Interaction of HeLa cell proteins with RNA

Interaction of HeLa cell proteins with RNA

J. Mol. Biol. (1970) 47, 263-273 Interaction DAVID of HeLa Cell Proteins with RNA BALTIMORE AND ALICE 8. HUANG Department of Biology, Massachwe...

819KB Sizes 0 Downloads 18 Views

J. Mol. Biol. (1970) 47, 263-273

Interaction DAVID

of HeLa Cell Proteins with RNA BALTIMORE

AND

ALICE

8.

HUANG

Department of Biology, Massachwetts Institute of Technology Cambridge, Mass. 02139, U.X.A. (Received 21 April

1969, and in revised form 20 September

1969)

When any RNA is added to a cytoplasmic extract from HeLa cells there is a large increase in .the rate of sedimentation of the RNA (Girard & Baltimore, 1966). Cytoplasmic extracts also cause RNA to be retained by a membrane filter. The same material appears to be responsible for these two effects; it has been called “binding factor.” A number of lines of evidence indicate that binding factor is a heterogeneous collection of soluble proteins which are bound to RNA by ionic forces and are probably not very basic : (1) centrifugation for 60 minutes at 105,000 g does not sediment the binding factor; (2) the active material can be bound to, and eluted from, DEAE-cellulose but fractions eluted at different salt concentrations have approximately the same specific activity ; (3) it is sensitive to pronase and trypsin; (4) it is not active in 0.5 M-NaCl; (5) it causes RNA to band at a buoyant density of 1.40 in CsCl; (6) the same material will bind to different RNA’s and (7) basic proteins will cause RNA to be retained by a filter but they precipitate the RNA rather than form a soluble complex like that formed by binding factor and RNA.

1. Introduction of the cytoplasm of mammalian cells cause a large increase in the sedimentation rate of purified RNA (Girard & Baltimore, 1966). Because an understanding of these factors is a necessaryprerequisite to any investigation of the interactions of RNA, protein and ribosomesin mammalian cells, we have studied in some detail the nature of these factors. As will becomeapparent, the increasedsedimentation rate of RNA is due to binding of macromoleculesto the RNA; we have therefore called the active material “binding factor.” To determine the nature of the binding factor it was necessaryto develop a quantitative assayfor binding factor. Assay by its ability to increasethe sedimentation rate of RNA, while ultimately the only test of activity, was too laborious for repeated use. The discovery that nitrocellulose filters would retain RNA complexed with binding factor afforded a more rapid and sensitive assay which has been exploited in the present study. Nitrocellulose filters have previously been usedto study specific recognition of nucleic acids by purified proteins (Jones & Berg, 1966; Yarus & Berg, 1967; Riggs, Bourgeois, Newby & Cohn, 1968) and to distinguish ribonucleoprotein from free RNA (Infante & Nemer, 1968). Components

2. Materials (a) Procedures for the described previously 18

General

and Methods methodology

growth of HeLa cells and their (Baltimore, Girard & Darnell,

263

infection 1966).

with poliovirus have Cytoplasmio extracts

been were

264

D.

BALTIMORE

AND

A.

S.

HUANG

prepared by homogenization of cells suspended in RS buffer?, (0.01 M-N&I; 0.01 M-TrisHCl, pH 7.4; 0.0015 M-Ndgc12) as described previously (Penman, Becker & Darnell, 1964). Purification of labeled poliovirus RNA and ribosomal RNA was carried out as previously described (Girard & Baltimore, 1966), except that gradients containing 0.01% sodium dodecyl sulfate were used to minimize the carryover of sodium dodecyl sulfate to the reaction mixtures. The “viral RNA” preparations contained about 100 pg of 28 s ribosomal RNA/ml. and about 5 x lo5 cts/min/ml. of 32P incorporated into less than 5 pg of viral RNA/ml. (b) (i)

Xeclimentation

Assays

assay

Up to 10 ~1. of viral RNA was mixed with the of RS buffer. The mixture was layered over a buffer and centrifuged in the SW25.3 rotor of gradients were collected directly into planchets (ii)

l?%ration

indicated preparations to be tested in 1 ml. 16-ml., 15 to 30% sucrose gradient in RS the Spinco model L ultracentrifuge. The and the fractions were dried and counted.

