- Email: [email protected]
Involvement of basic amino acids in the activity of a nucleic acid helix-destabilizing protein
42
Bioehimica et Biophysica Acta, 654 (1981) 42 51
Elsevier/North-HollandBiomedicalPres~ BBA 99864 INVOLVEMENT OF BASIC AMINO ACIDS IN THE ACTIVITY OF A NUCLEIC ACID HELIX-DESTABILIZING PROTEIN RICHARD L. KARPEL,DAVID J. MERKLER *, BRIAN K. FLOWERSand MARTHAD. DELAHUNTY Department of Chemistry, University o f Maryland Baltimore County, Catonsville, MD 21228 (U.S.A.)
(Received September 3rd, 1980) (Revised manuscript received February 2nd, 1981)
Key words: Helix-destabilizing protein; Chemical modification; RNAase A; DNA melting
Under conditions of low ionic strength, ribonuclease A, which binds more tightly to single- than to doublestranded DNA, lowers the melting temperature of DNA helices (Jensen and yon Hippel (1976) J. Biol. Chem. 251, 7198-7214). The effects of chemical modification of lysine and arginine residues on the helix-destabilizing properties of this protein have been examined. Removal of the positive charge on the lysine e-amino group, either by maleylation or acetylation, destroys the ability of RNAase A to lower the Tm of poly[d(A-T)]. However, reductive alkylation of these residues, which has no effect on charge, yields derivatives which lower the Tm by only about one-half that seen with unmodified controls. Phenylglyoxalation of arginines can largely remove the Tin-depressing activity of RNAase A. RNAase S, which is produced by cleavage of RNAase A between amino acids 20 and 21, possesses DNA helix-destabilizing activity comparable to that of the parent protein, whereas S-protein (residues 21-124) increases poly[d(A-T)] Tm and S-peptide ( 1 - 2 0 ) has no effect on Tin. These results suggest that the specific location of several basic amino acids situated on the surface of RNAase A is largely responsible for this protein's DNA melting activity.
Introduction A number of different classes of protein have been described which affect the conformation of nucleic acids [1,2]. Those proteins, which lower the melting temperatures of nucleic acids by selectively binding to single strands, are termed helix-destabilizing proteins [3]. They have been found in organisms ranging from bacteriophages (e.g., gene 32 protein from T4 phage [4] ) to higher species (e.g., UP1 from calf thymus [5,6]). Although the physiological roles of most helix-destabilizing proteins have not been determined, the T4 protein [7,8] as well as one isolated from Escherichia coli [9] have been shown by genetic
* Present address: Department of Microbiology,Cell Biology, Biochemistryand Biophysics,PennsylvaniaState University, University Park, PA 16802, U.S.A.
methods to be required for DNA replication and recombination. Although little is presently known about the structural basis for the binding specificity of helix-destabilizing proteins, their selective affinity for single strands presumably is the result of specific chemical interactions between particular amino acid residues and functional groups in the nucleic acid. The strong salt dependence of the binding between protein and nucleic acid is indicative of electrostatic interactions between basic amino acids and nucleic acid phosphates. Record et al. [I0] have shown that the slope of the linear dependence of binding constant on log [monovalent cation] can be interpreted in terms of the number of ion pairs, m', formed between each protein molecule and the length of polynucleotide chain to which it is bound. This approach has been applied to several helix-destabilizing proteins, including pancreatic ribonuclease A, which lowers the
0 005-2787/81/0000-0000/$02.50 © Elsevier/North-HollandBiomedicalPress
43 melting temperatures of DNA double helices [ 1 1 14]. In the case of RNAase A, m' = 7 for the interaction with denatured DNA and 4 for the (weaker) interaction with native DNA [10,14]. Basic amino acids are probably not the only residues involved (several tyrosines in the fd phage gene 5 protein appear to be involved in its binding to single-stranded DNA [2,15,16]); however, electrostatic forces clearly are an important component o f the affinity of helixdestabilizing proteins for single-stranded nucleic acids. Although this Tin-depressing activity of RNAase A may have no direct physiological relevance, the extensive body of knowledge of its structural and chemical properties [17-19] suggests that it is an excellent model for the study of protein-induced nucleic acid helix destabilization. In order to explore the involvement of basic amino acids in this process, we have studied the effects of specific chemical modification of lysine and arginine residues on the DNA Tmdepressing activity of ribonuclease A. RNAase A is especially well suited for this type of study, since a number of modification procedures have been carried out on this protein, and the resultant derivatives have been well characterized [20-23]. These include acetylation [24], amidination [25] and reductive alkylation [20,26,27], which have been shown to react specifically with RNAase lysines. Under the proper conditions [28], phenylglyoxalation of RNAase A leads to specific derivatization of arginines, with concomitant deamination of N-terminal a-amino group. We have also utilized maleylation, which in general is specific for lysines [29]. We also report the effects on DNA Tm of the cleavage product, RNAase S, and its two components, S-protein and S-peptide. The results obtained with these products indicate that the native tertiary structure of the complete protein is a necessary requirement for helix-destabilizing activity, and confirm the view that the interaction with single-stranded DNA involves binding of nucleic acid phosphates to a set of basic amino acids in a specific three-dimensional orientation. Materials and Methods
Reagents and buffers. Ethyl acetimidate hydrochloride and 2,3-butane-
dione were purchased from Aldrich Chemical Co.; 2,4,6-trinitrobenzenesulfonic acid was purchased from Sigma Chemical Co. and recrystallized from 5 M HC1 [30]; N-ethylmorpholine was obtained from Pierce Chemical Co. and phenylglyoxal from Sigma. All other chemicals were reagent or comparable grade. The pH of protein solutions was measured with nonbleeding pH strips (Merck No. 9583, graduated to 0.2 pH units). The buffering systems used throughout were: buffer A: 1.2 mM phosphate (K*)/8.8 mM TrisHC1/0.1 mM Na2EDTA, pH 8.0 +_0.1; buffer B; 10 mM phosphate (K*)/0.1 mM Na2EDTA/pH 7.0 +- 0.1; buffer C: 10 mM Tris-HC1/0.1 mM Na2EDTA, pH 8.0 +-0.1. Unless otherwise noted, the modified proteins and the controls were exhaustively dialyzed against buffer C.
Proteins Lyophilized, phosphate-free pancreatic ribonuclease A was purchased from Worthington Biochemical. Several of the lysine-modified derivatives (those subjected to maleylation, amidination and reductive methylation) were prepared from RNAase A which had been heated in buffer C to 60°C for 5 min, a procedure which breaks up aggregates [ 19,31 ]. However, no differences were observed between the Tin-depressing activity of heat treated RNAase A and that of material which had not been subjected to this treatment, suggesting either that the extent of aggregation was minimal, or that it has no effect on the Tmdepressing properties of the protein. Preliminary experiments with a less highly purified RNAase (Sigma type I-A) gave results very similar to that of the Worthington protein. RNAase A concentrations were measured spectrophotometrically at 277.5 nm, using an extinction coefficient of 9.56.103 M -~. cm -~ [14]. Concentrations of derivatized and control samples were determined by the Lowry method [32] with underivatized RNAase A as the standard. RNAase S and S-peptide were obtained from Sigma; S-protein was the gift of Dr. I. Chaiken. Concentrations of RNAase S were determined optically, using the same era as that of RNAase A [33]. The molar extinction coefficients used for S-protein and S-peptide were, respectively, 9.20.103 at 280 nm [33] and 2.9,102 at 258 nm [34]. Nucleic acids Poly[d(A.T)] was obtained from P-L Biochemicals,
44
Modification procedures
and dissolved in buffer B. Its c o n c e n t r a t i o n was determ i n e d s p e c t r o p h o t o m e t r i c a l l y in buffer B, using e262 = 6.65 • 103 M -1 • cm -1 [35]. Clostridium perfringens D N A was o b t a i n e d f r o m Worthington, and treated in the same way as the p o l y [d(A-T)].
The specific m o d i f i c a t i o n reactions study are summarized in Table I. F o r tions, control samples were subjected to procedures in the absence o f a reagent
TABLE I SPECIFIC MODIFICATION REACTIONS 1. Lysine a . + .--~ - -
Maleylation:
+0 / P-NH 2 \
0 II C-C-H
O pH 8.0 II --* R-NHCCH=CHCOO- + H +
C-C-H II O b.+~O O CCH3
Acetylation:
0
II , R-NHCCH3 + CH3COO- + I-I÷
P - N H 2 + O/\ CCH3 II O
C. + - +
+
Reductive Alkylation:
Oil pH 9 / R NaBH 4 /R P - N H 2 + RCR' ~ P-N=C\ , P-NHCH , R'
° 1NaBH4
II RCR' R
P-NH+(C H/ \
R'
Amidination:
pH<9
)2
R
/
P-N(CH
)2 R'
NH 2 +NH2 II pH>8.s II P-NH2 + CH3COCH2CH3 ~ P-NHCCH3 + CHaCH2OH
2. Arginine Phenylglyoxalation:
+
NH2
P-NH--C(N.
©
O=C
O=C I H
÷
©
0
N-C-OH
N-C-O
N CO.
N-C
I H
OII
I H
(one of the proposed structures, refs. 23 and 28)
used in this all modificathe identical essential for
45 derivatization. In this manner, it could be confirmed that all effects on the DNA Tin-depressing activity of RNAase A were due to specific amino acid modification, rather than to any other changes occurring during the derivatization procedures.
