Effect of Glycosylation on Yeast Invertase Oligomer Stability*

Vol. 262, No. 9,l w e of March 25, pp. 43954401,1987

THEJOURNAL OF B!OLOGIC?L CHEMIST~Y Q 1987by The Amencan Socrety of B~olotpealChemists, lnc.

Printed in U.S.A.

Effect of Glycosylation on Yeast Invertase Oligomer Stability* (Received for publication, October 20,1986)

Markku TammiS, LunBallou, Alice Taylor, and Clinton E. BallouQ From the Department of Biochemistry, University of California, Berkeley, California 94720

Yeast external invertase is a glycoprotein that exists nents which are also present in much smaller amounts are as a dimer that can associate to form tetramers,hex- enzymatically active and indicate the presence of an associaF., Watorek, W., and Maley,tion-~ssociationequilibrium.” This association was stated by amers, and octamers (Chu, F. (1983) Arch. Biochem. Biophys. 223, 543-555; Chu and Maley (6) to lead to octamer formation, a conclusion Esmon, P. C., Esmon,B. E., Schauer, I.E., Taylor, A., that was documented by Chu et al. (7). The tendency of and Schekman, R. (1987) J. Bioi. Chem., 262,4395- glycosylated invertase to form multimers had also been noted 4401), a process that is facilitated by the attached by Esmon et al. (8) in studies on the invertase trapped in the oligosaccharide chains. W e have studied this associa- endoplasmic reticulum as a result of a genetic defect in the tion by high performance liquid chromatography a on secretory pathway, and a more detailed analysis by Esmon et gel filtration matrix, by which procedure wild-type al.(9) has provided clear documentation by gelelectrophoresis bakers’ yeastinvertase gives two peaks, and invertase and electron microscopy for the interaction of dimeric glycofrom a core mutant (mnnl mnn9) of Saccharomyces sylated invertase to form tetramers, hexamers, and octamers. cereuisiae X2180 gives three peaks. Concentration of Although internal nonglycosylated invertase forms a stable an invertase solution by freezing drives the dimers into dimer, it yields unstable higher oligomers (lo), and the enhigher aggregates that, at 30 “C, re-equilibrate to a hanced ability of the glycosylated external invertase to form mixture of smaller forms, the composition of which such oligomers depends on the extent of glycosylation (7, 9), depends onpH, concentration, and time. The invertase from a mutant, mnnl mnn9 dpgf, which underglyco- as though the carbohydrate chains help to stabilize the mulsylates its glycoproteins and producesinvertase with timeric state. In this report, we present additiona~support for 4-7 oligosaccharide chains, forms oligomers ofmuch this view and demonstrate that theoligomerization is afreely lower stability than the mnnl mnn9 invertase, which reversible process that can be followed conveniently by high has 8-11 carbohydrate chains. Both ofthese mutants performance liquid chromatography. We have also studied the release external invertase from the periplasm into the effect of mannoprotein glycosylation defects on the release of medium during growth, but we conclude that defects invertase into the medium during growth of yeast mannoproin the cell wall structure may be more important in tein mutants (11)and assessed the possible role of oligomer this release than an altered tendency of the invertases formation in retaining invertase in the periplasm. to aggregate. Investigation of aggregate formation by electron microscopy revealed that all invertases, inEXPERIMENTALPROCEDURES cluding the internal nonglycosylated enzyme, form ocGeneral Methods-Chromatographic effluents were monitored for tamers under appropriate conditions. protein a t 206or280nm with an LKB Uvicord S 2138, and salt Internal, nonglycosylated bakers’ yeast invertase was shown by Gascon and Lampen (1)to have a molecular weight of 135,000, whereas Neumann and Lampen (2) reported that the external enzyme contained 50% carbohydrate and had a molecular weight of about 270,000, thus giving both enzymes about the same polypeptide molecular weight (3). Subsequently, Trimble and Maley (4) demonstrated that the carbohydrate-depleted external invertase produced by endoglucosaminidase H digestion wascomposed of two identical protein subunits of 60,000 daltons, whereas the amino acid sequence inferred from the DNA sequence of the cloned gene gives a polypeptide Mr of about 59,000 (5). In sedimentation velocity studies, Neumann and Lampen (2) observed that the external invertase contained a major component with an s&+ of 10.4 S , and that “heavier compo* This work was supported by National Science Foundation Grant PCM~-00251 andUnited States Public Health Service Grant AI12522. The costs of publication of this article were defrayed in part by the payment of page charges. This article musttherefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 5 Visiting professor from the University of Kuopio, Finland, and recipient of Fogarty International Fellowship 1 F 0 5 TW03668. 0 To whom correspondence should be addressed.

