purification to homogeneity of a β-galactoside α2→3 sialyltransferase and partial purification...

THE JOURNAL cm BKJLOCICAL CHEMISTRY Vol. 254, No. 11, Issue of June 10. pp. 4434-4443,1979 Printed in U.S.A.

Purification to Homogeneity of a /?-Galactoside cw2 --) 3 Sialyltransferase and Partial Purification of an a-N-Acetylgalactosaminide a2 + 6 Sialyltransferase from Porcine Submaxillary Glands*

(Received for publication, August 29, 1978)

J. Evan Sadler& James I. Rearick, James C. Paulson,@ and Robert L. Hill From the Department of Biochemistry, Duke University Medical Center, Durham, North Carolina 27710

Two different sialyltransferases (EC 2.4.99.1) have been resolved from T&on X-100 extracts of porcine submaxillary glands by affinity chromatography on CDP-hexanolamine agarose. The predominant sialyltransferase of this tissue, a CMP-N-acetylneuraminate: a-D-N-acetylgalactosaminide a2 + 6 sialyltransferase, has been obtained in a partially purified and stable form. A less abundant but highly active enzyme, a CMP- N-acetylneuraminate$-D-galactoside (~2 += 3 sialyltransferase, was purified over 90,000-fold to homogeneity. Chromatography of the latter enzyme on Sepha- dex G-200 separated two noninterconverting forms, designated A and B, with Stokes radii of 51 A and 31 A, respectively. Both forms have equal specific acitv- ity toward lactose and contain a single polypeptide with a molecular weight of about 50,000 as estimated by gel electrophoresis. Form A appears to bind 1.18 g of Triton X-100 per g of protein, or nearly an entire detergent micelle per polypeptide, while Form B binds little or no detergent. The enzymatic properties of both forms are similar (Rearick, J. I., Sadler, J. E., Paulson, J. C., and Hill, R. L. (1979) J. Biol. Chem 264, 4444- 4451) supporting the conclusion that Form A may represent the native sialyltransferase with an intact membrane-binding site, and Form B may be a large proteolytic fragment of Form A.

Sialic acids occur in glycoproteins and glycolipids in a wide variety of structures. Three linkage patterns, Siacu2 + 6Ga1,’

* This work was supported by Research Grant HL-06400 from the National Heart and Lung Institute, National Institutes of Health, and Grant GB-29334 from the National Science 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 “adver- tisement” in accordance with 18 USC. Section 1734 solelv to indicate this fact.

$ Supported by the Medical Scientist Training Program, Grant GM-07171, National Institute of General Medical Sciences. National Institutes of Health.

0 Present address, Department of Biological Chemistry, School of Medicine, University of California, Los Angeles, California 90024.

’ The abbreviations used are: Sia, any sialic acid; NeuAc, specifi- cally refers to the sialic acid, N-acetylneuraminic acid; CMP-NeuAc. cytidine 5’-monophospho-N-acetylneuraminic acid; GDP-Fuc, gua- nosine 5’-diphosphofucose; UDP-GalNAc, uridine 5’-diphospho-N- acetylgalactosamine; OSM, ovine submaxillary mucin; FaAc, trifluo- roacetyl; Mes, 2-(N-morpholino)ethanesulfonic acid; GalNAcot2 + 6 sialyltransferase, CMP-NeuAc:cu-N-acetylgalactosaminide (~2 + 6 sialyltransferase; Gala2 --f 3 sialyltransferase, CMP-NeuAc$-galactoside (~2 + 3 sialyltransferase; CDP-hexanolamine, P’-(6-amino-l- hexyl)-P*-(5’-cytidine) pyrophosphate. The prefix, asialo-, indicates

Siaa2 + 3Gal, and Siacu2 + GGalNAc, are commonly found in glycoproteins (l), and two, SiacuP + 3Gal and Siacu2 --+ 8Sia, occur frequently in gangliosides (2). However, with the excep- tion of the Siacr2 + 6GalNAc structure, which seem to be limited to glycoproteins, each of these structures has been described in both gangliosides and glycoproteins (2-4). Al- though a role for sialic acids has been proposed in the regu- lation of many biological phenomena, the purpose of this structural diversity remains largely obscure.

There appear to be many more sialyltransferase activities than there are sialic acid linkages; for example, at least three different enzymes that form the S&2 + 3Gal structure have been distinguished on the basis of their preferred acceptor substrates (5-8). This multiplicity of enzymes supports the proposal (9) that the distribution of sialic acid-containing compounds within a tissue is regulated in part by the strict substrate specificity of the enzymes in that tissue. However, the presence of multiple transferase activities in tissue extracts greatly complicates the study of enzymatic specificity, and extensive purification is a prerequisite to the unambiguous determination of substrate requirements. To date, only one sialyltransferase has been purified to homogeneity, a P-galactoside (~2 --, 6 sialyltransferase from bovine colostrum (10). Knowledge of its enzymatic properties has furthered the un- derstanding of sialylation patterns observed in the asparagi- nyl-linked oligosaccharides of glycoproteins (ll), and the enzyme has proved to be a valuable tool in the investigation of oligosaccharide structure and function (12, 13). Purified sialyltransferases with different acceptor substrate requirements are expected to be similarly useful as they become available.

This report describes the resolution of two sialyltransferases from porcine submaxillary glands (14) by means of affinity chromatography on CDP-hexanolamine agarose, an a-N-acetylgalactosaminide (~2 -+ 6 sialyltransferase and a ,&galactoside ~y2 + 3 sialyltransferase which catalyze Reactions 1 and 2, respectively.

CMP-NeuAc + GalNAccul + 0-Ser/Thr -

NeuAca2 -+ 6GalNAccul -+ 0-Ser/Thr f CMP (1)

CMP-NeuAc + Galpl -+ 3GalNAccrl+ 0-Ser/Thr -

NeuAccuP + 3Galj31+ 3GalNAccul+ 0-Ser/Thr + CMP (2)

The latter enzyme has been further purified to homogeneity, and some of its physical properties have been determined. The enzymatic properties of these transferases are reported in the succeeding paper (15). A preliminary account of this work has been presented (16).

that the glycoprotein has been treated with neuraminidase to remove sialic acid. All sugars mentioned are pyranoses and all except fucose have the D configuration.

