Studies on metrizamide-protein interactions

Biochimica et Biophysica Acta, 453 (1976) 176-184

© Elsevier/North-Holland Biomedical Press BBA 37476 STUDIES ON M E T R I Z A M I D E - P R O T E I N I N T E R A C T I O N S

ALOYS HIJTTERMANN and GERTRUD WENDLBERGER-SCHIEWEG Forstbotanisches Institut der Universitiit, D 34 Gi~ttingen, Biisgenweg 2 (G.F.R.)

(Received May 24th, 1976)

SUMMARY 1. The apparent density of catalase after isopycnic centrifugation in metrizamide gradients is dependent on the metrizamide concentration into which the enzyme is dissolved at the beginning of the centrifugation. 2. This different behaviour of the enzyme in metrizamide gradients is due to the formation of a metrizamide.protein complex which is more dense than the uncomplexed catalase. 3. A bimodal distribution of the catalase, with additional heavy bands, was only observed in metrizamide gradients in light water, where rather high metrizamide concentrations are needed even for a banding of the uncomplexed enzyme. 4. The half-life of the metrizamide.protein complex is less than 5 min. This was shown by spectroscopical measurements and band sedimentation analysis in an analytical ultracentrifuge.

INTRODUCTION Two different solvents have been used for the isopycnic centrifugation of proteins in metrizamide gradients: light water [1, 2], and heavy water [3]. In metrizamide gradients in light water, which were formed during the centrifugation, always a bimodal distribution was found for proteins: one band which corresponded to the density of a fully hydrated protein, and additional bands with much higher densities [1, 2]. If proteins, however, were layered on a preformed metrizamide/ZH20 gradient and centrifuged, always only one single band was found for the enzymes which have been studied so far in this system [3-6], and multiple bands only in cases of enzymes which were present in multiple forms [5]. The additional heavy bands which were observed in the metrizamide gradients in light water, have been interpreted as the result of a labile protein'metrizamide complex which occurs only at high metrizamide concentrations and which is completely reversible [2]. In a recent publication, however, a different interpretation of these heavy bands was presented [7]: the complex band pattern was interpretated to be the result of differences in the rate constants of the formation of the metrizamideprotein complex. On the base of computer simulations, a half-life of 1-1.5 h was postulated for this complex. Since the formation of any heavy satellite band may obscure the interpretation

177 of the density labeling experiments, the method for which the metrizamide/2H20 gradients have been used so far, we were rather interested in studying this phenomenon and finding the reasons for the different behaviour of proteins in the two solutes mentioned above. For our studies, we used catalase, the enzyme for which the formation of the heavy additional bands has been described before [2] in the gradient system with light water. Our studies revealed a dependency of the formation of these additional bands on the metrizamide concentration present in the solution. In addition, we tried to measure the half-life of the metrizamide.protein complex by two different methods: spectroscopically and by band sedimentation studies in the analytical ultracentrifuge. The results of these experiments show that the halflife of the metrizamide.protein complex is rather short, less than 5 min. MATERIALS AND METHODS

