- Email: [email protected]
Selective modification of Sendai virus hemagglutinin neuraminidase by pyridoxal 5′-phosphate: evidence for an allosteric modulation of neuraminidase activity
Biochimica et Biophysica Acta, 1161(1993)323-327
323
© 1993 Elsevier Science Publishers B.V. All rights reserved 0167-4838/93/$06.00
BBAPRO 34402
Selective modification of Sendai virus hemagglutinin neuraminidase by pyridoxal 5'-phosphate: evidence for an allosteric modulation of neuraminidase activity Tiziana Bellini a, Maurizio Tomasi b and Franco Dallocchio
a
a Istituto di Chimica Biologica, Universit~ di Ferrara, Ferrara (Italy) and t, Laboratorio di Biologia Cellulare, Istituto Superiore di Sanitd, Roma (Italy)
(Received 5 August 1992)
Key words: Hemagglutinin;Neuraminidase; Sendai virus; Allosteric inhibition Incubation of Sendai virus with pyridoxal 5'-phosphate (PLP) causes inhibition of hemolytic activity, a slight reduction of hemagglutinating activity, and an increase in neuraminidase activity. The effects on hemagglutination and neuraminidase are prevented by the presence in the incubation mixture of sialyl lactose, a substrate of hemagglutinin-neuraminidase. Incubation with PLP of the water-soluble enzymatic domain of the neuraminidase has no effect on enzymatic activity, while the allosteric inhibition (Dallocchio et al. (1991) Biochem. Int. 25, 663-668) disappears. Both virus-bound and solubilized neuraminidase are selectively modified by PLP at the lysine-553. Our data suggest that PLP inactivates a previously undetected inhibitory site on the viral neuraminidase, and that a physiological effector is present on the viral envelope.
Introduction Sendai virus envelope contains two glycoproteins, the fusion protein (F) and the hemagglutinin neuraminidase (HN). F protein causes the fusion of the viral envelope with host cell membrane [1,2], while H N protein is not only able to bind and hydrolyse the sialic acid receptors [1,3], but is also required for the fusion process [4-6]. These different activities raised the question whether H N has a single multifunctional sialic acid binding site, or hemagglutinating and neuraminidase activities involve two different sialic acid binding sites. Furthermore, it is totally unknown how HN is involved in virus-cell fusion. We have previously reported that the water soluble, C-terminal fragment of H N (cHN), prepared by selective proteolytic digestion of the viral glycoprotein [7], is uncompetitively inhibited by fetuin and asialofetuin [8]. This inhibition can be observed only on cHN at low temperature, while virus-bound H N is not inhibited either at 37°C or at low temperature. However, the presence on cHN of an inhibitory mechanism, which can shift the neuraminidase to an hemagglutinin, can
Correspondence to: F. Dallocchio, Istituto di Chimica Biologica,Via Borsari 46, 1-44100Ferrara, Italy. Abbreviations: HN, hemagglutinin neuraminidase; F, fusion protein; PLP, pyridoxal5'-phosphate; TFA, trifluoroaceticacid.
explain how a single sialic acid binding site can display different functions. In order to investigate whether a similar inhibition could occur not only on cHN but also on viral H N at physiological temperature, we looked for a reagent which could bind selectively to the inhibitory site. In the present report we show that pyridoxal 5'phosphate, a widely used reagent for selective modifications of proteins and enzymes, specifically modifies cHN and H N at the lysine-553. This modification protects cHN against inhibition, and increases the neuraminidase activity of viral HN. This suggests that, under physiological conditions, the activity of viral neuraminidase could be modulated by an inhibitory mechanism.
Materials and Methods Z strain Sendai virus was grown in the allantoic fluid of 10-day-old fertilized eggs. T h e virus was purified by differential centrifugation [9]. Fetuin, sialyl lactose, and pyridoxal 5'-phosphate were from Sigma. HPLC grade solvents were from Baker. Asialofetuin was prepared .as previously described [8]. Fresh human erythrocytes were washed three times with 10 mM phosphate buffer (pH 7.4), containing 145 mM NaCI (PBS).
324 Neuraminidase assay was carried out in 0.1 M acetate buffer (pH 5.0). cHN was incubated for 60 min at 4°C with either fetuin or sialyl lactose as substrate, and viral H N was incubated for 5 min at 37°C with sialyl lactose as substrate. The sialic acid released was assayed using the method of Warren [10]. Hernagglutination assay. Serial two-fold virus dilutions in PBS (50/~1) were prepared in V-shaped multiwell plates. An equal volume of 3% (v/v) erythrocytes in PBS was added to each well, and the minimum agglutinating concentration was determined after 1 h at 4°C. In control experiments, PLP was added to the erythrocyte suspension before mixing with virus, and did not affect hemagglutinating activity. Hemolysis assay. Virus (20 izg/ml) was mixed with an equal volume of 3% (v/v) erythrocytes in PBS at 37°C. At the desired times, 50 /zl of the reaction mixture was diluted with 1 ml of cold PBS, centrifuged, and the absorbance of the supranatant at 413 nm was determined. Modification of neuraminidases by PLP. cHN or H N were incubated with 5 mM PLP in PBS at room temperature. At the desired times, samples were assayed for neuraminidase activity. In control assays PLP was added to the assay mixture just before addition of the neuraminidase. All assays were in triplicate. Stoichiometry of PLP labelling. The stoichiometry of the labelling of viral glycoproteins by PLP was calculated as the ratio between fluorescence of the pyridoxyl fluorophore (excitation 325 nm, emission 395 nm), and fluorescence of protein (excitation 280 nm, emission 340 nm). Sendai virus or cHN were labelled by incubation with 5 mM PLP, in either the presence or absence of sialyl lactose. The modified proteins were purified by SDS-PAGE followed by electroelution, and fluorescence was measured on the electroeluate using electroelution buffer as a blank. To measure the effect of asialofetuin on the labelling of cHN by PLP, 0.1 mg of cHN was incubated in 150 Ixl of PBS with PLP in the presence of 13 m g / m l of asialofetuin, in either the presence or absence of 5 mM sialyl lactose. Samples were passed through small columns (0,5 x 3 cm) of D E A E cellulose, equilibrated at pH 6, which retains asialofetuin but not cHN. The eluates were lyophilized and cHN was further purified by SDS-PAGE and electroelution.