Lysine modification Maleylation [29,36]. 0.08 ml of 0.3 M maleic anhydride (twice recrystallized from chloroform [21]), in dioxane was added to 1.12 ml of 5.4 mg/ml RNAase A in 0.20 M phosphate buffer, pH 9.0. The reaction was allowed to continue for 2 h at 4°C, maintaining the pH at 8.0. A parallel control was run by adding 0.08 ml dioxane without maleic anhydride. Modification was also done using one-half of the above amounts. In order to eliminate any possibility of formation of o-maleyl adducts, maleylated RNAase samples were treated with 0.8 M hydroxylamine (pH 9.5) for 15 min at room temperature, then overnight at 4°C [37]. This treatment had no effect on the derivative's DNA Tin-depressing activity. Acetylation [21]. A 25 mg/ml solution of RNAase A was mixed with an equal volume of a solution of saturated sodium acetate, chilled in an ice bath and subjected over a period of 1 h to four additions of acetic anhydride, with the total mass of the latter equal to that of the protein. Controls were treated in the same manner with water substituted for acetic anhydride. The protein was dialyzed against buffer C, and the procedure was repeated. Although this method is quite selective for primary amino groups [21], the derivative was then treated with 1 M hydroxylamine/1 mM Na2EDTA, pH 8.0, for 20 min at 27°C to hydrolyze any o-acetyltyrosine which may have formed [24]. Reductive methylation [20,26]. A 1.0-2.5 mg/ml RNAase A solution in 0.2 M borate buffer, pH 9.0 at 0°C, was adjusted to 0.5 mg/ml NaBH4 by the addition of solid NaBH4. To the mixture, 0.5/.tl/rnl increments of formalin (37% aqueous formaldehyde solution) were added at 5-min intervals until a final concentration of 2.5 /~l/ml was reached. Controls were treated similarly, with water being substituted for the formalin.
Reductive isopropylation [20,26]. To 0.55 ml of 1.0 mg/ml RNAase A in 0.2 M borate buffer, pH 9.0, containing 1 I% (v/v) acetone at 0°C was added a 0.03
ml aliquot of 10 mg/ml NaBH4 every 5 min until a final concentration of 2.5 mg/ml was reached. Controis consisted of protein in borate buffer treated only with NaBH4 or protein in borate buffer treated only with acetone. Some samples and controls were subjected to three rounds of this procedure (see Results). Amidination [38]. 0.45 ml of 1.0 M NaOH was added to 0.111 g of ethyl acetimidate hydrochloride, then mixed with 0.75 ml of 10 mg/ml RNAase A in buffer C. The pH of the reaction mixture was adjusted and maintained at 8.3-8.6 for 2 h at 0°C with stirring. Controls were prepared in the same manner, but in the absence of ethyl acetimidate. Modification was also done using one-third of the above amounts.
Quantitation of free amino groups The extent of lysine modification was determined by quantitation of remaining free amino groups via reaction of modified derivatives with 2,4,6-trinitrobenzenesulfonic acid, as modified by Habeeb [39]. Parallel assays were run on unmodified RNAase A, and the extent of modification computed by comparison of the A a3s of modified with that of unmodified protein. The absorbance obtained with unmodified RNA. ase A was consistent with its content of 11 free amino groups and the range of extinction coefficients reported in the literature [20,21]. On the basis of results with reductively alkylated RNAase A derivatives, it appears that the trinitrobenzenesulfonic acid assay indicates a slightly higher number of free amino groups than that Obtained from amino acid analysis [26].
Arginine modification Phenylglyoxalation [28]. An aliquot of 15 mg]ml phenylglyoxal in 0.2 M N-ethylmorpholine acetate buffer, pH 8.0, was added to an equal volume of 10 mg/ml RNAase A in the same buffer. The mixture was allowed to incubate for 1 h at room temperature. Controls were prepared in the same manner in the absence of phenylglyoxal. Modified protein and the control were exhaustively dialyzed against 10 mM phosphate (K ~) buffer, pH 3.9, containing 0.1 mM Na2EDTA.
A bsorbance-temperatureprofiles Teflon-stoppered micro quartz cuvettes containing
46
100 #1 of test solutions were placed in a Gilford 2400-2 spectrophotometer designed to raise the temperature from 0 to 95°C at a constant rate, which was usually 25 K/h in these experiments. Generally, three nucleic acid/protein samples were melted simultaneously with a semi-micro reference cuvette containing buffer C. The concentration of nucleic acid was usually 5.95 • 10 -s M(p), with [protein] as indicated. Temperature was continually monitored by means of a calibrated thermistor (Yellow Springs Instruments) inserted through a narrow hole in the stopper of the reference cuvette; absorbance was measured at 260 nm. Tm values are defined as the inflection points of the transitions as determined by differentiating the measured absorbance-temperature profiles [40]. The reproducibility of the Trn values was -+1°C, except when profiles were very broad, where it was
1Z100 ZO
8 4
/
0
E
3
2.5
/
Pr0,e,n
S-Peptide o °
V"
~7
-12 -16 -20
281~- -
In order to assess conditions optimal for testing the effects of chemical modifications on the poly[d(A-T)] Tin-depressing activity of RNAase A, a range of protein concentrations (with constant [DNA]) in buffer A was explored. The results are shown in Fig. 1. Under the conditions of these experiments, the magnitude of the Tm depression (--ATm) sigmoidally increases with increasing [RNAase A ] : [poly[d(A-T)]]p. Beyond a ratio of 0.2, only a slightly larger IATml value is obtained, with precipitation apparent at [RNAase A] : [poly[d(A.T)]]p > 0.3. The poly[d(A-T)] Tin-depressing activity of derivatives was thus generally measured at [protein[: [DNA] p > 0.2. Although inorganic phosphate is an inhibitor of
4
-8
Circular dichroism experiments
Variation of poly[ d(A-T)] Tm with [ribonuclease A/
5
-4
-24
Results
7
D
±2°C.
Measurements of circular dichroism (CD) were taken with a JASCO J-40 spectropolarimeter. The cuvette holder of the instrument was adapted to hold rectangular semi-micro quartz cuvettes containing as little as 400 ;zl of sample, and was connected to an external circulating water bath. Cuvette temperature was determined via a calibrated thermistor probe (Yellow Springs Instruments).
10
RNa
_''--_ ~
1
i
al
0.~
a'3
-
0.4
[P.o.E,N Fig. 1. Dependence of poly[d(A-T)] Tm in buffer A on [RNAaseA], o o; [RNAase S], A__ _zx; S-protein, a-~; and S-peptide, z~ ~. [poly[d(A-T)]] = 5.95. 10-s M(p). In the absence of added protein, Tm = 38 -+ I°C. RNAase enzymatic activity [19], at the low levels of phosphate used in these experiments, no significant effect on Tm was observed. In 10 mM phosphate (pH 7.5)/0.1 mM Na2EDTA, we observed that at a [RNAase A] : [poly[d(A-T)]]p of 0.23, ATm = - 2 2 K, essentially identical to that obtained in buffer A. Unless otherwise stated, the Tm depression experiments were performed in buffer A, which contained 1.2 mM phosphate.
Effects of lysine modifications on RNAase A melting activity We have chosen several procedures which differ in their effect on the charge of the e-NH~ group: maleylation replaces the positive charge with a negative one, acetylation removes the charge, and reductive alkylation and amidination have no effect on the charge (Table I). The Tin-depressing effects of these derivatives, as well as those of arginine, are tabulated in Table II.
47 TABLE II SUMMARY OF Tm-DEPRESSING ACTIVITY OF DERIVATIVES OF RNAases Unless otherwise stated, experiments were performed in Buffer A, 1.2 mM phosphate (K*)/8.8 mM Tris-HCl/0.1 mM Na2EDTA, pH 8.0 Modification
[protein]
[poly[d(A-T)]lp
A. Formation of lysine derivatives Maleylation 0.21 Acetylation 0.23 Reductive methylation 0.23 0.33 Reductive isopropylation 0.23 0.45 Amidination 0.23 B. Formation of arginine derivatives Phenylglyoxalation 0.23
ATm (K)
,xTm, control (K)
0 0 -13.5 -12.5 -17 to -8 a -13 b - 19
- 18 -21 -23 -24 -22 ± 2
pH 7.1:-10 c pH7.5: - 5 d pH 8.3: 0e
_ c -23d -22 e
-22
a Derivatives varied between 32 and 70% modification (see text). b Derivative was 54% modified, and gave the same ATm at [protein] : [poly[d(A-T)] ]p = 0.23. e In 5.5 mM Tris-HCl]4.5 mM phosphate (K*)/0.1 mM Na2EDTA. Control samples precipitated. d In 7.1 mM Tris-HC1/2.9 mM phosphate/0.1 mM Na2EDTA. e Same as footnote d, but adjusted to pH 8.3.