concentration was measured in gradients with a Bio-Rad Conductivity Monitor equipped with a Standard Flow Cell. Protein was quantitated by the Lowry procedure (12) with bovine serum albumin as the standard. For electron microscopy, desalted samples were applied to carbon-coated Pelco 8HGC grids (400mesh), stained with uranyl acetate according to Kirschner et al. (13), and viewed in a Philips 201 electron microscope. Cell mass was estimated from the wet packed volume or by dry weight after lyophilization, and cell number was determined by counting in a hemocytometer. For the mnn9 strains, which are clumpy, it was not possible to count cells in the usual manner, so cell number was estimated by assuming there are 10” cells/g of lyophilized cells. Separation of invertase oligomers was done by gel electrophoresis on native (8) and SDS’ gels (14) or by gel filtration on Bio-Gel A-5m, and high performance liquid chromatography was carried out on a Du Pont GF-450 gel filtration column (0.94 X 25 cm). To follow the rates of equilibration between oligomers, samples were frozen to drive the purified invertase into the higher oligomeric forms owing to concentration of the solution and, after thawing, were analyzed by HPLC. For invertase assays, samples (5-50 pl) were incubated with 200 pl of 0.1 M sodium acetate, pH 5.1, for 2-20 min a t 37 “C after the addition of 20 a1 of 40% sucrose. Glucose standards (5-50 fig) were used for calibration. The reaction was stopped by boiling the tubes after addition of200 pl of 200 mM potassium phosphate, pH 7.0. .-

’ The abbreviations used are: SDS, sodium dodecyl sulfate; HPLC,

high performance liquid chro~atography;man, mutation that affects mannoprotein glycosylation;dpg, mutation that affects dolichol phosphoglucose synthesis.

4395

Glucose oxidase reagent (1.0 ml) was added, and the tubes were incubated a t 37 'C for 30 min. The absorbance was measured at 540 nm on a Spectronic 100 colorimeter after addition of 1.5 ml of 6 N HCl. One unit of enzyme activity released 1 pmol of glucose/min at 37 "C. To compare invertase release by whole yeast cells, 5-ml YEPD liquid cultures were inoculated and shaken at 23 "C for 48 h. Assay for glucose in the medium showed that it was all consumed by 24 h, a t which time invertase synthesisand secretion had begun. The 48-h culture was centrifuged to separate cells and medium, and a portion of each was taken for invertase assay. The cells were washed with distilled water, lyophilized, and weighed. Invertase release is expressed as thepercent of the total invertase in the culture that was found in the medium. Yeast Cultures-The yeast strains used were ~ ~ c ~ r o m ycereces visiaeX2180 (wild-type), mnnl mnn2 ( l l ) , mnnl mnn9 (15), and 4AL (mnnl mnn9 glsl dpgl) (16), and they were grown on a YEPD medium (1% yeast extract, 2% Bactopeptone, and 2% glucose). For large-scale growth of yeast, each 1-liter culturewas inoculated with a 10-ml culture grown as above, and the 200-liter fermenter was inoculated with 6 1-liter cultures. For small cultures, cells were isolated by centrifugation at 4000 X g for 10 min, whereas a Sharples continuous flow centrifuge was used for the 200-liter cultures. The yield of wet cell paste was 17-25g/liter from the mnnlmnn9 and 4AL strains. Isolation of Inuertase from Cells-The method of Lehle et al. (17) was used with minor modifications. The cell paste from a 6-liter culture was suspended in 400 ml of 10 mM sodium phosphate, pH 6.5, containing 1 mM phenylmethylsulfonyl fluoride, and homogenized either by a Braun homogenizer (5 times for 1min each of a 10%cell paste suspension in a 1:2 v/v ratio of liquid to glass beads) or by a Biospec Bead-Beater (5 times for 1 min each of a 50% cell paste suspension in a 1:1v/v ratio of liquid to beads). The homogenate was centrifuged for 90 min a t 14,000 X g, the precipitate was discarded, Streptomycin sulfate was added to the crude extract (1g/80 g of cell paste), and thesolution was stirred for 1h a t 4 "C. After centrifugation at 14,000 X g for l h, the precipitate was discarded, the solution was kept at 50 "C for 30 min and again centrifuged at 14,000 X g for 1h, and theprecipitate was discarded. Addition of ammonium sulfate to 60% of saturation precipitated internalnonglycosylated invertase, whereas addition to 80% saturation was required to precipitate 4AL external invertase, at which concentration the mnnl mnn9 external invertase remained in solution. Each ammonium sulfatefraction was dialyzed against 10 mM sodium phosphate, pH 6.5, and applied to a DEAE-Sephacel column (1.4 X 12 em) equilibrated in the same buffer. Elution of external invertases was accomplished with a 600-ml NaCl gradient (0-0.4 M), whereas a gradient from 0 to 0.6 M was used to elute internal invertase. A flow rate of 30 ml/h was used and 5-ml fractions were collected. The fractions with invertase activity were combined, dialyzed against 10 mM sodium citrate, pH 3.7, and applied to anSP-Trisacryl column (1.4 X 8 cm) equilibrated in the same buffer. External invertaseswere eluted with a 400-ml NaCl gradient (0-0.3 M ) at a flow rate of 30 ml/ h, whereas a gradient to 0.4 M salt was used to elute internal invertase. Combined peak fractions from the SP-Trisacryl column were concentrated on an Amicon PM-30 filter to about 2 ml and the solution was applied to a Bio-Gel A-5m column (2.5 X 92 em) and eluted at 13 ml/h with 10 mM sodium acetate, pH 5.0, containing 0.1 M NaCI. The highly associated invertase (tetramer to octamer) was obtained in a peak free from impurities. Concentration and rechromatography of the lower molecular weight invertase fraction increased the yield of the pure, highly associated form. The fractions that contained enzyme activity were combined and dialyzed against distilled water before l y o p h i ~ ~ t i oorn against 10 mM sodium phosphate, pH 6.5, before further purification on a hydroxylapatite column (1.4 X 8 cm). This column was pre-equilibrated with 10 mM phosphate, pH 6.5, and the elution was carried out with 200mlof a 10-200 mM sodium phosphate gradient of the same pH. Fractions containing invertase were combined, dialyzed against distilled water and lyophilized. The recovery of mnnl mnn9 invertase from the crude cell extract was about 20%, whereas the yield of internal invertase amounted to 16% of the corresponding purified external invertase. l ~ o ~ u t of ~ oInuertase n from Culture Media-For isolation of invertase from the medium of a 200-liter fermentation, sodium dihydrogen phosphate was added to a concentration of 10 mM and adjusted to pH 6.5 with NaOH. The cells were removed with a Sharples centrifuge, sodium azide was added to the effluent to a concentration of 0.02%, and thesolution was stored a t 4 "C until it could be processed. All subsequent steps were carried out at 4 "Cin the cold room. The