Identification of the amino acid labelled by PLP. Sendai virus (3 mg), or cHN (0.3 mg), were incubated for 10 min at room temperature with 5 mM PLP, and the reaction was blocked by addition of solid NaBH 4 to reduce the Schiff base between the inhibitor and the protein; the excess of borohydryde was destroyed by acidification. The protein sample was treated with 1 mM D T T (30 min), then with 2 mM iodoacetamide. After 30 min at room temperature the sample was boiled for a few minutes with 5% SDS containing 0.2
M mercaptoehthanol, and run on a 10% SDS-PAGE [11]. The gel was stained with zinc chloride [12]; the band corresponding to HN (or to cHN) was excised and destained with EDTA. The protein, electroeluted with a Biorad apparatus, was lyophilized, SDS was extracted with acetic acid-triethylamine-acetone [13], and the pellet was dissolved in 50 mM ammonium bicarbonate (pH 8.0). The protein was digested for 5 h at 37°C with chyrnotrypsin (4% w/w), and peptides were separated on a 200 x 4.6 mm C~8 reverse-phase column. Elution was performed at 1 m l / m i n with a linear gradient of acetonitrile in water, both containing 0.02% TFA. In order to detect fluorescence of the PLP-labelled peptides, column effluent was neutralized by mixing with 0.2 M sodium phosphate (pH 7.4), delivered at 0.2 m l / m i n with a Varian 2010 pump. The fluorescence of the neutralized effluent was measured with a Varian 2070 fluorescence detector (excitation 330 nm, emission 395 nm). The single fluorescent peak was collected, lyophilized, dissolved in 50 tzl of water and rechromatographed on the same H P L C apparatus with an isocratic elution with 82% water, 18% acetonitrile, 0.02% TFA. Amino-acid analysis. The fluorescent peak eluted from the second HPLC was lyophilized and hydrolyzed in HC1 vapor at 160°C for 8 h. The hydrolyzate was neutralized with a few microlitres of saturated NaHCO 3, reacted with o-phtalaldehyde/acethyl-cysteine reagent and analyzed by H P L C [14]. Results
Effects of PLP on the neuraminidase activity of cHN cHN is inhibited at 4°C by asialofetuin, and this causes an apparent substrate inhibition when neuraminidase activity is assayed using fetuin as substrate [8]. Other substrates, such as sialyl lactose and o~l-acidic glycoprotein, do not show substrate inhibition. Although incubation of cHN with PLP up to 30 min does not affect enzymatic activity on sialyl lactose, it nevertheless prevents inhibition by asialofetuin (Fig. 1A), and avoids substrate inhibition by fetuin (Fig. 1B). Loss of inhibition was not due to interaction between fetuin and PLP during the enzymatic assay: indeed, preincubation of cHN with PLP was required in order to remove the inhibition. As shown in Fig. 1B, PLP slightly reduced the enzymatic activity of cHN using fetuin as substrate; however, substrate inhibition was still present when PLP was added to the assay mixture without preincubation with cHN. The presence in the incubation mixture of either sialyl lactose or asialofetuin did not protect against loss of the allosteric inhibition, nor did it reduce the extent of PLP incorporation, as judged by the fluorescence intensity of cHN-bound pyridoxyl fluorophore (Table I). On the other hand, if
325 .-
40
A
30
...-'"
==
-
20
o
-:-'-'-'- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
"'-1
'
y
10 ~ " / t ~ / ~ /
~
'
I
3b
i5
Li5
6b
TIME (mln)
2
2
5
5
1/[SIALYL] (n~1-1 ) Fig. 1. Effect of PLP on the enzymatic activity of cHN. All assays were performed in 0.1 M acetate buffer (pH 5), for 60 min at 4°C. (A), Lineweaver-Burk plot of the activity of cHN with sialyl lactose as substrate. (o), untreated cHN; (e), untreated cHN in the presence of 20 m g / m l of asialofetuin; ( • ) , cHN preincubated 10 min with 5 mM PLP and assayed in the presence of 20 m g / m l of asialofetuin. (B), Lineweaver-Burk plot of the activity of cHN with fetuin as substrate. Concentrations of fetuin are expressed as concentrations of sialyl residues. (e), untreated cHN; (o), cHN preincubated 10 min with 5 mM PLP; ( • ), cHN and 5 mM PLP without preincubation.