Maleylation Reaction of maleic anhydride with ribonuclease A produced a derivative in which 75% of the primary amino groups were no longer sensitive to trinitrobenzenesulfonic acid, and thus were modified. This derivative possessed no helix-denaturing activity (Table II). The control RNAase depressed Tm by 18°C, nearly that of unaltered RNAase (Fig. I). Although maleylated derivatives slowly revert to primary amino groups at acid pH [36], this clearly was of no consequence, since no effect on Tm was observed. Substantially the same result was obtained with CI. perfringens DNA, where the Tm of 53°C (in buffer A) was depressed by 25°C in the presence of R N A a s e A ([protein]: [nucleic acid] p = 0.23) but was unaffected by maleylated RNAase. In all probability, the change of eight or nine positive to negative charges brings about a gross perturbation in isoelectric point. In a preliminary experiment, we failed to observe electrofocusing of maleylated RNAase A on a pH 3 - 1 0 isoelectrofocusing gel, suggesting that the pl of the modified protein was too acidic to be resolved on this system. The polyanionic
modified RNAase thus cannot interact with the polyanionic DNA, and has no effect on Tin. No gross alteration of the RNAase tertiary structure occurred upon maleylatlon, since its far-utraviolet CD spectrum was essentially unaltered from that of unmodified protein (data not shown).
Acetylation Acetylation of RNAase A completely inactivates its enzymatic activity [24]. A derivative that was 72% modified had no effect on poly [d(A-T)] Tin, whereas the control RNAase depressed T m comparably to untreated protein (Table II). Removal of the lysine positive charge thus completely destroys the interaction of the protein with single-stranded DNA.
R eductive alkylation Reductive methylation of RNAase A with formaldehyde and NaBH4, leading to dimethylated lysine derivative, destroys catalytic activity with little change in physical properties and only slight loss in 3'-cytidylic acid binding [20,26]. A similar result, with somewhat greater loss of CMP binding, was ob-
48 tained upon modification with acetone and NaBH4 (leading to monoisopropylated derivatives [26,41]). Both methylated and isopropylated RNAase A were tested for poly[d(A-T)] Tin-depressing activity. These derivatives showed no evidence on SDS-polyacrylamide gels for any protein-protein cross-linking resulting from the modification procedure. A methylated derivative with 92% of the amino groups modified depressed Tm by 13.5°C at the standard [protein] : [DNA]p of 0.23 (Table II). When the [methylated RNAase] : [DNA]p was increased to 0.33, the Tm depression was virtually unchanged (-12.5°C), indicating that T m depression by this derivative was maximal, and therefore less than that achieved by unmodified RNAase A. Several other methylated samples and controls were tested for poly[d(A-Y)] Tm-depressing activity, and all gave very similar ATm values (+1 K). Several isopropylated RNAase A derivatives were tested for poly[d(A-T)] Tm-depressing activity. Between 38% and 75% of primary amino groups were reacted, with the greatest degree of reaction resulting from three rounds of modification. ATm varied between - 1 7 K and - 8 K (Table II), with the greatest effect occurring with two derivatives that were both 75% modified. A 54% modified isopropylated derivative yielded a ATm of - 1 3 K over 2-fold range of [protein] : [DNA]p, so that, as with the methylated derivative, Tm depression was maximal. Isopropylated RNAase A/poly[d(A-T)] mixtures could be melted, cooled and re-melted to yield melting temperatures within 1°C of the initial values. The Tin-depressing activity of RNAase A is reduced to a larger degree by extensive isopropylation than by methylation, and probably reflects the greater bulkiness of isopropylated relative to dimethylated eiamino groups (see Discussion). Although the transition breadths (which we define as the widths at half-height of &A/AT vs. T p l o t s ) o f RNAase A - and alkylated RNAase A - poly [d(AT)] mixtures were broader ( 4 - 7 K) than that of poly[d(A-T)] alone (2.5 K), there was virtually no overlap among poly[d(A-T)], methylated (or isopropylated) RNAase-poly[d(A-T)], and unmodified RNAasepoly[d(A-T)] transitions (Fig. 2). This indicates that the population of atkylated protein molecules is reasonably homogeneous. Indeed, unmodified RNAase A effected somewhat broader transitions ( 6 - 7 K)
than alkylated protein ( 4 - 5 K). The substantial poly[d(A-T)] Tin-depressing activity of alkylated RNAases indicates that these derivatives retain their binding to single-stranded DNA. The effect of isopropylated and methylated RNAase A on the conformation of single-stranded DNA was assessed by monitoring the influence of these derivatives on the CD spectrum of poly [d(A-T)] at a temperature at which the nucleic acid is fully melted, under standard conditions ([RNAase A] : [poly[d(AY)]]p = 0.23). The maximum, minimum, and crossover point wavelengths of poly[d(A-Y)]/alkylated RNAase A mixtures measured at a temperature at which the nucleic acid is totally denatured (34°C)
- -
I
!
I
~
[
1
/ C
E e ~q
~q
10 f4
f8
2'6 0
3'4 3'8 4b
°C Fig. 2. Differentiated melting profiles of RNAaseA/poly[d(A-T)] mixtures. Poly[d(A-T)] in the absence of added protein, o m; poly[d(A-T)] + methylated RNAaseA, [protein] : [DNA]p = 0.3, o o; poly[d(A-T)] + control RNAase A (subjected to the reductive alkylation procedure in the absence of formaldehyde), [protein[ : [DNAlp = 0.3, A ~. lsopropylated RNAaseA/poly[d(A-T)[ mixtures showed profiles similar to that of the methylated derivative+ DNA mixture.
49 were virtually identical (within 2 nm) to that obtained from the sum of individual derivatized protein and heat-denatured poly[d(A-T)] CD spectra, although the ellipticity maxima and minima were slightly different (not shown). These results are similar to those of poly[d(A-T)] and unmodified RNAase A [ 14], and suggest that little distortion of the nucleic acid chain occurs when it interacts with the bulky isopropylated or dimethylated e-amino groups of the modified derivatives.
A midination RNAase A loses its enzymatic activity upon amidination [25]. Upon reaction with ethyl acetimidate, 82% of the amino groups were found to be modified. This derivative had only a slight effect on the protein's poly [d(A-T)] Tin-depressing activity, yielding a ATm o f - 1 9 K, compared to a ATm of - 2 2 K observed with the control sample (Table II). Effects of arginine modification on RNAase A melting activity Reaction of RNAase A with phenylglyoxal under the conditions described in the Materials and Methods results in the modification of three of the four arginine residues of this protein, deamination of the N-terminal lysine, and the loss of 89% of enzymatic activity [28]. This derivative displayed a pH-dependent poly [d(A-T)] Tm-depressing activity, which was always smaller than that of unmodified RNAase A (Table II). At pH 8.3, AT m was only -0.5 K (~0), whereas it was -5 K at pH 7.5. At still lower pH, 7.1, the melting profile was extremely broad, with ATm = -10 + 2 K. Phenylglyoxal derivatives are known to be labile at pH > 4 [26] ; however, remelting of phenylglyoxalated RNAase/poly[d(A-T)] mixtures at the highest and lowest pH conditions yielded essentially the same (-+I°C) Tin, indicating no change in the degree of modification during melting. These results suggest that the bully phenylglyoxal groups (Table I) effectively block close contact of lhe single-stranded DNA phosphates with the positive charge of derivatized arginines and/or unmodified lysines. Although the pKA of this derivative is not known, the pH dependence of the data indicates that the derivative is likely unprotonated at pH 8.3, and protonated at pH 7.1. Several other modification procedures for arginine
have been developed [23]. Preliminary experiments with a butanedione derivative of RNAase A, prepared by the arginine-specific method of Riordan [42], showed no effect on poly[d(A-T)] Tm depression.
RNAase S, S-protein, and S-peptide Cleavage of RNAase A by subtilisin between residues 20 and 21 produces a derivative, RNAase S, which has tertiary structure and catalytic activity similar to that of the uncleared protein [17-19]. We have observed that RNAase S in buffer A possesses slightly greater poly[d(A-T)] Tin-depressing activity than RNAase A (Fig. 1). However, S-protein, residues 21-124 derived from RNAase S, acts as a helix stabilizer, raising the Tin by 10.5°C at a [protein] : [poly[d(A-T)]p of 0.10 in buffer A (Fig. 1). Precipitation was observed at higher [S-protein] : [DNA]p. S-peptide, residues 1-20, had no effect on poly[d(AT)] Tm. The Tm of RNAase S- and S-protein/poly[d(A.T)] mixtures was reproduced upon melting, cooling and re-melting. When S-protein and S-peptide were mixed in equimolar amounts at 4°C to reconstitute enzymatically active material [25], the resultant protein (RNAase S') was found to depress poly[d(A-T)] Tm as well as (within 2°C) RNAase S. Thus, although S-protein contains eight of the ten lysine residues and three of the four arginines of the parent protein, it is clear that the tertiary structure of the complete protein, as well as, possibly, the missing basic amino acids are required for DNA melting activity (see Discussion). Discussion
The results presented herein clearly indicate that basic amino acid residues are necessary components of the interaction of the model helix-destabilizing protein, RNAse A, with single-stranded DNA. The removal of the positive charge on lysine e-amino groups destroys the DNA Tin-depressing activity of this protein, as shown by the inability of maleylated and acetylated RNAase A to affect poly[d(A-T)] melting. However, when methyl, isopropyl or acetimido groups are substituted on lysine, retaining the positive charge under the near neutral pH conditions utilized, substantial melting activity is still observed. Since these derivatives are known to be enzymatically inactive, we conclude that substituents on lysine
50 e-amino groups more critically affect RNAase A catalytic activity than its DNA helix-destabilizing activity. At least some of the residues that bind to DNA are involved in the enzymatic activity of RNAase A, since single-stranded DNA is an effective competitive inhibitor of the degradative action on RNA substrates [431. It seems likely that the role of the lysines is to provide neutralizing positive centers "for the negative charges of the single-stranded DNA chain. Substitution of alkyl or acetimido groups apparently permits sufficient electrostatic interaction to occur, suggesting that there may be some flexibility in the orientation of the basic side-chains and/or nucleic acid backbone (the CD results indicate no significant distortion of the latter). The binding free energy should be dependent on the phosphate oxygen-amino nitrogen distance which reflects the accessibility of the positive charge on the derivatized amino group. This accounts for the decrease in IATml with increasing bulkiness and hydrophobicity of the positive charged lysine derivatives: amidinated > dimethylated > isopropylated (Tables I and II). This variation may also reflect a decreasing ability to form hydrogen bonds with increasing size of the substituents [271. Likewise, the bulky phenylglyoxal groups substituted on arginines might make the distances too large for optimal interaction of these residues with single-stranded DNA. Alternatively, large substituents could make it difficult for the nucleic acid backbone to interact with unsubstituted residues. Of the modifications used in this study, this might most probably occur with phenylglyoxalation, where two bulky substituent groups bind per amino acid (Table I; [23,28]). The pH dependence of poly [d(A-T) Tm depression by the phenylglyoxalated arginine derivatives probably reflects a requirement for a positive charge on the modified amino acids, and is in line with the observations on lysine substitution. Additional evidence that lysines are involved in the interaction of RNAase A with single-stranded DNA has been provided by protection experiments with [all]formaldehyde. Excess denatured calf thymus DNA significantly reduces the level of reductive methylation (Karpel and Merkler, unpublished data). Double label experiments performed under DNAinteractive and non-interactive (high [Na+])conditions also indicate a high level of protection.
