Synthesis, characterization and properties of sialylated catalase

Bt{

ELSEVIER

Biochi ~ m i et BiophysicaA~ta Biochimica et Biophysica Acta 1293 (1996) 90-96

Synthesis, characterization and properties of sialylated catalase Ana I. Fernandes, Gregory Gregoriadis * Centre for Drug Delivery Research, School of Pharmacy, University of London, 29//39, Brunswick Square, London WC1N lAX, UK

Abstract

Colominic acid (CA), a a-(2 ~ 8) N-acetylneuraminic acid (sialic acid) polymer (average molecular weight of 10 kDa) was activated by periodate oxidation of carbon 7 at the non-reducing end of the saccharide. The oxidized CA was then coupled to catalase by reductive amination in the presence of sodium cyanoborohydride. The extent of sialylation of catalase, estimated by ammonium sulfate precipitation as 3.8 + 0.4 (mean + S.D.) moles of CA per mole of catalase, did not improve significantly when depolymerized CA was used in the coupling reaction. At the end of the coupling reaction, sialylated catalase exhibited a two-fold (70%) retention of initial activity compared to enzyme controls (29-35%) subjected to the same conditions. Formation of sialylated catalase was confirmed by ammonium sulfate or trichloroacetic acid precipitation, molecular sieve chromatography and SDS-PAGE electrophoresis. Enzyme kinetics studies revealed an increase in the apparent K m of the enzyme from 70.0 (native) to 122.9 mmol 1- ' H202 (sialylated catalase) indicating a reduction of enzyme affinity for the substrate (hydrogen peroxide) on sialylation. Compared to native enzyme, sialylated catalase was much more stable in the presence of specific proteinases, completely resisting degradation by chymotrypsin and losing only some of its activity in the presence of trypsin. The increased stability conferred to catalase by sialylation agrees with similar observations on enzymes modified by other hydrophilic molecules (e.g., monomethoxypoly(ethyleneglycol)) and suggests that steric stabilization with the biodegradable polysialic acid may prove an alternative means to render therapeutic proteins more effective in vivo. Keywords: Catalase; Polysialic acid; Protein delivery; Covalent coupling

1. Introduction

The use of proteins in intravenous therapy can be severely hampered by short half-lives in the circulation and proteolytic degradation [1]. Moreover, administration of large doses of proteins required for the maintenance of therapeutic efficacy can often provoke toxicity and also increase the possibility of adverse immune responses. Circumvention of such problems can be achieved with various degrees of success by covalent coupling of proteins to hydrophilic macromolecules such as albumin [2], dextrans [3,4] and monomethoxypoly(ethyleneglycol) (mPEG) [1,5]. Of these, mPEG is probably the most successful and comprehensively studied and its use has now been extended to particulate drug delivery systems such as lipo-

Abbreviations: mPEG, monomethoxypoly(ethyleneglycol); CA, colominic acid; MW, molecular weight; PBS, 0.15 M sodium phosphatebuffered saline (pH 7.4); TCA, trichloroacetic acid; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; K m, MichaelisMenten constant; Vm~~, maximum velocity. * Corresponding author. Fax: +44 171 7535820. 0167-4838/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved

SSDI 0 1 6 7 - 4 8 3 8 ( 9 5 ) 0 0 2 2 7 - 8

somes [6]. Polysialic acids have been recently suggested [7] as an alternative to the non-biodegradable PEG. These are biodegradable linear homopolymers or heteropolymers of N-acetylneuraminic acid which are not known to have a receptor in the body and their catabolic products are not toxic [7]. In the present work we have employed a low molecular weight, poorly immunogenic [8] and antigenic [9] polysialic acid, colominic acid (CA) (Fig. 1), as a means to render catalase more hydrophilic. Catalase (EC 1.11.1.6) was chosen as a model therapeutic protein because of its increasing use as an oxygen radical scavenger [10-13] or in enzyme replacement therapy [14]. The rationale of this work is that a hydrophilic surface should not only improve the pharmacokinetics and stability of the enzyme but also reduce its immunogenicity, similarly to what has been observed with other modified catalases [4,5]. Here we report on the preparation and characterization of watersoluble sialylated catalase. In spite of the relatively low degree of modification (CA molecules coupled per catalase molecule) obtained, the neoglycoprotein retains much of its enzymatic activity and its stability under a variety of conditions is superior to that of the native enzyme. Prelim-

