferative state of the
cells (4). When quiescent 3T3 cells were stimulated by the addi—
tion of serum, immunofluorescence studies with a polyclonal anti-
serum directed against CBP35 revealed prominent punctate intra-
nuclear staining, suggesting the association of the lectin with
subnuclear structures. Monospecific, affinity purified rabbit
anti—CBP35 was used to screen a gtll cDNA library derived from 3T3
fibroblasts. One identified clone, designated as clone 1, expressed
a fusion protein which, upon digestion with S. aureus V-8 protease,
yielded.21 polypeptide (Mr 30,000) that had carbohydrate-binding
activity (5). We now report the complete nucleotide sequence of

this cDNA clone and its deduced amino acid sequence.

58

MATERIALS AND METHODS
A: DNA cloning
CBP35 cDNA clone was identified as described before (5). cDNA
insert was isolated and cloned into Ml3mp18 in both orientations
(17).
B: DNA sequencing
The nucleotide sequences of the insert in M13 clones were determined
using the dideoxy chain termination method (18). The universal
primer was purchased from New England Biolabs. Six synthetic
oligonucleotide primers were synthesized on an Applied Biosystems

DNA synthesizer (Model 380 B).

59

RESULTS AND DISCUSSION

A restriction map and the sequencing strategy for clone 1 is shown
in Fig. 1, and the nucleotide sequence of the 883 residue cDNA
insert of clone 1 is shown in Fig. 2. The first amino acid of the
polypeptide encoded by this cDNA and its translational reading frame
were determined on the basis of: (i) the known nucleotide sequence
of the EcoRI cleavage site, which represents the juncture of the
b—galactosidase sequence and the cDNA insert in the lambda gtll
cloning vector (5); and.(ii) homology'between the deduced.amino acid
sequence and several b—D-galactoside specific lectins (see below).
The deduced amino acid sequence accounted for 263 amino acids (Fig.
2). Comparison of the sequence derived from clone 1 and another
cDNA clone indicates that the sequence shown.in.Fig.2 lacks only the
ATG initiation codon fronlthe full protein coding region (S. Jia and
J.L. Wang, unpublished observations). Therefore, the deduced amino
acid sequence is missing only the amino terminal methionine of the
CBP35 polypeptide. Previous Northern blotting analysis had yielded
1.3 kilobases as the transcript size corresponding to the clone 1
cDNA insert (5). Thus, although clone 1 contains essen-tially all
the protein structural information, it nevertheless represents a
less than full length cDNA clone in terms of the 5'- and

3'-untranslated regions.

Analysis of the deduced amino acid sequence of clone 1 in terms of
hydrophilicity or hydropathy value (6) clearly shows a two-domain
structure. Using the numbering system for the amino acid sequence
shown in Fig. 2, the amino terminal portion (residues 1-125)

exhibited neither a highly hydrophilic nor a hydrophobic region.

60

Figure 1: Restriction map of the cDNA insert of CBP35 clone 1 and

strategy for sequencing: ‘The CDNA insert was subcloned into»Ml3mp18

(17) in both orientations and nucleotide sequences were determined

using the dideoxy chain termination method (18). PNSJdATP (New

England Nuclear, 500 Ci/mmole) was used in place of [32P]dATP. The

arrows indicate the regions that were sequenced using one universal

primer and six synthetic oligonucleotide primers. The oligonucleo-

tides were synthesized on an Applied Biosystems DNA synthesizer
(Model 380 B).

 

61

a .mE

 

 

 

 

 

 

 

 

V. .
. (Y Y.
lily
.I lllllllllllll Aocozm .mAIJmV
1A.. AA nnnnnnnnnnnnnnnnnn .. uuuuuu
A L:
35:. .TJB
_ 58m Ham . ESL
_

 

B com com owe _ cow 4.1.28

62

Figure 2: Nucleotide and deduced amino acid sequence for CBP35 clone

1. The numbers at the right indicate the positions of the

nucleotide base pairs (top line) and the amino acids residues

(bottom line). The open reading frame in the nucleotide sequence

covers 263 amino acid residues. * represents the in-frame stop
codon. The amino acid sequence was deduced from the nucleotide

sequence using the program of Staden (19).

 

63

30

090

cam
o~9

o_m
omo

oo—
cam

om—
0m:

owl
00m

00
09m

00
00*

on
om

oo<

cm<
0<<

oca
<00

cn<
0<<
Lc9
<0<

L99
9<9

azo
poo

ad<
ooo

ado
ooo

000 000 00<

a»;
0<<

090
<<<

co<
0<<

xao
000

ota
900

one

son

aaa
ooo

oaa
900

out
09<

ago
ooo

ado
0<0

om:
09<

ado
ooo

L99
9<9

ota
<00

cH0
0<o

mL<
000

row
90<

cn<
9<<

oHH
09<

0<<
900

0H<
900

cam
900

cn<
0<<

<0<

0“:
9<0

2H0
0<0

era
099

ace
<o<

0H<
900

ado
loo

aa<
900

ado
000

000

co<
0<<

era
999

mL<
000

0am
09<

oca
900

oaa
900

0H0
0<o

aa9
009

000

I
<<9

cx9
0<9

ota
ooo

ora
ooo

zoo
090

L99
9<9

0<<
000

ado
«oo

ua<
<00

9<0

oAH
o9<

ado
0<o

one
099

co<
0<<

no:
09<

9H0
000

Lee
eo<

oaa
<00

sac

<0<

as:
oa<

ooo
09o

a~<
000

0:1
pee

mL<
ooo

9H0
<00

ota
<00

L99
o<9

ota

<90 990 990

aa<.n«: ca<

000

:00
<90

Low
<09

a“:
0<o

oaa
000

com
909

ado
ooo

ua<
000

cs9

.o<o

0“:
o<o

cao
0<0

one
ope

pox
09<

oxo
009

ora
900

9H0
000

5H0

.900 000 9<9 <00

h A.-
‘ d
H-514].

0<<

0H<
900

0L<
<0<

0H<
000

aa>
090

cao
0<0

L99
9<9

oca
<00

cao

09<

0H<
900

an<
9<0

sac
<<0

Am>
990

ado
<oo

oaa
ooo

0a<
900

oaa
900

oaa

a .0321

990 000

Low
00<

co<
0<<

3H0
<<0

an<
9<0

ado
<00

c>9
o<9

ado
<00

.ao<

900

cn<

any

.

oo<

H0>
090

can
0<<

co<
9<<

oca
900

aa<
000

oaa
900

CH0
0<0

oca

<<o 900 0<< <00

H.903;
ova-kw
.

R.Afll

D.,)..u
.

909

:00
090

~H<
000

ado

<00

3H0
000

zoo
099

mac
000

0a<
000

1&0
«oo

cn<

099

pro
oo<

Ha>
990

0L9
009

mc<
<0<

oaa
ooo

ora
900

Lee
pol

ora
<00

9H0

9<o

oAH
<9<

”>0
0<<

ca<
0<<

0L<
00<

zoo
09o

ma<
<00

oca
<00

r99
9<9

com

099

ao<
o<0

.21
ova

cu<
9<<

era
099

an<
0<0

ado
ooo

ado
ooo

ua<
ooo

ado

9o<

ado
900

no:
o<0

dn<
o<0

ao<
9<0

L99
9<9

oaa
900

02a
ooo

ado
ooo

aa<

0<< <00 909 000 900

. .
V‘u
luau»

<<0 009 0<0 0<9

com
90<

00<
o<0

ado
0<o

so;
<90

ora
000

CH0
<<o

099
o<9

oca
900

:00

one 12o goo
oa< ooo oeo

0<<
900

a»;
0<<

A0>
990

am>
090

ado
000

aa<
000

L99
9<9

0H<

<99 000

3H0

aa>

<<04990

are
oo<

oHH
99<

are
oo<

oaa
<00

com
90<

aa<
ooo
am<
9<0

cn<
0<<

mc<
00<

zoo
090

0:1
ob?

oca
900

0H<
900

cn<
0<<

900

CH0
<<o

:00
090

oxo
909

co<
0<<

ora
<00

0H<
000

02a
900

1H0
ooo

:00

090

com
00<

z~>
oao

Hm>
090

0H<
<00

loo
«oo

»Ho
«00

0H<
900

oaa
<00

com

909

0HH
09¢
Coo
<<o

ca“
99<

cm<
0<<

0H<
900

.02a

900

ado
0<0

a>9

.o<9

one

000

3H0
<<0

0AM
<9<

Hm>
090

oca
ooo

ota
000

0H<
ooo

11o
<00

ado
ooo

ooo

990 009 999 00<

00<

ma<
000

090
<<<

0L<
<0<

«>0
<<<

Hm>
090

02a
900

021
<00

9&0
000

<<0

ooo
09o

oca
099

0L<
<0<

Ho>
090

11o
Foo

CH0
<<0

cx9
9<9

0H<
<00

an< ma<
040 <00

64

This was quite distinct from the carboxyl terminal portion
(residues 126-263) , which contained both hydrophobic and hydrophilic
regions that are characteristic of many globular proteins. Each of
these two domains is homologous to distinct groups of proteins
previously identified on the basis of their function/activity. A
region containing 76 amino acids (residues 138—214) was found to be
homologous to a number of b—D-galactoside specific lectins (7—11)
(Fig. 3). For bovine heart lectin (10), chicken skin lectin (7,8),
and electric eel lectin (11), the complete amino acid sequences over
this 76-residue stretch were available. The extents of homology
between CBP35 with these lectins were: (a) 34% with lectin from
electric eel; (b) 36% with bovine heart lectin; and (c) 38% with
chicken skin lectimn Severa1.peptide sequences are highly conserved
in all the lectins sequenced so far. These include the sequence
His—Phe-Asn-Pro-Arg~Phe-Asn (residue 171—177) and. the sequence
Trp-Gly-Lys-Glu-Glu-Arg—GlnSer—Ala-Phe—Pro-Phe (residues 194-205).
The latter sequence, containing a tryptophan and two glutamic acid
residues, correspond to the peptide hypothesized to be present in

the b—D-galactoside binding site of the eel lectin (11).