assay

No more than 2.5 ~1. of RNA was mixed with the preparation to be tested in 1 ml. of binding buffer (0.01 m-Nacl; 0.01 M-Tris, pH 7.4) at 0 to 4°C. After 15 set to 15 min (different times of incubation give identical results), the sample was slowly filtered through a nitrocellulose membrane filter held on an all-glass Millipore filtration unit. The filtration rate was about 4 ml./min. The test tube was rinsed and the filter was washed with 2 portions of 2 ml. each of binding buffer. The filter was removed, glued to a planchet, dried and counted. To reduce the amount of RNA bound to filters in the absence of binding factor, the filters were soaked for at least 15 min in 1 mg of yeast RNA/ml. before use. Using such soaked filters, no more than 5% of the input RNA and usually less than 2% was bound in the absence of binding factor. Addition of either 0.02 M-EDTA or 0.01 M-Mgcl, to the binding buffer did not affect the ability of binding factor to attach labeled RNA to the filter. (c)

Buoyant

density determinations

The method published previously (Baltimore buoyant density of ribonucleoprotein particles fixation of fractions from a sucrose gradient with CsCl gradient and centrifugation of the sample (d)

& Huang, 1968) was used to determine the in &Cl. Briefly, the method involves glutaraldehyde, layering over a preformed to equilibrium.

&laterials

Millipore filters, type HA, 24 mm diameter were used for the filtration assay. Pronase was a product of Calbioohem and was used without preincubation. Trypsin end cytochrome c were products of Sigma Chemical Co. Poly-L-ornithine hydrobromide (mol. wt 62,000) was obtained from Pilot Chemicals Inc. Glutaraldehyde (33%) was purchased from Fisher Scientific Co. and CsCl was obtained from Harshaw Chemical Co.

3. Results (a) Assay of binding Cytoplasmic

extracts

factor

of either

by attachment

of RNA

uninfected

poliovirus-infected

or

to nitroceElubse cells

jilters cause

polio-

virus RNA to be retained by nitrocellulose filters (Table 1). This phenomenon has provided a quantitative assay because if varying amounts of extract are added to a constant amount of RNA, there is proportionality between the amount of added extract and the amount of RNA retained by the filter. Table 1, for instance, shows that 0.002 ml. of either an extract of uninfected or infected cells binds equivalent percentages of the input RNA to the filter. In Figure 1, the percentage of the input RNA t Abbreviation

used:

RS buffer,

reticulocyte

standard

buffer

(formerly

known

as RSB).

PROTEIN-RNA

INTERACTION TABLE

265

1

Effect of cytoplasmic extracts of uninjected and infected cells on binding of RNA to Jilter 0/0 of input RNA to filter (A) (B) (C) (D) (E) Cytoplasmic with poliovirus.

0.02 ml. of cytoplasmic of uninfected cells 0.002 ml. of cytoplasmic of uninfected cells 0.02 ml. of cytoplasmic of infected cells 0.002 ml. of cytoplasmic of infected cells No addition

bound

extract 99 extract 53 extract 99 extract 44 0.1

extracts were made from uninfected HeLa cells and from cells infected For each extract, 1.5 ml. of RS buffer was used for 4 x lo7 cells.

for 3.5 hr

which is not bound to a filter is plotted on a logarithmic scale against the amounts of different cell fractions added to the RNA. The curves show initial, variable shoulders followed by linear slopes. To compare extracts, one unit of activity has been defined as the amount of factor necessary to bind 50% of a given preparation of labeled RNA to a filter. Specific activities were calculated from protein concentrations determined by the method of Lowry, Rosebrough, Farr & Randall (1951). Once bound to a filter, RNA is not eluted by vigorous washing of the filter with RS buffer. The reaction between binding factor and RNA occurs too rapidly to be easily measured (less than 15 set). To test whether the same material which causes RNA to be retained by a filter also causes the increased sedimentation rate, samples of labeled viral RNA were added

FIG. 1. Factor activity in different amounts of various cellular fractions. The active fractions were prepared as described in Materials and Methods. The indicated amounts of the different preparations were added to standard assay mixtures and the percentage of total labeled RNA retained by a filter was determined. This was converted to a percentage of the input RNA not bound by the filter and plotted on a semi-logarithmic scale. -•c]--l-~--, Supernatant after centrifugation at 30,000 g for 20 min; --O--O-, supernatant after centrifugation at 105,000 g for 60 min; -X-X-, dialyzed 105,000 g supernatant ; -a-a-, 0.3 M-NaCl eluate from a DEAEcellulose column (not from the same experiment as the other samples and therefore not directly comparable with them).