sialyl lactose and asialofetuin were present in the same incubation mixture, a significant decrease in the extent of labelling was observed, indicating that PLP binds at or near the allosteric site of cHN. Following reduction of the Schiff base between PLP and cHN, the irreversibly-modified protein was subjected to chymotryptic digestion, and peptides were separated by HPLC. The elution profile revealed a single fluorescent peak (Fig. 2A), which was further chromatographed (Fig. 2B). Only the peptide eluted at 22.5 min showed the spectrum characteristic of alkyl pyridoxamine. This peptide, after total hydrolysis, was
15
30
45
TIME (min)
Fig. 2. HPLC purification of PLP-labelled peptides. Chymotryptic peptides were prepared as described under Materials and Methods and separated on a C18 reverse phase column. The fluorescence (excitation 330 nm, emission 395 nm) was recorded after post-column neutralization. (A) First HPLC separation with a gradient of acetonitrile ( - - - - - - ) ; (B) Second HPLC separation with isochratic elution.
subjected to amino-acid analysis. On the basis of the known amino-acid sequence of HN [15], the PLP-modified peptide was identified as peptide 546-555, with PLP bound to the lysine-553 (Table II).
Effects of PLP on the hemolytic activity of Sendai virus Incubation of Sendai virus with P L P caused r e m a r k able inhibition in hemolytic activity of the virus (Fig. 3),
and the labelling of both HN and F. The presence in the incubation mixture of sialyl lactose, a substrate of HN, did not protect against PLP action, nor did it modify the extent of labelling of F (Table I). These data suggest that PLP inhibits hemolytic activity of the virus by interaction with the F protein.
TABLE I
Effects of the substrate on the binding of PLP to viral proteins The relative amount of PLP bound to viral proteins was determined as the ratio between pyridoxyl an protein fluorescence. Proteins were labelled with PLP either in the presence or in the absence of 5 mM sialyl-lactose. Proteins were reduced by NaBH4, purified by SDSPAGE, and electroeluted from the gel in 50 mM ammonium hydrogen carbonate containing 0.1% SDS. Protein fluorescence was excited at 280 nm and recorded at 340 nm; pyridoxyl fluorophore was excited at 325 nm and fluorescence was recorded at 395 nm. Protein
Fluorescence ratio - sialyl
+ sialyl
%
cHN cHN+asialofetuin HN F
0.0357 0.0368 0.0528 0.0408
0.0399 0.0275 0.0360 0.0387
+ 11.7 -25.2 -31.8 -5,1
TABLE I1
Amino acid composition of PLP labelled peptides Values are expressed as mol/mol of leucine. Amino acid
Asx GIx Ser His lie Leu Lys Lys-Pyx
Peptide cHN
HN
546-555
0.98 2.20 0.82 0.67 2.41 1.0 0.75
1.09 1.97 0.71 0.74 2.34 1.0 1.23
1 2 1 1 3 1 1 -
326
Identification of the amino acid(s) labelled by PLP 100
//./.I j'1"I" 0
-J 50
PLP-labelled HN was reduced with NaBH4, purified by SDS PAGE, and digested with chymotrypsin. H P L C purification of the peptides gives only one labelled peptide, whose retention time on HPLC and amino acid composition (Table II) were identical to those of the peptide labelled by PLP on cHN. Hence, also in virus-bound HN, PLP selectively modifies lysine-553. Discussion
I0
20
30
TIME (rain)
Fig. 3. Effect of PLP on the hemolytic activity of SV. Sendai virus,
either native (o) or preincubated 10 min with 5 mM PLP (o), was mixed with human erythrocytes and the extent of hemolysis was measured as described in Materials and Methods.
Effects of PLP on hemagglutinating and neuraminidase actiuities of Sendai uirus Preincubation of the virus with PLP decreased hemagglutinating activity by less than 50%, while neuraminidase activity increased (Fig. 4). These two opposite effects can be correlated: an increase in the rate of sialic acid hydrolysis increases the release of the virus from erythrocytes, thus reducing hemagglutination. The presence in the incubation mixture of the substrate, sialyl lactose, prevents these effects, and causes a significant decrease in the labelling of H N (Table I). The effect of substrate appears very similar to that observed on cHN in the presence of asialofetuin.
140
>
120
loo-
80
~o
2'o
INCUBATION TIM[ (mln)
Fig. 4. Effect of PLP on the neuraminidase activity of SV. Sendai virus was incubated with 5 m M PLP in the absence ( o ) or in the presence (e) of 3 m M sialyl lactose. At the times indicated in abscissa aliquots were withdrawn from the incubation mixture and assayed for neuraminidase activity at 37°C using sialyl lactose as substrate.