Although the in vitro DNA Tin-depressing activity of RNAase A may not directly reflect a physiological role for this protein, this property is very likely the consequence of the specific placement of positively charged residues on the surface of the protein, and not simply the result of a nonspecific basic proteinacidic substrate interaction. The very different property of S-protein, which increases Tin, indicates a differential affinity for the higher charge density of the DNA double helix than for that of the singlestranded form. On the basis of circular dichroic [44] and other physical properties [19], the higher order structure of S-protein appears to be significantly diferent from that of RNAase A or RNAase S. S-protein undergoes a structural transition at 37°C [45], so it is likely that its conformation at the temperature at which the poly[d(A-T)] melts under the conditions utilized (over 37°C) is different than at lower temperatures. But since both RNAase A and RNAase S bring about poly[d(A-T)] melting at temperatures (under 20°C) well below the transition point of S-protein, there must be higher order structure and/or essential amino acids present in the complete protein and not in S-protein which are necessary for the specific interaction with single-stranded DNA. These results parallel those obtained with heat-denatured RNAase A, which raises the Tm of DNA [14]. The affinity of RNAase A for single-stranded DNA may be related to the temporal association of the enzyme with RNA substrates. We have observed that ribose-phosphate backbone devoid of heterocyclic bases reversibly inhibits the ability of RNAase A to depress poly[d(A-T)] Tm to a degree comparable to that observed with poly(dT) as an inhibitor (Karpel et al., unpublished data). This result also indicates that the heterocyclic bases of the DNA may have little if any involvement in the interaction with the protein. The involvement of lysine and arginine residues in the recognition of single-stranded nucleic acids by helix-destabilizing proteins with defined physiological roles has not been extensively investigated [2]. As more structural information on these proteins becomes available, the procedures and approaches presented in this study on a model DNA melting protein should prove useful for application to systems of established biological significance.
51
Acknowledgement This work was supported by research grant CA 21374 from the National Cancer Institute, National Institutes of Health, U.S. Public Health Service. We would like to thank Dr. Irwin Chaiken for useful discussions concerning this work.
References 1 Champoux, J.J. (1978) Annu. Rev. Biochem. 4 7 , 4 4 9 479 2 Coleman, J.E. and Oakley, J.L. (1980) CRC Crit. Rev. Biochem. 7,247-289 3 Alberts, B.M. and Sternglanz, R. (1977) Nature 269, 655-661 4 Alberts, B.M. and Frey, L. (1970) Nature 227, 13131318 5 Herrick, G. and Alberts, B. (1976) J. Biol. Chem. 251, 2124-2132 6 Karpel, R.L. and Burchard, A.C. (1980) Biochemistry 19, 4674-4682 7 Epstein, R.H., Bolle, A., Steinberg, C.M., Kellenberger, E., Boy de la Tour, E., Chevalley, R., Edgar, R.S., Susman, M., Denhardt, G.H. and Lielausis, A. (1963) Cold Spring Harbor Symp. Quant. Biol. 28,375-392 8 Tomizawa, J., Amaku, N. and Iwama, Y. (1966) J. Mol. Biol. 21,247-253 9 Meyer, R.R., Glassberg, J. and Kornberg, A. (1979) Ptoc. Natl. Acad. Sci. USA 76, 1702-1705 10 Record, M.T., Jr., Lohman, T.M. and de Haseth, P. (1976) J. Mol. Biol. 107,145-158 11 Felsenfeld, G., Sandeen, G. and yon Hippel, P.H. (1963) Proc. Natl. Acad. Sci. USA 50, 644-651 12 Kosaganov, Yu.N., Lazurkin, Yu.S. and Sidorenko, N.V. (1967) Mol. Biol. 1,301-309 13 Lacombe, C., Cattan, D. and Laigle, A. (1974) Bioehimie 56,649-657 14 Jensen, D.E. and yon Hippel, P.H. (1976) J. Biol~Chem. 251, 7198-7214 15 MacPherson, A., Jurnak, F.A., Wang, A.H.J., Molineux, I. and Rich, A. (1979) J. Mol. Biol. 134, 379-400 16 Anderson, R.A., Nakashima, Y. and Coleman, J.E. (1975) Biochemistry 14,909-917 17 Kartha, G., Bello, J. and Harker, D. (1967) Nature 213, 862-865 18 Wyekoff, H.W., Tsernoglou, D., Hanson, A.W., Knox, J.R., Lee, B. and Richards, F.M. (1970) J. Biol. Chem. 245,305-328 19 Riehards, F.M. and Wyckoff, H.W. (1971) in The Enzymes (Boyer, P.D., ed.), 3rd edn., Vol. 4, pp. 647896, Academic Press, New York
20 Means, G.E. and Feeney, R.E. (1971) Chemical Modification of Proteins, Holden-Day, San Francisco 21 Glazer, A.N., Delange, R.J. and Sigman, D.S. (1975) Chemical Modification of Proteins: Selected Methods and Analytical Procedures, Laboratory Techniques in Biochemistry and Molecular Biology (Work, T.S. and Work, E., eds.), Vol. 4, Part I, North-Holland/American Elsevier, New York 22 Glazer, A.N. (1976) in The Proteins (Neurath, H. and Hill, R.L., eds.), 3rd edn., Vol. 2, pp. 1-103, Academic Press, New York 23 Riordan, J.F. (1979) Mol. Cell. Biochem. 26, 71-92 24 Walter, B. and Wold, F. (1976) Biochemistry 15, 304310 25 Reynolds, J.H. (1968) Biochemistry 7, 3131-3135 26 Means, G.E. and Feeney, R.E. (1968) Biochemistry 7, 2192-2201 27 Means, G.E. (1977) Methods Enzymol. 47E, 469-478 28 Takahashi, K. (1968) J. Biol. Chem. 243, 6171-6179 29 Butler, P.J.G. and Hartley, B.S. (1972) Methods Enzyme1.25B, 191-199 30 Plapp, B.B., Moore, S. and Stein, W.H. (1971) J. Biol. Chem. 246,939-945 31 Matheson, T.T. and Scheraga, H.A. (1979) Biochemistry 18, 2347-2445 32 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 33 Sherwood, L.M. and Potts, T.T., Jr. (1965) J. Biol. Chem. 240, 3799-3805 34 Klee, W.A. (1968) Biochemistry 7, 2731-2736 35 Inman, R.B. and Baldwin, R.L. (1962) J. Mol. Biol. 5, 172-184 36 Butler, P.J.G., Harris, J.I., Hartley, B.S. and Leberman, R. (1969) Biochem. J. 112,679-689 37 Freedman, M.H., Grossberg, A.L. and Pressman, D. (1968) Biochemistry 7, 1941-1950 38 Wofsy, L. and Singer, S.J. (1963) Biochemistry 2, 104116 39 Habeeb, A.F.S.M. (1967) Arch. Biochem. Biophys. 119, 264-268 40 Karpel, R.L., Bertelsen, A.H. and Fresco, J.R. (1980) Biochemistry 19, 504-512 41 Fretheim, K., Iwai, S. and Feeney, R.E. (1979) Int. J. Peptide Protein Res. 14, 451-456 42 Riordan, J.F. (1973) Biochemistry 12, 3915-3923 43 Sekine, H., Nakano, E. and Sakaguchi, K. (1969) Bioehim. Biophys. Aeta 174,202-210 44 Pflumm, M.N. and Beychok, S. (1969) J. Biol. Chem. 244, 3973-3981 45 Tsong, Y.T., Hearn, R.P., Wrathall, D.P. and Sturtevant, J.M. (1969) Biochemistry 9, 2666-2677
Bioehimica et Biophysica Acta, 654 (1981) 42 51
Elsevier/North-HollandBiomedicalPres~ BBA 99864 INVOLVEMENT OF BASIC AMINO ACIDS IN THE ACTIVITY OF A NUCLEIC ACID HELIX-DESTABILIZING PROTEIN RICHARD L. KARPEL,DAVID J. MERKLER *, BRIAN K. FLOWERSand MARTHAD. DELAHUNTY Department of Chemistry, University o f Maryland Baltimore County, Catonsville, MD 21228 (U.S.A.)