266

D.

BALTIMORE

AND

A.

S. HUANG

TABLET Binding

of RNA from gradients with and without cytoplasm o/Q of input RNA retained by a filter

(A) (B)

RNA RNA

from from

gradient gradient

with cytoplasm without cytoplasm

97 2.4

The RNA’s were prepared by mixing 150,000 cts/min of 3aP-labeled viral RNA in 0.1 either 0.9 ml. of RS buffer or 0.8 ml. of RS buffer plus 0.1 ml. of s, fresh cytoplasmic treated with 1 y0 sodium deoxyoholate and 1 y0 BRIJ-58 (Girard & Baltimore, 1966). The were layered onto separate sucrose gradients and centrifuged for 17 hr at 20,000 rev./mm. sample with added cytoplasm, the RNA sedimented slightly ahead of the peak of ribosomes; other sample the RNA sedimented at about 50 s (Girard & Baltimore, 1966). The peak from eaoh sample was removed and tested for its retention by Millipore filters by mixing with 1 ml. of RX buffer and filtering.

ml. with extract mixtures In the in the fraction 0.01 ml.

either to RS buffer or to a cytoplasmic extract prepared in RS buffer and then centrifuged through sucrose-RS buffer gradients. As shown previously (Girard & Baltimore, 1966), the RNA sedimented at about 80 s in the presence of cytoplasm and at 50 s in its absence. Fractions containing RNA from the two gradients were then passed through membrane filters. Most of the RNA which had been mixed with cytoplasm was bound to the filter, whereas little of the untreated RNA was retained (Table 2). Thus, RNA which sediments rapidly due to interaction with cytoplasmic components is also retained by filters. (b) Puri$cation

of the factor

Using the filtration assay, the titer of binding factor has been determined after different treatments. Figure 1 shows that high-speed centrifugation removes little or TABLE 3 Elution of the factor from DEAE-cellulose Fraction

Supernatant after 30,000 g for 10 min Supernatrsnt after 105,000 g for 30 min DEAE eluate, 0.3 M-N&I DEAE eluate, 1 M-N&~

Total

units

87,000 87,000 44,000 570

0/0 of initial activity

100 100 50 0.7

Frozen HeLa cells (7 x lo3 cells in 20 ml. of RS buffer) were rapidly thawed and then chilled to 4°C. They were sonic&ted for 30 set at a setting of 4 on the Branson Sonifier and then centrifuged at 30,000 g for 10min in a Sorvall refrigerated centrifuge. The supernatant was then centrifuged at 105,000g for 30 min in the no. 40 rotor of the Spinco ultracentrifuge. The supernatant was dialyzed for 18 hr against 1 1. of 0.02 M-phosphate buffer, pH 7.4, and then passed through a DEAEcellulose column (10 cm x 0.9 cm) previously equilibrated with the same buffer. After elution with about 10ml. of this buffer, the eluant was changed to 0.02 m-phosphate buffer, pH 6.8, and elution was continued until the effluent was pH 6.8. Elution with O-3 M-N&~ in 0.02 M-phosphate, pH 6.8, was then begun and a sharp peak emerged. This material had a ratio of absorbance at 280 rnp to absorbance at 260 rnp of 1.3 to 1.4. Elution with 1-O M-N&~ in 0.02 m-phosph&e buffer, pH 6.8, then yielded another sharp peak with an absorbanoe ratio of about 0.6.

PROTEIN-RNA

INTERACTION

l

267

: I\

.-

.

1

.

t i

I\\

_

0..

,,..,* i

I-

II

IO Fraction

I

I

I

20 no.