Previous searches of data bases with the bovine heart lectin
sequence (10) and with the sequences of two hepatoma clones (9)
failed to reveal any significant homologies with other proteins.
As a result of these analyses, it was proposed that the soluble
vertebrate lectins with b-D—galactoside binding activity represent
a new protein family. CBP35 now joins this protein family on the
basis of its sequence homology with these lectins. Moreover, the
present sequence results also provide the structural basis for the
observation that the clone 1 fusion protein, upon V—8 protease

digestion, yielded a polypeptide with carbohydrate-binding activity

65

Figure 3: Comparison of the amino acid sequence of CBP35 clone 1
(residues 127 through 214 in the numbering system of Figure 2) With

the amino acid sequences of six other b—D-galactoside binding

lectins. BHL, bovine lung lectin (10), CSL, chicken skin lectin

(7,8), HLL, peptides from human lung lectin (9), HEP 1 and HEP 2,

sequences deduced from two cDNA clones derived from a human hepatoma

library (9), and EEL, lectin :from electric eel (ll). Dashes
indicate gap introduced for optimal alignment; complete gaps are

positions for which no sequence information is available; * denotes

uncertain residue assignments. Sequence identities between CBP35

clone 1 and the other lectins are highlighted by boxes. The

sequences were compared using the FASTP program (16).

 

 

K P N A N

 

N A K

 

 

 

127LTvaDLPLPGGvMIFRMLIUMF

CBP35

BHL
CSL

M S C Q G P V C T N L G

HLL

NMDMKflGSTLKmTGSIADGTU
NGVVDERMSFKAGQNLIVKpVPSIDST

HEP l
EEL

66

 

 

m

N P R D A H G D 0 0 A V

ELMH F

 

 

G L H F N P R F D A H G D V N L 1
' L
N L H F N P R F S G S T

L
L
L
L

HEVEFHFNPRFNENNRRv-fl
NLYLHFNPRFNAHGDVNLI

 

12:2
91:23am
:waP—¢
..ocncaw
mxxxoz
01004901903
LL._J._J_J——l>
szzzz
__J..JZ__.|'—‘""‘
>_.J>>><
p—ou_u.u..u_u—
ormwwoz

c3935
Bu
CSL
HLL
HEP 1
EEL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

       

 

 

 

 

 

 

   

  

   

 

<3

._—.

>2

>-9—LL

uJHCD

¥w<i ”0):".
(you) 00:3]
wmxgmmo
macaw max
ImLLu—U—u-u'
Q.Q_O->-‘-—’°-°'
u_u_u_u__Ju.|-l—
<<r>>==<99
mmt—<EQ°-°9
ammUJquU-l
(3:10.501: 05me
(140900333
[—JuJuJ-u LIJUJLBJ
¥<9—t Cat—:—
«.9 (some)
3; 3333
2 <2th
zoth-Dm ‘9
04:20:29 £9
9 E]

,J u.

m (I)

z 2?

U >

> >

 

 

 

 

 

 

 

y—mcf)

222

one

>>>

m r—‘N
m “J
33:: 09-4 ”JL“ 31
QmQIII

67

(5).

While the amino acid sequence toward the carboxyl terminal half of
CBP35 is homologous to the sequences of other lectins, the sequence
at the amino terminal end also showed interesting features. First,
the sequence between residues 40-112 showed eight internal sequence
homologies (Fig. 4): (I) residues 40-48; (II) residues 49—57;
(III) residues 58-66; (IV) residues 67-75; (V) residues 76—84; (VI)
residues 85-93; (VII) residues 94-102; and (VIII) residues 104-112.
Each of these homologous regions consists of a 9-residue repeat,
with a consensus sequence of Pro-Gly-Ala-Tyr-Pro-Gly, followed by
three additional amino acids. As a result, this stretch of the

sequence is characterized.by a high proportion of Pro (27%) and Gly

(24%).

In addition, the sequence at the amino terminal portion also showed
homology with the amino acid sequences of several proteins identi-
fied.as polypeptides of the heterogeneous nuclear ribonucleoprotein
complex (hnRNP) (Fig. 5). These include: (a) 25 identities over
108 residues with a glycine-rich protein (GRP33) of the hnRNP of

brine shrimp.Artemia salina (12); (b) 18 identities over 71 residues

 

with the deduced amino acid sequence of clone DL-4, identified from
a human hepatoma cDNA library on the basis of it expression of a
fusion protein reactive with chicken antibodies directed against
bovine hnRNP proteins (13); and (c) 11 identities over 42 residues
with human hnRNP protein C1 (14). There was no apparent homology
between the sequences of CBP35 and the rat hnRNP protein A1 (15)

when subjected to the same analysis with the FASTP program (16).

None of'tflua hnRNP proteins contained any copies of the striking

68

9-residue repeat sequence (Fig. 4) found in CBP35. However, the
extent of homology between CBP35 and the hnRNP proteins shown in
Fig.6 was 25%, comparable to the level of homology between the
hnRNP proteins themselves (12—15). More over, based on sequence-
scrambling comparison of 20 random sequences, the FASTP program
of Lipman and. Pearson (16) yielded. the following statistical
significance for the observed homologies: (a) The aligned score
between CBP35 and.GRP33 was 1.52 standard deviations above the mean
score; (b) The aligned score between CBP35 and hnRNP protein Cl was
0.96 standard deviations above the mean score; and (c) The aligned
score between CBP35 and.DL-4 sequences was 0.03 standard deviations
above the mean score. For comparison, the aligned score between
hnRNP proteins Cl and.A1 was 1.04 standard.deviations above the mean

score .

Finally, CBP35 shares with a number of hnRNP proteins what appears
to be one of their typical features, i.e. the.presence of distinct
domains with non-uniforntdistribution.of Gly and.Pro residues in the
polypeptide chains. In.the case~of CBP35, 86% of the 36 Gly residues
and 87% of the 39 Pro residues are in the amino terminal domain of
the molecule. For both GRP33, the hnRNP protein from brine shrimp
(12) and rat hnRNP protein A1 (15), approximately 77% of the total
Gly residues are located within the carboxyl terminal 124 amino acid
residues. An unequal distribution of Pro has also been.observed.for
human hnRNP protein C1; all of the Pro residues were found in the
amino terminal half of the polypeptide chain (14). Therefore, the
homology of CBP35 to the sequences of hnRNP proteins and the
conservation of certain structural features in distinct domains

suggest that the lectin might be one of the hnRNP proteins.

69

Figure 4: Alignment of the amino acid residues showing internal

sequence homologyu The amino terminal domain of CBP35 clone 1 shows

a repetitive sequence, each. consisting’ of nine amino» acids.

Numbers at the left and right indicate the amino acid residue in the

numbering system of Figure 2.

 

70

 

 

 
 
 
 
 

 

 

 

007650.322
455700901

111
PPPAAAGG
AAATTPPS
QQQPPQQC
066556689
PPPPP.D..D.P
YYYYYYEY
AAAAAAAA
GGGEGGGG
PPPPPPPP
090075548.
0.0.5678gm

Fig. 4

71

Figure 5: Comparison of the deduced amino acid sequence of CBP35

clone 1 (residues 14 through 120 in the numbering system of Figure

2) with the amino acid sequences of hnRNP proteins. GRP33,

glycine-rich protein of the hnRNP of Artemia salina (12), DL-4.

sequence deduced from a CDNA clone derived from a human hepatoma

library (13) , and human hnRNP C1 protein (14) . Dashes indicate gaps

introduced for optimal alignment. j Sequence identities between CBP35

clone 1 and the hnRNP proteins are highlighted by boxes. The

sequences were compared using the FASTP program (16).