invertase was concentrated by passing the medium through a DEAESephadex A-50 column (8 X 32 cm) a t a flow rate of 400 ml/h. The effluent was monitored daily for invertase activity and, when invertase began to appear, the column was eluted with a 3-liter NaCl gradient (0-0.7 M) in the same buffer. The medium from the mnnl mnn9 fermentation was processed in seven batches, whereas that from the 4AL strain required only four cycles owing to a higher affinity of the latter enzyme for the ion-exchanger. The major peak of enzyme activity from each batch, which was eluted at about 0.1 M salt, was collected separately from a later fraction that was eluted a t a higher salt concentration and was saved for isolation of internal invertase. External invertase from the mnnl mnn9 medium was recovered in the supernatant liquid after adding ammonium sulfate to the combined DEAE-Sephadex fractions to 80% of saturation a t 4 "C,whereas the 4AL invertase was found in the precipitate obtained between 55 and 90% of saturation of the corresponding DEAE-Sephadex fractions. Internal invertase was precipitated at 55% of saturation by ammonium sulfate from the latefractions of enzyme activity obtained from the DEAE-Sephadex column. The invertases obtained in the ammonium sulfate step were dialyzed against 10 mM sodium phosphate, pH 6.5, and purified as for the cellular invertases. From the medium of a 200-liter mnnl mnn9 culture, 34% ofthe invertase activity was recovered,19.1 mg of protein (12), whereas the 4AL medium yielded 31.6 mg of invertase protein, a 36% recovery. Hydroxylapatite Chromatography of Purqieied Invertases-A hydroxylapatite (high resolution, Behring Diagnostics) column (0.7 X 17 cm) was operated at a flow rate of 13 ml/h with a peristaltic pump located before the column. The column was equilibrated in 10 mM sodium p h ~ p h a t e pH , 6.5, samples of less than 2 mg of protein were introduced, and the invertases were eluted with sodium phosphate gradients, pH 6.5, up to 200 mM. Salt concentration and absorbance a t 280 nm wererecorded with flow cells installed between the column and fraction collector. The column was washed with 400 mM phosphate between samples, and fractions to be rechromatographed were equilibrated in the loading buffer on an Amicon PM30 membrane concentrator. Electrophoresis-A Hoefer vertical high-separating slab gel system was used (0.75 mm thick, 14 ern wide, and 10 cm high), with a 2.5-cm 3% stacking gel. The nondenatured multimeric forms of invertase were analyzed by the Laemmli SDS-polyacrylamide gel electrophoresis system (14) in a 4-12.5% gradient gel a t 4 "C. For separation of invertase monomers, the samples were mixed in a 1:l (v/v) ratio with buffer, containing 2.5% SDS and 6.25% mercaptoethanol, and were boiled before application. A constant current of 20-30 mA was used. For quantitation,gels werescanned ona Kratos Sp~trodensitometer Model SD3000 at 540 nm for the activity stain and at620 nm for the silver stain. All gels werecast on Gel-Bond PAG (FMC C o ~ r a t i o n ) plastic supports. The gel wascut intohalves so that thenondenatured part could be stained for invertase activity and the part with the denatured samples and molecular weight standards could be visualized with the silver stain of Morissey (18) but omitting the glutaraldehyde treatment. Protein molecular weight standards (Pharmacia P-L Biochemicals) were used for calibration. RESULTS