(Received September 3rd, 1980) (Revised manuscript received February 2nd, 1981)
Key words: Helix-destabilizing protein; Chemical modification; RNAase A; DNA melting
Under conditions of low ionic strength, ribonuclease A, which binds more tightly to single- than to doublestranded DNA, lowers the melting temperature of DNA helices (Jensen and yon Hippel (1976) J. Biol. Chem. 251, 7198-7214). The effects of chemical modification of lysine and arginine residues on the helix-destabilizing properties of this protein have been examined. Removal of the positive charge on the lysine e-amino group, either by maleylation or acetylation, destroys the ability of RNAase A to lower the Tm of poly[d(A-T)]. However, reductive alkylation of these residues, which has no effect on charge, yields derivatives which lower the Tm by only about one-half that seen with unmodified controls. Phenylglyoxalation of arginines can largely remove the Tin-depressing activity of RNAase A. RNAase S, which is produced by cleavage of RNAase A between amino acids 20 and 21, possesses DNA helix-destabilizing activity comparable to that of the parent protein, whereas S-protein (residues 21-124) increases poly[d(A-T)] Tm and S-peptide ( 1 - 2 0 ) has no effect on Tin. These results suggest that the specific location of several basic amino acids situated on the surface of RNAase A is largely responsible for this protein's DNA melting activity.
Introduction A number of different classes of protein have been described which affect the conformation of nucleic acids [1,2]. Those proteins, which lower the melting temperatures of nucleic acids by selectively binding to single strands, are termed helix-destabilizing proteins [3]. They have been found in organisms ranging from bacteriophages (e.g., gene 32 protein from T4 phage [4] ) to higher species (e.g., UP1 from calf thymus [5,6]). Although the physiological roles of most helix-destabilizing proteins have not been determined, the T4 protein [7,8] as well as one isolated from Escherichia coli [9] have been shown by genetic
* Present address: Department of Microbiology,Cell Biology, Biochemistryand Biophysics,PennsylvaniaState University, University Park, PA 16802, U.S.A.
methods to be required for DNA replication and recombination. Although little is presently known about the structural basis for the binding specificity of helix-destabilizing proteins, their selective affinity for single strands presumably is the result of specific chemical interactions between particular amino acid residues and functional groups in the nucleic acid. The strong salt dependence of the binding between protein and nucleic acid is indicative of electrostatic interactions between basic amino acids and nucleic acid phosphates. Record et al. [I0] have shown that the slope of the linear dependence of binding constant on log [monovalent cation] can be interpreted in terms of the number of ion pairs, m', formed between each protein molecule and the length of polynucleotide chain to which it is bound. This approach has been applied to several helix-destabilizing proteins, including pancreatic ribonuclease A, which lowers the
0 005-2787/81/0000-0000/$02.50 © Elsevier/North-HollandBiomedicalPress
43 melting temperatures of DNA double helices [ 1 1 14]. In the case of RNAase A, m' = 7 for the interaction with denatured DNA and 4 for the (weaker) interaction with native DNA [10,14]. Basic amino acids are probably not the only residues involved (several tyrosines in the fd phage gene 5 protein appear to be involved in its binding to single-stranded DNA [2,15,16]); however, electrostatic forces clearly are an important component o f the affinity of helixdestabilizing proteins for single-stranded nucleic acids. Although this Tin-depressing activity of RNAase A may have no direct physiological relevance, the extensive body of knowledge of its structural and chemical properties [17-19] suggests that it is an excellent model for the study of protein-induced nucleic acid helix destabilization. In order to explore the involvement of basic amino acids in this process, we have studied the effects of specific chemical modification of lysine and arginine residues on the DNA Tmdepressing activity of ribonuclease A. RNAase A is especially well suited for this type of study, since a number of modification procedures have been carried out on this protein, and the resultant derivatives have been well characterized [20-23]. These include acetylation [24], amidination [25] and reductive alkylation [20,26,27], which have been shown to react specifically with RNAase lysines. Under the proper conditions [28], phenylglyoxalation of RNAase A leads to specific derivatization of arginines, with concomitant deamination of N-terminal a-amino group. We have also utilized maleylation, which in general is specific for lysines [29]. We also report the effects on DNA Tm of the cleavage product, RNAase S, and its two components, S-protein and S-peptide. The results obtained with these products indicate that the native tertiary structure of the complete protein is a necessary requirement for helix-destabilizing activity, and confirm the view that the interaction with single-stranded DNA involves binding of nucleic acid phosphates to a set of basic amino acids in a specific three-dimensional orientation. Materials and Methods
Reagents and buffers. Ethyl acetimidate hydrochloride and 2,3-butane-
dione were purchased from Aldrich Chemical Co.; 2,4,6-trinitrobenzenesulfonic acid was purchased from Sigma Chemical Co. and recrystallized from 5 M HC1 [30]; N-ethylmorpholine was obtained from Pierce Chemical Co. and phenylglyoxal from Sigma. All other chemicals were reagent or comparable grade. The pH of protein solutions was measured with nonbleeding pH strips (Merck No. 9583, graduated to 0.2 pH units). The buffering systems used throughout were: buffer A: 1.2 mM phosphate (K*)/8.8 mM TrisHC1/0.1 mM Na2EDTA, pH 8.0 +_0.1; buffer B; 10 mM phosphate (K*)/0.1 mM Na2EDTA/pH 7.0 +- 0.1; buffer C: 10 mM Tris-HC1/0.1 mM Na2EDTA, pH 8.0 +-0.1. Unless otherwise noted, the modified proteins and the controls were exhaustively dialyzed against buffer C.
Proteins Lyophilized, phosphate-free pancreatic ribonuclease A was purchased from Worthington Biochemical. Several of the lysine-modified derivatives (those subjected to maleylation, amidination and reductive methylation) were prepared from RNAase A which had been heated in buffer C to 60°C for 5 min, a procedure which breaks up aggregates [ 19,31 ]. However, no differences were observed between the Tin-depressing activity of heat treated RNAase A and that of material which had not been subjected to this treatment, suggesting either that the extent of aggregation was minimal, or that it has no effect on the Tmdepressing properties of the protein. Preliminary experiments with a less highly purified RNAase (Sigma type I-A) gave results very similar to that of the Worthington protein. RNAase A concentrations were measured spectrophotometrically at 277.5 nm, using an extinction coefficient of 9.56.103 M -~. cm -~ [14]. Concentrations of derivatized and control samples were determined by the Lowry method [32] with underivatized RNAase A as the standard. RNAase S and S-peptide were obtained from Sigma; S-protein was the gift of Dr. I. Chaiken. Concentrations of RNAase S were determined optically, using the same era as that of RNAase A [33]. The molar extinction coefficients used for S-protein and S-peptide were, respectively, 9.20.103 at 280 nm [33] and 2.9,102 at 258 nm [34]. Nucleic acids Poly[d(A.T)] was obtained from P-L Biochemicals,
44
Modification procedures
and dissolved in buffer B. Its c o n c e n t r a t i o n was determ i n e d s p e c t r o p h o t o m e t r i c a l l y in buffer B, using e262 = 6.65 • 103 M -1 • cm -1 [35]. Clostridium perfringens D N A was o b t a i n e d f r o m Worthington, and treated in the same way as the p o l y [d(A-T)].
The specific m o d i f i c a t i o n reactions study are summarized in Table I. F o r tions, control samples were subjected to procedures in the absence o f a reagent
TABLE I SPECIFIC MODIFICATION REACTIONS 1. Lysine a . + .--~ - -
Maleylation:
+0 / P-NH 2 \
0 II C-C-H
O pH 8.0 II --* R-NHCCH=CHCOO- + H +
C-C-H II O b.+~O O CCH3
Acetylation:
0
II , R-NHCCH3 + CH3COO- + I-I÷
P - N H 2 + O/\ CCH3 II O
C. + - +
+
Reductive Alkylation:
Oil pH 9 / R NaBH 4 /R P - N H 2 + RCR' ~ P-N=C\ , P-NHCH , R'
° 1NaBH4
II RCR' R
P-NH+(C H/ \
R'
Amidination:
pH<9
)2
R
/
P-N(CH
)2 R'
NH 2 +NH2 II pH>8.s II P-NH2 + CH3COCH2CH3 ~ P-NHCCH3 + CHaCH2OH
2. Arginine Phenylglyoxalation:
+
NH2
P-NH--C(N.
©
O=C
O=C I H
÷
©
0
N-C-OH
N-C-O
N CO.
N-C
I H
OII
I H
(one of the proposed structures, refs. 23 and 28)
used in this all modificathe identical essential for
45 derivatization. In this manner, it could be confirmed that all effects on the DNA Tin-depressing activity of RNAase A were due to specific amino acid modification, rather than to any other changes occurring during the derivatization procedures.