Materials. Catalase (EC 1.11.1.6) three times recrystallized, was obtained from Boehringer-Mannheim G.m.b.H. (Mannheim-Waldhof, G.F.R.), metrizamide from Nygaard (Oslo, Norway), and 2H20 from Sharp and Dohme G.m.b.H. (Mfinchen, G.F.R.). All other chemicals were reagent grade and purchased from E. Merck (Darmstadt, G.F.R.). Sephadex G-25 medium came from Pharmacia (Uppsala, Sweden). Density gradient eentrifugation. Density gradient sedimentation was always performed in preformed step gradients according to the general procedure described earlier [3]. Tris/acetate buffer, 0.1 M, pH 7.1, always was used as solvent, prepared either in heavy or light water. In the experiments with light water, a metrizamide concentration range from 30 to 40 % (w/w) was used, in those with heavy water one from 15 to 35 %. The catalase (12 mg for each gradient, except where noted otherwise) was usually dissolved in one of the different metrizamide solutions forming the gradient, as indicated in Results and Discussion. After centrifugation for 18 h at 350 000 x g, 4 °C, in a swinging bucket rotor (SW 65 from Spinco, Beckman Instruments) 5 ml volume, length of the liquid column 5 cm, fractions of five drops each were collected. Of every fifth fraction, the refractive index was determined in an Abb6 refractometer (Zeiss). To each of the remaining fractions, 0.9 ml of 0.1 M Tris/acetate buffer, pH 7.1 was added. After buzzing in a Vortex mixer, the A405 ,m in the fractions was measured in a Zeiss PMQ II spectrophotometer. Desalting of the eatalase solutions in metrizamide. 12 mg catalase were dissolved in 1 ml 60 % (w/w) metrizamide in light water and stored for about 20 h at 4 °C. The solution was then desalted on a 7 ml Sephadex G-25 medium column (0.5 cm diameter, 25 cm long), equilibrated with 0.05 M Tris/acetate buffer and eluted with the same buffer. The catalase which came in the void volume was taken for the spectroscopical determination of the A28o nm/A4o5 nm ratio and band sedimentation analysis. Band sedimentation analysis. Band sedimentation studies were performed according to the procedure given by Chervenka [8]; with a double-sector bandforming centerpiece using a Model E analytical ultracentrifuge equipped with ultraviolet optics and scanning system. 0.01 ml sample solution containing about 20-50/~g catalase was layered on 1 M NaC1 at~a speed of 5000-8000 rev./min. After attainment of 44 000 rev./min, 8-10 scans were taken at 8-min intervals. For the determination of the s value, the peak position was measured in the scans and from these data the s value was calculated using a computer program.

178 RESULTS AND DISCUSSION

Dependency of the apparent density of catalase in metrizamide gradients on the initial present metrizamide concentration For a test whether the initial present metrizamide concentration has any influence on the apparent density of catalase in the gradient, a step gradient of 3 0 - 4 0 ~ (w/w) metrizamide in light water was formed and 12 mg of catalase were either added to the top, the middle or the bottom of the gradient. The gradient was then centrifuged and analysed subsequently. As Fig. 1 shows, the apparent density of the enzyme was dependent on the metrizamide concentration in which it was added initially. Only when the enzyme was at the beginning at the bottom of the tube, no heavy satellite band was formed, whereas in the two other cases always a heavy band was observed at a density of 1.33 g/ml. The bulk of the enzyme, however, banded as a broad peak near the position where it was placed at the beginning of the run, always with a small shift to a higher density.

A,~.5

1,4

1.0.

1,3 ,.-.,-~ii~ j

~--~ 50

,24,23 1.0 ~

~

1.2 ~ No.

t.1.4

~

t 1.3

35"/.

1.2

4O%

.1

sb

~; No.

1,

1.3 I

"'"

1.0

1.2 .1

o 5'0

F,.' No.

Fig. 1. lsopycnic banding of catalase in metrizamide gradients made with light water. A preformed gradient, made in five steps in the range from 30 to 40~ (w/w) was always used. The enzyme, 12 nag, was dissolved either in the 30~ (upper curve), the 35~ (middle curve) or the 40~ (lower curve) metrizamide solution at the beginning of the centrifugation. 0----0, density slope in the gradient; 0--0,

A40s .m.

179 A405

1.255

tl.40

0.3-

10

25°/o~ 15"/o~

35 °/o

30

50 Fr. No.

1.0

/

I/ /

/

/

\1 \

1(3 1.318

11"°

°~°~.

"\"

I1"20

\ 3t3

'

50 Fr.No.

1.0. ~ "x,~...o~.°.,~..o...o...°

/40 .3o -1.20

%

\

lb

3b

'