Isohtion of Invertase from Cells and Growth Medium of mnnl mnn9 and4AL Strains andSeparation of Oligomers by Gel Filtration-Although wild-type strains of S. cereuisiae, such as X2180, retain most of the secreted invertase in the periplasm, strains containing the mnn9 mannoprotein mutation release about 20%of the external invertase into the medium (Table I). The invertase in the medium from a 200liter culturewas concentrated by passing the solution through a large DEAE-Sephadex A-50 column. The bound invertase was eluted with a NaCl gradient and subjected to ammonium sulfate fractionation followed by chromato~aphyon DEAESephacel and SP-Trisacryl and gel filtration on Bio-Gel A5m to yield the pure enzyme. In some instances, an additional hydroxylapatite step was employed. Details and differences in the behavior of invertases from different strains are given under "Experimental Procedures." The recovery of invertase was about 35%. Similar steps were used to isolate invertase

Invertase Oligomer Stability

4397

TABLE I Release of external invertase into the medium

I . _ _ _ "

Strain

Dry cell

Invertase activity

0.8

-

Released

Cells" mg

unitslg dry cells

X2180 290 90 mnnl mnn2 90 mnnl mnnZmnn.9 301090 80 4AL ."

12 66 790 676

864 20.7 2075

%

3.8 7.1 24.6

"

Activity was assayed with whofe cells and is normalized for differences in cell mass. The higher activity with cells containing mnn mutations is not due to a greater ease of diffusion by substrate and products across the cell wall because cell homogenates gave similar values. * Native gel electrophoresis confirmed that all of the activity in the medium was due to glycosylated invertase, which eliminated cell lysis as a mechanism for enzyme release. a

from the homogenized cell paste, which led to a 20% recovery of the external invertase in the cells and some internal nonglycosylated invertase, which equaled 16% of the external form retained by the cells. Only minor amounts of nonglycosylated invertase were found in the medium, which suggests that lysis was not involved in release of invertase from the cells. The Bio-Gel A-5m chromatography step was importantly affected by the pH and invertase concentration,owing to their influence on aggregation of the dimeric enzyme to larger oligomers. In some instances, the mnnl mnn9 invertase was eluted in a major part consisting of the larger oligomers, whereas in other experiments three peaks of activity were observed (Fig. 1). External invertase from mnnl mnn9 cells gave three peaks when chromatographed on Bio-Gel A-5m, the fastest of which migrated with thyroglobulin (MI 670,000) and the slowest with ferritin half-unit (MI 220,000). These probably represent an octamer-hexamer mixture and dimer, respectively, whereas an intermediate peak ( M , 400,000) corresponds to tetramer. The expected sizes of the oligomers are 180,000 (dimer), 360,000 (tetramer), 540,000 (hexamer) and 720,000 (octamer), and why four peaks were not observed is unclear. Rechromatography of the fastest and slowest eluting material gave a similar pattern of three peaks, demonstrating the ease with which re-equilibration of the separated oligomers can occur. Gel filtration of the 4AL external invertase gave a less well-defined pattern, with a major peak corresponding to dimer (calculated M, 150,000) and other material eluting in the region for higher oligomeric forms (data not shown). Observance of Invertase Oligomers by Gel ElectrophoresisWild-type external invertase is not resolved into oligomeric bands by gel electrophoresis, apparently owing to the large and heterogeneous carbohydrate component (4). The first demonstration of such resolution came from studies on the invertase accumulated in the see18 mutantin which the carbohydrate chains have a uniform Man8GlcNAc2-structure (8).Subsequently, the invertase of the mnn9 mutant, which chains, was found possesses Manlo_13GlcNAcz-oligosaccharide to give a similar gel electrophoretic pattern of four enzymatically active bands (9). In the present study, we have confirmed this observation using the mnnl mnn9 invertase, which has uniform ManloGlcNAc2-chains,and compared its gel pattern with that of the 4AL strain. Freshly thawed invertase samples were electrophoresed at 4 "C on a gradient gel, in the presence of 0.1% SDS but without boiling, and the enzyme was visualized either with an activity stain (8) or a silver stain (18). Four bands were observed for the mnnl

c-)l

0

.t

IO

d Y

0.6

>.