4434

by guest, on July 24, 2011w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org/

Purification of Porcine Submaxillary Sialyltransferases 4435

RESULTS’

Specific Assays for Porcine Submaxillary Sialyltransferases

Initially, the purification of these enzymes was complicated by the lack of a truly specific acceptor substrate for the GalNAccu2 + 6 sialyltransferase. A presumed natural substrate, porcine submaxillary asialo-mucin, is a very good acceptor for the Gala2 -+ 3 enzyme as well as the GalNAccv2 + 6 enzyme due to the large fraction of Gal/?1 + 3GalNAccvl + 0-Thr/Ser among its oligosaccharides (36), and is, therefore, a nonspecific substrate.

The predominant oligosaccharide of ovine submaxillary mucin is NeuAco2 + 6 GalNAccrl + 0-Thr/Ser (37); thus, the asialomucin should be a specific acceptor for the GalNAccv2 + 6 sialyltransferase. However, small amounts of Gal/31 ---, 3GalNAc and Fuccvl + 2GalPl + 3GalNAc structures, even in very pure mucin, can be quantitated enzymati- tally utilizing the specificity of other purified glycosyltransferases, as shown in Table I. The A blood group-specific o-N- acetylgalactosaminyltransferase from porcine submaxillary glands (18) has been used as a probe for Fuccul + 2Gal structures, and the H blood group-specific fucosyltransferase from the same source (17) has been used to detect nonreducing terminal P-galactosides. The acceptors for the latter enzyme are removed by the endo-a-N-acetylgalactosaminidase from Streptococcus pneumoniae, demonstrating that their structure is Galpl + 3GalNAcol+ 0-Thr/Ser (38). These chains enable ovine submaxillary asialo-mucin to serve as an acceptor substrate for both sialyltransferases present in the porcine submaxillary gland, and this lack of specificity is aggravated by kinetic differences between the two enzymes which favor transfer by the Gale2 + 3 sialyltransferase (16), complicating the assessment of transferase purity. For example, under the conditions of standard assay I, but using 85 pg of asialomucin as the acceptor, over half of the sialic acid transfer by the crude Triton extract (Step 1 below) is due to the Gala2 + 3 sialyltransferase, even though it comprises only -25% of the total sialyltransferase units. Fortunately, digestion of ovine submaxillary asialo-mucin with the endo-cY-N-acetylgalactosaminidase from S. pneumoniae produces a substrate that is quite specific for the GalNAccv2 + 6 sialyltransferase, partic- ularly under the conditions of standard assay II.

For the routine assay of the Galru2 + 3 sialyltransferase, lactose was chosen over other potential acceptors because it is easy to obtain, even though it is a very poor acceptor compared to some other oligosaccharides (15). The bovine colostrum sialyltransferase (9) will also use lactose as a substrate, but a specific assay of the Galo2 + 3 enzyme is possible using lactose because there is no detectable P-galactoside a2 + 6 sialyltransferase to interfere (15). Units of enzyme have been expressed in terms of a better acceptor, antifreeze glycoprotein, by means of an arithmetic conversion factor (Table II).

Purification and Resolution of Porcine Submaxillary Sialyltransferases

The purification over 90,000-fold to homogeneity of the Gala2 + 3 sialyltransferase from 4 kg of porcine submaxillary

* Portions of this paper (including “Experimental Procedure” and “Appendix” on the calculation of some physical parameters of proteins from sucrose density centrifugation data) are presented in miniprint at the end of this paper. Some references are to be found in the miniprint. Miniprint is easily read with the aid of a standard magnifying glass. Full-size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, Maryland 29014. Request Document No. 78M-1543, cite author(s), and include a check or money order for $2.10 per set of photocopies.

TABLE I

Oligosaccharides of ovine submaxillary mucin and derivatives

Native ovine submaxillary mucin was digested with neuraminidase to produce asialo-OSM, which was further digested with the endo-o- N-acetylgalactosaminidase from S. pneumotkze to produce endosidase-treated asialo-OSM. as described under “Experimental Proce- dure.” The N-acetylgalactosamine (25, 26) and si-alit acid contents (24) of each derivative were determined calorimetrically after release by mixed Clostridium perfringens glycosidases (neuraminidase and a-N-acetylgalactosaminidase). This enzyme mixture contains no glycosidases active with Galpl + 3GalNAcol+ 0-Thr/Ser. Thus, total N-acetylgalactosamine is equal to the sum of enzymatically released N-acetylgalactosamine plus structures III and IV. Quantitation of structures III and IV was performed by enzymatic glycosylation using purified glycosyltransferases of appropriate specificity as described under “Experimental Procedure.” The units of N-acetylgalactosaminyltransferase employed represent a 2000-fold excess of enzyme rel- ative to the acceptors quantitated. Similarly, the units of fucosyltransferase represent a 20,000-fold excess. The time course of incorporation indicated that saturation of available acceptors had been achieved.

Endosi-

structure Native OSM

Asialo- OSM

dase- treated asialo- OSM

mol % of total N-acetylgalactosamine

I. NeuAco2 -+ 6GalNAcal + 0- 86” 1.6 1.6 Thr/Ser

II. GalNAcol -+ 0-Thr/Ser 14” 96 96 III. Fucol + 2Galbl + 3GalNAccul 2.2 2.2

-+ 0-Thr/Ser IV. Gal/?1 -+ 3GalNAcol + 0-Thr/ 0.5 0.1

Ser

a Data of H. D. Hill et al. (21).

glands is summarized in Table II. Despite the large quantity of starting material employed and the numerous column chromatography steps required to purify this enzyme as described, a single practiced worker can complete the entire procedure in less than 2 weeks. The partial purification of the GalNAccu2 + 6 sialyltransferase is summarized in Table III, up to the step at which the two purifications diverge.