A.L Fernandes, G. Gregoriadis / Biochimica et Biophysica Acta 1293 (1996) 90-96

8

boiling water for 5 and 8 h to obtain hydrolysed colominic designated here as CA~ and CA 2 respectively. Gel filtration chromatography on a Sephadex G-50 column (0.9 × 55.0 cm) confirmed that the order of size was CA > CA 1 > CA 2. CA1 and CA 2 were then oxidized and coupled to catalase as described in Section 2.4 for intact CA.

o~----t--?~[,~_~_cH31

9 CH2OH~o~

5

In

L 1 7~on s

J on

9

cn--o,--4,~-~-cn3 I

91

2.4. Preparation of sialylated catalase

2

9 CH2OHoo~O~ 5

I

9 CH2OH Fig. 1. Structure of colominic acid. N-acetylneuraminic acid units are linked via ~-(2 ~ 8) glycosidic linkages [17]. Arrow indicates the carbon atom (C 7) at the non-reducing end ,3f the sugar where periodate oxidation introduces an aldehyde group.

inary accounts of this work have been presented elsewhere [15,16].

2. Materials and m e t h o d s

Bovine liver catalase Chydrogen peroxide: hydrogen peroxide oxidoreductase, EC 1.11.1.6) twice crystallized with a specific activity of :58000 U / m g protein (E10~ = 13.5), colominic acid (CAll, a a-(2---> 8) N-acetylneuraminic acid polymer [17] from Escherichia coli KI (sodium salt, average MW 10 kDa as determined by low angle laser light scattering), trypsin (from bovine pancreas), a-chymotrypsin (from bovine pancreas) and pronase E (from Streptomyces griseus) were obtained from Sigma (Poole, Dorset, UK). Sephadex and Sepharose were from Pharmacia-LKB Biotechnology (St. Albans, Herts, UK). All other reagents were of analytical grade. A Wallac Compuspec UV-visible spectrophotometer was used in all spectrophotometric determinations.

2. I. Assay of catalase activity Catalase and sialylated-,zatalase activity was measured [18] by monitoring spectrophotometrically at 240 nm the disappearance of hydroge~ peroxide. One unit (U) of catalase is defined here as the amount of enzyme required to decompose 1 mmol of hydrogen peroxide, per min at pH 7.0 and 25°C.

2.2. Assay of colominic acid Total sialic acid content was estimated by the resorcinol method [19].

2.3. Depolymerization of colominic acid A 10 mg ml -l solution of CA in 0.15 M sodium phosphate-buffered saline of pH 7.4 (PBS) was placed in