Fro. 2. Effect of increasing concentrations of factor on the sedimentation rate of viral RNA. Different amounts of 0.3 ~-N&cl eluate from a DEAE-cellulose column were added to a constant amount of 32P-labeled viral RNA in RS buffer and the mixtures were sedimentad for 15.3 hr at 17,000 rev./min. For sample F, the gradient also contained 0.5 na-NaCl. A sample with no factor would sediment at the same position as the peak in sample F. The first fraction in each pattern represents the pellet on the bottom of the centrifuge tube and centrifugation was from right to left. Samples A, 160 units; B, 65 units; C, 32 units; D, 10 units; E, 3.2 units; F, 32 units plus 06 M-Nacl.

no activity and that dialysis against 0.02 M-phosphate buffer, pH 7.4, causes a slight reduction. The effect of dialysis is due to precipitation of protein; the specific activity of dialyzed material is the same as that of the high speed supernatant. Most of the activity in the dialyzed preparation is held by DEAE-cellulose, is eluted with O-3 M-NaGl (Table 3), and can be concentrated with 60% ammonium sulfate. Active material purified by these methods has approximately the same specific activity as the supernatant after high-speed centrifugation. Gradient elution of material bound to DEAE-cellulose has shown that the active material is heterogeneous; fractions eluted at different salt concentrations differ no more than twofold in their specific activities. Material removed by 1 M-NaGl from a column previously eluted with 0.3 i+.r-NaG1 contains mostly RNA and has very little activity (Table 3). The effect of the O-3 ivi-NaGl eluate from a DEAE-cellulose column on the sedimentation properties of RNA is shown in Figure 2. Labeled poliovirus RNA was incubated with different amounts of factor and the mixtures were analyzed on sucrose gradients. A maximal increase in sedimentation rate was produced by 65 units of factor; a twofold greater amount of factor had no further effect. By comparison with ribosomes, viral RNA sediments at about 110 s when a maximum amount of factor is bound.

268

D.

BALTIMORE

AND

r

A.

x .k! .5

S. HUANG

.’ \

8 200 0 z d IOO-

i

(j a a.....,....

e -.-.-.-.

.

-cd

i

!

\

:\“%%...lc \

/” o -+ .. ..-.-.-. _. ..-. +.-.-.-.-.--~* I IO Fraction

3. Effect of pronase samples of 3ZP-labeled viral made in 1 ml. of RS buffer. sample were incubated for gradients and centrifuged plus factor, 5 min at 37%; addition. FIG.

I

\, *h..*-. I 20

D I 30

_ _

no.

on the sedimentation rate of factor-viral RNA complexes. Three RNA plus 100 units of the 0.3 M-NaCl eluate from DEAE-cellulose were To one sample was added 50 pg of pronase and this and an untreated 5 min at 37°C. They were chilled and all three were layered on sucrose for 16 hr at 18,500 rev./min. Sample A, RNA plus factor; B, RNA C, RNA plus factor with pronase, 5 mm at 37%; D, RNA with no

(c) Inactivation

of binding factor

The binding factor is sensitive to the action of proteolytic enzymes. As determined by the filtration assay, both trypsin and pronase are able to inhibit most of the activity in cytoplasmic extracts (Table 4). Also, sedimentation analysis shows that pronase destroys the binding-factor activity of material eluted from a DEAE-cellulose column (Fig. 3). Other treatments which abolish activity are overnight incubation of cytoplasmic extracts with 1 o/osodium deoxycholate or treatment with 1 o/osodium dodecyl sulfate for less than one minute. The binding of protein, as measured by the filtration assay, is sensitive to salt concentration (Table 5). As the concentration of NaCl is increased, there is less RNA retained by filters; O-5 M-NaCl is sufficient to completely prevent binding. Sedimentation analysis also shows that the binding of purified protein does not occur in 0.5 M-NaCl (Fig. 2). (d) Nature of the interaction of binding factor with RNA To measure the amount of protein binding to RNA, W-labeled cellular proteins were prepared by elution from DEAE-cellulose with 0.3 M-NaCl and the maximum amount of 32P-labeled ribosomal RNA which this protein would bind to a filter was determined. A large excess of ribosomal RNA was added to a portion of the protein and the mixture was sedimented through a sucrosegradient. Lessthan 10% of

PROTEIN-RNA

INTERACTION

269

TABLE 4 Effect of pronase and trypsin on binding o/0 of input RNA retained by a filter Experiment (A) (B) (C)

Experiment (A) (B) (C) (D) For added

1 0.05 ml. cytoplasmic 0.05 ml. cytoplasmic 100 pg pronase 0.05 ml. cytoplasmic 50 pg trypsin

extract extract

plus

63

extract

plus

8 10

2 Factor at O’C Factor at 37% for 5 min Factor plus pronase at 37% 5 min No addition

experiment 2, O.l-ml. samples were removed to 1 ml. RS buffer and fLltered as usual.