Lysine modification Maleylation [29,36]. 0.08 ml of 0.3 M maleic anhydride (twice recrystallized from chloroform [21]), in dioxane was added to 1.12 ml of 5.4 mg/ml RNAase A in 0.20 M phosphate buffer, pH 9.0. The reaction was allowed to continue for 2 h at 4°C, maintaining the pH at 8.0. A parallel control was run by adding 0.08 ml dioxane without maleic anhydride. Modification was also done using one-half of the above amounts. In order to eliminate any possibility of formation of o-maleyl adducts, maleylated RNAase samples were treated with 0.8 M hydroxylamine (pH 9.5) for 15 min at room temperature, then overnight at 4°C [37]. This treatment had no effect on the derivative's DNA Tin-depressing activity. Acetylation [21]. A 25 mg/ml solution of RNAase A was mixed with an equal volume of a solution of saturated sodium acetate, chilled in an ice bath and subjected over a period of 1 h to four additions of acetic anhydride, with the total mass of the latter equal to that of the protein. Controls were treated in the same manner with water substituted for acetic anhydride. The protein was dialyzed against buffer C, and the procedure was repeated. Although this method is quite selective for primary amino groups [21], the derivative was then treated with 1 M hydroxylamine/1 mM Na2EDTA, pH 8.0, for 20 min at 27°C to hydrolyze any o-acetyltyrosine which may have formed [24]. Reductive methylation [20,26]. A 1.0-2.5 mg/ml RNAase A solution in 0.2 M borate buffer, pH 9.0 at 0°C, was adjusted to 0.5 mg/ml NaBH4 by the addition of solid NaBH4. To the mixture, 0.5/.tl/rnl increments of formalin (37% aqueous formaldehyde solution) were added at 5-min intervals until a final concentration of 2.5 /~l/ml was reached. Controls were treated similarly, with water being substituted for the formalin.
Reductive isopropylation [20,26]. To 0.55 ml of 1.0 mg/ml RNAase A in 0.2 M borate buffer, pH 9.0, containing 1 I% (v/v) acetone at 0°C was added a 0.03
ml aliquot of 10 mg/ml NaBH4 every 5 min until a final concentration of 2.5 mg/ml was reached. Controis consisted of protein in borate buffer treated only with NaBH4 or protein in borate buffer treated only with acetone. Some samples and controls were subjected to three rounds of this procedure (see Results). Amidination [38]. 0.45 ml of 1.0 M NaOH was added to 0.111 g of ethyl acetimidate hydrochloride, then mixed with 0.75 ml of 10 mg/ml RNAase A in buffer C. The pH of the reaction mixture was adjusted and maintained at 8.3-8.6 for 2 h at 0°C with stirring. Controls were prepared in the same manner, but in the absence of ethyl acetimidate. Modification was also done using one-third of the above amounts.
Quantitation of free amino groups The extent of lysine modification was determined by quantitation of remaining free amino groups via reaction of modified derivatives with 2,4,6-trinitrobenzenesulfonic acid, as modified by Habeeb [39]. Parallel assays were run on unmodified RNAase A, and the extent of modification computed by comparison of the A a3s of modified with that of unmodified protein. The absorbance obtained with unmodified RNA. ase A was consistent with its content of 11 free amino groups and the range of extinction coefficients reported in the literature [20,21]. On the basis of results with reductively alkylated RNAase A derivatives, it appears that the trinitrobenzenesulfonic acid assay indicates a slightly higher number of free amino groups than that Obtained from amino acid analysis [26].
Arginine modification Phenylglyoxalation [28]. An aliquot of 15 mg]ml phenylglyoxal in 0.2 M N-ethylmorpholine acetate buffer, pH 8.0, was added to an equal volume of 10 mg/ml RNAase A in the same buffer. The mixture was allowed to incubate for 1 h at room temperature. Controls were prepared in the same manner in the absence of phenylglyoxal. Modified protein and the control were exhaustively dialyzed against 10 mM phosphate (K ~) buffer, pH 3.9, containing 0.1 mM Na2EDTA.
A bsorbance-temperatureprofiles Teflon-stoppered micro quartz cuvettes containing
46
100 #1 of test solutions were placed in a Gilford 2400-2 spectrophotometer designed to raise the temperature from 0 to 95°C at a constant rate, which was usually 25 K/h in these experiments. Generally, three nucleic acid/protein samples were melted simultaneously with a semi-micro reference cuvette containing buffer C. The concentration of nucleic acid was usually 5.95 • 10 -s M(p), with [protein] as indicated. Temperature was continually monitored by means of a calibrated thermistor (Yellow Springs Instruments) inserted through a narrow hole in the stopper of the reference cuvette; absorbance was measured at 260 nm. Tm values are defined as the inflection points of the transitions as determined by differentiating the measured absorbance-temperature profiles [40]. The reproducibility of the Trn values was -+1°C, except when profiles were very broad, where it was
1Z100 ZO
8 4
/
0
E
3
2.5
/
Pr0,e,n
S-Peptide o °
V"
~7
-12 -16 -20
281~- -
In order to assess conditions optimal for testing the effects of chemical modifications on the poly[d(A-T)] Tin-depressing activity of RNAase A, a range of protein concentrations (with constant [DNA]) in buffer A was explored. The results are shown in Fig. 1. Under the conditions of these experiments, the magnitude of the Tm depression (--ATm) sigmoidally increases with increasing [RNAase A ] : [poly[d(A-T)]]p. Beyond a ratio of 0.2, only a slightly larger IATml value is obtained, with precipitation apparent at [RNAase A] : [poly[d(A.T)]]p > 0.3. The poly[d(A-T)] Tin-depressing activity of derivatives was thus generally measured at [protein[: [DNA] p > 0.2. Although inorganic phosphate is an inhibitor of
4
-8
Circular dichroism experiments
Variation of poly[ d(A-T)] Tm with [ribonuclease A/
5
-4
-24
Results
7
D
±2°C.
Measurements of circular dichroism (CD) were taken with a JASCO J-40 spectropolarimeter. The cuvette holder of the instrument was adapted to hold rectangular semi-micro quartz cuvettes containing as little as 400 ;zl of sample, and was connected to an external circulating water bath. Cuvette temperature was determined via a calibrated thermistor probe (Yellow Springs Instruments).
10
RNa
_''--_ ~
1
i
al
0.~
a'3
-
0.4
[P.o.E,N Fig. 1. Dependence of poly[d(A-T)] Tm in buffer A on [RNAaseA], o o; [RNAase S], A__ _zx; S-protein, a-~; and S-peptide, z~ ~. [poly[d(A-T)]] = 5.95. 10-s M(p). In the absence of added protein, Tm = 38 -+ I°C. RNAase enzymatic activity [19], at the low levels of phosphate used in these experiments, no significant effect on Tm was observed. In 10 mM phosphate (pH 7.5)/0.1 mM Na2EDTA, we observed that at a [RNAase A] : [poly[d(A-T)]]p of 0.23, ATm = - 2 2 K, essentially identical to that obtained in buffer A. Unless otherwise stated, the Tm depression experiments were performed in buffer A, which contained 1.2 mM phosphate.
Effects of lysine modifications on RNAase A melting activity We have chosen several procedures which differ in their effect on the charge of the e-NH~ group: maleylation replaces the positive charge with a negative one, acetylation removes the charge, and reductive alkylation and amidination have no effect on the charge (Table I). The Tin-depressing effects of these derivatives, as well as those of arginine, are tabulated in Table II.
47 TABLE II SUMMARY OF Tm-DEPRESSING ACTIVITY OF DERIVATIVES OF RNAases Unless otherwise stated, experiments were performed in Buffer A, 1.2 mM phosphate (K*)/8.8 mM Tris-HCl/0.1 mM Na2EDTA, pH 8.0 Modification
[protein]
[poly[d(A-T)]lp
A. Formation of lysine derivatives Maleylation 0.21 Acetylation 0.23 Reductive methylation 0.23 0.33 Reductive isopropylation 0.23 0.45 Amidination 0.23 B. Formation of arginine derivatives Phenylglyoxalation 0.23
ATm (K)
,xTm, control (K)
0 0 -13.5 -12.5 -17 to -8 a -13 b - 19
- 18 -21 -23 -24 -22 ± 2
pH 7.1:-10 c pH7.5: - 5 d pH 8.3: 0e
_ c -23d -22 e
-22
a Derivatives varied between 32 and 70% modification (see text). b Derivative was 54% modified, and gave the same ATm at [protein] : [poly[d(A-T)] ]p = 0.23. e In 5.5 mM Tris-HCl]4.5 mM phosphate (K*)/0.1 mM Na2EDTA. Control samples precipitated. d In 7.1 mM Tris-HC1/2.9 mM phosphate/0.1 mM Na2EDTA. e Same as footnote d, but adjusted to pH 8.3.