Colominic acid (or its hydrolysed derivatives) was activated by controlled periodate oxidation as described by Jennings and Lugowski [20]. At the end of the reaction, oxidized CA was isolated from reagents by gel filtration on a Sephadex G-25 column equilibrated with deionized water. The sialic acid positive [19] fractions were pooled and freeze-dried. The oxidized polysialic acids were coupled to catalase (at varying molar ratios) by reductive amination [20] in the presence of sodium cyanoborohyride (NaCNBH 3) in 5 ml of 0.75 M dipotassium hydrogen phosphate (pH 9.0). Controls of catalase alone (in the absence of CA and reagents) or mixed with NaCNBH 3 as above were included. The mixtures were magnetically stirred in sealed vials for 48 h at 35-40°C. The reaction was followed by the concomitant precipitation of protein and covalently linked CA by trichloroacetic acid (TCA) (in preliminary experiments it was established that CA is not precipitable by TCA): aliquots taken at time intervals were mixed with TCA (10% w / v final concentration), allowed to stand at 4°C for 30 min, centrifuged at 3000 × g for 30 min and pellets and supernatants assayed for CA [19]. CA linked to the enzyme was expressed in terms of CA: catalase molar ratio, assuming that all enzyme molecules had been modified and that precipitation of protein by TCA was quantitative. In a separate experiment the stability of both the sialylated catalase and catalase controls under the coupling conditions employed was evaluated in terms of enzyme residual activity. In another experiment zero time and 48 h aliquots (500 /xl) of the reaction mixture were centrifuged at 3000 × g for 30 min, the supernatants were chromatographed on a Sephadex G-100 column (1.1 × 45.0 cm) and eluted fractions (1 ml) were assayed for CA [19] and catalase [21] contents. Fractions eluted in the void volume were subjected to polyacrylamide gel electrophoresis.

2.5. Polyacrylamide gel electrophoresis (PAGE) Prior to electrophoresis, samples were desalted on Sephadex G-25 mini-columns (Pasteur pipettes equilibrated with deionized H20) and boiled for 1 min in the presence of sodium dodecyl sulfate (SDS) and /3-mercaptoethanol. Electrophoresis was performed on Phast System (Pharmacia) using Phast Gel homogeneous 12.5% polyacrylamide and Phast Gel SDS buffer strips according to the manufacturer's instructions. Gels were stained with Coomassie brilliant blue. High molecular weight markers

92

A.L Fernandes, G. Gregoriadis / Biochimica et Biophysica Acta 1293 (1996) 90-96

(myosin, 212 kDa; a-2-macroglobulin, 170 kDa; flgalactosidase, 116 kDa; transferrin, 76 kDa and glutamate dehydrogenase, 53 kDa) were also from Pharmacia. Gels were evaluated by computerized image analysis.

2.6. Isolation of the sialylated catalase Catalase conjugates with CA (using a CA: catalase molar ratio of 50:1 in reaction mixture) were precipitated by salting-out with ammonium sulfate. In short, (NH4) 2SO 4 was slowly added to an aliquot of the reaction mixture to 70% ( w / v ) saturation and the suspension was stirred for 1 h at 4°C. It was then centrifuged at 3000 × g for 40 min and the pellet, containing the conjugate, was washed with a saturated (NH4)2SO 4 solution to remove free, non-covalently linked CA and centrifuged again at 3000 × g for 10 min. The pellet was redissolved in PBS, dialysed extensively against the same buffer and its protein ( A 4 0 5 ) a n d CA [19] contents determined. The isolated conjugate was then subjected to the kinetic and stability studies described below. As with TCA (see above), in preliminary experiments it was established that CA is not precipitable with ( N H 4 ) 2 S O 4.

2.7. Enzyme kinetics Initial velocity rates for native and sialylated catalase were determined spectrophotometrically (A24o), at 20°C in 0.05 M sodium phosphate buffer (pH 7.0) using increasing H202 concentrations. The apparent K m and Vmax were estimated by the Hanes-Woolf plot [22], using the reaction kinetics software (Pharmacia LKB Biochrom, Cambridge, UK, version 2.1, 1994) of the spectrophotometer.

2.8. Exposure of catalase to proteinases Native and sialylated catalase were subjected to proteolysis at 37°C in the presence of trypsin, chymotrypsin or pronase E in PBS. Each reaction mixture (total volume 500 /xl) contained 1 mg ml - l catalase and 0.2 mg m1-1 proteinase (0.5 mg ml- 1 in the case of pronase E). Aliquots removed at time intervals, were diluted appropriately and assayed for catalase activity.