Step 1: Triton Extraction-Before Triton extraction, solu- ble proteins including mucin are removed by washing the crude submaxillary membrane fraction that contains most of the glycosyltransferases. Manganese chloride is added to the second and third tissue suspensions before centrifugation to induce aggregation of the membrane fraction, permitting the large volume of homogenate to be centrifuged efficiently at low speed in a high capacity rotor (39). Subsequent extractions with 1% (w/v) Triton X-100 solubilize virtually all of the membrane-bound transferases. The pH of the extraction buffer was chosen such that after addition of manganese chloride the extract would be pH 5.9 for the following ion exchange step. By these methods, using two Sorvall RC-3 centrifuges, up to 4 kg of submaxillary glands can be processed conveniently in 2 days, preparing washed tissue pellets on the fist day and extracting them with Triton X-100 on the second.

Step 2: Separation of Fucosyl- and Sialyltransferases on SP-Sephadex-The porcine submaxillary /3-galactoside clll + 2 fucosyltransferase is both quickly inactivated when sol- ubilized with Triton X-100 and unstable when adsorbed to SP-Sephadex; however, if adsorbed and rapidly eluted with detergent-free buffer it can be stored for many months in 50% glycerol at -20°C with almost no loss of activity.3 Adsorption with several small batches of SP-Sephadex is preferred to treatment with a single larger aliquot of ion exchange resin because the former method removes the fucosyltransferase more efficiently. In contrast, at this level of purity the sub-

3 T. A. Beyer and R. L. Hill, unpublished observations.


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/

4436 Purification of Porcine Submaxillary Sialyltransferases

TABLE II Purification of the porcine submaxillary P-galactoside a2 + 3 sialyltransferase

Results are shown for one preparation of enzyme from 4 kg of submaxillary glands. Details of the purification are described under “Experimental Procedure.” Representative column profiles are presented in Fig. 1. (Step 4), Fig. 2 (Step 5), and Fig. 3 (Step 6).

step Volume Total protein Tota!tyactiv- Specific sc- Step purifi- tivity Yield cation

Total purification

ml m&? units” units/mg % Homogenate ‘2fAooo 338,000 39 0.00011 100 1 1

1. Triton extract 15,300 26,600 21 0.ooo80 55 7 7 2. SP-Sephadex 17,ooa 23,000 23 0.0010 60 1.3 9 3. CDP-Sepharose I (NaCl pulse) w@ 3,500 18 0.0051 46 5.1 44 4. CDP-Sepharose II (NaCl gradient) 240 93 6.4 0.068 16 13 600 5. CDP-Sepharose III (CTP-gradient) 26 0.41 2.74 6.6 7 97 57,800 6. Sephadex G-200 18 0.19 2.00 10.6 5 1.6 92,200

Form A 7.4 0.132 1.41 10.6 3.6 1.6 92,200 Form B 11.0 0.055 0.59 10.6 1.5 1.6 92,200

’ One unit corresponds to 1 pmol of product formed per min of incubation at saturating concentrations of CMP-NeuAc and antifreeze glycoprotein. Values obtained in Assay I (pm01 per min per assay) with lactose as acceptor were multiplied by the factor, 155, to obtain units with antifreeze glycoprotein per assay.

TABLE III Purification of the porcine submaxillary a-N-acetylgalactosaminide a2 -+ 6 sialyltransferase

Results are presented for the same preparation that is summarized in Table II.

SkP Volume Total protein Total activity Specific activity Yield step purifi-

cation Total purifi-

cation ml mg units” units/mg 5%

Homogenate 20,000 338,000 102 0.00030 100 1 1 1. Triton extract 15,300 26,600 77 0.0029 75 10 10 2. SP-Sephadex 17,000 23,000 71 0.0031 70 1 10 3. CDP-Sepharose I (NaCl pulse) WOO 3,500 110 0.0032 108 10 104 4. CDP-Sepharose II (NaCl gradient) 179 820 69 0.083 66 2.7 280

’ One unit corresponds to 1 pmol of product formed per min of incubation at saturating concentrations of CMP-NeuAc and endosidase- treated asialo-ovine submaxillary mucin, prepared as described under “Experimental Procedure.” Values obtained in Assay II (pm01 per min per assay) under subsaturating conditions were multiplied by the factor, 5.5, to obtain unite of enzyme per assay.

maxillary sialyltransferases will adsorb neither to SP-Sepha- dex at pH 5, nor to QAE-Sephadex at pH 10. The failure of these enzymes to adsorb to ion exchange media necessitates the use of a rather large affinity column at an early state of the purification. However, treatment of the Triton extract with SP-Sephadex removes some proteins that interfere with adsorption of the sialyltransferases to CDP-Sepharose, thereby improving both the yield and purification obtained in the next step.

Step 3: Concentration of Sialyltransferases on CDP-Seph- arose-The submaxillary sialyltransferases have a lower affinity for CDP-hexanolamine Sepharose than the sialyltransferase from bovine colostrum which was purified using the same adsorbent (lo), and a high ligand concentration (- 14 pmol/ml of gel) is required for efficient adsorption. There is correspond- ingly more ionic adsorption of contaminating proteins to the highly charged gel, and only a modest purification is achieved on elution with 1 M NaCl. I f the Triton X-100 concentration in the elution buffer is decreased to O.l%, the yield of Gala2 + 3 sialyltransferase may fall to less than 15%. In contrast. to the bovine colostrum sialyltransferase, which has markedly different pH optima for activity and stability (lo), both submaxillary sialyltransferases are maximally active and stable between pH 6.0 and pH 6.5 (data not shown). Because nonspecific protein interaction with CDP-Sepharose is reduced with increasing pH, this and subsequent chromatographic steps have been performed within the upper range of the activity and stability versus pH profiles, at pH 6.5

Step 4: Resolution of Sialyltransferases-The different affinities of the submaxillary sialyltransferases for CDP-Seph- arose have been exploited to separate the Gala2 + 3 and GalNAca2 + 6 enzymes by NaCl-gradient elution. A representative column profile is shown in Fig 1. The GalNAccu2 + 6 sialyltransferase emerges slightly behind the major protein

Volume (ml)

FIG. 1. Resolution of the two porcine submaxillary gland sialyltransferases by NaCl gradient elution from CDP-Sepharose. As described under “Experimental Procedure,” dialyzed enzyme from Step 3 was adsorbed to a column of CDP-Sepharose II (2.2 x 21.0 cm) and eluted at the arrow (i) with a linear NaCl gradient as shown. Fractions (14 ml) were assayed for GalNAccu2 + 6 sialyltransferase (A), Gala2 + 3 sialyltransferase (O), protein (0), and NaCl (by conductivity, +-), then pooled as indicated (I and II).