Maleylation Reaction of maleic anhydride with ribonuclease A produced a derivative in which 75% of the primary amino groups were no longer sensitive to trinitrobenzenesulfonic acid, and thus were modified. This derivative possessed no helix-denaturing activity (Table II). The control RNAase depressed Tm by 18°C, nearly that of unaltered RNAase (Fig. I). Although maleylated derivatives slowly revert to primary amino groups at acid pH [36], this clearly was of no consequence, since no effect on Tm was observed. Substantially the same result was obtained with CI. perfringens DNA, where the Tm of 53°C (in buffer A) was depressed by 25°C in the presence of R N A a s e A ([protein]: [nucleic acid] p = 0.23) but was unaffected by maleylated RNAase. In all probability, the change of eight or nine positive to negative charges brings about a gross perturbation in isoelectric point. In a preliminary experiment, we failed to observe electrofocusing of maleylated RNAase A on a pH 3 - 1 0 isoelectrofocusing gel, suggesting that the pl of the modified protein was too acidic to be resolved on this system. The polyanionic
modified RNAase thus cannot interact with the polyanionic DNA, and has no effect on Tin. No gross alteration of the RNAase tertiary structure occurred upon maleylatlon, since its far-utraviolet CD spectrum was essentially unaltered from that of unmodified protein (data not shown).
Acetylation Acetylation of RNAase A completely inactivates its enzymatic activity [24]. A derivative that was 72% modified had no effect on poly [d(A-T)] Tin, whereas the control RNAase depressed T m comparably to untreated protein (Table II). Removal of the lysine positive charge thus completely destroys the interaction of the protein with single-stranded DNA.
R eductive alkylation Reductive methylation of RNAase A with formaldehyde and NaBH4, leading to dimethylated lysine derivative, destroys catalytic activity with little change in physical properties and only slight loss in 3'-cytidylic acid binding [20,26]. A similar result, with somewhat greater loss of CMP binding, was ob-
48 tained upon modification with acetone and NaBH4 (leading to monoisopropylated derivatives [26,41]). Both methylated and isopropylated RNAase A were tested for poly[d(A-T)] Tin-depressing activity. These derivatives showed no evidence on SDS-polyacrylamide gels for any protein-protein cross-linking resulting from the modification procedure. A methylated derivative with 92% of the amino groups modified depressed Tm by 13.5°C at the standard [protein] : [DNA]p of 0.23 (Table II). When the [methylated RNAase] : [DNA]p was increased to 0.33, the Tm depression was virtually unchanged (-12.5°C), indicating that T m depression by this derivative was maximal, and therefore less than that achieved by unmodified RNAase A. Several other methylated samples and controls were tested for poly[d(A-Y)] Tm-depressing activity, and all gave very similar ATm values (+1 K). Several isopropylated RNAase A derivatives were tested for poly[d(A-T)] Tm-depressing activity. Between 38% and 75% of primary amino groups were reacted, with the greatest degree of reaction resulting from three rounds of modification. ATm varied between - 1 7 K and - 8 K (Table II), with the greatest effect occurring with two derivatives that were both 75% modified. A 54% modified isopropylated derivative yielded a ATm of - 1 3 K over 2-fold range of [protein] : [DNA]p, so that, as with the methylated derivative, Tm depression was maximal. Isopropylated RNAase A/poly[d(A-T)] mixtures could be melted, cooled and re-melted to yield melting temperatures within 1°C of the initial values. The Tin-depressing activity of RNAase A is reduced to a larger degree by extensive isopropylation than by methylation, and probably reflects the greater bulkiness of isopropylated relative to dimethylated eiamino groups (see Discussion). Although the transition breadths (which we define as the widths at half-height of &A/AT vs. T p l o t s ) o f RNAase A - and alkylated RNAase A - poly [d(AT)] mixtures were broader ( 4 - 7 K) than that of poly[d(A-T)] alone (2.5 K), there was virtually no overlap among poly[d(A-T)], methylated (or isopropylated) RNAase-poly[d(A-T)], and unmodified RNAasepoly[d(A-T)] transitions (Fig. 2). This indicates that the population of atkylated protein molecules is reasonably homogeneous. Indeed, unmodified RNAase A effected somewhat broader transitions ( 6 - 7 K)
than alkylated protein ( 4 - 5 K). The substantial poly[d(A-T)] Tin-depressing activity of alkylated RNAases indicates that these derivatives retain their binding to single-stranded DNA. The effect of isopropylated and methylated RNAase A on the conformation of single-stranded DNA was assessed by monitoring the influence of these derivatives on the CD spectrum of poly [d(A-T)] at a temperature at which the nucleic acid is fully melted, under standard conditions ([RNAase A] : [poly[d(AY)]]p = 0.23). The maximum, minimum, and crossover point wavelengths of poly[d(A-Y)]/alkylated RNAase A mixtures measured at a temperature at which the nucleic acid is totally denatured (34°C)
- -
I
!
I
~
[
1
/ C
E e ~q
~q
10 f4
f8
2'6 0
3'4 3'8 4b
°C Fig. 2. Differentiated melting profiles of RNAaseA/poly[d(A-T)] mixtures. Poly[d(A-T)] in the absence of added protein, o m; poly[d(A-T)] + methylated RNAaseA, [protein] : [DNA]p = 0.3, o o; poly[d(A-T)] + control RNAase A (subjected to the reductive alkylation procedure in the absence of formaldehyde), [protein[ : [DNAlp = 0.3, A ~. lsopropylated RNAaseA/poly[d(A-T)[ mixtures showed profiles similar to that of the methylated derivative+ DNA mixture.
49 were virtually identical (within 2 nm) to that obtained from the sum of individual derivatized protein and heat-denatured poly[d(A-T)] CD spectra, although the ellipticity maxima and minima were slightly different (not shown). These results are similar to those of poly[d(A-T)] and unmodified RNAase A [ 14], and suggest that little distortion of the nucleic acid chain occurs when it interacts with the bulky isopropylated or dimethylated e-amino groups of the modified derivatives.
A midination RNAase A loses its enzymatic activity upon amidination [25]. Upon reaction with ethyl acetimidate, 82% of the amino groups were found to be modified. This derivative had only a slight effect on the protein's poly [d(A-T)] Tin-depressing activity, yielding a ATm o f - 1 9 K, compared to a ATm of - 2 2 K observed with the control sample (Table II). Effects of arginine modification on RNAase A melting activity Reaction of RNAase A with phenylglyoxal under the conditions described in the Materials and Methods results in the modification of three of the four arginine residues of this protein, deamination of the N-terminal lysine, and the loss of 89% of enzymatic activity [28]. This derivative displayed a pH-dependent poly [d(A-T)] Tm-depressing activity, which was always smaller than that of unmodified RNAase A (Table II). At pH 8.3, AT m was only -0.5 K (~0), whereas it was -5 K at pH 7.5. At still lower pH, 7.1, the melting profile was extremely broad, with ATm = -10 + 2 K. Phenylglyoxal derivatives are known to be labile at pH > 4 [26] ; however, remelting of phenylglyoxalated RNAase/poly[d(A-T)] mixtures at the highest and lowest pH conditions yielded essentially the same (-+I°C) Tin, indicating no change in the degree of modification during melting. These results suggest that the bully phenylglyoxal groups (Table I) effectively block close contact of lhe single-stranded DNA phosphates with the positive charge of derivatized arginines and/or unmodified lysines. Although the pKA of this derivative is not known, the pH dependence of the data indicates that the derivative is likely unprotonated at pH 8.3, and protonated at pH 7.1. Several other modification procedures for arginine
have been developed [23]. Preliminary experiments with a butanedione derivative of RNAase A, prepared by the arginine-specific method of Riordan [42], showed no effect on poly[d(A-T)] Tm depression.
RNAase S, S-protein, and S-peptide Cleavage of RNAase A by subtilisin between residues 20 and 21 produces a derivative, RNAase S, which has tertiary structure and catalytic activity similar to that of the uncleared protein [17-19]. We have observed that RNAase S in buffer A possesses slightly greater poly[d(A-T)] Tin-depressing activity than RNAase A (Fig. 1). However, S-protein, residues 21-124 derived from RNAase S, acts as a helix stabilizer, raising the Tin by 10.5°C at a [protein] : [poly[d(A-T)]p of 0.10 in buffer A (Fig. 1). Precipitation was observed at higher [S-protein] : [DNA]p. S-peptide, residues 1-20, had no effect on poly[d(AT)] Tm. The Tm of RNAase S- and S-protein/poly[d(A.T)] mixtures was reproduced upon melting, cooling and re-melting. When S-protein and S-peptide were mixed in equimolar amounts at 4°C to reconstitute enzymatically active material [25], the resultant protein (RNAase S') was found to depress poly[d(A-T)] Tm as well as (within 2°C) RNAase S. Thus, although S-protein contains eight of the ten lysine residues and three of the four arginines of the parent protein, it is clear that the tertiary structure of the complete protein, as well as, possibly, the missing basic amino acids are required for DNA melting activity (see Discussion). Discussion
The results presented herein clearly indicate that basic amino acid residues are necessary components of the interaction of the model helix-destabilizing protein, RNAse A, with single-stranded DNA. The removal of the positive charge on lysine e-amino groups destroys the DNA Tin-depressing activity of this protein, as shown by the inability of maleylated and acetylated RNAase A to affect poly[d(A-T)] melting. However, when methyl, isopropyl or acetimido groups are substituted on lysine, retaining the positive charge under the near neutral pH conditions utilized, substantial melting activity is still observed. Since these derivatives are known to be enzymatically inactive, we conclude that substituents on lysine
50 e-amino groups more critically affect RNAase A catalytic activity than its DNA helix-destabilizing activity. At least some of the residues that bind to DNA are involved in the enzymatic activity of RNAase A, since single-stranded DNA is an effective competitive inhibitor of the degradative action on RNA substrates [431. It seems likely that the role of the lysines is to provide neutralizing positive centers "for the negative charges of the single-stranded DNA chain. Substitution of alkyl or acetimido groups apparently permits sufficient electrostatic interaction to occur, suggesting that there may be some flexibility in the orientation of the basic side-chains and/or nucleic acid backbone (the CD results indicate no significant distortion of the latter). The binding free energy should be dependent on the phosphate oxygen-amino nitrogen distance which reflects the accessibility of the positive charge on the derivatized amino group. This accounts for the decrease in IATml with increasing bulkiness and hydrophobicity of the positive charged lysine derivatives: amidinated > dimethylated > isopropylated (Tables I and II). This variation may also reflect a decreasing ability to form hydrogen bonds with increasing size of the substituents [271. Likewise, the bulky phenylglyoxal groups substituted on arginines might make the distances too large for optimal interaction of these residues with single-stranded DNA. Alternatively, large substituents could make it difficult for the nucleic acid backbone to interact with unsubstituted residues. Of the modifications used in this study, this might most probably occur with phenylglyoxalation, where two bulky substituent groups bind per amino acid (Table I; [23,28]). The pH dependence of poly [d(A-T) Tm depression by the phenylglyoxalated arginine derivatives probably reflects a requirement for a positive charge on the modified amino acids, and is in line with the observations on lysine substitution. Additional evidence that lysines are involved in the interaction of RNAase A with single-stranded DNA has been provided by protection experiments with [all]formaldehyde. Excess denatured calf thymus DNA significantly reduces the level of reductive methylation (Karpel and Merkler, unpublished data). Double label experiments performed under DNAinteractive and non-interactive (high [Na+])conditions also indicate a high level of protection.