3. R e s u l t s a n d d i s c u s s i o n

3.1. Preparation and characterization of sialylated catalase The periodate-activated CA was coupled to catalase by reductive amination in the presence of NaCNBH 3 at a 10:1 and 50:1 CA:enzyme molar ratios. The reaction was followed by TCA precipitation of the protein and results, expressed in terms of CA: catalase molar ratio found in the precipitates, are presented in Fig. 2B. The initially rapid

4.S

(100:11 A

.o.,..,,,3.0

:~

" 1.5 "3 E

0

•~

3-

N

2

ISO:~l

- f

91~ (..)

110:11

1

0

I

B

I

I

I

I

I

e

i

0 6 12 18 24 30 36 42 48 Time (h)

Fig. 2. Colominic acid (CA): catalase molar ratios in the conjugates as estimated by TCA precipitation. (A) Depolymerized CA 1 (11) and CA 2 ([]); (B) intact CA. Molar ratios of colominic acid: catalase used in the coupling reaction are given in parentheses. Values denote mean +S.D. from three separate experiments. CA~ and CA 2, denote depolymerized CA with molar mass lower than CA in the following order: CA > CA i > CA 2.

reaction, observed with both ratios, reaches a plateau after about 24 h indicating termination of the reaction. Previously, conjugation of sugars to proteins had been reported [20,23] as a very slow process that took 11-13 days to completion. The degree of modification of catalase was found to be low (up to 3 . 8 _ 0.4 mol of CA per mole of catalase), although in agreement with values obtained for other proteins modified by the same method [20]. The possibility that the low yield of the reaction was due to the high molecular weight of the reacting moieties, was examined using CA of reduced weight. CA was depolymerized as above into two smaller derivatives designated as CA~ and CA 2 (order of size CA l > CA 2) which were then coupled to catalase. As anticipated, the reaction was faster (Fig. 2A) than with the intact CA (Fig. 2B), presumably because of the increased mobility of the smaller CA molecules and accessibility to binding sites of the protein. However, the molar ratio of CA:catalase in the conjugates obtained with the depolymerized CA was only slightly improved, with no apparent correlation between degree of modification and size (Fig. 2). For instance, when CA 1 and CA 2 were used at the same (100:1) colominic acid to catalase molar ratio in the reaction mixture, there was no difference in the number of CA 1 or CA 2 molecules linked to the enzyme (Fig. 2A). Moreover, CA 1 used at a molar ratio of 100:1 during a 48 h reaction period, gave the same degree of coupling as did intact CA at one-half the ratio (50:1). This suggests that factors other than CA size are involved in the coupling yield. It is apparent, however, that the extent of sialylation is dependent on the molar ratio CA:enzyme used in the coupling reaction.

A.L Fernandes, G. Gregoriadis / Biochimica et Biophysica Acta 1293 (1996) 90-96 100

T.

=o

93

ing the beads of the gel. Nonetheless, the appearance of a small shoulder corresponding to the conjugated CA eluting together with the enzyme in the void volume (V0) of the column is evident (Fig. 4A). To further confirm the presence of the conjugate, the

80 60

•,~m 40 t-

~t 20

0

I

6

I

"I

I

I

I

I

V0

I

12 18 24 30 36 42 48 1.5

^

-

1.5

Time (hi Fig. 3. Stability of catalase during the coupling procedure. Native catalase (O); sialylated catalase (©); native catalase mixed with NaCNBH 3 ([]). The sialylated enzyme was prepared with a CA: catalase molar ratio of 50:1. Results are mean 5: S.D. of lhree to four separate experiments.