 

R
G

GEQEGAMVAAT

 

G R G R G R G G F S

GNflPGAGGYP

YTAN
- QGR
20(QSA

(E)

NPO
(39m

i
P
P60
LEK
GSG

llIG'N

200 G G

CBP35
GRP33
DL-A

 

fiGA”GQ-A-PPSAYP
e FD

0R

m

GPGAEERLEALE]

 

PPGAYLPJGQAP
RMNTSETMDP‘

CBP35
GRP33
DL-A
HNRNP C1

72

0c:
0......J
.0:
<0
00:
Cl.<C
<03

....

cram

RRGKSGflNS

 

EB
GA
A0
3

 

CBP35
GRP33
DL-A
HNRNP C1

 

 

 

 

G P Y 120
A P Y 308

A
A

 

 

 

 

<00
0.03
>—o:
«.902
OLD
mo:
u<r
cam

CBP35
GRPSS

HNRNP C1

Fig. 5

73

This notion is supported by recent observations that have identified
CBP35 in hnRNP (20). Nucleoplasm derived from 3T3 cells was
fractionated on a cesium sulfate gradient (1.25—1.75 g/ml);
immunoblotting analysis with anti-CBP35 localized the lectin in
fractions with densities of 1.30-1.35 g/ml. This range of densities
corresponds to the density of hnRNP on cesium sulfat gradients.
Conversely, affinity chromatography on the basis of saccharide—
binding could also be used to isolate hnRNP. When nucleoplamn
derived from [35$]methionine-labeled 3T3 cells was fractionated on
a column of Sepharose derivatized with N-e-aminocaproyl-D-galac-

tosamine, the bound and eluted fraction yielded CBP35 as well as a
set of polypeptides whose molecular weights matched those reported
for the core particle of hnRNP. Moreover, the complex isolated in
the saccharide affinity column also contained RNA. These results,
along with the present structural data, both suggest that CBP35 is

a component of hnRNP.

74
REFERENCES

1: Roff, C.F., Rosevear, P.R., Wang, J.L., and Barker,R. (1983)
Biochem. J. 211, 625-629.

2: Roff, C.F., and.‘Wang, J.L. (1983) J1 Biol. Cheat 258,
10657-10663.

3: Moutsatsos, I.K., Davis, J.M., and Wang, J.L. (1986) J. Cell
Biol. 102, 477-483.

4: Moutsatsos, I.K., Wade, M., Schindler, M., and.Wang, J.L. (1987)
Proc. Natl. Acad. Sci. U.S.A. 84, 6452-6456.

5: Jia, S., Mee, M.P., Morford, G., Agrwal, N., Voss, P.G.,
Moutsatsos, I.K., and Wang, J.L., (1987) Gene 60, 197-204.

6: HOPP. T.P., and Woods, K.R. (1981) Proc. Natl. Acad. Sci.
U.S.A. 78, 3824-3828. ‘

7: Ohyama, Y., Hirabayashi, J., Oda, Y., Ohno, 8., Kawasaki, H.,
Suzuki, K., and Kasai, K. (1986) Biochem. Biophys. Res. Commun.
134, 51-56.

8: Hirabayashi, J., Kawasaki, H., Suzuki, K., and Kasai, K. (1987)
J. Biochem. (Tokyo) 101, 775-787.

9: Gitt, M.A., and Barondes, S.H. (1986) Proc. Natl. Acad. Sci.

U.S.A. 83, 7603-7607.
10: Southan, C., Aitken, A., Childs, R.A., Abbott, W.M., and Feizi,

T. (1987) FEBS Lett. 214, 301-304.
11: Paroutaud, P., Levi, G., Teichberg, V.I., and Strosberg, A.D.
(1987) Proc. Natl. Acad. Sci. U.S.A. 84, 6345-6348.

12: Cruz—Alvarez, M., and.Pellicer, A. (1987) J. Biol. Chem. 262,

13377-13380.

13: Lahiri, D.K., and Thomas, J.O. (1986) Nucleic Acids Res. 14,

4077-4094.

75
14: Swanson, S., Nakagawa, T.Y., LeVan, K., and.Dreyfuss, G. (1987)
Mol. Cell Biol. 7, 1731—1739.
15: Cobianchi, F., SenGupta, D.N., Zmudzka, B.Z., and Wilson, S.H.
(1986) J. Biol. Chem. 261, 3536-3543.
16: Lipman, D.J., and.Pearson, W.R. (1985) Science 227, 1435-1441.
17: Messing, J. (1983) Methods in Enzymol. 101, 20-78.
18: Sanger, F., Nicklen, S., and Coulson, A.R. (1977) Proc. Natl.
Acad. Sci. U.S.A. 74, 5463-5467.
19. Staden, R. (1979) Nucleic Acids Res. 6, 2601-2610.

20. Laing, J. G., and Wang, J. L. (1988) Biochem. 27, 5329-5334.

76

CHAPTER IV

Nucleotide Sequence of the Murine Gene

for Carbohydrate Binding Protein 35*
Shizhe Jia and John L. Wang
Department of Biochemistry

Michigan State University

East Lansing, MI 48824

Running Title: Genomic Sequence of Carbohydrate Binding Protein 35

(This chapter, with some revisions and additions, will be submitted

for publication)

77

FOOTNOTES

This work was supported by grants GM—38740 and GM-27203 from

the National Institutes of Health.
The abbreviations used. are: CBP35, Carbohydrate Binding
Protein (Mr ~ 35,000); kb, kilobase(s); bp, base pairs; IVS,

intervening sequence; SRE, serum responsive element.

Heng-Yin Yang, unpublished results.

78

S UMMARY

We have isolated a cloned segment of the gene for mouse Carbohydrate
Binding Protein 35 (CBP35). Nucleotide sequence analysis of the
cloned gene, as well as the results of genomic Southern blot
hybridization, revealed tflun: the gene is ‘unique and. spans 9
kilobases of genomic DNA. The mRNA for the lectin is encoded by
five exons separated by four introns. Intron I is located in the
5' transcribed but untranslated region of the primary transcript.
Examination of the nucleotide sequence 5’ to the transcription
initiation site revealed characteristic TATA and CCAAT boxes. A
'putative serum-responsive element has also been identified.about 200
nucleotides upstream.from.the site for initiation of transcription.
This may serve as a binding site for a class of transcription
factors responsible for the serum-stimulated expression of the CBP35

gene.

79

INTRODUCT ION

Carbohydrate Binding Protein 35 (CBP35; Mr 35,000), a galactose-
'specific protein found in many cells and tissues (1,2), belongs to
the S-type family of lectins (3). In mouse 3T3 fibroblasts, this
protein has been identified in several subcellular locations,
predominantly in the cytoplasm.and nucleus (4). When quiescent 3T3
cells were stimulated by the addition of serum, there was an
increase in the overall level of CBP35, as well as a dramatic rise

in the amount of the polypeptide in the nucleus (5).

A cDNA clone for CBP35 has been isolated and sequenced (6). The
deduced amino acid sequence showed that the polypeptide consisted
of two domains. The NHz—terminal half was proline- and glycine-rich
and showed some structural similarity to proteins of the heteroge-
neous nuclear ribonucleoprotein complex (hnRNP). The carboxyl
terminal half was homologous to many other B-galactoside specific

lectins (7-15).

On the basis of structural information, a number of other proteins,
identified independently'in.different laboratories and for distinct
reasons, have been found either to be identical to CBP35 or to be
a homologue of CBP35 in another species/cell type. These include:
(a) L-34, a. metastasis-associated. lectin isolated. front murine
fibrosarcomas (15); (b) Mac-2, a cell surface marker linked to
macrophage differentiation (16); (c) LBP, the major non-integrin
Laminin Binding Protein of murine macrophages (17); and (d) EBP, an
IgE-binding protein identified in rat basophilic leukemia cells

(18,19). Although only L-34 was originally isolated on the basis

80
of carbohydrate-binding activity, all the other proteins listed

above have now been shown to bind saccharides (15-19).

The majority of the CBP35 of mouse 3T3 fibroblasts is intracellular
(4). Although a large proportion of L-34, Mac-2, and EBP are also
found inside their respective cells, they have been studied, for the
most part, because of their extracellular localization (15,16,20).
For these reasons, it was important to compare the structural
features of the proteins that might lead to the possibility of dual
localization. Thus far, the coding regions of the mature proteins
do not offer hints on the mechanism of externalization (6,15,16,18) .
Nor do the cDNA sequences implicate an obvious signal sequence for
targeting to the lumen of the endoplasmic reticulum, a classical
endomembrane pathway for secretion. In their studies on Mac-2,
Cherayil et a1. (16) demonstrated the possible existence of two
alternatively spliced Mac-2 cDNAs, one of which has the potential
to encode a colinear, EUR terminally extended. Mac-2 protein
containing a signal peptide-like sequence. An analysis of the
genomic region encoding these cDNAs would provide confirmation of
these postulated splicing events. In the present paper, we report
the structure of the murine CBP35 gene, paying particular attention
to the region S'to the translation initiation codonn In the course
of these studies, we also found.a possible structural basis for the

observed serum-stimulated increase in expression of the CBP35 gene.

81

EXPERIMENTAL PROCEDURES

Screening of the Genomic Libraries

The cDNA for CBP35 (6) was labeled with.[92P]dCTP using a random
priming kit (Boehringer Mannheim) and the method of Feinberg and
Vogelstein (21). The specific activity of the labeled cDNA was ~
108 cpm/ug DNA. This probe was used to screen a A EMBL4 library
derived from mouse liver chromosomal DNA (Clone Tech Laboratory)
using procedures described previously (22). Three positive clones
were identified” .Since these three clones appeared.to have the same
inserts, only one clone, designated as 1G3, was characterized by

restriction mapping and DNA sequence analysis (see Fig. 1).