TABLE

100 96 for

10 3 from

the four

mixtures

Binding Binding Binding Binding Binding Binding

buffer buffer buffer buffer buffer buffer

in Figure

3,

5

Effect of salt concentration on the forrrzation of factor-RNA oh input

Solution

described

complexes

RNA retained by a filter 100

+ + + + +

0.02 M-NaCl 0.05 M-NaCl 0.10 M-NaCl 0.25 M-Nacl 0.5 rd-NaCl

100 96 23 20 4

The binding reaction was carried out with 4 units of the 0.3 M-NaCl cellulose column in the appropriate solutions. The filters were also washed

eluate from a DEAEwith the same solution.

the protein was carried into the gradient with the RNA, indicating that only a small fraction of the protein had binding-factor activity. It has previously been shown that ribosomal RNA’s, poliovirus RNA and even DNA, are bound to the factor as measured by the rate of sedimentation (Girard & Baltimore, 1966). Table 6 showsthat binding factor also causesribosomal RNA’s to be retained by a filter; transfer RNA is similarly affected and its sedimentation rate is also increased (unpublished results). Table 6 further demonstrates that ribosomal RNA’s and viral RNA compete with each other for binding factor. However, if factor is added to Fl?-labeled RNA and an excessof unlabeled RNA is then added, competition is not evident (Table 6). The binding is therefore effectively irreversible. The binding factor not only increasesthe sedimentation rate but decreasesthe buoyant density of RNA. Pure RNA has a buoyant density in CM31of about 1.9

270

D.

BALTIMORE

AND TABLE

Binding of factor to different RNA’s

Labeled

RNA

28 s ribosomal 28 s ribosomal 28 s ribosomal 28 s ribosomal 35 s viral 35 s viral

(O-012) (O-012) (0.012) (0.012)

HUANG

and competition between RNA’8 Amount

RNA

of added factor

o/0 of input labeled RNA retained by a filter

(0.22) (0.18)

5 5 5

5 100 27 20

(0.09)

7 7

33

(0.09) (O-09)

5 5 5

28 s ribosomal 18 s ribosomal

18 s ribosomal (0.017) (0.017) (0.017)

S.

6

-

(0.015t) (0.015t)

18 s ribosomal 18 s ribosomal 18 s ribosomal

Unlabeled

A.

18 s ribosomal 18 s ribosomalt

99

92

25 78

The units of factor were determined in each case relative to the indicated amounts of each of the labeled RNA’s. The numbers in parentheses are the absorbance units at 260 mp of either labeled or unlabeled RNA added to the reaction mixtures. t Absorbance is due to 28 s ribosomal $ In this tube, the factor was added

RNA (see Materials and Methods). just before the competing unlabeled

RNA.

(Bruner & Vinograd, 1965); however, RNA added to cytoplasmic extracts bands at l-40 (Baltimore & Huang, 1968; Huang & Baltimore, 1970). The density of RNA with a small amount of binding factor is extremely heterogeneous (Fig. 4(c)), whereas the density with a maximum amount of added factor is about l-39 (Fig. 4(d)).

Fraction

no.

Fra. 4. Buoyant density of complexes of RNA with cytochrome c or binding factor. Equal samples of 32P-labeled viral RNA were mixed with: (a) 50 pg cytochrome o; (b) 500 pg cytochrome o; (c) 0.05 ml. of a dialyzed 105,000 g supernatant and (d) 0.5 ml. of a dialyzed 105,000 g supernatant. The preparations in (c) and (d) are similar to those described in Table 3 and were dialyzed against RS buffer. Each mixture was made up to 0.5 ml. with RS buffer, they were fixed with 0.5 ml. of 13.2% neutralizedglutaraldehyde and centrifuged for 4 hr at 37,000 rev./min in the Spinco SW41 rotor in a preformed gradient of CsCl (Baltimore & Huang, 1968). The descending lines in each panel represent the buoyant densities. The arrows in panels (b) and (d) represent the positions of visible bands; the heavier band is the RNA-protein complex, the lighter band is free protein.