PLP selectively modifies cHN and viral HN at the same amino acid, lysine-553. However, the effects are quite different: (i), cHN enzymatic activity is not affected, while H N activity is increased; (ii), the substrate 'protects' HN while it has no effect on cHN, unless asialofetuin is present in the incubation mixture; (iii), the substrate-dependent inhibitory site present on cHN is suppressed, while a similar inhibitory site cannot be detected on HN. PLP appears to selectively modify the allosteric site of cHN, although it is not possible to determine whether lysine-553 is at or near the allosteric site. In fact PLP-treated cHN loses inhibition by both fetuin and asialofetuin, and protection against PLP binding is obtained by sialyl lactose together with asialofetuin, while sialyl lactose or asialofetuin alone provide no protection. These protection experiments are consistent with the inhibition mechanism of cHN, which binds asialofetuin only as an enzyme-substrate complex [8]. The substrate alone does not affect PLP binding, suggesting that the chemical reactivity a n d / o r accessibility of lysine-553 is not modified by conformational changes induced by the substrate. The effects of the modification of lysine-553 on viral HN are a decrease of hemagglutinating activity, and an increase of neuraminidase activity. If HN has two sialic acid binding sites, inhibition of hemagglutination means that PLP modifies the hemagglutinating site. However, this hypothesis does not fit experimental data obtained with cHN, where the PLP site is protected by asialofetuin, and is not protected by sialyl lactose. Thus PLP does not bind, at least on cHN, at a sialic acid binding site. Conversely, if HN has only one sialic acid binding site which can act as both neuraminidase and hemagglutinin, an increase in rate of breakdown of sialic acid will result in reduced hemagglutinating activity. The fact that both activities are perturbed by the modification of a single amino-acid residue, which has been located at the allosteric site of cHN, indicates that the two activities involve a single site. It is noteworthy that activation of cHN by PLP can be observed at high concentrations of fetuin as substrate (Fig. 1B); however, this activation is simply a release of preexisting inhibition. Experimental results can be fitted by suggesting that
327 virus bound HN possesses a substrate-dependent inhibitory site, and that a physiological inhibitor is present on the viral surface. This means that neuraminidase activity measured in intact virus is always partially inhibited. Treatment with PLP specifically inactivates the inhibitory site of both cHN and viral HN. This chemical modification has no effect on cHN assay when it is done in the absence of inhibitor, while it activates viral HN, as the inhibitor is always present in the assay. The presence of sialyl lactose does not modify the labelling of cHN unless asialofetuin is present in the incubation mixture, while the substrate causes binding of the inhibitor to viral HN, thus protecting the inhibitory site from PLP. Activation of Sendai virus neuraminidase activity has already been reported [5,16]. It has been shown that some monoclonal antibodies against HN increase neuraminidase activity and inhibit hemagglutination [16,17]. Such observations can be easily explained, assuming that these antibodies bind at or near an inhibitory site; whereas binding to hemagglutinating site and an indirect effect on the catalytic properties of HN appears more questionable. Other data suggest that enzymatic activity of HN is only partially displayed in virus-cell interactions. The rate constant for release of the virus from erythrocyte ghosts is 0.003 s -1 [18], at 37°C (pH 7.4). However, under identical temperature and pH conditions, the catalytic rate constant of HN is 2.1 s -1 (Dallocchio F. and Martuscelli, G., unpublished results). If lowering of the rate constant is due to the multiple binding of HN to the cell surface, then 27oo molecules of fully active HN should be involved in the interaction of a single viral particle with the host cell. This quantity greatly exceeds the number of HN molecules present on a viral particle (about 1000). Although great care is required in quantitative comparisons of data obtained in different laboratories, under similar but not identical conditions, a difference of several orders of magnitude between expected and experimental release rate constants cannot be ignored. This discrepancy can be explained either by assuming hemagglutinating site different from the catalytic site, or assuming that the neuraminidase activity of cell-bound virus is partially inhibited.
Our results strongly suggest that virus-bound HN is allosterically inhibited by a substrate-dependent mechanism similar to the inhibition previously shown on cHN, and that a physiological inhibitor of HN may be present on the surface of the virus. This hypothesis also agrees with previously reported experimental data, which were interpreted as proof of the two-site hypothesis. In order to obtain more conclusive evidence, studies are in progess to identify the inhibitor of HN, and to clarify the physiological significance of the allosteric site.
Acknowledgement This work was supported by grants from Italian M.U.R.S.T. (fondi 40% and 60%).
References 1 Poste, G. and Pasternak, C.A. (1978) Cell Surf. Rev. 5, 305-367. 2 White, J., Kielian, M. and Helenius, A. (1983) Q. Rev. Biophys. 16, 151-195. 3 Choppin, P.W. and Scheid, A. (1980) Rev. Infect. Dis. 2, 40-61. 40zawa, M., Asano, A. and Okada, Y. (1979) Virology 99, 197-202. 5 Miura, N., Uchida, T. and Okada, Y. (1982) Exp. Cell Res. 141, 409-420. 6 Heat, T.D., Martin, F.J. and Macher, B.A. (1983) Exp. Cell Res. 149, 163-175. 7 Thompson, S.D., Laver, W.G., Murti, K.G. and Portner, A. (1988) J. Virol. 62, 4652-4660. 8 Dallocchio, F., Bellini, T., Martuscelli, M., Baiocchi, M. and Tomasi, M. (1991) Biochem. Int. 25, 663-668. 9 Moscufo, N., Gallina, A., Schiavo, G., Montecucco, C. and Tomasi, M. (1987) J. Biol. Chem. 262, 11490-11496. 10 Warren, L. (1959) J. Biol. Chem. 234, 1971-1974. 11 Laemmli, U.K. (1979) Nature 227, 680-685. 12 Dzandu, J.K., Johnson, J.F. and Wise, G.E. (1988) Anal. Biochem. 174, 157-167. 13 Konigsberg, W.H. and Henderson, L. (1983) Methods Enzymol. 91, 254-249. 14 Aswad, D.W. (1984) Anal. Biochem. 137, 405-409. 15 Blumberg, B., Giorgi, C., Roux, L., Raju, R., Dowling, P., Challet, A. and Kolakofsky, D. (1985) Cell 41, 269-278. 16 Portner, A., Scroggs, R.A. and Metzer, D.W. (1987) Virology 158, 61-68. 17 Thompson, S.D. and Portner, A. (1987) Virology 160, 1-8. 18 Nit, S., Klappe, K. and Hoekstra, D. (1986) Biochemistry 25, 2155-2161.