~b ~r.No

Fig. 2. Isopycnic banding of catalase in metrizamide gradients made with 2H20 as solvent. A preformed gradient made in five steps in the range from 15 to 35 % (w/w) was always used. The enzyme was either placed as a solution of 3 mg in 0.25 ml buffer (made from light water) on top of the gradient (upper curve), or 12mg catalase were dissolved in the 25% solution (middle curve) or 35 % solution (lower curve). 0-----0, density slope of the gradient; 0 - - 0 , A405~m. A similar result was obtained if the same type of experiment was carried out in metrizamide gradients made with heavy water. The apparent density was again dependent on the metrizamide concentration (Fig. 2). In this case, however, no heavy satellite bands were observed in all gradients made. When the enzyme was layered on top of the gradient, a narrow symmetrical peak was obtained, which banded highly reproducibly (six gradients analysed) at a mean density of 1.2554 g/ml. It should be noted that the viscosities of the gradients in both solvents were rather low and would permit always a flotation of the protein to a lighter area in the gradient, as was observed for D N A in metrizamide gradients in light water [9]. The speed and time of centrifugation, 350 000 × g, 18 h, was high enough to expect a sedimentation of the proteins to their isodensity point [3]. In a series of similar experiments carried out in metrizamide gradients in ZH20, the initial metrizamide concentration in which the enzyme was dissolved, was varied, and the refractive index was determined both in the solution to which the enzyme was added and in the peak fraction after centrifugation. As Table I shows, the enzyme never flotated except for the region close to the bottom of the tube, but banded always

180 TABLEI THE A P P A R E N T DENSITY OF CATALASE IN METRIZAMIDE/2HzO G R A D I E N T S AS A

FUNCTION OF THE INITITAL METRIZAMIDE CONCENTRATION lnitial metrizamide concentration (w/w)

Initial density of the solution

Peak density after centrifugation

Sample layered on top 15% 19~ 23 ~ 27 ~ 31 ~ 35~

1.0 1.186 1.210 1.235 1.269 1.290 1.327

1.2554 1.255 1.255 1.283 1.282 1.299 1.318

at a higher density. No heavy satellite bands were obtained in this experiment again. It could be argued that the time of centrifugation was not sufficient to band the enzyme at the isodensity point. To test this, catalase was layered on top of the preformed gradient according to our usual procedure and centrifuged at 350 000 × g for 10, 14, 18, and 20 h. After l0 h of centrifugation, the enzyme banded at a density of 1.236 g/ml, after 14 h and longer time periods, the enzyme banded at the usually found density of 1.255 g/ml. This indicates that with our routine procedure the isodensity region in the gradient is reached by the enzyme at a reasonable time before the run is analysed. On the base of the results reported here, a simple interpretation for the different behaviour of catalase in the two gradient systems: metrizamide in either light or heavy water could be offered: if metrizamide gradients in light water have been used [1, 2, 9] for the banding of proteins, the gradients always were formed during the centrifugation in a fixed angle rotor. At the beginning of the run, protein and metrizamide are evenly distributed in the liquid column. During the subsequent formation of the metrizamide gradient, the protein is exposed to a variety of metrizamide concentrations. In the upper part, the gradient becomes lighter and the protein will move down to its isodensity point. In the lower part, however, the protein will become denser due to the high metrizamide concentration and shift to an even higher density. The procedure for protein banding in heavy water is to layer the sample on top of the preformed gradient [3]. During centrifugation it moves to its isodensity point according to its hydration in the heavy water. Since the solvent itself has an initial density of 1.1 g/ml, significantly less metrizamide is needed to reach the density necessary for protein banding. Thus the metrizamide-protein interaction is too weak to result into a stable complex which would then have a higher density. Thus even after a prolonged centrifugation the position of the protein remains unchanged [3] and no bimodal distribution of the proteins is'observed as is the case in gradients in light water with the same setup (Fig. 1). An interesting result of our study is the observation that catalase flotates at the very bottom of the centrifuge tube. The explanation for this is the following: in this region, a very high hydrostatic pressure builds up which reaches values higher

181 than 1400 atm [I0]. Considering the very high dissociation constant for the metrizamide-protein complex which is greater than 10 -2 [2] and a positive reaction volume which can be expected in this kind of complex, a pressure of 1000 atm would be enough to dissociate this labile complex completely.