t 1.

l-

2

0.4

W

cn 4

l-

aW: >

z

0.2

70

80

90

100

110

FRACTION

FIG.1. Purification and fractionation of external invertase by gel filtrationon a Bio-Gel A-5m column. Invertase-containing fractions, from chromatography of purified mnnl mnn9 external invertase on a Bio-Gel A-5m column (2.5 X 93 cm) eluted with 100 mM NaCl in 10 mM sodium acetate buffer, pH 5, were combined as shown in the inset, and the forms of highest and lowest molecular weight (peaks I and I I I ) were concentrated on an Amicon PM-30 filter and rechromatographed separately on a Bio-Gel A-5m column (1 X 195 cm) at 4 "C. Open circles, peak I; closed circles, peak 111; A, elution position of thyroglobulin (M, 670,000);B, elution position of ferritin half-unit (Mr 220,000); C, elution position of aldolase (M, 158,000).

mnn9 invertase with both reagents, the fastest moving having an apparent M, 200,000 and corresponding to thedimer, with the other bands at Mr 360,000, 560,000, and 700,000 (Fig. 2). In contrast, the 4AL invertase showed one major band that moved slightly faster than the mnnl mnn9 invertase dimer (apparent Mr 160,01)0), a slower migrating minor band (Mr 320,000), and two barely visible bands that correspond to higher oligomers. Under these same conditions, internal nonglycosylatedinvertase showed an enzymically inactive protein band of the monomer at MI 60,000 and a second diffuse band that ranslightly slower than themonomer and showed invertase activity. Thus, absence of carbohydrate chains destabilizes the dimeric invertase so that it dissociates under these conditions. The enzymatic activity may reflect reassociation of monomer in the gel to yield active dimer, since published evidence suggests that themonomer is inactive (7). As observed in the gel filtration experiments, the pretreatment of an invertase sample had a major effect on the pattern of oligomers observed following gel electrophoresis. Freezing the sample a t low pH increased the amount of higher oligomers, whereas warming the sample at higher pH led to dissociation to thedimer. Interconversion of Invertase Oligomers Observed by HFLCPurified invertase shows multiple peaks when analyzed by HPLC on a Du Pont GF-450 column that includes mouse IgM (Mr 900,000) and resolves proteins down to M , 17,500 (myoglobin) when eluted with 0.2 M sodium phosphate, pH 7.5. Most invertase preparations show three peaks that correspond in molecular weights to dimer, tetramer and hexamer + octamer, the latter two apparently failing to separate from each other. Because the oligomeric composition of the invertase does not change perceptively during chromatography,


4398

H

I

DISTANCE ALONG GEL

FIG.2. Gel electrophoresis of nativeexternalinvertase multimers. Samples were applied in a 10-fold dilution of the gel buffer with 10%glycerol and electrophoresed on a 4-12.5% gradient 8.3, containing 0.1% SDS (14). gel at 4 "Cin a Tris-glycine buffer, pH The invertase multimers were detected with the activity stain and were compared quantitatively by densitometer scan at 540 nm. Top panel, men1 men9 invertase in which the four peaks correspond,left to right, to octamer (28%), hexamer (27%), tetramer (29%), and dimer (16%). Bottom panel, 4AL invertase, inwhich the two distinct peaks correspond to tetramer (25%) and dimer (56%),with smaller amounts of octamer (8%)and hexamer (11%). The faster migration and predominance of the dimeric form in the 4AL invertase reflect its lower carbohydratecontent.Molecularmassmarkers ( m ) are indicated in the middle panel