323
© 1993 Elsevier Science Publishers B.V. All rights reserved 0167-4838/93/$06.00
BBAPRO 34402
Selective modification of Sendai virus hemagglutinin neuraminidase by pyridoxal 5'-phosphate: evidence for an allosteric modulation of neuraminidase activity Tiziana Bellini a, Maurizio Tomasi b and Franco Dallocchio
a
a Istituto di Chimica Biologica, Universit~ di Ferrara, Ferrara (Italy) and t, Laboratorio di Biologia Cellulare, Istituto Superiore di Sanitd, Roma (Italy)
(Received 5 August 1992)
Key words: Hemagglutinin;Neuraminidase; Sendai virus; Allosteric inhibition Incubation of Sendai virus with pyridoxal 5'-phosphate (PLP) causes inhibition of hemolytic activity, a slight reduction of hemagglutinating activity, and an increase in neuraminidase activity. The effects on hemagglutination and neuraminidase are prevented by the presence in the incubation mixture of sialyl lactose, a substrate of hemagglutinin-neuraminidase. Incubation with PLP of the water-soluble enzymatic domain of the neuraminidase has no effect on enzymatic activity, while the allosteric inhibition (Dallocchio et al. (1991) Biochem. Int. 25, 663-668) disappears. Both virus-bound and solubilized neuraminidase are selectively modified by PLP at the lysine-553. Our data suggest that PLP inactivates a previously undetected inhibitory site on the viral neuraminidase, and that a physiological effector is present on the viral envelope.
Introduction Sendai virus envelope contains two glycoproteins, the fusion protein (F) and the hemagglutinin neuraminidase (HN). F protein causes the fusion of the viral envelope with host cell membrane [1,2], while H N protein is not only able to bind and hydrolyse the sialic acid receptors [1,3], but is also required for the fusion process [4-6]. These different activities raised the question whether H N has a single multifunctional sialic acid binding site, or hemagglutinating and neuraminidase activities involve two different sialic acid binding sites. Furthermore, it is totally unknown how HN is involved in virus-cell fusion. We have previously reported that the water soluble, C-terminal fragment of H N (cHN), prepared by selective proteolytic digestion of the viral glycoprotein [7], is uncompetitively inhibited by fetuin and asialofetuin [8]. This inhibition can be observed only on cHN at low temperature, while virus-bound H N is not inhibited either at 37°C or at low temperature. However, the presence on cHN of an inhibitory mechanism, which can shift the neuraminidase to an hemagglutinin, can
Correspondence to: F. Dallocchio, Istituto di Chimica Biologica,Via Borsari 46, 1-44100Ferrara, Italy. Abbreviations: HN, hemagglutinin neuraminidase; F, fusion protein; PLP, pyridoxal5'-phosphate; TFA, trifluoroaceticacid.
explain how a single sialic acid binding site can display different functions. In order to investigate whether a similar inhibition could occur not only on cHN but also on viral H N at physiological temperature, we looked for a reagent which could bind selectively to the inhibitory site. In the present report we show that pyridoxal 5'phosphate, a widely used reagent for selective modifications of proteins and enzymes, specifically modifies cHN and H N at the lysine-553. This modification protects cHN against inhibition, and increases the neuraminidase activity of viral HN. This suggests that, under physiological conditions, the activity of viral neuraminidase could be modulated by an inhibitory mechanism.
Materials and Methods Z strain Sendai virus was grown in the allantoic fluid of 10-day-old fertilized eggs. T h e virus was purified by differential centrifugation [9]. Fetuin, sialyl lactose, and pyridoxal 5'-phosphate were from Sigma. HPLC grade solvents were from Baker. Asialofetuin was prepared .as previously described [8]. Fresh human erythrocytes were washed three times with 10 mM phosphate buffer (pH 7.4), containing 145 mM NaCI (PBS).
324 Neuraminidase assay was carried out in 0.1 M acetate buffer (pH 5.0). cHN was incubated for 60 min at 4°C with either fetuin or sialyl lactose as substrate, and viral H N was incubated for 5 min at 37°C with sialyl lactose as substrate. The sialic acid released was assayed using the method of Warren [10]. Hernagglutination assay. Serial two-fold virus dilutions in PBS (50/~1) were prepared in V-shaped multiwell plates. An equal volume of 3% (v/v) erythrocytes in PBS was added to each well, and the minimum agglutinating concentration was determined after 1 h at 4°C. In control experiments, PLP was added to the erythrocyte suspension before mixing with virus, and did not affect hemagglutinating activity. Hemolysis assay. Virus (20 izg/ml) was mixed with an equal volume of 3% (v/v) erythrocytes in PBS at 37°C. At the desired times, 50 /zl of the reaction mixture was diluted with 1 ml of cold PBS, centrifuged, and the absorbance of the supranatant at 413 nm was determined. Modification of neuraminidases by PLP. cHN or H N were incubated with 5 mM PLP in PBS at room temperature. At the desired times, samples were assayed for neuraminidase activity. In control assays PLP was added to the assay mixture just before addition of the neuraminidase. All assays were in triplicate. Stoichiometry of PLP labelling. The stoichiometry of the labelling of viral glycoproteins by PLP was calculated as the ratio between fluorescence of the pyridoxyl fluorophore (excitation 325 nm, emission 395 nm), and fluorescence of protein (excitation 280 nm, emission 340 nm). Sendai virus or cHN were labelled by incubation with 5 mM PLP, in either the presence or absence of sialyl lactose. The modified proteins were purified by SDS-PAGE followed by electroelution, and fluorescence was measured on the electroeluate using electroelution buffer as a blank. To measure the effect of asialofetuin on the labelling of cHN by PLP, 0.1 mg of cHN was incubated in 150 Ixl of PBS with PLP in the presence of 13 m g / m l of asialofetuin, in either the presence or absence of 5 mM sialyl lactose. Samples were passed through small columns (0,5 x 3 cm) of D E A E cellulose, equilibrated at pH 6, which retains asialofetuin but not cHN. The eluates were lyophilized and cHN was further purified by SDS-PAGE and electroelution.