peak, followed by the Gala2 + 3 sialyltransferase. The adsorbed enzymes are recovered quantitatively in the gradient, and the lower yields indicated for this step in Tables II and III reflect a compromise between maximizing purification and yield. Usually fractions can be pooled to achieve 40 to 70% yield with lo- to 20-fold step purification of the Gala2 + 3 enzyme which will still contain roughly an equal number of units of the GalNAca2 + 6 sialyltransferase. The GalNAca2 + 6 enzyme pool contains between 0.4% and 4% Gala2 + 3 sialyltransferase at this step and can be stored with 0.02% NaNa for at least 6 months at 4°C with less than 10% loss of activity.

Step 5: Specific El&ion of the Gala.2 + 3 Sialyltransfer-


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/


use-Further resolution from the contaminating GalNAccu2 + 6 sialyltransferase and considerable purification is obtained by CTP-gradient elution of the Gala2 + 3 enzyme from CDP- Sepharose, as shown in Fig 2. Following a salt wash step, the Gala2 + 3 sialyltransferase is eluted in a sharp peak at the start of the gradient ahead of a broader peak of GalNAccu2 + 6 enzyme. The remainder of the adsorbed GalNAccv2 + 6 sialyltransferase elutes with the 1 M NaCl pulse, together with the bulk of the other proteins.

Among cytidine nucleotides, CTP is the most potent inhib- itor of both the Gala2 + 3 sialyltransferase (& = 0.22 pM)

and the GalNAccu2 + 6 sialyltransferase (KI = 5.8 PM). For each enzyme the KI for CDP and CMP is about 10 and 100 times larger, respectively. Thus, CTP is the best choice for specific elution, and the two sialyltransferases emerge from the affinity resin in the order of increasing KI. Although a limit concentration of 4 mM CTP was arbitrarily chosen for the purification reported here, the Gala2 + 3 sialyltransferase will elute at CTP concentrations below 0.2 mM, and a more shallow gradient (i.e. 0 to 1 mu) will give better purification with no sacrifice in yield. Some NaCl is necessary in the gradient buffers to prevent weak nonspecific interaction of the Gala2 + 3 sialyltransferase with CDP-Sepharose, even in the presence of CTP. To avoid an increase in ionic strength during CTP gradient elution, the concentration of NaCl in the limit buffer is lower than that of the starting buffer. I f glycerol is omitted, nearly all of the GalNAccr2 + 6 enzyme elutes with the 1 M NaCl pulse, giving slightly better resolution of the two sialyltransferases. However, the specific activity of the recovered Gala2 + 3 enzyme is reduced in the absence of glycerol. Because removal of CTP by dialysis is inefficient, gel filtration on Sephadex G-50 was chosen to desalt the enzyme. Adsorption on a small column of CDP-Sepharose and elution with 1 M NaCl serves to concentrate the sialyltransferase for the next step.

Step 6: Separation of Two Gala.2 + 3 Sialyltransferase Forms by Gel Filtration-Chromatography on Sephadex G- 200 (superfine) as shown in Fig. 3 yields three protein peaks, two of which possess Gala2 -+ 3 sialyltransferase activity. The specific activity across both peaks is roughly constant and is equal to 10.6 units/mg for the pooled fractions from Peak A

Fraction Number

FIG. 2. CTP gradient elution of porcine submaxillary sialyltransferases from CDP-Sepharose. Dialyzed enzyme from Step 4 was chromatographed on a column (1.5-x 8.5 cm) of previously unused CDP-Sepharose equilibrated with 10 mru sodium cacodylate, pH 6.5, 25% (w/v) elvcerol. 1% (w/v) Triton X-100, and 50 mM NaCl as descibed’under “Experimental Procedure.” At A, the column was washed with buffer containing 175 mM NaCl, followed by buffer containing 100 mu NaCl at B. A gradient of 0 to 4 mM CTP was begun at C, and after a brief wash with 106 InM NaCl buffer, the gradient was followed with buffer containing 1.0 M NaCl at D. Frac- tions 1 to 55 (13.5 ml) and Fractions 56 to 135 (2.8 ml) were monitored for GalNAccu2 + 6 sialyltransferase (A), Gala2 --) 3 sialyltransferase (O), protein (0), and CTP (by absorbance at 271 nm, pH 12, - - -).

0.5 100

0.4 60 y

0.3 60 E 0, 3

0.2 40 i is

0.1 20

0 0 ,O 20 00 40 50 60 70 60 90 Froct~on Number (0.93ml)

FIG. 3. Separation of two Gala2 --f 3 sialyltransferase forms by gel filtration. Concentrated enzyme from Step 5 was chromatographed on a column (1.6 x 43 cm) of Sephadex G-200 (superfine) as described under “Experimental Procedure.” Fractions (0.93 ml) were monitored for GalNAccY 2 + 6 sialyltransferase (A), Gala2 -+ 3 sialyltransferase (O), and protein (0).

Volume (ml, J

FIG. 4. Concentration and desalting of purified Gala2 + 3 sialyltransferase in a single step. Approximately 0.52 unit of the B enzyme form was adsorbed to a small column (0.3 X 25 cm) containing 4.5 cm of CDP-Sepharose layered above 20.5 cm of Sephadex G-50 (fine) equilibrated with 10 mM sodium cacodylate, pH 6.5,50 mu NaCl, 1% (w/v) Triton X-100, and 25% (w/v) glycerol. After washing with 2 column volumes of equilibration buffer, the column was developed with the same buffer but containing 1 M NaCl. Fractions (0.13 ml) were monitored for Gala2 + 3 sialyltransferase (0) and NaCl (by conductivity, 0).

(Ko = 0.1) and Peak B (Ko = 0.4). In addition, most of the GalNAca2 + 6 sialyltransferase elutes near the void volume with the remaining major protein contaminant. After concentration on a small column of CDP-Sepharose overlayered on Sephadex G-50 (fine) as shown in Fig. 4, both enzyme forms can be stored at -20°C in 50% (w/v) glycerol for over 6 months with no loss of activity.

