Two Lactose Binding Lectins from Chicken Tissues PURIFIED LECTIN FROM INTESTINE IS DIFFERENT FROM THOSE IN LIVER AND MUSCLE*

(Received for publication, November 16, 1979, and in revised form, January 15, 1980)

Eric C. Beyer,f Stephen E. Zweig, and Samuel H. Barondesg From the Departments of Psychiatry and Biology, University of California at Sun Diego, La Jolla, California 92093

Endogenous lactose-binding proteins from adult chicken liver and intestine have been purified to ho- mogeneity by affinity chromatography on asialofetuin- derivatized Sepharose, followed by isoelectric focusing. Since these carbohydrate-binding proteins are assayed as hemagglutinins, they are referred to by the opera- tional t e rm lectins. Although the hemagglutination ac- tivity of these lectins is inhibited by similar concentra- tions of inhibitory saccharides, they are not the same. They differ in specific hemagglutination activity, sub- unit molecular weight, and isoelectric point. They have very different peptide m a p s and show no detectable immunological cross-reactivity. Based on gel filtration studies, the liver lectin behaves as a dimer with appar- ent M, = 31,000 +- 1100, whereas the intestinai lectin behaves as a monomer with apparent M, = 14,000 +- 1700. Although clearly different from the intestinal lec- tin, the lectin from adult liver appears identical with the lectin previously purified from embryonic skeletal muscle.

Extracts of a number of embryonic and adult chicken tissues contain carbohydrate-binding proteins which resemble plant lectins in that they can be assayed as saccharide-inhibitable hemagglutinins. Because of these similarities the operational term lectin is also used for the chicken proteins. The lectin from embryonic muscle has been purified to homogeneity (1, 2) and shown to be a dimer composed of subunits with a molecular weight of about 15,000 which reacts with lactose and related saccharides. The lactose-binding lectins from em- bryonic brain and liver (3), and adult liver (4), appear to be identical in that they are dimers of the same subunit molecular weight and isoelectric point and appear to be antigenically identical. Recently, we found that adult chicken intestine also contains a lactose-binding lectin which we assumed to be identical with the others (4). In the present report, we show that the purified lactose-binding lectin from chicken intestine differs strikingly from the lactose-binding lectins of both adult chicken liver and embryonic chick muscle. In contrast, further studies with the latter two lectins failed to turn up any differences.

* These studies were supported by Research Grant MH 18282 from the United States Public Health Service and by a grant from the McKnight Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “uduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. + Medical Science Training Program Trainee supported by United States Public Health Service Training Grant GMO 7198.

5 This manuscript was prepared while S. H. B. was a Fogarty- Scholar-in-Residence at the National Institutes of Health, Bethesda, Md. Please address all correspondence to Dr. Barondes at: Depart- ment of Psychiatry (M-003), University of California at San Diego, La Jolla, Calif. 92093.

EXPERIMENTAL PROCEDURES

Livers or intestines from adult White Leghorn hens were homoge- nized for 1 min in a Sorvall Omni-Mixer at 4°C in 9 volumes of mercaptoethanol/NaCI/P,’ containing 0.3 M lactose. The homogenate was centrifuged a t 100,OOO X g for 1 h and the supernatant was dialyzed and applied to an asialofetuin-Sepharose affinity column prepared as described previously (4).

Hemagglutination activity of extracts and purified proteins was assayed by serial 2-fold dilutions with trypsin-treated, glutaraldehyde- fixed rabbit erythrocytes by methods similar to those described pre- viously (1). Protein was determined by the method of Bradford (5). Lectin specific activity is defined as the reciprocal of the titer of a sample divided by milligrams of protein/ml of the sample. Isoelectric focusing was performed for 4 days a t 4OC in an LKB isoelectric focusing apparatus as described previously (1) . Gel fdtration on Seph- adex G-100 was performed on a column (1.5 X 30 cm) equilibrated with mercaptoethanol/NaCl/P, containing 0.3 M sucrose. Polyacryl- amide gel electrophoresis in SDS was performed with a 15% poly- acrylamide sample gel using buffer systems and procedures like those described previously (1). Production of antiserum to the liver lectin and demonstration of its specificity have been described previously (4). For the present studies, a y-globulin fraction was prepared by precipitation of the serum in 33% saturated ammonium sulfate. Map- ping of ‘”I-labeled tryptic peptides was performed by iodination of bands from polyacrylamide-SDS gels and tryptic digestion according to the methods of Elder et al. (6) as modified by Zweig and Singer (7).