A 1.7 kb PstI-SalI fragment at the left end of the 1G3 clone
(highlighted by a hatched rectangle in Fig. 1) was isolated and
labeled with [°2P]dCTP as described above. This was then used as
a hybridization probe to screen a I. FIXII library, also derived from
mouse liver chromosomal DNA (Strategene). The KGH clone was

identified and isolated from this screening.

Cloning and Sequencing

The 6.4 1d) SalI—EcoRI fragment of 1G3 clone (Fig. 1) was first
cloned into the PUC-l3 plasmid. After digestion with PstI and
HindIII, the individual fragments were subcloned.into the sequencing
vectors M13mp18 and M13mp19 (23) . The 1.7 kb SalI-PstI fragment and
parts of the 2.5 kb PstI-HindIII fragment were sequenced using the

dideoxy method (24) (Sequenase Kit, United States Biochemical Co.).

 

Figure 1:

82

Restriction map and sequencing strategy for the murine

CBP35 gene. The scale at the top indicates the length

in kilobases. Exons are indicated by dark lines/boxes

and are numbered I-V. The intervening sequences are

designated. IVS and. are numbered. 1-4. The numbers

between restriction enzyme sites indicate the size of

the restriction fragment. The length and direction of

sequencing reactions for the 1GB and 1G3 clones are

shown by the arrows. The hatched rectangle in the AGJ

clone indicates the fragment that was used as a hybrid-

ization probe to identify the 1GB clone.

83

 

 

 

 

 

 

 

 

 

 

V .-
1 fig
'0 1; °
N
‘2
¢
("IHWOG- -
O
N
'0.
0
-§
‘50
O i— 18033 t "'
C?
N
-IIIP“IH- - ..
Ignore
9
jgtmma
IIIUOXO
1100a:
. s
SAI (s. ’
N
IUOXO
-— Iaooa- —
'9.“- IHWDG" )-
‘9
J I I

 

 

 

kb
DNA
>391
NH

Figure 1

84

The 6.4 kb EcoRI-EcoRI fragment and the 6.7 kb EcoRI-SalI fragment
from the AGH clone (Fig. 1) were cloned into the PUC-18 plasmid.
After HindIII digestion, these fragments were further subcloned into
M13mp18 and.M13mp19 vectors (23). Both single-stranded and double-
stranded DNA sequencing analyses were performed with the Sequenase
Kit. Eight oligonucleotide primers, synthesized on the basis of the
CDNA. sequence (6), were used, in carrying out the sequencing
reactions. The oligonucleotides were synthesized.cx12nn Applied
Biosystems DNA synthesizer (Model 380B) in the Macromolecular
Structure Facility at Michigan State University. The areas and
directions of the sequencing analysis are shown by the arrows in

Figure 1.

Directional Deletion of a 2.7 kb Fragment for DNA Sequencing

The method used followed that described in Sambrook et a1. (25).
The 2.7 kb EcoRI-HindIII fragment of XGH clone (Fig. 1) was digested
by Nuclease Sl (Boehringer Mannheim) to blunt end.the DNA fragment.
This was then cloned into the SmaI site of the M13mp18 vector. The
orientation of the insert was determined by restriction analysis.
A clone with the original HindIII end toward the SalI site of the
M13mp18 multiple cloning site was selected. Approximately 10 ug of
M13 replicative form DNA were isolated from this clone and were
digested with SalI and PstI. Nuclease $1 (40 U; Boehringer
.Mannheinn and Exonuclease III (200 U; Boehringer Mannheim) were then
added to the double-digested DNA at 37°C. Samples were withdrawn
from this digestion mixture at 90, 135, 180, and 225 seconds. The
samples were placed at 30°C for 30 minutes. Then, the Klenow

fragment of DNA polymerase I (10 U; Boehringer Mannheim) was added

85

to each sample, along with the dNTPs (0.5 mM each) and incubated at
25°C for 30 minutes to create blunt ends. The DNA samples were
finally ligated and transformed into E. coli. Twenty-five clones
were selected from each transformation and sequenced using universal
primer and the Sequenase .Kit. A total of about 96 clones were

analyzed.
Primer Extension Assay

Quiescent, serum-starved cultures of 3T3 fibroblasts were stimulated
by the addition of serum (5) . After 18 hours, cytoplasmic RNA was
isolated following the protocol of Henikoff (26). The cells were
isolated and resuspended on ice for 5 minutes in 10mM Tris Buffer
(pH 8.6) containing 0.14 M NaCl, 1.5 mM MgC12, 0.5% Triton X-100,
and 1000 U/ml of Rnasin. The nuclei were pelleted by centrifugation
at 500 x g for 5 minutes. The supernatant was then made 2% in'
sodium dodecyl sulfate and 50 (lg/ml in proteinase K. After 1 hour
at 37°C, the supernatant was extracted with phenol: chloroform (v/v,
1:1) once and the RNA was collected by ethanol precipitation. The
DNA in the sample was removed by digestion with 2 ug/ml deoxyribonu-

clease I at 37°C for 1 hour.

The general method for the primer extension assay has been described
(25). A 22— nucleotide primer, 5’-CGAAAAGCTGTCTGCCATTTTC-3’ was
synthesized and 32PO4-labeled at the 5' end with T4 polynucleotide
kinase (2 U/ul) at 37°C for 30 minutes. The labeled primer (107
cpm/llg) was isolated by ethanol precipitation and was then allowed
to hybridize for 12 hours at 30°C to 30 (lg of cytoplasmic RNA
isolated from 3T3 cells. Reverse transcriptase (60 U; Boehringer

Mannheim) was added at 37°C for two hours. The radioactive product

 

86

of the extension reaction was electrophoresed on a 10% polyacryl-
amide sequencing gel, followed by autoradiography. The length of
the extended cDNA was estimated by comparison to a sequencing

reaction using a 20-nucleotide primer.
Southern Analysis of Genomic DNA

Mouse liver nuclei were isolated from a homogenate of 15 grams of
liver tissue by centrifugation at 500 x g for 51ninutes. The nuclei
were suspended in 10 mM Tris (pH 8.0) containing 0.14 M NaCl, 1.5
mM MgC12, 0.5% sodium dodecyl sulfate and 100 ug/ml proteinase K at
37°C for 16 hours. DNA was isolated by ethanol precipitation and
purified by CsCl gradient centrifugation. The DNA was then digested
with EcoRI, HindIII and BamHI. 'The initial digestion used 20 U
enzyme per ug of DNA and was carried.out for 20 hours. An addition-
al 10 U of enzyme per ug of DNA was added and incubated for 4 more
hours. The digested DNA.was separated by agarose gel electrophore—
sis (0.8%), subjected to Southern blotting (27) using 32P-labeled

CBP35 cDNA (108 cpm/ug) as a hybridization probe.

87

RESULTS
Isolation and Characterization of Genomic Clones

IX X EMBL4 library derived from House liver chromosomal DNA was
screened using the cDNA for CBP35 (6) as a hybridization probe. Of
the 106 plaques screened, three gave reproducible positive signals.
Oneof these, designated as AG3, was further cloned by lower density
rescreening and plaque purification. Digestion of AG3 with PstI and
sequence analysis in both directions (Fig. 1) revealed the single
PstI site in the nucleotide sequence of the cDNA (6). Comparison
of the genomic sequence with the cDNA sequence defined exon V,

corresponding to the 3'end.of the cDNA clone.

The 1.7 kb fragment, derived from the "left" end of 163 by PstI
digestion, was sequenced in its entiretyz, without finding any
additional sequence matching the cDNA. This suggested.that the 1.7
kb fragment was inside an intron of the genomic DNA” Thus, this 1.7
kb fragment of 163 (highlighted by a hatched rectangle in Fig. 1)
was used as a hybridizatirxxprobe to search for other genomic clones
from.the A FIXII library, also derived.fronlmouse liver chromosomal

DNA. Clone KGH was selected from such a screening.

Analysis of restriction enzyme cleavages and of the partial
nucleotide sequence showed that the XGH and 1G3 clones overlapped
over a region of approximately 6 kb, corresponding to the "right"
end of AGH and the "left" end of 163 (Fig. 1). When the cDNA for
CBP35 (6) was digested with PstI, a fragment corresponding to the

3’ end (nucleotides 696-883) of the cDNA was isolated. This

Figure 2 :

0

88

Southern blot hybridization of mouse genomic DNA with

the cDNA clone for CBP35. Chromosomal DNA was isolated

from mouse liver nuclei. The DNA was digested with

restriction enzymes, separated on agarose gels (0.8%)

and subjected to Southern blot analysis with 32P-labeled

CBP35 CDNA (108 cpm/lig). Panel A: lane 1, BamHI

digest; lane 2, HindIII digest; lane 3, EcoRI digest;

and lane 4. undigested DNA. In panel B, the same EcoRI

digest sample that was used for lane 3 of panel A was

electrophoresed for a longer time to better resolve the

bands of a doublet.