Characterization of the Two Gala2 + 3 Sialyltransferase Forms

Gel filtration gives two different sialyltransferase forms, designated A and B, but as shown in Fig. 5, each migrates as a single diffuse band with a molecular weight of about 50,000 upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis under either reducing or nonreducing conditions. The bands corresponding to the B form (Fig. 5, Gels c and d) are 10 to 15% broader than those of the A form (Fig. 5, Gels a and f-4.

In contrast to their behavior on gel electrophoresis, the A and B forms differ markedly on analytical gel fdtration. As shown in Fig. 6, Form A elutes much earlier than B and behaves as a globular protein with a Stokes radius of 51 A, which is consistent with a protein of molecular weight 220,000, or 4 to 5 times the molecular weight estimated by gel electrophoresis. Form B has a Stokes radius of 31 A, which is consistent with the molecular weight of 50,000 estimated by gel electrophoresis. During rechromatography neither enzyme peak gives rise to the other, ruling out a simple equilibrium


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/


7769 60 43 26 h!&culor Weight X10-’

and DzO can be analyzed to obtain estimates of the .~z,,,~ and the partial specific volume for both sialyltransferase forms, as described under “Appendix.” Together with the Stokes radii (Fig. 6), these quantities then permit calculation of molecular weights and frictional coefficients. As shown in Table IV, the molecular weight of 44,000 for transferase B, calculated by this method, is close to the estimate of 50,000 obtained by gel electrophoresis. Furthermore, Form B has the frictional ratio (1.3) and partial specific volume (0.72 cm3/g) of a globular protein that binds little or no lipid. On the other hand, the

FIG. 5. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the purified Gala2 + 3 sialyltransferase forms. Fractions obtained by gel filtration on Sephadex G-200 (Step 6) were treated with Bio- Beads SM-2 (Bio-Rad) to remove Triton X-100 (40) and dialyzed against H20. Aliquots containing -12 pg of protein (30) were lyophi- lized and dissolved with heating in sample buffer (5 min at 11O’C) in the presence or absence of 2% fi-mercaptoethanol, then electrophoresed on 7.5% polyacrylamide gels (41, 42). After staining with Coo- massie blue, gels were scanned at 525 nm. Gel CL, Form A, reduced with &mercaptoethanol; Gel b, Form A, nonreduced, Gel c, Form B, reduced; Gel cl, Form B, nonreduced. The molecular weights of standard proteins electrophoresed under the same condition (with reduction) are indicated on the abscissa: transferrin, 77,ooO; serum albumin, 69,ooO; catalase, 60,OCQ ovalbumin, 43,ooO; and o-chymo- trypsinogen, 26,000.

between monomer and a descrete aggregated state. The width at half-height of Peak A is equal to that of adjacent standards, while the width of Peak B is roughly 1.3 times that of the adjacent malate dehydrogenase and ovalbumin peaks, sug- gesting some size heterogeneity in Form B that is also reflected in the gel electrophoresis patterns.

Upon centrifugation through sucrose density gradients prepared in H20, both Form A and Form B sediment close to the ovalbumin standard (Fig. 7). Apparent sedimentation coefficients of 3.1 S and 3.5 S, respectively, can be calculated by interpolation between the standard proteins, all of which have partial specific volumes of 0.73 to 0.75 cm3/g. If Form B has a partial specific volume close to that of the standards, then the sedimentation coefficient obtained by this method is consistent with the molecular weight estimated by gel electrophoresis and the Stokes radius obtained by gel filtration. These results suggest that the active species of Form B contains a single globular polypeptide of 50,000 daltons. On the other hand, the unusual combination of a large Stokes radius (51 A) and a small apparent sedimentation coefficient (3.1 S) suggests that the A form contains a single 50,000-dalton polypeptide that binds a large amount of low density lipid or detergent. This interpretation is supported by centrifugation through sucrose gradients prepared in DzO. Under these conditions, Form B still sediments near ovalbumin (Fig. 7d) and must have a similar partial specific volume, while Form A shifts to a position near myoglobin (Fig. 7~) consistent with a very high partial specific volume due to the binding of lipid or detergent. This particular sample of the B form is seen to contain a small amount of the A form (Fig. 7d).

I f major changes in conformation and composition do not occur as a function of sucrose concentration, the data obtained by sedimentation through sucrose density gradients in Hz0

60

IO r / / I / I I / I 0 0.2 0.4 0.6 0.6 I.0 1.2 1.4

erf-’ (I-KJ

FIG. 6. Determination of Stokes radii by gel filtration. Samples (1 ml) containing &IyItransferases, standard proteins, and 1 M NaCl were chromatographed on Sephadex G-200 (fme, 1.6 x 50 cm) equilibrated with 10 nw~ sodium cacodylate, pH 6.5, 250 mM NaCl, and 0.05% (w/v) Triton X-100. Fractions (-1 ml, 40 drops) were weighed to determined elution volumes. In a separate experiment, void and included volumes were determined with blue dextran 2000 (Pharma- cia) and NaCl, respectively. Standard proteins and their Stokes radii were: /%galactosidase, 68.6 A (31); lactate dyhydrogenase (LDH), 42.0 A (31); malate dehydrogenase (JWW), 35.1 A (43); ovalbumin, 27.5 A (32); and myoglobin, 18.9 A (31). /?-Galactosidase, lactate dehydrogenase, and malate dehydrogenase were assayed enzymatically. Oval- bumin was detected by protein assay (30). Myoglobin was quantitated spectrophotometrically at 409 nm. Stokes radios is plotted uersus erf-’ (1 - Ko) according to Ackers (44), where erf’ is the inverse error function and KD is the partition coefficient of a protein.