At the end of the coupling procedure, the sialylated catalase showed a two-foht retention of its stability (70.8 ___1.3% of initial activity) compared to the controls of native catalase (29.1 4- 0.'9%) and catalase mixed with NaCNBH 3 (35.5 _ 3.0%) (Fig. 3). Ideally, a control in the presence of CA but without the reducing agent should have also been included to determine whether the sugar (CA) itself has a protective effect, similarly to that described for other polyols [24]. However, at the pH (9.0) and temperature (35-40°C) of the coupling reaction, the intermediate Schiff' s bases formed can undergo an Amadori rearrangement to form stable ketoamines [25], rendering such a control a false one. Moreover, the retention (70.8 41.3%) of enzyme activity by the conjugated catalase observed when a CA-to-catalase molar ratio of 50:1 was used in the reaction mixture (Fig. 3) was not significantly different from that (72.1 :t- 1.5%) measured at the end of the reaction for the conjugated enzyme when a lower molar ratio (10:1) was used (data not shown). Thus, activity retention by the sialylalced catalase seems to be independent of the amount of coupled CA. Evidence for the presence of sialylated catalase in the reaction mixture was also obtained by size exclusion chromatography on Sephadex G-100. Reaction mixtures were centrifuged to remove arty insoluble material, chromatographed and the collected fractions measured for protein (A595) and total sialic acid. An aliquot taken immediately after adding all the reagents (zero time) was regarded as a control (Fig. 4B). Although in theory the difference in molecular weight of catalase (240 kDa; [26]) and CA (average of 10 kDa) should have been enough to achieve separation, chromatography under the conditions described proved to be a low resolution technique as shown by the overlapping of the two peaks (Fig. 4B). However, this is not surprising as, in addition to CA being polydisperse, highly sialylated macromolecules (CA in this case) are known to behave abnormally on gel filtration [27], probably because the negative charges prevent them from enter-

o

1.0

1.0

0.5

0.5

i 0

,~

0 0

10

20

30

40

50

<: -

"o 1.5

1.5

B

m _~

0"~

U .~.

1.0

1.0

oE o

_

0

0.5

0.5

0

|

o

10

20

Fraction

30

50

40

number

- 212 - 170 - 116 - 76

- 53

a

b

c

d

c

f

g

Fig. 4. Molecular sieve chromatography of catalase and colominic acid, (A) After reaction for 48 h; (B) zero time, Colominic acid: catalase molar ratio was 50:1. Samples were chromatographed on Sephadex G-100 (column: 1.1 )<45.0 cm; sample volume, 500 p,l; eluent, PBS; flow rate, 14 ml h - l ) . Arrow indicates the elution of the conjugate in the void volume (V0) of the column. Panel (C) shows SDS-PAGE (12.5% polyacrylamide, stained with Coomassie brilliant blue) of the following: eluted fractions 12 and 13 (a, b); high molecular weight markers (c); eluted fractions 14 and 15 (d, e); native catalase (f); zero time reaction mixture (g). Numbers in the inset indicate the molecular mass (kDa) of the markers. For other details see Section 2.

94

A.L Fernandes, G. Gregoriadis/Biochimica et Biophysica Acta 1293 (1996) 90-96

sialic acid and catalase positive fractions (fractions 12-15) obtained as above, were electrophoresed (Fig. 4, inset; lanes a,b,d and e, respectively) on SDS-PAGE together with controls of native enzyme (lane f) and zero time reaction mixture (lane g). As anticipated [26], both controls exhibit sharp bands of dissociated protomers of about 60 kDa under denaturing conditions. In contrast, the enzyme patterns in the void volume fractions appeared broad in shape suggesting [28] microheterogeneity as a result of the carbohydrate chains, thus supporting the presence of a neoglycoprotein. It is well established [29] that because of the weak affinity of SDS for the carbohydrate chains of glycoproteins, estimation of their molecular mass by SDS-PAGE is not accurate. However, the negatively charged sialic acid residues are expected [29] to partially compensate for the reduced SDS binding thus rendering the measurement of molecular mass more precise. This was estimated to be in the region of 54.6+ 1.1 to 96.1 ___3.0 kDa for the sialylated monomers. Observed (Fig. 4, inset) faint higher molecular mass bands (-- 140 kDa) on the other hand, are likely to represent cross-linked catalase monomers formed through the function of CA molecules as dialdehydes. In this respect, a-t2 ~ 8) linkages are, under the mild conditions used here, resistant to periodate oxidation [30] and the activated CA would therefore be expected to have essentially one terminal aldehyde group at C 7 of the non-reducing end (see Fig. 1). However, the other ketonic group at the reducing end, although less reactive because of its involvement in the hemiketal linkage, is in a tautomeric equilibrium and could potentially react with the amino groups of the enzyme. This possibility is supported by the formation of a conjugate (at much lower yields than observed with activated CA) when non-oxidized CA is used in the reaction (data not shown).