89

23-6 '

9.4-
vzth 6.6' a '23
Litéi 4.4. i i
...-I a .9.4
2.5% 2.3- __

. I - . -66
1.3- :
1.0"
1 2 3 4

Figure 2

90

fragment hybridized to both XGH and 1G3 clones. In contrast, the
other fragment of the cDNA generated by PstI (nucleotides 1-695)
hybridized to the AGH clone, but not the 1G3 clone. Therefore, it
appears that the overlapping region between 1GB and 1G3 contained
the 3' end of the cDNA, while the 5’ end of the cDNA was found in
the remainder of the XGH clone (Fig. 1). Nucleotide sequence and
primer extension analyses (see below) indicate that the gene for
CBP35, including several consensus sequences for 5’ regulatory

elements, can be entirely mapped onto the AGH clone.

DNA isolated from nuclei of mouse liver was digested with several
restriction enzymes, separated by gel electrOphoresis and then
subjected to Southern blot analysis with the cDNA for CBP35. A
single fragment (19-20 kb) was observed after BamHI digestion (Fig.
2A, lane 1). HindIII digestion yielded fragments of 4.2 kb, 3.2 kb,
and 1.3 kb (Fig. 2A, lane 2). Although the EcoRI digested.material
yielded a broad.band in the experiment shown in Figure 2A (lane 3),
it actually contained two fragments, which could be resolved into
a doublet upon longer electrophoresis of the same material (Fig.
2B). The positions of migration of the bands corresponded to DNA
molecules of ~ 6.4 kb and ~ 7.0 kb. The lengths of each of these
genomic DNA fragments that hybridized with the cDNA probe were in
good agreement with those derived from the genomic clones isolated
above (Fig. 1). No other bands could be observed. These results
indicate that the CBP35 gene is a single gene in the normal mouse

genome.
Sequence Analysis of the Genomic Clones

Using.a strategy similar to that described for the analysis of the

‘-

 

 

91

PstI fragment of clone 1G3 (highlighted by a hatched rectangle in
Fig. 1), nucleotide sequences were determined for parts of the AGH
clone (see Fig. 1). Figure 3 reports the nucleotide sequences of
the 5' and 3’ flanking regions, the exons, and the exon/intron
boundaries of the CBP35 gene. The beginning of exon I was identi—
fied by carrying out directional deletion on the 2.7 kb EcoRI-
HindIII fragment of 1GB, cloning each of the resulting fragments,
and sequencing approximately the first 100 nucleotides of each of
the 90 or so clones obtained (highlighted by dotted line in Fig. 1) .
This revealed a sequence CTTCCG, fitting the consensus for a mRNA
cap site (rectangle highlighted by a wavy underline in Fig. 3). On
the basis of this landmark, the initiation.site of transcription was

deduced.to be an adenine residue (circled.and labeled +1 in Fig. 3).

The partial sequences of the 2.7 kb EcoRI—HindIII fragment (high-
lighted by dotted line in Fig. 1) also revealed a 2.3 kb intron
(IVSl in Fig. 3). On thei? side of this IVSl, exon II starts with
the sequence GAAAATGG. This ATG triplet (solid triangle in Fig. 3)
falls within the consensus translation initiation sequence A/GNN-
ATG-G (28). On the basis of comparing the nucleotide sequence
obtained in this study with that of the cDNA (6), the present
assignment of methionine to this initiation codon revealed: (a) the
cDNA clone coded for the entire CBP35 polypeptide chain except for
the NHZ-terminal Met residue; and (b) the second amino acid of the
polypeptide should be Ala rather than the Arg that was reported
previously (6). These conclusions are consistent with the results
of the nucleotide and amino acid sequences reported for Mac-2 (l6),

L-34 (15), and SBP (18).

Given these assignments of the start of exon I and the translation

Figure 3 :

92

Nucleotide sequence of the murine CBP35 gene. The
numbering system, shown on the left, is based on the
transcription initiation site; this is the A residue

highlighted by a circle and whose position is labeled

+1. The nucleotide sequences of the introns, denoted

IVS 1—4, are not shown except for the 5’ and 3’ ends to
highlight the conservation of the donor and acceptor

splice site consensus sequences. The sequence

CCAATTAAGG, highlighted by a rectangle with a double

underline, represents a putative Serum Response Element.
The sequence CCAAT, highlighted by a rectangle with a
single solid underline, and the sequence AATATATAT.
highlighted by a rectangle, represent sequences that fit

the consensus sequences and locations of CCAAT and TATA

boxes, respectively. The sequence CTTCCG, highlighted

by a rectangle with a wavy underline, represents a

putative mRNA cap site. The translation initiation

codon, ATG. is highlighted by a triangle. The open

reading frame extends to a termination codon, TM:

highlighted by an inverted triangle. The sequence

AATAAA, highlighted by a rectangle with a dotted under-

line. represents the polyadenylation signal

 

( D
(,1)

‘ 99999

M.A.-

- 280
~220
- IGO
— ‘00

—40'

20

'00
ISO
220
280!

34()
‘400

‘460
520
58C)
($40

700
760
820
880
940
1000

93

CATCTCATGAGATGCTGATCTCGTAGCTGAAGTCTGATCTAGATAGATGTGTGTTACAAC
GTGTGCI-TACAACTTACTCGGGTCCAATGACTGTTGTAACCTCCGTTTC
GCCGAATTCCTGTGGATCTGTAGGGGTCTCGCCAGAGGGACAGGAéhCCAGAGGAGAAAT
ACTTCAACCACCAT§§§§§ACGACAGAGGGTTTTcncccccTAGAGAACGACTGTAGACG

+1
TAAGAc—ACCTCTTCAACGAGGTCACCCAGccyrTTGAc-GGAAC

IVSI
GTACCCATACTCTAGGGTCCTCAGGGATGGGGTA[étaaa ----- (2-3 kb) ———————
---——---—————-—-—-¥ ————————————— cctag]GAAAAC§§CAGACAGCTTTTCG

Ivsz
Etaaa ——————— (0.5 kb) ---------- cttaé}CTTAACGATGCCTTAGCTGGCTC

TGGAAACCCAAACCCTCAAGGATATCCGGGTGCATGGGGGAACCAGCCTGGGGCAGGGGG
CTACCCAGGGGCTGCCTATCCTGGGGCCTATCCAGGACAGGCTCCTCCAGGGGCCTACCC
AGGACAGGCTCCTCCAGGGGCCTATCCAGGACAGGCTCCTCCTAGTGéCTACCCéGGCCC
AACTGCCCCTGGAGCTTATCCTGGCCCAACTGCCCCTGGAGCTTATCCTGGTCAACCTGC
CCCTGGAGCCTTCCCAGGGCAACCTGGGGCACCTGGGGCCTACCCCCAGTGCTCTGGAGG
CTATCCTGCTGCTGGCCCTTATGGTGTQCCCGCTGGACCACTG éYziig -----------

----- (3.1 kb) --

 

 

cttag ACGGTGCCCTA‘I‘GACCT
GCCCTTGCCTGGAGGAGTCATGCCCCGCATGCTGATCACAATCATGGGCACAGTCAAACC
CAACGCAAACAGGATTGTTcTAGATTTCAGGAGAGGGAATGAIGTTGCCTTCCACTTTAA
CCCCCGCTTCAATGAGAACAACAGAAGAGTCATTGTGTGTAACACGAAGCAGGACAATAA
CTGGGGAAAGGAAGAAAGACAGTCAGCCTTCCCCTTTGAGAGTGGCAAACCATTCAAA
élfgat; ------------------- (1 . Bkb) -------------------- cctag AT
ACAAGTCCTGGTTGAACCTGACCACTTCAAGGTTGCGGTCAACGATGCTCACCTACTGCA
GTACAACCATCGGATGAAGAACCTCCGGGAAATCAGCCAACTGGGGATCAGTGGTGACAT
AACCCTCACCAGCGCTAACCACGCCATGATC§§§CCCAGAAGGGGCGGCACCGAAACGCC
CTGTGTGCCTTAGGACTGGGAAACTTGGCATTTCTCTCTCCTTATCCTTCTTGTAAGACA
TCCCCCATTflEZ:E§ZBTCTCATGGGAGAGAGAGCCATGTTTTGGGGGTTTTTATGATAT

GGGTTCAAATTCTTTAGGAC

Figure 3

94
initiation codon in exon II, there must necessarily be an intron in
the 5’ transcribed but untranslated region of the primary tran-
script, a somewhat unusual occurrence. Tfidr;conclusion.is supported
by preliminary experiments that indicate an oligonucleotide probe
synthesized on the basis of a nucleotide sequence in IVSl failed to
hybridize to the 1.3 kb mRNA, previously identified with our cDNA
clone (29,30). The sequence results also predict that the 5’
untranslated region of the CBP35 mRNA will be 57 nucleotides long.
A primer extension experiment was performed to ascertain this
length. A 22-mer oligonucleotide complementary to the entire exon
II (Fig. 3) was synthesized, labeled with 32P, and was used as the
primer for reverse transcription of the mRNA (Fig. 4A) . The results
showed that the product of theextension reaction contained 72
nucleotides (Fig. 4B), consistent with the notion that the tran-
scription initiation site was located ~ 54 bases upstream from the

initiation ATG codon in the mRNA.
Structural Features of the CBP35 Gene

The CBP35 gene spans ~ 9 kb of genomic DNA and contains five exons
(Fig. 1): (a) exon I, 53 bp; (b) exon II, 22 bp; (c) exon III, 366
bp; (d) exon IV, 255 bp: and (e) exon V, ~ 323 bp (the exact size
of last exon is not known; see below). These exons are interrupted
by four introns: (a) IVSl, ~ 2.3 kb: (b) IVSZ, ~ 0.5 kb: (0) IVSB,
~ 3.1 kb; and (d) IVS4, ~ 1.8 kb. The RNA donor and acceptor splice
sites are characterized by GTAAA/G ---- CC/TTAG sequences (Fig. 3),

conforming to the GT - AG rule (31).