RESULTS

Lectin Purification-Lectin activities from liver or intestine homogenates bound quantitatively to the asialofetuin-seph- arose column and were quantitatively eluted with 0.3 M lac- tose. Purifications were on the order of 1600- and 400-fold for the liver and intestine, respectively (Table I). Upon polyacryl- amide gel electrophoresis in an SDS system, each lectin gave rise to a single major protein band (Fig. 1). The liver prepa- rations had either one or two very minor protein bands of lower molecular weight (Fig. 1). Some intestinal preparations had some trace contaminants of higher molecular weight. Further purification was achieved by isoelectric focusing. After this procedure, each lectin preparation showed a single peak of activity (Fig. 2). The lectin from liver had an isoelectric point of approximately 4.1, similar to that reported for the lectin from embryonic muscle (1). The minor bands seen in Fig. la were no longer seen after isoelectric focusing (Fig. lb; overloaded gels not shown). All lectin activity was associated with the peak consisting of the single, major band, suggesting that the minor bands which were removed are inactive. The lectin from intestine clearly differed from the liver lectin since it had an isoelectric point of approximately 6.3 (Fig. 2). During isoelectric focusing there was some loss of lectin activity, as observed previously with the lectin from embryonic chick muscle (1).

Molecular Weight Estimation-The subunit molecular ’ The abbreviations used are: mercaptoethanol/NaCl/P,, 75 mM

NaC1, 75 mM Na2HP04/KH4P0, pH 7.2, containing 4 mM P-merCaP- toethanol, 2 mM EDTA; SDS, sodium dodecyl sulfate.

4236

Different Lactose Binding Lectins in Chicken Tissues 4237

TABLE I Lectin purification

Specific activ- Purification itv Recovery

Liver Crude extract After affinity chro-

matography After isoelectric fo-

cusing Intestine

Crude extract After affinity chro-

matography After isoelectric fo-

cusing " ~~~~ ~~~

100 1 100 160,000 1.600 100

250.000 2.500 75

16 1 100 6,400 400 100

10,OOO 625 40

FIG. 1. Polyacrylamide gel electrophoresis in SDS of puri- fied lectins from adult chicken liver and intestine. n, liver lectin after affinity chromatography on asialofetuin-Sepharose. Arrows in- dicate faster migrating faint minor hands. h. liver lectin after further purification by isoelectric focusing. c, intestine lectin purified hV affinity chromatography on asialofetuin followed by isoelectric focus- ing. d , intestine lectin after affinity chromatography. For molecular weight determinations the standards used were: lysozyme, M , = 14,300, a-lactoglobulin, M , = 18.400; trypsinogen. M , = 24,000, pepsin, M , = 34,700, ovalbumin, M, = 45,000, bovine serum albumin, M , = 66,OOO.

weight.. of the lectins based on four independent SDS-poly- acrylamide gel electrophoresis studies, using two different preparations of each purified lectin, were estimated to be 15,900 f 400 and 14,200 & 300 (8 f S.E.) for the liver and intestine, respectively. In other studies, the liver lectin was found to co-migrate with purified lectin from embryonic mus- cle when the two were mixed in equal amounts.

The liver and intestine lectins could also be distinguished by their behavior during gel filtration (Fig. 3). The protein concentration in these fractions roughly paralleled lectin ac- tivity. All protein found was associated with the peak of lectin activity. The fact that this parallelism was not precise as well as the asymmetrical pattern of the lectin activity (for example, the intestinal lectin pattern seen in Fig. 3) is due to the inherent error in the hemagglutination assay which estimates activity by 2-fold dilutions. Based on four experiments using two different preparations, the apparent molecular weight of the native liver lectin was 31,000 2 1100 and the apparent molecular weight of the native intestinal lectin was 14,000 f 1700 (8 f S.E.). Identical results were found in one gel Wtration experiment using Bio-Gel P-60. These results suggest that the liver lectin is present in solution primarily in dimeric