3.2. Purification of sialylated catalase Further characterization of the sialylated catalase necessitated its purification. To this end, because of the poor resolution obtained with gel filtration, ammonium sulfate precipitation was used instead. The salting-out conditions were optimized to 70% salt saturation and the pellets, after being washed with ( N H 4 ) 2 S O 4 and dissolved in PBS, were analysed for enzyme and CA contents. The degree of enzyme modification (moles of CA per mole of catalase) following sialylation was calculated to be 3.80 + 0.35 (mean + S.D.; n = 3), this being a greater value than that (2.59_ 0.08) estimated after TCA precipitation. Pellets obtained with ammonium sulfate precipitation were resuspended in PBS and reprecipitated with TCA as described. It was found that a significant proportion (23.14 + 2.17 of the total CA; n = 3) of CA did not precipitate and was recovered in the supernatant. It is conceivable that such CA represents catalase species, perhaps highly sialylated, that are not precipitable in the lower pH of TCA. Ammo-

I_= - 300

150

-125

-100

-75

-50

-25

0

I

I

25

50

[H202] (raM) Fig. 5. Hanes Woolf plot for native (O) and sialylated (11) catalase. Results are mean + S.D. of three independent experiments, performed at 20°C in 0.05 M sodium phosphate buffer (pH 7.0). V0 denotes initial velocity. K m values of native and sialylated catalase, obtained by extrapolation to the abscissa ,were compared by a Student's t-test ( p <

0.0001).

nium sulfate precipitates of sialylated catalase were used for the enzyme kinetics and stability studies following extensive dialysis against PBS.

3.3. Enzyme kinetics Chemical modification of enzymes is known [31] to often adversely affect their Michaelis-Menten constant ( K m) and thus limit their efficacy in vivo [31]. It was therefore deemed appropriate to study the kinetic properties of native and sialylated catalase with their apparent K m calculated by the method of Hanes and Woolf: the substrate concentration divided by the reaction rate ([S]/V) was plotted against substrate concentration ([S]) with the abscissa intercept, ordinate intercept and slope being - K i n , Km//Vrnax and 1/Vmax, respectively. Values (mean __+S.D.) of 69.96 + 5.99 (native) and 122.88 + 4.33 (sialylated catalase) mmol 1-1 H202 respectively (Fig. 5) indicate a statistically significant ( p < 0.0001; Student's t-test) reduction of enzyme affinity to the substrate upon sialylation. Vmax values estimated from the same plot, were 0.390 AA240 min -1 and 0.644 AA240 min -1, respectively.

3.4. Effect of sialylation on the stability of catalase The stability of sialylated catalase was investigated in terms of retention of enzyme activity in the presence of proteolytic enzymes (which is relevant to the effective function of catalase in therapy). Fig. 6B shows that sialylated catalase (degree of modification 2.59 + 0.08; TCA precipitation) was nearly completely protected against chymotrypsin even after 3 h of exposure, presumably because the hydrophilic layer formed around the protein prevents this endopeptidase from approaching the hy-

A.L Fernandes, G. Gregoriadis / Biochimica et Biophysica Acta 1293 (1996) 90-96 100

A

80 60 40 20 0 I00

~ 6o ~

4O

'E 2o 0 100 BO

4O 2O 0

i

0.0 0.5

i

1.0 1.5 2.0 2.5 3.0

Time {h) Fig. 6. Exposure of catalase to proteases. Native ( 0 ) and sialylated (O) catalase were incubated in the presence of trypsin (A), a-chymotrypsin (B) and pronase E (C) at 37°C. Samples removed at time intervals were assayed spectrophotometrically. Values are mean:l: S.D. (three separate preparations). For other details s,~e Section 2.