The Met initiation codon is located in exon II (residues 58-60,

solid triangle, Pig. 3), characterized by the consensus initiation

Figure 4:

95

Identification of the transcription initiation site of
the marine CBP35 gene by primer extension assay. A) The
primer was synthesized to be complementary to the
sequence of exon II (Figure 3), including four nucleo-

tides of the S'untranslated region, the AUG initiation

codon, anui 15 additional nucleotides of the coding

region. In The product of the primer extension reaction

was separated on a 10% polyacrylamide gel, followed by

autoradiographyu jLane 1, jproduct of the extension

reaction using primer synthesized as shown in panel A.

Lane 2, product of the extension reaction using tRNA as

a control. The four lanes marked G, C, T, and A on the

right are size markers from a sequencing reaction. The
numbers on the right indicate the length of the nucleo-
tide that would migrate to the position on the gel. The
number on the left highlights the length of the nucleo-

tide determined for the primer extended product.

 

v'
~~~~~

fl.
L3?

30'

96

 

 

A 0 3‘20 1990 nucleotides
5' 3
mRNA {ll—Tl COWNG REGION j UT DOW A

 

     

GAAAAUGGCAGACAGCUUUUCG
CTTTTACCGTCTGTCGAAAAGC

 

4‘v‘v“
prime:
8 12 GCTA
.. '11. ‘
If” i" 1
t: '
:fl '2‘; 4 10 0
3...:
“hf
f'!
72 .' . a” ‘ 70
,7 -<6o
Figure 4

 

97

sequence A/GNN-ATG-G. From this initiation codon through the
termination codon at positions 851-853 (inverted triangle, Fig. 3) ,
there is a translation Open reading frame coding for a polypeptide
of 264 amino acids. Thus, the coding sequence accounts for 795
nucleotides. As discussed above, the 5’ untranslated region of the
mRNA contains 57 nucleotides. In the 3’ untranslated region, there
is an AATAAA sequence (rectangle highlighted by dotted underline in
Fig. 3), characteristic of the consensus polyadenylation signal
(32), located 103 nucleotides downstream from the translation
termination codon. Although we have an additional 64 nucleotides
of sequence information, in terms of the genomic sequence, we have
not determined the precise site of polyadenylation. Assuming the
typical distance between the polyadenylation signal and the site of
poly A addition to be ~ 30 nucleotides, the size of exon V is

estimated to be 290-323 bp.

As indicated previously, the 5' end of the CBP35 gene was an adenine
residue (circled and labeled +1 in Fig. 3), deduced to be the
initiation of transcription site on the basis of a consensus
sequence for mRNA capping, CTTCCG, located six residues downstream
(rectangle highlighted by a wavy underline in Fig. 3). At -34
through -26, a TATA box-like sequence, AATATATAT (rectangle in Fig.
3), was found. This agreed well with the customary location of such
a promoter sequence (31,33) . Approximately 80 nucleotides upstream
from the adenine residue described above, the sequence CCAAT {-86
to ~82) was found (rectangle highlighted by a single solid underline

in Fig. 3).

In previous studies, we had demonstrated that the expression of the

gene for CBP35 was stimulated upon addition of serum to serum-

;'.

98

starved, quiescent cultures of 3T3 fibroblasts (29,30). This
stimulation resulted in an increase in the nuclear transcription of
the CBP35 gene and in the accumulation of its mRNA, early in the
activation process. Moreover, both of these increases were not
dependent on de novo protein synthesis, inasmuch as they occurred
even in the presence of cycloheximide. From studies on the genomic
sequences of several serum-stimulated genes, a regulatory element
designated as Serum Responsive Element (SRE) had been identified
(34,35) . The SRE is characterized by the consensus sequence, C—C-A-
A/T-A-T—A/T-A/T-G—G, to which certain transcription factors bind.
A search for a.possible SRE in the genomic sequence of CBP35 derived
from the present study resulted in a candidate; this sequence
CCAATTAAGG is located at positions -213 through -204 (rectangle

highlighted by a double underline in Fig. 3).
Comparisons of the Sequences of CBP35 with L-34, Mac-2, and eBP

We had previously shown (19) that the amino acid sequences of CBP35
and.eBP were identical in 223 out of the 262 positions compared (85%
identity). Analyses of antibody cross-reactivity and.demonstration
of carbohydrate-binding activity in sBP established that CBP35 and
SBP were mouse and rat homologous, respectively. Since that report,
the sequences of two other mouse proteins have been published, L-34
(15) and.Mac-2 (16). The nucleotide sequences of CBP35 and.L-34 are
compared in Figure 5, starting at the 5’<end with the translation
initiation codon through all the nucleotide sequence information
available at the 3Quui. There were very few differences between
CBP35 and L-34 at the nucleotide level. However, because the
sequences are best aligned by a few insertions/deletions (Fig. 5),

the reading frame changes between the proteins. This results in

 

Figure 5:

99

Comparison of the nucleotide sequence of the coding
region of the CBP35 gene with the corresponding sequence
reported for L-34. The sequence for CBP35, in capital
letters, was taken from the cDNA sequence (reference 6)
and was revised on the basis of the present work on the
genomic sequence. The sequence for L-34 was taken from
reference 15. A dot at any position indicates that the
sequence of L—34 was identical to that of CBP35; where
the sequence of L-34 differed from that of CBP35, the L-
34 sequence is shown in lower case lettering. Certain
deletions in either the CBP35 sequence (hyphen) or the
L-34 sequence (underline) were included in this compari—

son in order to maximize the alignment of identities.

 

100

CBP35 ATGGCAGACAGCTTTTCGCTTAACGATGCCTTAGCTGGCT

L—34

.......... cg
CTGGAAACCCAAACCCTCAAGGATATCCGGGTGCATGGGG

........................................
........................................
........................................
........................................

........................................

CCTGGCCCAACTGCCCCTGGAGCTTATCCTGG- TCAACCT
................................ C..--
GCCCCTGGAGCCTTCCCAGGGCAACCTGGGGCACCTGGGG

........................................

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
........................................

oooooooooooooooooooooooooooooooooooooooo
........................................
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

oooooooooooooooooooooooooooooooooooooooo
.........................................
........................................

oooooooooooooooooooooooooooooooooooooooo

AGAAGGGGCGGCACCGAAAC-—GCCCTGTGTCCTTAGGAT

.................... cg
GGGAAACTTGGCATTTCTCTCTCCTTATCCTTCTTGTAAG

........................................

ACATCCCCCATTTAATAAAGTCTCATGG—GAGAGAGAGCC
............................ t....a.ag
ATGTTTTGGGGGTTTTTATGATATGGGTTCAAATTCTTT

GGAC

Figure 5

 

101

clusters of differences at the level of the amino acid sequences,

as summarized in Table I.

In the study on the Mac-2 antigen (16), it was reported that the
sequence was identical to that of CBP35 in the coding region, with
the exception of the second amino acid (Ala for Mac—2 and Arg for
the previous report on CBP35). This difference was attributed to
a sequencing error (16). Our present genomic nucleotide sequence
confirms this notion and, therefore, our revision of the second
amino acid to an Ala residue makes the CBP35 and Mac-2 coding
sequences identical. We also infer, then, that the Mac-2 sequence
differs from.the sequence of L-34 at the same positions as summa—
rized in Table I. In the coding regions of all four structures
(CBP35, Mac—2, L—34, and EBP), there was no sequence obviously
corresponding to a signal peptide for sequestration into the
endoplasmic reticulum (36). ZNo potential-N-linked carbohydrate

attachment sites (Asn-XXXeSer/Thr) were found.

In contrast to the high degree of identity/homology observed in the
coding regions Iof CBP35, L-34, Mac-2, and SBP, the nucleotide
sequence of the CBP35 S'untranslated region is quite distinct from
those of the corresponding region in the mRNA of L-34, Mac-2, and
eBP (Fig. 6). These latter three mRNAs share a high degree of
identity for the sequences available in the 5'untranslated.regions.
The reason for this difference between the CBP35 mRNA and the other
three mRNAs is not clear. Oligonucleotides synthesized on the basis