Stability of the metrizamide'protein complex The explanation for the formation of the heavy satellite bands in metrizamide gradients in light water and their absence in those with heavy water which is given above, is based only on a concentration dependency of the complex formation. This complex can be expected to have weak ligand interactions which should rapidly dissociate upon removal of the excess metrizamide. This interpretation, however, is not compatible with the theory presented by Skerrett [7], who explained the formation of the additional bands on the base of different rate constants of the complex formation. On the basis of his computer simulations, he postulated a half-life of the metrizamide.protein complex between 1 and 1.5 h [7]. We measured therefore the rate of dissociation of the metrizamide.protein complex in two ways, on the base of spectroscopic measurement and by analysis in the analytical ultracentrifuge. Our starting material always was a solution of 12 mg catalase in 1 ml 60% (w/w) metrizamide in light water, which was stored before use for about 20 h at 4 °C. According to the data and calculations of Rickwood et al. [2] and the experiments described in this study, the catalase present in this solution can be expected to be completely complexed to metrizamide. It therefore has a density higher than 1.5 g/ml and contains about 40-50 % metrizamide bound as a complex. I f this complex would have a half-life of more than 1 h [7] this should be detectable by either a measurement of the A2a0 nm/A405 nm ratio or by sedimentation analysis in the ultracentrifuge. For the measurement of the A280,m/A405 n m ratio, the catalase solution was desalted in a Sephadex G-25 column and the enzyme present in the void volume was taken and the absorption both at 405 and 280 nm was measured in a Zeiss P M Q II spectrophotometer. The results shown in Table II indicate that no significant difference was found for the A2s o nrn/A4o 5 n m ratio in this fraction compared to the native enzyme. As controls, the spectra of catalase dissolved in buffer and in 60 7o metrizamide were measured in the range from 650 to 360 nm and no difference was found in TABLE II ABSORPTION OF CATALASE AT 405 AND 280 nm WITH AND WITHOUT TREATMENT WITH METRIZAMIDE Catalase was dissolved in 60 % (w/w) metrizamide solution in water, stored for about 20 h at 4 °C, desalted of Sephadex G-25 and the absorption at 405 and 280 nm measured in a spectrophotometer. Enzyme preparation

A2ao n m

A280 nm

"4405 nm

"4280 nm/'440S nm

¢alcd a

Native catalase Treated catalase Exp. 1 Exp. 2

0.588 0.396 0.435

1.6 1.65

-4280 nm/A405 am calcd a

0.509

1.155

0.388 0.373

1.020 1.17

a If the metrizamide, protein complex still is present.

5 5

182 the molar extinction nor in the position of the 405 nm absorption band of the enzyme in both solvents. The gel chromatography of the catalase in 60 % metrizamide always took less than 5 min. According to this experiment, the half-life of the metrizamide. protein complex must be much less than this time period. The stability of the metrizamide.protein complex was measured in addition in a different way: by the band sedimentation studies in the analytical ultracentrifuge. The catalase in the 60% metrizamide solution has a density which is higher than 1.5 g/ml. This density difference, compared to 1.27 mg/ml of the native enzyme, should increase the apparent sedimentation coefficient by a factor of at least 1.5, according to the Pedersen-Svedberg equation (cf. ref. 11). This calculation accounts only for the differences in the apparent partial specific volumes and does not consider the additional increase in molecular weight due to the complexed metrizamide. Thus it could be expected that the metrizamide.protein complex which is supposed to have a half-life of 1-1.5 h [7] should dissociate during the sedimentation analysis, and the sedimentation rate should decrease with an increase of time. In a band sedimentation study, the dissociation of the complex should result into at least wo bands travelling with significantly different rates.

L [;

; s

.,

log

0.8

003"

082.

081

OaO

8

16

24

32

40

48

56

64

72 ......

Fig. 3. Band sedimentation of catalase treated first with metrizamide and desalted prior to the centrifugation. Catalase was stored for 20h in a 60% (w/w) metrizamide solution, desalted and analysed immediately afterwards in the ultracentrifuge. The upper part shows every second scan of the run, which were taken at 8-min intervals, the upper scan being the first one taken after attainment of 44 000 rev./min. The lower part gives the plot log x versus the time of the sedimentation.