Although the in vitro DNA Tin-depressing activity of RNAase A may not directly reflect a physiological role for this protein, this property is very likely the consequence of the specific placement of positively charged residues on the surface of the protein, and not simply the result of a nonspecific basic proteinacidic substrate interaction. The very different property of S-protein, which increases Tin, indicates a differential affinity for the higher charge density of the DNA double helix than for that of the singlestranded form. On the basis of circular dichroic [44] and other physical properties [19], the higher order structure of S-protein appears to be significantly diferent from that of RNAase A or RNAase S. S-protein undergoes a structural transition at 37°C [45], so it is likely that its conformation at the temperature at which the poly[d(A-T)] melts under the conditions utilized (over 37°C) is different than at lower temperatures. But since both RNAase A and RNAase S bring about poly[d(A-T)] melting at temperatures (under 20°C) well below the transition point of S-protein, there must be higher order structure and/or essential amino acids present in the complete protein and not in S-protein which are necessary for the specific interaction with single-stranded DNA. These results parallel those obtained with heat-denatured RNAase A, which raises the Tm of DNA [14]. The affinity of RNAase A for single-stranded DNA may be related to the temporal association of the enzyme with RNA substrates. We have observed that ribose-phosphate backbone devoid of heterocyclic bases reversibly inhibits the ability of RNAase A to depress poly[d(A-T)] Tm to a degree comparable to that observed with poly(dT) as an inhibitor (Karpel et al., unpublished data). This result also indicates that the heterocyclic bases of the DNA may have little if any involvement in the interaction with the protein. The involvement of lysine and arginine residues in the recognition of single-stranded nucleic acids by helix-destabilizing proteins with defined physiological roles has not been extensively investigated [2]. As more structural information on these proteins becomes available, the procedures and approaches presented in this study on a model DNA melting protein should prove useful for application to systems of established biological significance.
51
Acknowledgement This work was supported by research grant CA 21374 from the National Cancer Institute, National Institutes of Health, U.S. Public Health Service. We would like to thank Dr. Irwin Chaiken for useful discussions concerning this work.
References 1 Champoux, J.J. (1978) Annu. Rev. Biochem. 4 7 , 4 4 9 479 2 Coleman, J.E. and Oakley, J.L. (1980) CRC Crit. Rev. Biochem. 7,247-289 3 Alberts, B.M. and Sternglanz, R. (1977) Nature 269, 655-661 4 Alberts, B.M. and Frey, L. (1970) Nature 227, 13131318 5 Herrick, G. and Alberts, B. (1976) J. Biol. Chem. 251, 2124-2132 6 Karpel, R.L. and Burchard, A.C. (1980) Biochemistry 19, 4674-4682 7 Epstein, R.H., Bolle, A., Steinberg, C.M., Kellenberger, E., Boy de la Tour, E., Chevalley, R., Edgar, R.S., Susman, M., Denhardt, G.H. and Lielausis, A. (1963) Cold Spring Harbor Symp. Quant. Biol. 28,375-392 8 Tomizawa, J., Amaku, N. and Iwama, Y. (1966) J. Mol. Biol. 21,247-253 9 Meyer, R.R., Glassberg, J. and Kornberg, A. (1979) Ptoc. Natl. Acad. Sci. USA 76, 1702-1705 10 Record, M.T., Jr., Lohman, T.M. and de Haseth, P. (1976) J. Mol. Biol. 107,145-158 11 Felsenfeld, G., Sandeen, G. and yon Hippel, P.H. (1963) Proc. Natl. Acad. Sci. USA 50, 644-651 12 Kosaganov, Yu.N., Lazurkin, Yu.S. and Sidorenko, N.V. (1967) Mol. Biol. 1,301-309 13 Lacombe, C., Cattan, D. and Laigle, A. (1974) Bioehimie 56,649-657 14 Jensen, D.E. and yon Hippel, P.H. (1976) J. Biol~Chem. 251, 7198-7214 15 MacPherson, A., Jurnak, F.A., Wang, A.H.J., Molineux, I. and Rich, A. (1979) J. Mol. Biol. 134, 379-400 16 Anderson, R.A., Nakashima, Y. and Coleman, J.E. (1975) Biochemistry 14,909-917 17 Kartha, G., Bello, J. and Harker, D. (1967) Nature 213, 862-865 18 Wyekoff, H.W., Tsernoglou, D., Hanson, A.W., Knox, J.R., Lee, B. and Richards, F.M. (1970) J. Biol. Chem. 245,305-328 19 Riehards, F.M. and Wyckoff, H.W. (1971) in The Enzymes (Boyer, P.D., ed.), 3rd edn., Vol. 4, pp. 647896, Academic Press, New York
20 Means, G.E. and Feeney, R.E. (1971) Chemical Modification of Proteins, Holden-Day, San Francisco 21 Glazer, A.N., Delange, R.J. and Sigman, D.S. (1975) Chemical Modification of Proteins: Selected Methods and Analytical Procedures, Laboratory Techniques in Biochemistry and Molecular Biology (Work, T.S. and Work, E., eds.), Vol. 4, Part I, North-Holland/American Elsevier, New York 22 Glazer, A.N. (1976) in The Proteins (Neurath, H. and Hill, R.L., eds.), 3rd edn., Vol. 2, pp. 1-103, Academic Press, New York 23 Riordan, J.F. (1979) Mol. Cell. Biochem. 26, 71-92 24 Walter, B. and Wold, F. (1976) Biochemistry 15, 304310 25 Reynolds, J.H. (1968) Biochemistry 7, 3131-3135 26 Means, G.E. and Feeney, R.E. (1968) Biochemistry 7, 2192-2201 27 Means, G.E. (1977) Methods Enzymol. 47E, 469-478 28 Takahashi, K. (1968) J. Biol. Chem. 243, 6171-6179 29 Butler, P.J.G. and Hartley, B.S. (1972) Methods Enzyme1.25B, 191-199 30 Plapp, B.B., Moore, S. and Stein, W.H. (1971) J. Biol. Chem. 246,939-945 31 Matheson, T.T. and Scheraga, H.A. (1979) Biochemistry 18, 2347-2445 32 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 33 Sherwood, L.M. and Potts, T.T., Jr. (1965) J. Biol. Chem. 240, 3799-3805 34 Klee, W.A. (1968) Biochemistry 7, 2731-2736 35 Inman, R.B. and Baldwin, R.L. (1962) J. Mol. Biol. 5, 172-184 36 Butler, P.J.G., Harris, J.I., Hartley, B.S. and Leberman, R. (1969) Biochem. J. 112,679-689 37 Freedman, M.H., Grossberg, A.L. and Pressman, D. (1968) Biochemistry 7, 1941-1950 38 Wofsy, L. and Singer, S.J. (1963) Biochemistry 2, 104116 39 Habeeb, A.F.S.M. (1967) Arch. Biochem. Biophys. 119, 264-268 40 Karpel, R.L., Bertelsen, A.H. and Fresco, J.R. (1980) Biochemistry 19, 504-512 41 Fretheim, K., Iwai, S. and Feeney, R.E. (1979) Int. J. Peptide Protein Res. 14, 451-456 42 Riordan, J.F. (1973) Biochemistry 12, 3915-3923 43 Sekine, H., Nakano, E. and Sakaguchi, K. (1969) Bioehim. Biophys. Aeta 174,202-210 44 Pflumm, M.N. and Beychok, S. (1969) J. Biol. Chem. 244, 3973-3981 45 Tsong, Y.T., Hearn, R.P., Wrathall, D.P. and Sturtevant, J.M. (1969) Biochemistry 9, 2666-2677