PROTEIN-RNA

271

INTERACTION

(e) Binding of2)uri$ed basic proteins to RNA A basic protein, like cytochrome c, will bind to RNA and cause it to be retained by a filter. However, sedimentation analysis of the complex shows that binding of cytochrome c to RNA precipitates the RNA, rather than causing the formation of a discrete soluble complex like that formed with binding factor (Table 7). Similarly, TABLE

Effect of cytochrome c on RNA

and

7 its

reversal by cytoplasmic extract % of total radioactivity in pellet

(A) (B) (C) (D) (E)

No addition Cytochrome Cytochrome Cytoplasmic Cytoplasmic

c e followed by cytoplasmic extract extract followed by cytochrome c extract

1 80 8 3 4

% of total radioactivity in gradient fractions

99 20 92 97 96

Mixtures of about 700 cts/min of 3aPWlabeled viral RNA, 0.5 ml. freshly prepared cytoplasmic extract with 1 o/0 sodium deoxycholate and 1 o/0 BRIJ-58, and 50 pg of oytochrome c were prepared as described above, in 1 ml. each of RS buffer. The mixtures were layered over sucrose gradients and centrifuged for 5 hr at 25,000 rev./min. The total cts/min in the gradient fractions and in the pellets on the bottom of the centrifuge tube were tabulated and the percentages calculated. In (C), (D) and (E) the oytoplasmic extracts caused the RNA to sediment at about 80 s (Girard & Baltimore, 1966). Radioactivity in the pellet represents all material sedimenting faster than 400 s.

RNA complexed to poly-L-ornithine is retained by a filter but sedimentation analysis shows that the RNA is precipitated. The precipitated complex with cytochrome c has a buoyant density of 1.40 to I.60 g/cc, depending on the concentration of oytochrome c (Fig. 4). Similarly, labeled RNA bands at a buoyant density of I.58 when precipitated from a solution of 1 tug of poly-L-ornithine/ml. and at a buoyant density of 1.40 when precipitated from a solution of 10 pg of poly-L-ornithinelml. When a mixture of cytochrome c and cytoplasmic extract is tested for its effect on RNA, a soluble complex is found which is identical to that formed with cytoplasm alone (Table 7). The cytochrome c can even be added to the RNA before addition of the cytoplasmic extract and a soluble complex will be formed.

4. Discussion Two assays for binding-factor activity have been used in this study: retention of RNA by nitrocellulose filters and increased rate of sedimentation of RNA. They appear to measure the same interaction because (1) RNA whose sedimentation rate was increased by interaction with cytoplasm does bind to a filter (Table 2), and (2) both effects are abolished by pronase and 0.5 M-NaCI. Much less protein is required for the filtration assay; it is likely that only a few molecules of protein are necessary to cause retention of the RNA (Fig. 1). Calculating from its density of l-40, RNA

272

D.

BALTIMORE

AND

A.

S.