183 Catalase was treated the same way as outlined above and sedimentation analysis was conducted immediately afterwards using a band-forming centerpiece. The first scan was taken always in less than 40 min after the beginning of the desalting of the enzyme. Fig. 3 shows the scans of a run with catalase which was stored in 60 metrizamide overnight and then desalted prior to the run. No apparent molecular heterogeneity was detectable and during the run no slower moving component appeared. The plot of log x versus the time was linear, too (Fig. 3). Table I I I shows the sedimentation coefficients of the different runs which were conducted. Always an s value of 10.5 4- 0.4 was obtained, and no differences were observed for the different treated enzyme preparations. Our studies indicate that the metrizamide, protein complex has in fact a halflife which is less than 5 min. We therefore conclude that the protein-metrizamide TABLE III APPARENT SEDIMENTATION COEFFICIENTS OF CATALASE WITHOUT AND AFTER TREATMENT WITH METRIZAMIDE Catalase was treated with metrizamide as described under Table II and analysed by band sedimentation in a Model E analytical ultracentrifuge (Beckman). Catalase preparation Native Exp. 1 Exp. 2 Desalted immediately after dissolution in metrizamide Treated with metrizamide Exp. 1 Exp. 2"

Sapp. 25 °C, 1 M NaC1 10.9 10.1 10.8 10.4 10.2

The scans and the plot log x versus t of this experiment are shown in Fig. 3. interactions are rather labile as was suggested by the N M R studies where the dissociation constant for this complex was found to be greater than 10 -2 (R. A. Badley quoted in ref. 2). They seem to be caused by van der Waals' affinity between the benzene ring in metrizamide and the aromatic amino acids of the proteins. This assumption has been strengthened by the observation that proteins with a low content of aromatic amino acids (e.g. histones) show no complex formation even at metrizamide concentrations above 60 ~ (w/v). (Gilhuus-Moe, C. Chr., personal communication). These interactions effect the density of proteins which have been studied so far only if the complex is stabilized by rather high metrizamide concentrations. Only in the case of proteins which have an extremely high content of aromatic amino acids, the complex might cause problems in the interpretation of density labeling experiments analysed by subsequent isopycnic banding of the protein under study in metrizamide/ZH20 gradients. ACKNOWLEDGEMENTS This work was supported by a grant from the Deutsche Forschungsgemeinschaft, Bad Godesberg. We would like to thank Dr. Diebler, Max-Planck-Institut fiir

184 Biophysikalische Chemie, G/Jttingen, Abt. M. Eigen, for valuable discussions regarding the complex-chemistry aspect of this work. We are indebted to Dr. H. Lentz, Physikalisch-Chemisches I n s t i t u t der Universit/it Karlsruhe, Abt. E. U. F r a n c k , for help with the i n t e r p r e t a t i o n of the influence of high pressure on the metrizamide. protein complex. REFERENCES 1 Birnie, G. D., Rickwood, D. and Hell, A. (1973) Biochim. Biophys. Acta 331, 283-294 2 Rickwood, D., Hell, A., Birnie, G. D. and Gilhuus-Moe, C. Chr. (1974) Blochim. Biophys. Acta 342, 367-371 3 H0ttermann, A. and Guntermann, U. (1975) Anal. Biochem. 64, 360-366 4 Hfittermann, A., Gebauer, M., Wessel, I. and Hoffmann, W. (1975) Biochim. Biophys. Acta 384, 493-500 5 H0ttermann, A. and Wendlberger, G. (1976) in The Use of lodinated Density Gradient Media for Biological Separations (Rickwood, D., ed.), Information Retrieval Ltd., London, pp. 15-25 6 Guntermann, U., Tan, I. and HiJttermann., A. (1975) J. Bacteriol. 124, 86-91 7 Skerrett, R. J. (1975) Biochim. Biophys. Act,h. 385, 28-35 8 Chervenka, C. H. (1969) A Manual of Methods for the Analytical Ultracentrifuge, p. 35, Beckman Instruments Inc. Palo Alto 9 Rickwood, D. and MacGillivray, A. J. (1975) in Symposium on the Cell Nucleus, Proceedings of the 9th FEBS Meeting, Budapest 1974. Akademiai Kiado, Budapest and North Holland Publishing Co., Amsterdam and London, in the press 10 McEwen, C. R. (1967) Anal. Biochem. 19, 23-39 11 Hagen, U. (1973) in Experimental Methods in Biophysical Chemistry (Nicolau, C., ed.), p. 173, John Wiley and Sons Ltd., New York