Identification of the amino acid labelled by PLP. Sendai virus (3 mg), or cHN (0.3 mg), were incubated for 10 min at room temperature with 5 mM PLP, and the reaction was blocked by addition of solid NaBH 4 to reduce the Schiff base between the inhibitor and the protein; the excess of borohydryde was destroyed by acidification. The protein sample was treated with 1 mM D T T (30 min), then with 2 mM iodoacetamide. After 30 min at room temperature the sample was boiled for a few minutes with 5% SDS containing 0.2
M mercaptoehthanol, and run on a 10% SDS-PAGE [11]. The gel was stained with zinc chloride [12]; the band corresponding to HN (or to cHN) was excised and destained with EDTA. The protein, electroeluted with a Biorad apparatus, was lyophilized, SDS was extracted with acetic acid-triethylamine-acetone [13], and the pellet was dissolved in 50 mM ammonium bicarbonate (pH 8.0). The protein was digested for 5 h at 37°C with chyrnotrypsin (4% w/w), and peptides were separated on a 200 x 4.6 mm C~8 reverse-phase column. Elution was performed at 1 m l / m i n with a linear gradient of acetonitrile in water, both containing 0.02% TFA. In order to detect fluorescence of the PLP-labelled peptides, column effluent was neutralized by mixing with 0.2 M sodium phosphate (pH 7.4), delivered at 0.2 m l / m i n with a Varian 2010 pump. The fluorescence of the neutralized effluent was measured with a Varian 2070 fluorescence detector (excitation 330 nm, emission 395 nm). The single fluorescent peak was collected, lyophilized, dissolved in 50 tzl of water and rechromatographed on the same H P L C apparatus with an isocratic elution with 82% water, 18% acetonitrile, 0.02% TFA. Amino-acid analysis. The fluorescent peak eluted from the second HPLC was lyophilized and hydrolyzed in HC1 vapor at 160°C for 8 h. The hydrolyzate was neutralized with a few microlitres of saturated NaHCO 3, reacted with o-phtalaldehyde/acethyl-cysteine reagent and analyzed by H P L C [14]. Results
Effects of PLP on the neuraminidase activity of cHN cHN is inhibited at 4°C by asialofetuin, and this causes an apparent substrate inhibition when neuraminidase activity is assayed using fetuin as substrate [8]. Other substrates, such as sialyl lactose and o~l-acidic glycoprotein, do not show substrate inhibition. Although incubation of cHN with PLP up to 30 min does not affect enzymatic activity on sialyl lactose, it nevertheless prevents inhibition by asialofetuin (Fig. 1A), and avoids substrate inhibition by fetuin (Fig. 1B). Loss of inhibition was not due to interaction between fetuin and PLP during the enzymatic assay: indeed, preincubation of cHN with PLP was required in order to remove the inhibition. As shown in Fig. 1B, PLP slightly reduced the enzymatic activity of cHN using fetuin as substrate; however, substrate inhibition was still present when PLP was added to the assay mixture without preincubation with cHN. The presence in the incubation mixture of either sialyl lactose or asialofetuin did not protect against loss of the allosteric inhibition, nor did it reduce the extent of PLP incorporation, as judged by the fluorescence intensity of cHN-bound pyridoxyl fluorophore (Table I). On the other hand, if
325 .-
40
A
30
...-'"
==
-
20
o
-:-'-'-'- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
"'-1
'
y
10 ~ " / t ~ / ~ /
~
'
I
3b
i5
Li5
6b
TIME (mln)
2
2
5
5
1/[SIALYL] (n~1-1 ) Fig. 1. Effect of PLP on the enzymatic activity of cHN. All assays were performed in 0.1 M acetate buffer (pH 5), for 60 min at 4°C. (A), Lineweaver-Burk plot of the activity of cHN with sialyl lactose as substrate. (o), untreated cHN; (e), untreated cHN in the presence of 20 m g / m l of asialofetuin; ( • ) , cHN preincubated 10 min with 5 mM PLP and assayed in the presence of 20 m g / m l of asialofetuin. (B), Lineweaver-Burk plot of the activity of cHN with fetuin as substrate. Concentrations of fetuin are expressed as concentrations of sialyl residues. (e), untreated cHN; (o), cHN preincubated 10 min with 5 mM PLP; ( • ), cHN and 5 mM PLP without preincubation.