~

K

which takes only a few minutes, the interconversion of multimers can be followed conveniently. Freezing a n invertase solution leadsto formation of the higher oligomers, probably owing to concentration of the glycoprotein. Upon thawing and chromatography, the sample shows one or more peaks T-rT" "T-r 9 10 1% 12 9 to 11 12 that re-equilibrate to a new mixture of oligomers at a rate TIME AFTER INJECTION (rnin) dependent on concentration, pH, temperature, and carbohyFIG.3. Time course for re-equilibrationof mnnl mnni) indrate structure of the invertase. The invertase from the mnnl mnn9 strain has 8-11 car- vertase multimers by HPLC. The invertase solution,3 mg/ml in 50 mM sodium phosphate, pH 5, was stored frozen. After thawing it 10 mannose units, and the HPLCwas kept at 30 "C and 1.4-fll samples were taken at different times bohydrate chains, each with pattern (Fig. 3A) of a freshly thawed sample, 3 mg/ml, shows for analysis by HPLC on the Du Pont GF-450 column by elution M , 600,000) that slowly changes with 0.3 M sodium phosphate, pH 5.5, at 1 mi/min. In the top panel one main peak (apparent over time at pH 5 and 30 "C toa mixture of three peaks (Fig. of six figures, incubationat 30 "C was for 0 min (A),15 min ( B ) ,and 4 ( E ) ,and 7 h (F).The same experiment doneat pH 8.2 3, B-F). At equilibrium, the predominant peak (41% of the 1 (C), 2 (D), gave the results in the lower panelof six figures,at 0 min (GI, 30 min t~otal) is the dimer along with tetramer(33%) and hexamer- (H),and 1 (I), 2 (J),4 ( K ) , and 22 h (L).The small peak of protein octamer (26%) (Fig. 3F). All peaks show the same specific eluting at 10.2 min inL did not haveinvertase activity and is assumed enzyme activity. When the same invertasewas analyzed in a to be an impurity or denatured enzyme. At pH 5, the equilibrium similar manner but at a concentration of 0.03 mg/ml, the mixturehad the composition shownin F, whereas at pH 8.2 the dimer eventually reached85% of the total invertase.At 3 mg/ invertase was converted totallyto the dimer shown in L. ml and pH 8.2, the dimeric form makes up 75% of the total invertase at equilibrium (Fig. 3, G-L). The ratesat which the 10 mannose units, and the HPLC pattern obtained for the poorly resolvedpeaks that, relative amounts of the three peaks change are plottedFig. in freshly thawed sample shows three 5 and 30 "C, equilibrate to a mixture inwhich the dimer a t p H 4, where the decrease in the octamer-hexamer peak(Fig. 4A) can be compared with the changes in the tetramer peak (Fig. predominates (51%)but witha significant amountof tetramer 4B) andsteadyincreasesinthedimerpeak (Fig. 4C) a t (33%)still present (data not shown). The rates at which the (Fig. 5), compared differentconcentrationsand p H values. The rates of re- changes occur for the three different peaks mnnl mnn9 invertase under the same equilibration are clearly lower at p H 8.2 (ctosed circles) than with those for the at pH5.0 (open circles) and at a concentration of 0.03 mg/ml conditions, show that re-equilibration has a strong inverse relationship to thecarbohydrate content of the protein. (open triangles) than at 3 mg/ml (open circles), the latter result reflecting the reversibility of the reaction. Wild-type invertase(Sigma) shows twopeaks when freshly The 4AL invertase has 4-7 carbohydrate chains, each with thawed, which equilibrate so that thelower molecular weight

4399


HOURS AT 30°C

1

2

I

I

I

I

3

4

5

6

HOURS AT 3OoC

FIG. 4. Rates of re-equilibration of mnnl mnn9 invertase multimers by HPLC. The invertase solution was frozen to enhance formation of higher oligomers, and thenthawed and assayed by HPLC at various times after incubation at 30 'C to follow relative changes in the three protein peaks. A is peak I, B is peak 11, and C is peak 111. The invertase solutions were at 3 mg/ml and pH 5 (open circles), 3 mJm1 and pH 8.2 (closed circles), or 0.03 mg/ml and pH 5 (open triangles). 40L

FIG. 6. Rates of re-equilibration of wild-typeinvertase multimers by HPLC. The conditions are as in Fig. 4, except that only two peaks were observed by HPLC and theinvertase solutions were 1.5 mg/ml and pH 5 (open circles) or 0.015 mg/ml and pH 5 (open triangles).

D

A

J ~

2 0 4 0 k z

TIMEAFTER INJECTION ( m i d

FIG. 7. Dependence of oligomeric composition on external invertase concentration. The invertase was dissolved in 50 mM sodium phosphate, pH 5.0, and kept at least 12 h a t room temperature (20-22 "C)before analysis by HPLC. A-D, mnnl mnn9 invertase at 0.3 (A}, 3 (B), 10 (C), and 20 mg/ml (D). E-H, 4AL invertase a t 0.3 ( E ) ,10 (F), 20 (GI, and 40 mg/ml (H). HOURS AT 30°C

FIG. 5. Rates of re-equilibration of 4AL invertase multimers by HPLC. The conditions and symbols are as in Fig. 4.

species predominates. The rates of these changes are shown in Fig. 6. Nonglycosylated invertase showed only one peak, corresponding to dimer in size, probably because equilibration was too rapid to allow detection of different intermediates even with HPLC. The concentration dependence of the oligomeric equilib-

rium composition of mnnl mnn9 and 4AL external invertases are compared in Fig. 7. A t pH 5 and a concentration of 20 mg/ml or higher, the mnnl mnn9 invertaseina solution equilibrated overnight at 20 "C exists exclusively as the octamer (Fig. 7 D ) , whereas the 4AL invertase is still mainly dissociated to dimer and tetramer at a concentration of 40 mg/ml {Fig. 7H). Confirmationof Invertase Oligomer Formation by Electron Microscopy-The procedure of Esmon et ai. (9) was used to observe oligomer formation by electron microscopy (Fig. 8).