HUANG

maximally complexed with protein has about three weights of protein per weight of RNA?; this would correspond to about 120 molecules of protein having a molecular weight of 50,000 daltons each binding to one molecule of poliovirus RNA which has a molecular weight of about 2 x lo6 daltons. The binding of protein to RNA which is being measured by the assays is effectively irreversible because the great dilution of the reactants in the sucrose gradients or during the washing of the filters would dissociate any weakly bound protein. There is no evidence to indicate that the binding factor is anything but a heterogeneous class of soluble proteins which fortuitously and non-specifically loind to RNA at low ionic strength. It can be concluded that binding factor is soluble protein from its lack of sedimentation at 105,000 g, its behavior on DEAE-cellulose and its sensitivity to trypsin and pronase. There seems to be no specificity to the interaction; binding factor attaches to all types of RNA and ribosomal RNA competes with viral RNA. Furthermore, the titer of binding-factor activity against viral RNA is the same in uninfected and infected cells. The nature of the interaction between the proteins and the RNA is unclear but available evidence suggests that it involves an electrostatic interaction between a basic region on an otherwise neutral or acidic protein and the phosphates of the RNA. Inhibition of the reaction by 0.5 M-Nacl suggests that the interaction is electrostatic. The fact that the active material binds to DEAE-cellulose at pH 6.8 implies that it does not consist of very basic proteins. Furthermore, excess basic protein precipitates RNA while excess binding factor does not. Precipitation is presumably due to the multiple basic sites on a molecule like cytochrome c; the lack of precipitation by binding factor suggests that it may have only one exposed basic region. The binding factor forms an extremely tight association with RNA at low ionic strength. It is able to reverse the precipitation by cytochrome c, and unlabeled RNA added after the formation of the complex between binding factor and labeled RNA does not lead to dissociation of the complex (Table 6). In all of its properties, binding factor acts like the purified proteins which Hofstee (1962,1964) has shown to form soluble complexes with DNA and RNA. At about their isoelectric point, many proteins bind to DNA forming a complex with up to 20 times more protein than nucleic acid (Hofstee, 1964). When the pH is much less than the isoelectric point of the protein, insoluble precipitates rather than soluble complexes are formed. These effects are reversed by increasing the ionic strength and therefore appear to be due to electrostatic forces. It should be noted that a large amount of protein must bind to an RNA molecule in order to cause a change in the sedimentation rate of the RNA. Therefore, to demonstrate the existence of binding factor in a cytoplasmic extract, only a small amount of RNA can be added to the extract (e.g. less than 2 ,ug of RNA added to the cytoplasm of 2 x lo7 cells). If a larger amount of RNA is added, the protein molecules will t Using the formula given in Table 2 of Perry & Kelly (1966) it is possible to calculate the ratio of RNA to protein from buoyant densities. Using Bruner & Vinograd’s (1965) determination of the buoyant density of RNA as 1.90 in C&l and our measured buoyant density for the majority of HeLa cell protein as l-29 (Huang & Baltimore, unpublished data), a particle of buoyant density 1.40 would have 3 weights of protein per weight of RNA. The accuracy of such a calculation is questionable, since poliovirus, with a buoyant density of 1.34 (Huang & Baltimore, 1970) is 25 to 30% RNA (Baltimore, 1969). Furthermore, the density of those proteins which are bound to viral RNA to form viral ribonucleoprotein is not known. The calculation does indicate that a large amount of protein is bound to RNA in the presence of cytoplasm and helps to explain the great effect of cytoplasm on the sedimentation rate of RNA (Girard & Baltimore, 1966).

PROTEIN-RNA

INTERACTION

273

distribute over all of the RNA and no increasein the sedimentation rate of the RNA will be evident. Although the role of binding factor in cellular metabolism remains an open question, there is no indication of any specific function. In fact, in the ionic environment of the cell, which includes more than 0.15 M-pOtaSSiUm ion (Davson, 1964), the tight binding which we have discussed here probably does not occur but freely reversible interactions of protein

and RNA

may be present.

This work was supported by United States Public Health Service grants no. CA-07592 and no. AI-08388 and American Cancer Society Reaserch Grant no. E-512. One of us (A.S.H.) was a recipient of a postdoctoral fellowship from the United States Public Health Service (F2-AI-35110). The other author (D.B.) received a Faculty Research Award from the American Cancer Society. We thank Laura1 Gonsalves and Virginia Lee for technical assistance. Part of this work was carried out at the Salk Institute for Biological Studies. REFERENCES Baltimore, D. (1969). In The Biochelnistry of VirzLses, ed. by H. Levy, p. 101. New York: Marcel Dekker. Baltimore, D., Girard, M. & Darnell, J. E. (1966). virology, 2, 179. Baltimore, D. & Huang, A. S. (1968). Science, 162, 572. Bruner, R. & Vinograd, J. (1965). Bio&m. biophye. Acta, 108, 18. Davson, H. (1964). A Textbook of General Physiology, Third edition. London: J. & A. Churchill Ltd. Girard, M. & Baltimore, D. (1966). Proc. Nat. Acad. Sci., Wash. 56, 999. Hofstee, B. H. J. (1962). Biochim. biophys. Acta, 55, 440. Hofstee, B. H. J. (1964). Biochim. biophys. Acta, 91, 340. Huang, A. S. & Baltimore, D. (1970). J. Mol. Biol. 4’7, 275. Infante, A. A. & Nemer, M. (1968). J. Mol. Biol. 32, 543. Jones, 0. W. & Berg, P. (1966). J. Mol. BioZ. 22, 199. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951). J. BioZ. Chem. 193, 265. Penman, S., Becker, Y. & Darnell, J. E. (1964). J. MoZ. BioZ. 8, 541. Perry, R. P. Sr;Kelly, D. E. (1966). J. Mol. BioZ. 16, 255. Riggs, A. D., Bourgeois, S., Newby, F. F. & Cohn, M. (1968). J. Mol. BioZ. 34, 365. Yarus, M. & Berg, P. (1967). J. Mol. BioZ. 28, 479.