"F a

8

LIVER

1

6 I E5

4

3

7

I

FRACTION

.i INTESTINE

A 7 - I

1 6 - AA

AA 5

= 5 - /AAAA I

4 - MAAAAA I I I

3 - A

I10 - 110 - IW - * - lo-

10 - 80-

"

60- 10 - IO - 10 - 0 2

5 1 0 15 20 25 30 35 40 45 50 55

FRACTION

FIG. 2. Isoelectric focusing of (a) liver lectin eluted from asialofetuin-Sepharose and (6) intestine lectin eluted from asialofetuin-Sepharose.

600 700 [

g 200

I t 100

1 5 9 13 17 21 25 29 33 31 41 45

FRACTION

FIG. 3. Sephadex G-100 gel filtration of purified lectins from liver (0) and intestine (0). For molecular weight determinations the standards used were: cytochrome c, M , = 12.400: lysozyme. M , = 14.300, chymotrypsinogen. M , = 25 ,oW ovalbumin, M , = 45.000; bovine serum albumin, M , = 66,OOO.

form, while the intestinal lectin is primarily a monomer. The lectin from embryonic muscle has previously been shown to have a molecular weight of approximately ~10,000, based on gel filtration studies (1).

Comparative Specific Activities-The purified intestine

4238 Different Lactose Binding Lectins in Chicken Tissues

lectin was found to he a much less potent hemagglutinin than the liver lectin (Table I). We repeatedly observed that the specific activity of the purified liver lectin was 20 to 30 times greater than that of the purified intestine lectin. Pure prepa-

TAHLE I1 Kc4atirt~ potency of sctccharides in inhibiting hemagglutination

crctitlity ofpurified lectins Hemagglutination activity of lectins purified by affinity chroma-

tography was determined in the presence of a range of concentrations of the indicated compounds. The concentration that inhibited hem- agglutination activity by 50% is shown.

~~ ~~ ~ ~~~ ~~ ~~~ ~ ~ - ~. Concentration that inhibits 5Or; in:

Saccharide

Lactose Thiodigalactositie Galactose Methyl-c~-l)-galactoside Methyl-/~-I)-galactoside Melibiose h'-Acet.l-o-galactosamine N-Acetyl-l)-glucosamine Glucose Mannose