drophobic amino acids residues where it exerts its catalytic activity. The modest modification of the lysine residues (sites of tryptic action) of the sialylated catalase probably accounts for its less marked stability observed on exposure to trypsin (Fig. 6A). As expected, however, both native and sialylated catalase completely lost their activity within 2 h in the presence of the non-specific protease from Streptomyces griseus (Fig. 6C). Catalase is known [321] to lose much of its activity on lyophilization. As lyophilized enzyme would be a preferred form of storage and transportation, it was therefore of interest to see whether sialylation of catalase improves its stability after freeze-drying. Results (not shown) indicate that whereas native catalase retains only 20.1 _ 3.2% (3 preparations) of its initial activity, the sialylated enzyme exhibits improved stabiliLy with 37.3 + 4.9% (3 preparations) activity retention. This can be attributed to the fact that, even in the lyophilized state, glycoproteins are never devoid of the hydrophilic microenvironment, necessary for the maintenance of their conformation. In other experiments where solutions of sialylated catalase were stored at - 2 0 ° C for six months, no appreciable loss of enzymatic activity was observed (results not shown). The stability of the bond between enzyme and CA under these conditions was confirmed by the quantitative and concomitant reprecipitation of catalase and CA with ammonium sulfate.

95

The increase in enzyme stability via sialylation has also been observed when other hydrophilic macromolecules (e.g., mPEG of varying molecular weights [1,3,5,33,34], albumin [2], cellulose [35] and dextrans [4]) are grafted to enzymes. In the case of proteins conjugated to mPEG for instance, the chains of the polymer are thought [34] to move freely in the water surrounding the proteins thus necessitating a high degree of modification in order to achieve protection and stability. The considerable promotion of catalase stability observed with the low number (3.8) of colominic acid residues per molecule of enzyme is difficult to explain at present. However, it may be that by virtue of its highly hydrophilic nature and negative charge density, CA, once attached to the enzyme, interacts ionically with neighbouring amino-acid residues of opposite charge to form an extensive hydrophilic layer which protects the inner regions of the enzyme. On the other hand, it is also conceivable that the tertiary structure of the sialylated catalase is altered in a way that it becomes less protease sensitive. Work is in progress to establish whether steric stabilization of enzymes by sialylation will render these less immunogenic and also increase their half-lives in the circulation thus contributing to a more effective use of enzymes and other proteins in therapy.

Acknowledgements We thank Prof. W. Gibbons and Dr. A. Nicolaou, Department of Pharmaceutical Chemistry, for their help with the electrophoresis experiments. This work was supported by a grant (BD/2158/92-ID) from JNICT (Portugal).

References [1] Nucci, M.L., Shorr, R. and Abuchowski, A. (1991) Adv. Drug Deliv. Rev. 6, 133-151. [2] Poznansky, M.J. (1986) in Methods of Drug Delivery (Ihler, G.M., ed.), Section 120, pp. 59-82, Pergamon Press, Oxford. [3] Benbough, J.E., Wiblin, C.N., Rafter, T.N.A. and Lee, J. (1979) Biochem. Pharmacol. 28, 833-839. [4] Marshall, J.J., Humphreys, J.D. and Abramson, S.L. (1977) FEBS Lett. 83, 249-252. [5] Abuchowski, A., McCoy, J.R., Palczuk, N.C., Van Es, T. and Davis, F. (1977) J. Biol. Chem. 252, 3582-3586. [6l Lasic, D. and Martin, F. (eds.) (1995) in Stealth Liposomes, CRC Press, Boca Raton, FL. [7] Gregoriadis, G., McCormack, B., Wang, Z. and Lifely, R. (1993) FEBS Lett. 315, 271-276. [8] Wyle, F.A., Artenstein, M.S., Brandt, B.L., Tramont, E.C., Kasper, D.L., Altieri, P.L., Berman, S.L. and Lowenthal, J.P. (1972) J. Infect. Dis. 126, 514-522. [9] Mandrell, R.E. and Zollinger, W.D. (1982) J. Immunol. 129, 21722178. [10] Zimmerman, J.J. (1991) Chest 100(Suppl. 3), 189S-192S. [lll Jones, J.B., Cramer, H.M., Inch, W.R. and Lampe, H.B. (1990) J. Otolaryngol. 19(5), 299-306.