sialyl lactose and asialofetuin were present in the same incubation mixture, a significant decrease in the extent of labelling was observed, indicating that PLP binds at or near the allosteric site of cHN. Following reduction of the Schiff base between PLP and cHN, the irreversibly-modified protein was subjected to chymotryptic digestion, and peptides were separated by HPLC. The elution profile revealed a single fluorescent peak (Fig. 2A), which was further chromatographed (Fig. 2B). Only the peptide eluted at 22.5 min showed the spectrum characteristic of alkyl pyridoxamine. This peptide, after total hydrolysis, was
15
30
45
TIME (min)
Fig. 2. HPLC purification of PLP-labelled peptides. Chymotryptic peptides were prepared as described under Materials and Methods and separated on a C18 reverse phase column. The fluorescence (excitation 330 nm, emission 395 nm) was recorded after post-column neutralization. (A) First HPLC separation with a gradient of acetonitrile ( - - - - - - ) ; (B) Second HPLC separation with isochratic elution.
subjected to amino-acid analysis. On the basis of the known amino-acid sequence of HN [15], the PLP-modified peptide was identified as peptide 546-555, with PLP bound to the lysine-553 (Table II).
Effects of PLP on the hemolytic activity of Sendai virus Incubation of Sendai virus with P L P caused r e m a r k able inhibition in hemolytic activity of the virus (Fig. 3),
and the labelling of both HN and F. The presence in the incubation mixture of sialyl lactose, a substrate of HN, did not protect against PLP action, nor did it modify the extent of labelling of F (Table I). These data suggest that PLP inhibits hemolytic activity of the virus by interaction with the F protein.
TABLE I
Effects of the substrate on the binding of PLP to viral proteins The relative amount of PLP bound to viral proteins was determined as the ratio between pyridoxyl an protein fluorescence. Proteins were labelled with PLP either in the presence or in the absence of 5 mM sialyl-lactose. Proteins were reduced by NaBH4, purified by SDSPAGE, and electroeluted from the gel in 50 mM ammonium hydrogen carbonate containing 0.1% SDS. Protein fluorescence was excited at 280 nm and recorded at 340 nm; pyridoxyl fluorophore was excited at 325 nm and fluorescence was recorded at 395 nm. Protein
Fluorescence ratio - sialyl
+ sialyl
%
cHN cHN+asialofetuin HN F
0.0357 0.0368 0.0528 0.0408
0.0399 0.0275 0.0360 0.0387
+ 11.7 -25.2 -31.8 -5,1
TABLE I1
Amino acid composition of PLP labelled peptides Values are expressed as mol/mol of leucine. Amino acid
Asx GIx Ser His lie Leu Lys Lys-Pyx
Peptide cHN
HN
546-555
0.98 2.20 0.82 0.67 2.41 1.0 0.75
1.09 1.97 0.71 0.74 2.34 1.0 1.23
1 2 1 1 3 1 1 -
326
Identification of the amino acid(s) labelled by PLP 100
//./.I j'1"I" 0
-J 50
PLP-labelled HN was reduced with NaBH4, purified by SDS PAGE, and digested with chymotrypsin. H P L C purification of the peptides gives only one labelled peptide, whose retention time on HPLC and amino acid composition (Table II) were identical to those of the peptide labelled by PLP on cHN. Hence, also in virus-bound HN, PLP selectively modifies lysine-553. Discussion
I0
20
30
TIME (rain)
Fig. 3. Effect of PLP on the hemolytic activity of SV. Sendai virus,
either native (o) or preincubated 10 min with 5 mM PLP (o), was mixed with human erythrocytes and the extent of hemolysis was measured as described in Materials and Methods.
Effects of PLP on hemagglutinating and neuraminidase actiuities of Sendai uirus Preincubation of the virus with PLP decreased hemagglutinating activity by less than 50%, while neuraminidase activity increased (Fig. 4). These two opposite effects can be correlated: an increase in the rate of sialic acid hydrolysis increases the release of the virus from erythrocytes, thus reducing hemagglutination. The presence in the incubation mixture of the substrate, sialyl lactose, prevents these effects, and causes a significant decrease in the labelling of H N (Table I). The effect of substrate appears very similar to that observed on cHN in the presence of asialofetuin.
140
>
120
loo-
80
~o
2'o
INCUBATION TIM[ (mln)
Fig. 4. Effect of PLP on the neuraminidase activity of SV. Sendai virus was incubated with 5 m M PLP in the absence ( o ) or in the presence (e) of 3 m M sialyl lactose. At the times indicated in abscissa aliquots were withdrawn from the incubation mixture and assayed for neuraminidase activity at 37°C using sialyl lactose as substrate.