-~ ~-

~ ~- ~

Liver Intestine m .n

0.29 0.29 0.15 0.15 38 38 38 19 I J 75 19 19

>75 >75 >7:i >75 >75 >75 >75 >75

-r

~~~~ ~ ~ ~~. ~ ~ " . "~

' Ab I "

r

a

Flc:. 4. Immunodiffusion studies with antiserum raised against purified liver lectin. a, immunodiffusion of purified lectins from embryonic muscle ( M ) and adult liver ( L ) with the y-globulin fraction of antiserum raised against purified liver lectin (Ab). h. immunodiffusion of purified lectins from adult chicken liver ( L ) and intestine (0 with antiserum raised against the purified liver lectin (Ah). Well contents were: Ah. lop1 of y-globulin fraction of antiserum directed against purified liver lectin; I,, 0.2 p g of liver lectin after isoelectric focusing; I, 25 pg of intestine lectin after isoelectric focusing. These gels were equilibrated with 50 mM lactose to block binding of the lectins to the agarose or serum glycoproteins. Gels were stained with Amido black t o improve visualization of protein bands.

rations of the lectin from embryonic muscle had specific activities similar to that of the adult liver lectin.

Effect of Saccharides on Hemagglutination Activity of Purified Lectins-The most potent inhibitors of hemaggluti- nation of both lectins were lactose and thiodigalactoside (Ta- ble 11). The inhibitory potency of the sugars tested was similar for both proteins. These inhibitory concentrations are also similar to those reported for the embryonic muscle lectin (1).

Antibody Studies-We have previously raised a potent antibody to the purified liver lectin (4). Upon immunodiffu- sion, this antibody precipitated pure liver lectin (Fig. 4a). Embryonic muscle and adult liver lectin, both purified by affinity chromatography and isoelectric focusing, gave a re- action of immunological identity, each forming a single pre- cipitin line, fusing in a smooth arc, with no spurs (Fig. 4a). In contrast, the intestinal lectin failed to form a precipitin line with this antiserum, even when large amounts of the pure protein were applied to the gel (Fig. 4h) . The antiserum also discriminated between these lectins as determined by its effect on their hemagglutination activities (Table 111).

Examination for Artifactual Proteolytic Degradation- We tested the possibility that the intestinal lectin represented an artifactual proteolytic degradation product of a protein identical with the liver lectin by mixing pure liver lectin with a crude intestine homogenate and carrying the mixture through the initial purification steps. Presumably proteolysis could account for the much lower specific activity of the intestinal lectin. We found no loss of activity even when the mixture was allowed to stand for 1 h a t room temperature. We also found that purification of the intestine lectin in the presence or absence of the protease inhibitor phenylmethane- sulfonyl fluoride gave an identical product.

TABLE 111 Effect of antiserum raised against pure liL1er lectin on

hemagglutination activity of purified lectins A y-globulin fraction was prepared from immune Serum raised

against pure liver lectin and from nonimmune rabbit serum and added to wells of the hemagglutination assay plate at a dilution of 1:100. After addition of serial dilutions of lectin and test erythrocytes, the hemaeelutination activitv of the lectin was observed.

Specific agglutination activity

Saline control Immune y-globu- Nonimmune y- lin globulin

titer"/rngprotein/ml Liver 128,000 8,000 96,000 Intestine 3,200 3,200 3,200 Embryo muscle 85,000 2,000 85,000

-

a h

t c V W

W J

- CHROMlTOGRAPHY FIG. 5. 12511-labeled tryptic peptide maps of 12611-labeled purified lectins from adult intestine (a), adult liver (b), and embryonic

muscle (c). Note the identity of the maps for liver and muscle, both of which are very different from that for intestine. The maps were all prepared under identical conditions. The nrrou1.s indicate the directions of chromatography and electrophoresis. The origin was located at the lower right hand corner of the maps.

Different Lactose Binding Lectins in Chicken Tissues 4239

1251-labeled Tryptic Peptide Maps-We sought to compare the structural similarity of the lectins by comparing their "'I- labeled tryptic peptides. Maps of the iodinated tryptic pep- tides of adult liver and embryonic muscle lectins were identi- cal, but the intestinal lectin was completely different (Fig. 5 ) . Identity of the muscle and liver lectins was also shown by mixing the tryptic peptides and running them together. We also used this technique to examine the two minor, lower molecular weight bands present in SDS-polyacrylamide gel electrophoresis of the liver lectin purified only by affinity chromatography (Fig. la). The band with apparent M , = 14,700 gave rise to peptides which were strikingly similar to those of the pure liver lectin and completely different from those found with the pure intestinal lectin. This suggests that this band is derived from the major liver lectin protein either in uiuo or during isolation. In contrast, the faint third band, with apparent M , = 14,300, present in some preparations of the liver lectin gave a map which was faint and complex, suggesting that it was a mixture of unrelated proteins.

DISCUSSION