of specific sequences in L-34 and in Mac—2 mRNA (Fig. 6) failed to
hybridize to the 6.4 kb EcoRI—EcoRI fragment and to the 2.7 kb

EcoRI-HindIII fragment of the KGH clone (Fig. l).

102

 

 

 

 

 

 

 

 

 

 

Table 1: Differences in the Amino Acid Sequencer of CBP35 Versus L-34
Residue Number ._ CBP35! V . ., L-3 41,-: .. _
4 3.. ' m ‘
92 Gln Ser
93 . Pro Thr
110 Gln Ser
“1 Cys Ala
112 Set pm
219 01“ Ala
221 A51’ Giu
222 His Pro

 

 

 

 

 

' The sequence for CBP35 was taken from reference (6) and was revised by adding NHZ-
terminal Met and by subsrituu'ng Ala for Arg in the second residue.

b The sequence for L-34 was taken from reference (15).

 

Figure 6 :

103

Comparison of the nucleotide sequences of the 5’ un-
translated region of L-34, Mac-2, EBP, and CBP35. The
numbering system, shown on the right, is based on the
ATG translation initiation codon; thus, the A residue
immediately upstream ix) the 5’ untranslated region is
numbered -1. The reference structure is the sequence
reported for L-34 (15), which extends 144 nucleotides 3
to the translation initiation codon. The end points of
the sequence information available from the Mac-2 (16)
and eBP (18) cDNA clones are indicated by the brackets
and their position numbers. In those positions for
which sequence information is available for Mac-2 and
EBP, a. blank space indicates that the sequence is
identical to that of L—34. At any position, only
nucleotides that differed from the L-34 reference are
shown. Certain deletions (hyphens) were introduced to
maximize the alignment of identities. The A sign
between residues —4 and -5 in the Mac-2 sequence indi-
cates the position where a 27 nucleotide insert has been
found in the 5' untranslated region of certain Mac-2
cDNAs. The shaded region denotes the sequence on the
basis of which oligonucleotide probes specific to the 5'
untranslated region of L-34 and Mac—2 mRNAs were synthe-
sized. The nucleotide sequence of the 5"untranslated
region of the CBP35 mRNA, deduced from the genomic
sequence obtained.in this study, is considerably differ”
ent from.those of L—34, Mac-2, and sBP, and is displayed

in its entirety.

104

o wusmam

 

«aeuaeeeoeeeoeoeoaooaeeoeeoaoe mmmmo
mmw

< NIn§3a

H- «edema. emu
edoooaeeodaeoeoooeeooaeaeefi5.. 3.80

I 5 .3: $8
vooooooooooxxxxxgfi o «-32

am. oeaeoeedodeeoauudoooeoaeooedeo emua
...H .2... «now:

3. oooooaeoeeoaaeeodeeaeoeoooooee «mun
3- ooooaoeeeooooeeoeedoeoaoooodoe «mad

.5.
HS- eoooaeeooodooeeeooooeaeou3? emuq

. .5 ... . a ”a . i
. \ . -

fit .3 . " .... 11- f. 2 r5; 24
\in . a .1 n h .~ . A

105

DISCUSSION

The results documented in this study indicate that one functional
murine CBP35 gene spans ~ 9 kb of genomic DNA. Five exons,
separated by four intervening sequences, have been identified" The
nucleotide sequences of exons II through V provide the amino acid
sequence for the coding region and are in good agreement with the
structure derived from the cDNA clone for CBP35 (6), as well as the
corresponding information for L-34 (15), Mac-2 (16), and.EBP (18).
The complete nucleotide sequence of IVS4 has been determined.2 In
addition, partial sequences of IVSl were obtained by sequencing the
first 100 nucleotides, or so, of directionally deleted fragments of
the 2.7 kb EcoRI—HindIII fragment of XGH. The sequences of these
introns failed.to show any identity/homology to each other or to any

structures in the gene bank.

Besides confirming the coding region structure, the genomic sequence
also provided important information on the regulatory elements
upstream from the site of initiation of transcription of the CBP35
gene. Both a TATA box—like sequence and a CCAAT box were found, at
positions expected for these structures on the basis of a knowledge
of previously characterized genes (31,33). The genomic sequence
also revealed a candidate for a SRE-like structure and, therefore,
a.possible binding site for serum-responsive transcription factors.
Such SRE and associated factors have been defined in certain genes,
such as c-fos and B-actin, that are activated in quiescent cells
upon serum addition (34,35). Addition of serum to serum-deprived

3T3 fibroblasts increased.the expression of CBP35 as shown by: (a)

106
analysis at the single cell level by immunofluorescence (5), (b)
analysis at the protein level by Western blotting (5) ; (c) analysis
at the mRNA level by Northern blotting (29,30); and (d) analysis of
the transcription of the CBP35 gene in nuclear run-off experiments
(30). The sequence in CBP35, CCAATTAAGG, differs from the consensus
SRE sequence, C-C-A-A/T—A—T-A/T~A/T-G-G (34,35), at one position.
An A residue in position 5 of the 10 nucleotide consensus is

replaced by a T residue in the CBP35 gene.

Of particular interest in this genomic sequence is the 5’ untrans—
lated region of the mRNA. This is because of the fact that,
although CBP35, L—34, Mac-2, and EBP are all predominantly intracel-
lular proteins, there is also evidence that L-34 and Mac—2 are found
on the outside of the cell. In this context, the situation is very
similar to that reported for probasin, a rat prostatic protein that
was found both in secretions and in the nucleus of prostatic
epithelial cells (37) . Only one probasin mRNA could be detected by
primer extension and $1 nuclease protection assays. In vitro
translation of this mRNA demonstrated that a protein containing a
signal sequence (36) and a protein lacking a signal sequence were
synthesized by initiation at different AUG codons. Since the coding
regions of C8935, L-34, Mac-2, and eBP were all the same, it was
hoped that the 5’ untranslated region might offer a hint concerning
the mechanism of dual localization of the protein, intracellularly

and extracellularly.

Indeed, Cherayil et a1. ' (16) have demonstrated the possible
existence of two alternatively spliced Mac—2 cDNAs, one of which can
use CTG (amino acid Leu) as an initiation codon and can encode an

additional 13 amino acids before the classical Met translation

107
initiator. This NHz-terminally extended segment contains uncharged
residues and could possibly serve as a signal peptide (36). Our
present genomic sequence results have not generated sufficient
information either -to confirm or to refute the hypothesized
alternative splicing events. ‘This is because, unlike the high
degree of identity/homology observed in the coding regions of CBP35,
L-34, Mac-2, and.EBP, the sequence of the 5'untranslated region of
the CBP35 mRNA, predicted from the presently available genomic
information, is totally different fronlthose reported for L-34, Mac-
2, and eBP. It does not appear that the sequences in the 5'
untranslated regions of these latter cDNAs are in our 1GB clone,
inasmuch as oligonucleotide probes synthesized according to these

sequences failed to hybridize to the clone.

In light of these results, it is important to consider whether the
genomic clone that we have sequenced actually codes for a functional
CBP35 gene. The evidences that are brought to bear include: (a)
The amino acid sequence is identical to those reported for Mac-2;
no substitutions were found in the coding region. (b) All
exon/intron boundary sequences are in good agreement with the GT-AG
rule. (c) The relative positions of the putative TATA.box and CCAAT
box correspond well with similar sequences of other genes. (d) A
polyadenylation signal was found in the 3’untranslated.region. “fl
There is a putative SRE-like sequence, which may account for the
observed behavior of the CBP35 gene upon serum stimulation of
quiescent cells. (f) Most importantly; we have obtained preliminary
evidence that an oligonucleotide complementary to nucleotide
residues 44-53 and residues 54-63 (i.e. it bridges exon 1 and<exon 2
in Figure 3) hybridizes to the 1.3 kb mRNA identified by our cDNA

clone in Northern blots (29,30). This is consistent with the

108
results of our primer extension assay and lends strong support to
the notion that the 2.3 kb IVSl is spliced out in the processing of

the mRNA.

10.

11.

12.

13.

14.

109

REFERENCES

Roff, C.F., and Wang, J.L. (1983) J. Biol. Chem. 258, 10657-
10663.

Crittenden, S.L., Roff, C.F., and Wang, J.L. (1984) Mol.
Cell. Biol. 4, 1252-1259.

Drickamer, K. (1988) J. Biol. Chem. 26;, 9557-9560.
Moutsatsos, I.K., Davis, J.M., and Wang, J.L. (1986) J. Cell
Biol. 192, 477-483.

Moutsatsos, I.K., Wade, M., Schindler, M., and Wang, J.L.