96

A.I. Fernandes, G. Gregoriadis / Biochimica et Biophysica Acta 1293 (1996) 90-96

[12] Greenwald, R.A. (1990) Free Radic. Biol. Med. 8, 201-209. [13] Greenwald, R.A. (1991) Free Radic. Res. Commun. 12-13, 531-538. [14] Scott, M.D., Lubin, B.H., Zuo, L. and Knypers, F.A. (1991) J. Lab. Clin. Med. I18(1), 7-16. [15] Fernandes, A.I. and Gregoriadis, G. (1994) J. Pharm. Pharmacol. 46(Suppl. 2), 1037. [16] Fernandes, A.I. and Gregoriadis, G. (1994) Eur. J. Pharm. Sci. 2, 111. [17] McGuire, E.J. and Binkley, S.B. (1964) Biochemistry 3, 247-251. [18] Beers, R.F., Jr. and Sizer, I.W. (1952)J. Biol. Chem. 195, 133-140. [19] Svennerholm, L. (1957) Biochim. Biophys. Acta 24, 604-611. [20] Jennings, H.J. and Lugowski, C. (1981) J. Immunol. 127, 10111018. [21] Bradford, M. (1976) Anal. Biochem. 72, 248-254. [22] Palmer, T. (1985) Understanding Enzymes, 2nd Ed., pp. 126-128, Ellis Horwood, Chichester. [23] Gray, G.R., Schwartz, B.A. and Kamicker, B.J. (1978) in Progress in Clinical and Biological Research (Marchesi, V.T. et al, eds.), Vol. 23, pp. 583-594, Alan R. Liss, New York. [24] Combes, D., Graber, M. and Ye, W.N. (1990) Ann. New York Acad. Sci. 613, 559-563.

[25] Isbell, H.S. and Frush, H.L. (1958) J. Org. Chem. 23, 1309-1319. [26] Aebi, H.E. (1983) in Methods of Enzymatic Analysis (Bergmeyer, H.U., Bergmeyer, J. and Gral31, M., eds.), Vol. III, pp. 273-286, Verlag Chemie, Weinheim. [27] Alhadeff, J.A. (1978) Biochem. J. 173, 315-319. [28] Carlson, S.R. (1993) in Glycobiology (Fukuda, M. and Kobata, A., eds.), pp. 14-15, Oxford University Press, Oxford. [29] Segrest, J.P. and Jackson, R.L. (1972) Methods Enzymol. 28, 54-63. [30] Lifely, M.R., Nowicka, U.T. and Moreno, C. (1986) Carbohydr. Res. 156, 123-135. [31] Abuchowski, A. and Davis, F.F. (1981) in Enzymes as Drugs (Holcenberg, J.S. and Roberts, J., eds.), pp. 367-383, Wiley, New York. [32] Deisseroth, A. and Dounce, A.L. (1970) Physiol. Rev. 50(3), 319375. [33] Beckman, J.S., Minor, R.L, White, C.W., Repine, J.E., Rosen, G.M. and Freeman, A.B. (1988) J. Biol. Chem. 263, 6884-6892. [34] Veronese, F.M., Sartore, L., Schiavon, O. and Caliceti, P. (1990) Ann. New York Acad. Sci. 613, 468-474. [35] Somers, P.J. and Barker, S.A. (1968) Carbohydr. Res. 8, 491-497.