M~ll!lllers from top of Gradlent (I.3S3cnVmll

FIG. 7. Sucrose density gradient centrifugation of the purified Gala2 + 3 sialyltransferase forms. Aliquota containing marker proteins and either sialyltransferase Form A (Panels a and c) or Form B (Panels b and d) were centrifuged through linear 5 to 20% (w/v) sucrose gradients prepared in either Hz0 (Panels a and b) or D20 (Panels c and d) as described under “Experimental Procedure.” The percentage of the applied sialyltransferase recovered from both Hz0 and D20 gradients was 75 to 80% of the A form and 85% or more of the B form. The marker proteins employed were: Mb, myoglobin; Ou, ovalbumin; MDH, malate dehydrogenase; u, transferrin. From these data, values for the partial specific volume, i, and ~20,~ were calculated for each sialyhmnsferase form as described under “Appendix.”


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/


TABLE IV Physical properties of the purified Gala2 + 3 sialyltransferase forms

The two sialyltransferase forms A and B were separated by gel Sucrose density gradient centrifugation of both enzyme forms was filtration chromatography on Sephadex G-200 as shown in Fig. 3. The performed as shown in Fig. 7. The values for the standard sedimen- Stokes radius (R,) of each form was determined as ill~trated in Fig. tation coefficient ( SZ&, partial specific volume (?), grams of Triton 6, and the polypeptide molecular weight of each form was estimated X-100 per g of protein, frictional ratio (f/fmin) and molecular weight by sodium dodecyl sulfate-gel electrophoresis as described in Fig. 5. for each form were calculated as described under “Appendix.”

Tranaferase &., R, v T&en/protein f/fmin M,, complex M, protein moi- M,, by gel elec- ety trophoresis

S A cm3/g g/g Form A 3.3 51 0.82 1.18 1.5 106,000 49,000 50,006 Form B 3.5 31 0.72 1.3 ‘we 5o.ooo

I %:*‘.& d, ;,,014 llo ,I . . . . Percent WV) Tnton X - 100

FIG. 8. Dependence of Gala2 + 3 sialyltransferase activity upon Triton X-100 concentration. Triton X-100 (1% w/v) was removed from each purified sialyltransferase form A and B by incubation with Bio-Beads EM-2 (40) in the presence of 10 mM sodium cacodylate, pH 6.5,50 mM NaCl, and 7 mg/ml of bovine serum albumin. SimiIar treatment of buffer containing 1% (w/v) Triton X-100 but no protein reduced the concentration of detergent to 0.016% assayed by absorbance at 274 nm (43). Aliquots (5 ~1) of each transferase form were diluted lo-fold into a volume of 50 ~1 containing 50 mru sodium cacodylate, pH 6.5, 7 mg/ml of bovine serum albumin, 2.6 pM CMP- [“C]NeuAc (356,000 cpm/nmol), 160 mM lactose, and the concentration of T&on X-100 indicated on the abscissa. After incubation at 37°C for 15 min, transfer to lactose was quantitated as described under “Experimental Procedure.” Transfer by either Form A (0) or Form B (0) is expressed as a fraction of that observed in the presence of 1% (w/v) Triton X-100, which is the detergent concentration in the standard assay for this enzyme. The approximate critical micellar concentration (UK) of Triton X-100,0.016% (47), is indicated by the arrow (t).

molecular weight of 106,006 calculated for transferase A is roughly twice that of the Form A polypeptide of 50,006 estimated by gel electrophoresis. Because the polypeptides of Forms A and B are similar in size and presumably are closely related, the partial specific volume of the protein moiety of Form A must be approximately equal to that of Form B, 0.72 cm3/g. The calculated partial specific volume of 0.82 cm3/g for the A form indicates that it must contain a large amount of lipid or detergent. Since the enzyme has been isolated under conditions that are sufficient to delipidate many membrane proteins (45), it is likely that most of the bound lipid is Triton X-100, which has a partial specific volume of 0.908 cm3/g (46). Knowledge of these partial specific volumes is sufficient to calculate that about 1.18 g of T&on X-100 are bound per g of protein in the sialyltransferase A complex; thus, the protein moiety of Form A has a molecular weight of approximately 49,000, in good agreement with the value estimated by gel electrophoresis.

For each sialyltransferase form, the dependence of catalytic activity upon Triton X-100 concentration also reflects the extent of lipid or detergent binding. As shown in Fig. 8, the A form is inhibited -60% by lowering the Triton X-100 concen-

tration to less than the critical micellar concentration, while the B form is unaffected. Both enzyme forms are stable in the absence of detergent, and the A form can be reactivated by increasing the detergent to greater than the critical micellar concentration. At Triton concentrations above O-4%, both transferase forms have approximately the same specific activity.

DISCUSSION

A single affinity adsorbent, CDP-hexanolamine Sepharose, has been used to resolve two porcine submaxillary sialyltransferases through exploitation of differences in their affinity for the adsorbent in the presence of NaCl and CTP. One enzyme, an ol-N-acetylgalactosaminide (~2 + 6 sialyltransferase, has been obtained in a partially purified stable form that is essen- tially free of other sialyltransferase activities. Its purification to homogeneity will be reported subsequently.* The other enzyme, a /?-galactoside 012 + 3 sialyltransferase, is the second sialyltransferase to be purified to homogeneity on this adsorbent. The fit such enzyme was a /3-galactoside o2 --, 6 sialyltransferase from bovine colostrum (10). The related adsorbent, UDP-hexanolamine Sepharose, has been used in the purification to homogeneity of three different glycosyltransferases, a galactosyltransferase from bovine milk (27), a fucosyl cul --, 2 gala&side a-N-acetylgalactosaminyltransferase from porcine submaxillary glands (18), and a glucuronyltrans- ferase from rat liver (48, 49). It is also useful for the partial purification of an ~-mannoside-~-N-acetylglucosaminyltrans- ferase from rabbit liver.5 Similarly, GDP-hexanolamine Seph- arose has been used in the purification to near homogeneity of an H blood group-specific fucosyltransferase from porcine submaxillary glands and a /3-N-acetylglucosaminide (~1 + 3 fucosyltransferase from human milk (17). The purification to homogeneity of the former fucosyltransferase will be reported elsewhere.’ The great variety of glycosyltransferases that have been resolved and purified on nucleotide-hexanolamine-sepharose conjugates amply illustrates the versatility and power of such adsorbents, even in the separation of enzymes that catalyze quite similar reactions using related substrates.