4400


B i

A

C

FIG.8. Negatively stained electron micrographs of invertase multimers. Samples in distilled water were frozen and then thawed just before spreading on the grids. A, commercial bakers' yeast invertase; B, mnnl mnn9 external invertase; C, 4AL external invertase; and D, 5'. ceTeuisiue internal invertase. B shows a preponderance of octamers, whereas C shows dimers, tetramers, hexamers, and octamers. The distribution of multimers observed on the micrographs paralleled that seen in HPLC patterns of the same sample. Note that the octamers in A, B, and C remain open on one side, whereas many of those in D have a closed conformation. Magnification is X 172,000.

The photographs for wild-type (Fig. 8A) and mnnl mnn9 (Fig. 8B) external invertasesare similar to those alreadypublished (9), whereas the results for the 4AL invertase (Fig. 8C) and for nonglycosylated internal invertase (Fig. 8D)are novel, although evidence that the latter does associate a t a sufficiently high concentration has been reported (10). In our investigation, all invertase preparations formed octamers to various degrees depending on conditions. Unfortunately, we were unable to assess oligomer composition in fractions obtained directly from the HPLC column because elution required the presence of salt, which had tobe removed before electron microscopy could be performed. Preparations of the mnnl mnn9 and 4AL external invertases, which had been allowed to equilibrate under similar conditions, were photographed and then counted to determine the oligomeric

composition. The results were consistent with those obtained by HPLC. DISCUSSION

The association of yeastexternalinvertasedimers into higher oligomers, postulated by Neumann and Lampen (1) and documented by Chu et al. (7), Esmon et al. (9), and the present work, can be observed by gel electrophoresis, gel filtration, sedimentation velocity, HPLC, and electron microscopy. Oligomer formation is enhanced by low pH, high ionic strength, anda high concentration as well as by glycosylation of the polypeptide chains. Here we have shown that freezing of samples is a convenientway to drive the invertase intothe octameric form andthatHPLC is an excellent procedure for assessing the oligomeric composition of a prep-