(1987) Proc. Natl. Acad. Sci. USA 84, 6452-6456.

Jia, S., and Wang, J.L. (1988) J. Biol. Chem. 263, 6009-
6011.

Gitt, M.A., and Barondes, S.H. (1986) IProc. Natl. Acad. Sci.
USA 83, 7603-7607.

Southan, C., Aitken, A., Childs, R.A., Abbott, W.M., and
Feizi, T. (1987) FEBS Lett. 214, 301-304.

Hirabayashi, J., Kawasaki, H., Suzuki, K., and Kasai, K.
(1987) J. Biochem. (Tokyo) 191, 775—787.

Paroutaud, P., Levi, G., Teichberg, V.I-, and Strosberg, A.D.
(1987) Proc. Natl. Acad. Sci. USA 84, 6345—6348.
Hirabayashi, J., and Kasai, K. (1988) J. Biochem. (Tokyo)
194, 1-4.

Clerch, L.B., Whitney, P., Hass, M., Brew, K., Miller, T.,
Werner, R., and.Massaro, D. (1988) Biochemistry 21, 692~699.
Wilson, T.J.G., Firth, M.N., Powell, J.T., and Harrison, F.L.
(1989) Biochem. J. 261, 847—852.

Couraud, P—O., Casentini-Borocz, D., Bringman, T.S., Griffith,

J., McGrogan, M., and Nedwin, G.E. (1989) J. Biol. Chem.

15.

l6.

l7.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.
29.

30.

110
264, 1310—1316.

 

Raz, A., Pazerini, G., and Carmi, P. (1989) Cancer Res. 48,
3489-3493.

Cherayil, B.J., Weiner, S.J., and Pillai, S. (1989) J. Exp.
Med. 110, 1959-1972.

Woo, H-J., Shaw, L.M., Messier, J.M., and Mercurio, AM.
(1990) J. Biol. Chem. 265, 7097-7099.

Albrandt, K., Orida, N., and Liu, F—T. (1987) Proc. Natl.
Acad. Sci. USA.§4, 6859-6863.

Laing, J.G., Robertson, M.W., Gritzmacher, C.A., Wang, J.L.,
and Liu, F-T. (1989) J. Biol. Chem. 264, 1907-1910.
Gritzmacher, C.A., Robertson, M.W., and Liu, F—T. (1988) J.
Immunol. 141, 2801—2806.

Feinberg, A.R., and Vogelstein, B. (1983) Anal. Biochem.
13;, 6—13.

Woo, S.L.C. (1979) Methods Enzymol. 68, 389-401.

Messing, J. (1983) Methods Enzymol. 191, 20—78.

Sanger, P., Nicklen, S., and Coulson, A.R. (1977) Proc.
Natl. Acad. Sci. USA 14, 5463-5467.

Sambrook, J., Fritsch, E.F., and. Maniatis, T. (1989)
Molecular Cloning: A Laboratory Manual, Cold Spring Harbor
Laboratory, Cold Spring Harbor, NY.

Henikoff, S. (1984) Gene (Amst.) 28, 351-362.

Southern, E.M. (1975) J. Mol. Biol. _8, 503-508.

Kozak, M. (1986) Cell 44, 283-292.

Jia, S., Mee, R.P., Morford, G., Agrwal, N., Voss, P.G.,
Moutsatsos, I.K., and Wang, J.L. (1987) Gene (Amst.) g9,
197—204.

Agrwal, N., Wang, J.L., and‘Voss, P.G. (1989) J. Biol. Chem.

264, 17236-17242.

 

31.

32.

33.

34.
35.

36.

37.

111

Breathnach, R., and Chambon, P. (1981) Ann. Rev. Biochem.

£2. 349-383.

Proudfoot, N.J., and Brownlee, G.G. (1976) Nature (London)

263, 211-214.

Cordon, J., Wasylyk, B., Buchwalder, A., Sasson—Corsi, P.,

Kedinger, C., and Chambon, P. (1980) Science 209, 1406-1414.

Treisman, R. (1986) Cell 46, 567-574.

Gustafson, T.A., Miwa, T., Boxer, L.M., and Kedes, L. (1988)

Mol. Cell Biol. g, 4110-4119.

Briggs, M.S., and Gierasch, L.M. (1986) Adv. Protein Chem.

is, 109—180.

Spence, A.M., Sheppard, P.G., Davie, J.R., Matuo, Y., Nishi,

N., McKeehan, W.L., Dodd, J.G., and Matusik, R.J. (1989).

Proc. Natl. Acad. Sci. USA 86, 7843—7847.

112

CHAPTER V

CONCLUDING REMARKS

When this thesis project was initiated, CBP35 had been purified and
characterized in terms of its carbohydrate-binding specificity.
There were also new and potentially interesting data pertaining to
the proliferation-dependent expression of the protein and its sub-
nuclear localization. In quiescent 3T3 fibroblasts, there was a low
level of CBP35, predominantly in the cytoplasm; stimulation of the
same cells resulted in an increased level of the protein and an
apparent translocation into the nucleus. On the basis of these
results, molecular cloning of CBP35 was undertaken: (a) to obtain
a reliable and. abundant source of the material for chemical
characterization (e.g. post-translational mmdification); (b) to
develop a probe for studying the expression of the CBP35 gene; and

(c) to obtain information on the primary structure of the protein.

The identification of the cDNA clone for CBP35 (Chapter II of the
thesis) provided the key stepping stone to the accomplishment of the
above three objectives. Recombinant CBP35 has been expressed in g.
.ggli, purified to homogeneity in milligram quantities, and has been
particularly important in determining the isoelectric variants of
CBP35, which in turn have shed light on the regulation of nuclear
versus cytoplasmic localization of the lectin (1). The cDNA clone
for CBP35 has also served as an invaluable probe for the analysis
of the regulation of the expression of CBP35 gene, in terms of rate
of transcription, accumulation_ofrmuUM and "super induction":fi1the

presence of cycloheximide (2).

 

 

113

The amino acid sequence of CBP35, deduced from the nucleotide
sequence of the cDNA clone (Chapter III of the thesis), indicated
that the polypeptide could be delineated into two domains: an NHf-
terminal domain that was proline- and glycine-rich and a C00!!-
terminal CRD. This result provided the first example of a S-type
lectin having two domains, a structural motif that had been well-
demonstrated in the C-type lectins (3). Thus, the primary structure
of CBP35 added an important dimension to the picture of the S-type
lectins: that the CRD of a S—type lectin can be coupled to another
functional domain (such as has been found for C-type lectins). In
the particular case of CBP35, the effector domain showed a highly
unusual internal repeat of 9-amino acid.residues ( Pro-Gly-Ala-Tyr-
Pro-Gly-Xxx-Xxx-Xxx), as a result of which there is an uneven
distribution of Pro and Gly residues in the CBP35 polypeptide. Both
features (internal repeat.and uneven distribution of Pro and Gly)
have been found in other nuclear proteins. Of particular interest
were the available structures of the hnRNP core proteins because
immunochemical studies parallel to my structural determination had

implicated the associa-tion of CBP35 with nuclear hnRNP (4).

Upon completion of the sequence studies on the cDNA clone, there
were discussions on carrying out mutational analysis on CBP35 to
address certain questions related to structure, activity, and/or
localization. For example, deletional mutation might be carried.out
to express only the NHz—terminal domain and to test whether this
determines the fate of localization of the polypeptide (cytoplasm,
nucleus, run? complex). Alternatively, tine COOH—terminal domain
might be expressed to test whether it formed dimers, since all the
14—16 kD group of S—type lectins had been reported to dimerize in

non-denaturing buffers while lectins of the 29—35 kD group were

114

reported to remain monomers.

The focus of my research, however, remained.at the structural level,
mainly as a result of two developments in the field. First, struc-
tural studies, antisera cross—reactivity, and tests of carbohy-
drate-binding activity on Mac-2, and eBP showed that these other
proteins, studied for their possible cell surface function, were
identical to CBP35 (5, 6). Second, analysis of cDNA sequences 5’
to the position where our cDNA clone began failed to identify an
obvious signal sequence for the endomembrane pathway for secretion.
Thus, the question in the field.at this juncture was: how to explain
the dual (intracellullar and extracellular) localization of CBP35,
L34, Mac-2, and eBP, all which are identical (homologous) in coding
region sequences. One possibility was the existence of two (or
more) alternatively spliced.RNAs, one of which has the potential to

code for a signal sequence containing polypeptide(6).

The genomic sequence analysis (Chapter IV of the thesis) was under-
taken with the hope of shedding light or providing confirmation of
these splicing events. The presently available genomic sequence
information does not, however, provide sufficient information to
either confirm or refute the hypothesis. This must remain as a

challenge for future studies.

 

 

 

115

References

l. Cowles, E. A., Agrwal N., Anderson, R. L., and Wang, J. L. (1990)
J. Biol. Chem. in press.

2. Agrwal, N., Wang, J. L., and Voss, P. G. (1989) J. Biol. Chem.
264, 17236-17242.

3. Drickamer, K. (1988) J. Biol. Chem. 263, 9557—9560.

4. Laing, J. G., and Wang, J. L. (1988) Biochem. 27, 5329-5334.
5. Laing, J. G., Robertson, M. W., Gritzmacher, C. A., Wang, J. L.,
and Liu, F.-T. (1989) J. Biol. Chem. 264, 1907-1910.

6. Cherayil, B. J., Weiner, S. J., and Pillai, S. (1989) J. Exp.

Med. 170, 1959-1972.

 

NSTRT E UNIV

||TlllWlm\IIHIIIIWWIDINIIN”IMIVIHIHMinl