The purified Gala2 + 3 sialyltransferase consists of two enzymatically similar (15) forms that are separable by gel filtration and have polypeptide molecular weights of -50,000. The occurrence of multiple forms of another purified sialyltransferase (10) and a galactosyltransferase (27) has been described before. In this case, one transferase form (A) appears to bind a large amount of detergent or lipid and may represent the native enzyme with an intact membrane-binding site. A second enzyme form (B) binds little or no detergent and may be a proteolytic fragment of the A form. It is possible that Form A is isolated with bound phospholipid, but this is

4 J. E. Sadler, J. I. Rearick, and R. L. Hill, (1979) J. Biol. Chem., 254, in press.

’ C. L. Oppenheimer, unpublished observations. ’ T. A. Beyer, J. E. Sadler, and R. L. Hill, manuscript in preparation.


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/


unlikely because the high concentration of Triton present in the extraction buffers would be expected to displace all but the most tightly bound lipids (45). Thus it has been assumed for the sake of discussion that the low density component of the sialyltransferase A complex is Triton X-100. A comparison of Form A and Form B by gel electrophoresis suggests that they may differ by as little as 1,000 to 2,000 daltons, or by 10 to 20 amino acids, although further studies will be required to substantiate this point. I f so, since the calculated molecular weight of the lipid component of Form A (57,000) is tw,o-thirds that of the Triton micelle (86,000) (50), it is reasonable to suggest that Form A binds nearly an entire Triton micelle through a short length of polypeptide that is missing from Form B. Furthermore, Form A is only 40% as active below the critical micellar concentration of Triton X-100, while the activity of Form B is independent of detergent concentration, Thus, the membrane-binding site and associated lipid of detergent do not play an essential role in catalysis, although the membrane may regulate or segregate the sialyltransferase in vivo.

The successful purification described here was greatly facil- itated by the use of other purified glycosyltransferases and glycosidases to construct a specific substrate for the GalNAcar2 + 6 sialyltransferase. Although very small quantities of acceptor substrate sites for the Gala2 + 3 sialyltransferase were found in ovine submaxillary asialomucin, the number of sites was sufficient to interfere with the GalNAcol2 + 6 sialyltransferase assay, thereby preventing the accurate assessment of the purity of the Gala2 + 3 sialyltransferase. Prior to the complete resolution of these two sialyltransferases, the contaminating acceptors were quantitated by saturation with the porcine submaxillary H blood group-specific fucosyltransferase (17). Subsequent digestion with the endo-cY-N-acetylgalactosaminidase from 5’. pneumoniae both demonstrated the structure of these acceptors (38) and showed that they could be completely removed. In addition, the product of these enzymatic modifications proved to be a very specific substrate for the GalNAccu2 + 6 sialyltransferase. Thus, the specificity of glycosyltransferases and glycosidases can be used to char- acterize and selectively modify minor but important fractions of complex oligosaccharide mixtures with a precision and economy of materials that may be very difficult to achieve by direct chemical methods.

The availability of several highly purified sialyltransferases with strict and well defined substrate specificity permits the study of the function of specific sialic acid-containing structures by a similar enzymatic approach. The porcine submaxillary gland sialyltransferases can be used to synthesize all of the common threonine or serine-linked sialylated oligosaccharides found in membrane glycoproteins such as glycophorin (51) or serum glycoproteins such as fetuin (52), namely, NeuAcaP + 3Ga461+ 3[NeuAca2 + G]GalNAcol --, OThr/ Ser and the two related but incomplete structures containing only a single sialic acid residue (15, 16). The &galactoside a2 + 6 sialyltransferase from bovine colostrum can be used to form the structure, NeuAca2 -+ 6Galpl + 4GlcNAc, wherever the appropriate disaccharide acceptor sequence appears in the asparagine-linked oligosaccharides of glycoproteins (9, 11). Together, these three enzymes can form all of the common sialic acid structures of glycoproteins except one, the NeuAca2 + 3G@l+ 4GlcNAc sequence that has been described in fetuin (53) and ribonuclease (54). The enzymatic specificity of these enzymes has recently been employed to study the role of sialic acids in the expression of the MN blood group antigens and myxovirus receptors on human red blood cells by restoring specific sialic acid-containing structures to the red cell surface (55, 56). These and other purified glycosyl-

transferases are expected to be increasingly useful in other investigations of oligosaccharide structure and function.

Acknowledgments-We wish to thank T. A. Beyer and L. R. Glasgow (Department of Biochemistry, Duke University Medical Center) for the preparation of porcine submaxillary gland fucosyltransferase and glycosidases from 5’. pneumoniae, respectively. We also wish to thank M. S. Schwyzer (Department of Molecular Biology, University of Geneva, Switzerland) for the preparation of porcine o- N-acetylgalactosaminyltransferase.

REFERENCES

1. Kornfeld, R., and Kornfeld, S. (1976) Annu. Rev. Biochem. 45, 217-237

2. Yg, S-S., and Dain, J. A. (1976) in Biological Roles of Sialic Acid (Rosenberg, A., and Schengrund, C.-L., eds) pp. 59-102, Plenum Press, New York

3. Wiegandt, H. (1973) Hoppe Seyler’s 2. Physiol. Chem. 354,1049- 1056

4. Finne, J., Krusius, T., Rauvala, H., and Hemminki, K. (1977) Eur. J. Biochem. 17,319-323

5. Carlson, D. M., Jourdian, G. W., and Roseman, S. (1973) J. Biol. Chem. 248,5742-5750

6. Den, H., Kaufman, B., and Roseman, S. (1970) J Biol. Chem. 246,6607-6615

7. Kaufman, B., Basu, S., and Roseman, S. (1966) in Proceedings of the Third International Symposium on Cerebral Sphingolipi- dosis (Aronson, S. M., and Volk, B. W., eds) pp. 193-213, Pergamon Press, New York