4401

aration. S. cerevisiae strains containing the mnnl mnn9 mu- thickness of 15 nm. This concentration of invertase would tationsmakeinvertasewith uniformly sized carbohydrate assure complete association of the mnnl mnn9 invertase, chains, which facilitates its study, although such invertase is although the4AL invertase might besignificantly dissociated. still heterogeneous in that individualpolypeptide chains carry This point is important in assessing whether the release of 8-11 oligosaccharide units (16). A mutation in dolichol phos- invertase into themedium in these mutants results primarily phoglucose synthesis superimposed on themnnl mnn9 back- from changes inwall organization orfrom a reducedtendency ground (mnnl mnn9 d p g l ) leads to the strain we call 4AL of the invertase to form higher oligomers. In addition to its (16), which produces external invertase with4-7 oligosaccha- effect on invertase, the mnn9 mutation is known to decrease ride chains. Our quantitative comparison of the stability of the size of the carbohydrate chains on the cell wall mannoinvertase oligomers from the mnnl mnn9 and 4AL strains proteins (19) and to alter the structural integrity of the cell provides support for the report (7) that glycosylation stabilizes wall itself (24). Our results showing that mnnl mnn9 and oligomer formation. In addition, our electron micrograph of 4AL external invertases are released to similar extents,even though they differ in oligomeric stability, suggest that wall nonglycosylated invertaseconfirmsthat even the internal invertase will form octamers if sufficiently concentrated. The structural defectsmay be more responsible for enzyme release fact that protein glycosylation can enhance subunit interac- from the mutant thana reduced oligomer stability. tion may have wide biological significance. An interesting feature of the electron micrographs of the REFERENCES external invertases is the shape of the octamer, which uniS., and Lampen, J. 0.(1968) J.Bid. Chem. 2 4 3 , 1567Gascon, 1. formly has one openside. This could result if the association 1572 sites on each dimer were oriented in such a manner that the 2. Neumann, N. P.,and Lampen, J. 0.(1967) Biochemistry 6,468angle between dimers is greater than 90". In contrast, the 475 octamers formed by the nonglycosylated internal invertase 3. Gascon, S., Neumann, N. P., and Lampen, J. 0.(1968) J. Bid. Chem. 2 4 3 , 1573-1577 appear to be more nearly symmetrical, so the unusual shape 4. Trimble, R. B., and Maley, F. (1977) J. Bwl. Chem. 252,4409of the glycosylated invertase octamer might result from the 4412 presence of carbohydrate chains, which could, through steric 5. Taussig, R., and Carlson, M.(1983) Nucleic Acids Res. 1 1 , 1943interference, prevent formation of the closed conformation 1954 even though they stabilize oligomerization. 6. Chu, F., and Maley, F. (1981) Fed. Proc. 40, 1556 7. Chu, F., Watorek, W., and Maley, F. (1983) Arch. Biochem. Of broader interest is the relevance of oligomerization of Biophys. 223,543-555 invertase in a test tube to that of invertase in the periplasm 8. Esmon, B., Novick, P., and Schekman, R. (1981) Cell 2 5 , 451of the cell, a behaviorthat will be most importantly influenced 460 by pH and concentration. Owing to the permeability of the 9. Esmon, P. C., Esmon, B. E., Schauer, I. E., Taylor, A., and cell wall tosmall molecules (%I), itisprobablethatthe Schekman, R. (1987) J. Bwl. Chem. 262,4387-4394 periplasmic pH is the same as thatof the medium, about 5.5 10. Williams, R. S., Trumbly, R. J., MacColl, R., Trimble, R. B., and Maley, F. (1985) J. Biol. Chem. 2 6 0 , 13334-13341 ina culture growing in YEPD (21) and a pH that favors multimer formation (7). With regard to concentration, Arnold 11. Raschke, W. C.,Kern, K. A., Antalis, C., and Ballou, C. E. (1973) J. Bid. Chem. 248,4660-4666 (22) has calculated the volume occupied by an unhydrated 12. Lowry, 0.H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. mm3) andthehydrated form ( 5 invertasedimer (2.8 X (1951) J. Biol. Chem. 193,265-275 x mm3), which are close towhat we calculate (3.4 x 13. Kirschner, M., Honig, L., and Williams, R. (1975) J. Mol. Biol. 99,263-276 mm3) from the apparent diameter (7ofnm) the invertase Laemmli, U. K. (1970) Nature 227,680-685 dimer seen in electron micrographs. From these volumes and 14. 15. Ballou, L., Cohen, R. E., and Ballou, C. E. (1980) J. Biol. Chem. the amount of invertase assayable in intactcells, Arnold (22) 255,5986-5991 estimated the concentration of the invertase (20%) if the 16. Ballou, L., Gopal, P., Krummel, B., Tammi, M., and Ballou, C. E. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , 3081-3085 periplasmic space had a thickness of 15 nm. At this concentration, we expect from our data that the invertasewould be 17. Lehle, L., Cohen, R. E., and Ballou, C . E. (1979) J. Bwl. Chem. 254,12209-12218 completely associated into octamer and that even a t a peri- 18. Morrissey, J. H. (1981) Anal. Biochem. 1 1 7 , 307-310 plasmic thicknessof 50 nm the invertasewould still be highly 19. Tsai, P.-K., Frevert, J., and Ballou, C. E. (1984) J. B i d . Chem. associated. Thus, there seems little question that,wild-type in 259,3805-3811 cells induced to make and secrete invertase, the enzyme is 20. Arnold, W. N. (1981) in Yeast Cell Enuelopes: Biochemistry, Biophysics and Ultrastructure (Arnold, W. N., ed) Vol. 11, pp. present in the periplasm as large aggregates and might even 25-47, CRC Press, Boca Raton, FL appear toform a crystalline array such as is seen in published 21. Suomalainen, H., and Oura, E. (1971) in The Yeasts (Rose, A. electron micrographs of fractured freeze-etched yeast cell H., and Harrison, J. S., eds) Vol. 2, pp. 3-74, Academic Press, New York envelopes that show organized particles of this same dimen22. Arnold, W. N. (1973) Physiol. Chem. Phys. 5 , 117-123 sion (23). In the mnnl mnn9 and 4AL strains, which release 20% of 23. Matile, P., Moor, H., and Robinow, C. F. (1969) in The Yeasts (Rose, A., H., and Harrison, J. S., eds) Vol. 1, pp. 219-302, the external invertase into the growth medium, we find about Academic Press, New York 40 mg/ml of invertase assuming a periplasmic space with a 24. Frevert, J., and Ballou, C. E. (1985) Biochemistry 24,753-759