In this report we show that lactose-binding lectins from adult chicken liver and intestine have the following differ- ences: 1) the purified liver lectin has a much greater specific hemagglutination activity than the intestinal lectin; 2) the lectins have different subunit molecular weights; 3) on gel filtration the liver lectin behaves like a dimer, whereas the intestinal lectin behaves like a monomer; 4) the lectins have very different isoelectric points; 5 ) the tryptic peptide maps of the two lectins are strikingly different; 6 ) the two lectins are immunologically distinct. Whereas the intestinal and liver lectins are clearly different proteins, the lectins purified from adult liver in these studies and from embryonic muscle in previous studies (1) appear to be identical by all criteria we have applied.

The relatively low specific hemagglutination activity of the intestinal lectin might be related to the fact that it is isolated as a monomer. To act as an agglutinin, it may be necessary for two subunits to combine, but the dimeric form is appar- ently not stable. Alternatively the agglutinating form could be the monomer if it contained two carbohydrate-binding sites. The agglutination activity of the intestinal lectin that we observe is not due to contamination of the purified material with the liver type of lectin since the agglutination activity associated with the intestinal lectin is distinct from the liver lectin on gel filtration and isoelectric focusing.

In previous immunohistological studies (4) using antibody directed against purified liver lectin, we found staining of the secretory granules of intestinal goblet cells, suggesting that the liver form of lectin, or a material that is immunologically cross-reactive, is present in intestine. To our surprise, the present studies failed to show detectable quantities of the liver form of the lectin after purification of intestinal material through the isoelectric focusing step. However, we did find immunologically cross-reactive material (a band of identity without spurring by double immunodiffusion) upon reaction of antibody to the liver lectin with highly concentrated crude extracts of intestine. Since the immunohistological studies

showed striking localization of the antigen in secretory gran- ules, it is possible that this material was not significantly solubilized by the present homogenization procedure or re- mains tightly associated with an inhibitor. Attempts are pres- ently being made to investigate this possibility and to identify the immunologically cross-reactive material in intestine.

The present report extends to four the number of lectins identified in chicken tissues. In addition to the two discussed here, a lectin activity that reacts with both N-acetyl-D-galac- tosamine residues (8) and heparin (9) has been described and a lectin from chicken liver that reacts with N-aCetyl-D-glUCO- samine residues (10) and differs strikingly from those dis- cussed here has been purified. Establishing the role of these lectins in functions as diverse as cellular association (11) and glycoprotein degradation (12), with which they have been implicated, will require much more investigation. Given the diversity of chicken lectins found EO far, it is notable that lectins purified from several bovine tissues were all judged to be identical (13), as is the case with the lactose-binding lectins from chick muscle and liver. The existence of distinct proteins with apparently similar active sites, as shown in the present study, adds yet another intriguing property of vertebrate lectins which begs functional explanation.

The differences also raise a problem of nomenclature. It does not seem appropriate to use the term electrolectin (14) for all vertebrate lectins that react with lactose. This name was proposed for the related lectin from electric organ of the eel (14) and is sometimes also used for the lactose-sensitive lectin from chicken muscle. To distinguish the two lactose- sensitive lectins from chickens, we tentatively propose the term chicken-lactose-lectin-I for the form from embryonic muscle and adult liver and the term chicken-lactose-lectin-I1 for the intestinal protein described in this report.

REFERENCES

1. Nowak, T. P., Kobiler, D., Roel, L. E., and Barondes, S. H. (1977)

2. Den, H., and Malinzak, D. A. (1977) J . Bioi. Chenz. 252, 5444-

3. Kobiler. D., Beyer, E. C., and Barondes, S. H. (1978) Dell. Biol.

4. Beyer. E. C., Tokuyasu, K. T.. and Barondes, S. H. (1979) J. Cell.

5. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 6. Elder, J . H., Pickett, P. A,, 11, Hampton, J., and Lerner, K. A.

7. Zweig, S. E., and Singer, S. J. (1979) Biochern. Biophys. Res.

8. Mir-Lechaire, F. J., and Barondes, S. H. (1978) Nature 272,256-

9. Kobiler, D., and Barondes. S. H. (1979) Febs. Lett. 101, 257-261 10. Kawasaki, T., and Ashwell, G. (1977) J . Bid. C'hena. 252, 6Wi-

6543 11. Barondes, S. H. (1980) Cell Adhesion andMotility (Curtis, A. S.

G., and Pitts, J. D., eds) pp. 309-324, Cambridge Universit,v Press, Cambridge, England

12. Ashwell, G., and Morell, A. G. (1977) A&). Enzynol. Relnt. Areas Mol. Biol. 41, 99-128

13. Briles, E. B., Gregory, W., Fletcher, P., and Kornfeld, S. (1979) <J , Cell Bid. 81, 528-537

14. Teichberg, V. I. (1978) Methods h'nzymol. 50C, 291-301

J. B i d . Chem. 252,6026-6030

5448

64,265-272

Biol. 82, 565-571

(1977) J. Biol. Chern. 252, 6510-6515

Cornmun. 88, 1147-1152

258