8. Kaufman, B., Basu, S., and Roseman, S. (1968) J. Biol. Chem. 243,5804-5807

9. Paulson, J. C., Rearick, J. I., and Hill, R. L. (1977) J. Biol. Chem. 262,2363-2371

10. Paulson, J. C., Beranek, W. E., and Hill, R. L. (1977) J. Biol. Chem. 252,2356-2362

11. Paulson, J. C., Prieels, J.-P., Glasgow, L. R., and Hill, R. L. (1978) J. Biol. Chem. 263,5617-5624

12. Paulson, J. C., Hill, R. L., Tanabe, T., and Ashwell, G. (1977) J. Bill. Chem. 262,8624-8628

13. Sodetz, J. M., Paulson, J. C., Pizzo, S. V., and McKee, P. A. (1978) J. Biol. Chem. 253,7202-7206

14. Carlson, D. M., McGuire, E. J., Jourdian, G. W., and Roseman, S. (1973) J. Biol. Chem. 248,5763-5773

15. Rearick, J. I., Sadler, J. E., Paulson, J. C., and Hill, R. L. (1979) J. Biol. Chem. 254,4444-4451

16. Sadler, J. E., Rearick, J. I., Paulson, J. C., and Hill, R. L. (1978) Fourth International Symposium on Glycoconjugates, Woods Hole, Mass., September, 1977 (Jeanloz, R. W., ed) Academic Press, New York, in press

17. Beyer, T. A., Prieels, J.-P., and Hill, R. L. (1978) Fourth Inter- national Symposium on Glycoconjugates, Woods Hole, Mass., September, 1977 (Jeanloz, R. W., ed) Academic Press, New York, in press

18. Schwyzer, M., and Hill, R. L. (1977) J. Biol. Chem. 262, 2338- 2345

19. Geisow, M. J. (1975) Biochem. J. 161, 181-183 29. Glasgow, L. R., Paulson, J. C., end Hill, R. L. (1977) J. Biol.

Chem. 262,8615-8623 21. Hill, H. D., Jr., Reynolds, J. A., and Hi& R. L. (1977) J. Biol.

Chem. 252,3791-3798 22. DeVries, A. L., Komatsu, S. K., and Feeney, R. E. (1970) J Biol.

Chem. 246,2901-2908 23. Hi, R. L., and Delaney, R. (1967) Methods Enzymol. 11, 339-

351 24. Jourdian, G. W., Dean, L., and Roseman, S. (1971) J Biol. Chem.

246,430-435 25. Reissig, J. L., Strominger, J. L., and Leloir, L. F. (1955) J. Biol.

Chem. 217,959-986 26. Levvy, G. A., and McAUan, A. (1959) B&hem. J. 73,127-132 27. Barker, R., Olsen, K. W., Shaper, J. H., and Hill, R. L. (1972) J.

Biol. Chem. 247,7135-7147 28. Hoard, D. E.. and Ott. D. G. (1965) J. Am. Chem. Sot. 87.1785-

1788 29. Cuatrecasas, P. (1970) J. Biol. Chem. 2453059-3065 30. Schaffner. W.. and Weissman. C. (1973) Anal. Biochem. 56.502-

514 31. Sober, H. A. (ed.) (1968) in Handbook of Biochemistry, 2nd Ed,

pp. C-10-11, Chemical Rubber Publishing Co., Cleveland, Ohio


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/


32. 33. 34. 35.

36. 37.

40. 41.

42. 43. 44. 45.

46.

Castellino, F. J., and Barker, R. (1968) Biochemistry 7,2207-2217 Thorne, C. J. R. (1962) Biochim. Biophys. A&u 59,624-633 Mann, K. G., and Vestling, C. S. (1969) Biochemistry 8,1105-1109 Roberts, R. C., Makey, D. G., and Seal, U. S. (1966) J. Biol.

Chem. 244,4907-4913

Biochemistry 13,2369-2379 47. Simons, K., Helen& A., and Garoff, H. (1973) J. Mol. Biol. 80,

119-133 48. Gorski, J. P., and Kasper, C. B. (1977) J. Biol. Chem. 252, 1336-

1343 Carlson, D. M. (1968) J. Biol. Chem. 243,616-626 Gottschalk, A., Bhargava, A. S., and Murty, V. L. N. (1972) in

Glycoproteinsz Their Composition, Structure, and Function (Cottschalk, A., ed) pp. 810-815, Elsevier, New York

Endo, Y., and Kobata, A. (1976) J. Biochem. (Tokyo) 80, l-8 Kamath, S. A., and Rubin, E. (1972) B&hem. Biophys. Res.

Commun. 49,52-59

49. Burchell, B. (1977) FEBS Lett. 78, 101-104 50. Yedgar, S., Barenholz, Y., and Cooper, V. G. (1974) Biochim.

Biophys. Actu 363,98-111 51. Thomas, D. B., and Winder, R. J. (1969) J. Biol. Chem. 244,

5943-5946 52. Spiro, R. G., and Bhoyroo, V. D. (1974) J. Biol. Chem. 249,5704-

5717 Holloway, P. W. (1973) Anal. B&hem. 63,304-306 Schwartz, M. L., Pkzo, S. V., Hill, R. L., and McKee, P. A. (1971)

J. Biol. Chem. 246,5851-5854 Weber, K., and Osborn, M. (1969) J. Biol. Chem. 244,4406-4412 Clarke, S. (1975) J. Biol. Chem. 260, 5459-5469 Ackers, G. K. (1967) J. Biol. Chem. 242,3237-3236 Helenius, A., and Simons, K. (1975) Biochim. Biophys. Acta 415,

53. Spiro, R. G. (1973) Ada Protein Chem. 27,349-467 54. Schut, B. L., Dorland, L., Haverkamp. J., Vliegenthart, J. F. G.,

and Fournet, B. (1978) Biochem. Biophys. Res. Commun. 82, 1223-1228

55. Sadler, J. E., Paulson, J. C., and Hill, R. L. (1979) J. Biol. Chem. 254,2112-2119

29-79 56. Paulson, J. C., Sadler, J. E., and Hill, R. L. (1979) J. Biol. Chem.

264,2120-2124 Tanford, C!., Nozaki, Y., Reynolds, J. A., and Makino, S. (1974) Additional references are found on p. 4443.


ww

.jbc.orgD

ownloaded from

http://www.jbc.org/



ww

.jbc.orgD

ownloaded from

http://www.jbc.org/

purification to homogeneity of a β-galactoside α2→3 sialyltransferase and partial purification...

Documents