PLP selectively modifies cHN and viral HN at the same amino acid, lysine-553. However, the effects are quite different: (i), cHN enzymatic activity is not affected, while H N activity is increased; (ii), the substrate 'protects' HN while it has no effect on cHN, unless asialofetuin is present in the incubation mixture; (iii), the substrate-dependent inhibitory site present on cHN is suppressed, while a similar inhibitory site cannot be detected on HN. PLP appears to selectively modify the allosteric site of cHN, although it is not possible to determine whether lysine-553 is at or near the allosteric site. In fact PLP-treated cHN loses inhibition by both fetuin and asialofetuin, and protection against PLP binding is obtained by sialyl lactose together with asialofetuin, while sialyl lactose or asialofetuin alone provide no protection. These protection experiments are consistent with the inhibition mechanism of cHN, which binds asialofetuin only as an enzyme-substrate complex [8]. The substrate alone does not affect PLP binding, suggesting that the chemical reactivity a n d / o r accessibility of lysine-553 is not modified by conformational changes induced by the substrate. The effects of the modification of lysine-553 on viral HN are a decrease of hemagglutinating activity, and an increase of neuraminidase activity. If HN has two sialic acid binding sites, inhibition of hemagglutination means that PLP modifies the hemagglutinating site. However, this hypothesis does not fit experimental data obtained with cHN, where the PLP site is protected by asialofetuin, and is not protected by sialyl lactose. Thus PLP does not bind, at least on cHN, at a sialic acid binding site. Conversely, if HN has only one sialic acid binding site which can act as both neuraminidase and hemagglutinin, an increase in rate of breakdown of sialic acid will result in reduced hemagglutinating activity. The fact that both activities are perturbed by the modification of a single amino-acid residue, which has been located at the allosteric site of cHN, indicates that the two activities involve a single site. It is noteworthy that activation of cHN by PLP can be observed at high concentrations of fetuin as substrate (Fig. 1B); however, this activation is simply a release of preexisting inhibition. Experimental results can be fitted by suggesting that
327 virus bound HN possesses a substrate-dependent inhibitory site, and that a physiological inhibitor is present on the viral surface. This means that neuraminidase activity measured in intact virus is always partially inhibited. Treatment with PLP specifically inactivates the inhibitory site of both cHN and viral HN. This chemical modification has no effect on cHN assay when it is done in the absence of inhibitor, while it activates viral HN, as the inhibitor is always present in the assay. The presence of sialyl lactose does not modify the labelling of cHN unless asialofetuin is present in the incubation mixture, while the substrate causes binding of the inhibitor to viral HN, thus protecting the inhibitory site from PLP. Activation of Sendai virus neuraminidase activity has already been reported [5,16]. It has been shown that some monoclonal antibodies against HN increase neuraminidase activity and inhibit hemagglutination [16,17]. Such observations can be easily explained, assuming that these antibodies bind at or near an inhibitory site; whereas binding to hemagglutinating site and an indirect effect on the catalytic properties of HN appears more questionable. Other data suggest that enzymatic activity of HN is only partially displayed in virus-cell interactions. The rate constant for release of the virus from erythrocyte ghosts is 0.003 s -1 [18], at 37°C (pH 7.4). However, under identical temperature and pH conditions, the catalytic rate constant of HN is 2.1 s -1 (Dallocchio F. and Martuscelli, G., unpublished results). If lowering of the rate constant is due to the multiple binding of HN to the cell surface, then 27oo molecules of fully active HN should be involved in the interaction of a single viral particle with the host cell. This quantity greatly exceeds the number of HN molecules present on a viral particle (about 1000). Although great care is required in quantitative comparisons of data obtained in different laboratories, under similar but not identical conditions, a difference of several orders of magnitude between expected and experimental release rate constants cannot be ignored. This discrepancy can be explained either by assuming hemagglutinating site different from the catalytic site, or assuming that the neuraminidase activity of cell-bound virus is partially inhibited.
Our results strongly suggest that virus-bound HN is allosterically inhibited by a substrate-dependent mechanism similar to the inhibition previously shown on cHN, and that a physiological inhibitor of HN may be present on the surface of the virus. This hypothesis also agrees with previously reported experimental data, which were interpreted as proof of the two-site hypothesis. In order to obtain more conclusive evidence, studies are in progess to identify the inhibitor of HN, and to clarify the physiological significance of the allosteric site.
Acknowledgement This work was supported by grants from Italian M.U.R.S.T. (fondi 40% and 60%).
References 1 Poste, G. and Pasternak, C.A. (1978) Cell Surf. Rev. 5, 305-367. 2 White, J., Kielian, M. and Helenius, A. (1983) Q. Rev. Biophys. 16, 151-195. 3 Choppin, P.W. and Scheid, A. (1980) Rev. Infect. Dis. 2, 40-61. 40zawa, M., Asano, A. and Okada, Y. (1979) Virology 99, 197-202. 5 Miura, N., Uchida, T. and Okada, Y. (1982) Exp. Cell Res. 141, 409-420. 6 Heat, T.D., Martin, F.J. and Macher, B.A. (1983) Exp. Cell Res. 149, 163-175. 7 Thompson, S.D., Laver, W.G., Murti, K.G. and Portner, A. (1988) J. Virol. 62, 4652-4660. 8 Dallocchio, F., Bellini, T., Martuscelli, M., Baiocchi, M. and Tomasi, M. (1991) Biochem. Int. 25, 663-668. 9 Moscufo, N., Gallina, A., Schiavo, G., Montecucco, C. and Tomasi, M. (1987) J. Biol. Chem. 262, 11490-11496. 10 Warren, L. (1959) J. Biol. Chem. 234, 1971-1974. 11 Laemmli, U.K. (1979) Nature 227, 680-685. 12 Dzandu, J.K., Johnson, J.F. and Wise, G.E. (1988) Anal. Biochem. 174, 157-167. 13 Konigsberg, W.H. and Henderson, L. (1983) Methods Enzymol. 91, 254-249. 14 Aswad, D.W. (1984) Anal. Biochem. 137, 405-409. 15 Blumberg, B., Giorgi, C., Roux, L., Raju, R., Dowling, P., Challet, A. and Kolakofsky, D. (1985) Cell 41, 269-278. 16 Portner, A., Scroggs, R.A. and Metzer, D.W. (1987) Virology 158, 61-68. 17 Thompson, S.D. and Portner, A. (1987) Virology 160, 1-8. 18 Nit, S., Klappe, K. and Hoekstra, D. (1986) Biochemistry 25, 2155-2161.