Eur. J. Biochem. 165,309-314 (1987) 0 FEBS 1987

Interactions of the high-mobility-group-like Ceratitis capitata C1 proteins with DNA Gabriel MARQUEZ, Ana T. RODRIGUEZ, Blanca A. FERNANDEZ and Francisco MONTERO Departamento de Bioquimica, Facultad de Ciencias Quimicas, Universidad Complutense, Madrid

(Received December 1, 1986/February 17, 1987) - EJB 86 1278

We have studied the interactions of the high-mobility-group-like proteins (Clal , Cla2 and Clb) from the fruit fly Ceratitis capitata with DNA. Nitrocellulose filter binding assays, thermal denaturation studies and spectrofluorimetry of the complexes revealed the existence of specific and nonspecific interactions. Thermal denaturation curves showed that the three proteins stabilized the DNA, thus suggesting a preferential binding to double-stranded DNA. The calculation of the thermodynamic parameters of the interactions showed that th'e nonspecific bindings were characterized by low association constants (K,) with values ranging from 2.7 x lo4 M-' to 2.0 x lo6 M-I. Also, the cooperativity of these interactions was relatively high (cooperativity factor, w, values ranging over 20 - 35), and the number of nucleotides involved was low (1 - 3 base pairs). On the other hand, the existence of specific interactions between C1 proteins and DNA was suggested by two facts: (a) the retention of C. capitata [3H]DNA in nitrocellulose filters was only a low percentage of total input DNA and (b) there was a marked size dependence of the binding (25% retention of a 40-kb DNA and only 3% retention with a DNA of 1 kb). The specific bindings had higher K, values than the nonspecific ones, and they also were cooperative. Some differences were observed between Clb and the Cla proteins about the way they interact with C. capitata DNA.

High mobility group (HMG) proteins are a class of non- histone chromosomal proteins present in chromatin at rela- tively high concentration [l, 21. They can be extracted from chromatin with 0.35 M NaCl[3] and their amino acid compo- sitions moderately high in both basic and acidic amino acids [3, 41. The HMG proteins have been reported to occur in all the eukaryotes examined to date [3, 5 - 71.

HMG proteins can be subdivided into two families based on their physical and chemical properties and the way they are released from chromatin after its digestion with nucleolytic enzymes. The first family (HMG1 and 2 in mammals, HMG1, 2 and E in chicken erythrocytes and HMG-T in trout testis) is composed of proteins with M , of about 26000. HMGl and 2 from mammals are able to fold in salt solutions [8 - 101 and they seem to be located in the linker DNA of nucleosomes [ll], although a number of different functions has been suggested for these HMG proteins [12- 151. The interaction of HMGl and 2 with DNA has been exhaustively studied [14 - 171 and the thermodynamic parameters of the associ- ation of HMGl to DNA have been determined [18]. The second family of HMG proteins is composed of low-M, polypeptides (8000-10000) such as HMG14 and 17 in mammals and H6 in trout testis. These proteins are not able to fold in salt solutions and they remain as aperiodical-coiled polypeptides [17, 191. Their presence seems to be essential for the DNase 1 sensitivity of active chromatin [20-221, thus suggesting a structural role for the low-M, HMG proteins in transcriptionally active nucleosomes.

Correspondence to F . Montero, Departamento de Bioquimica, Facultad de Ciencias Quimicas, Universidad Complutense, Madrid 3, Spain

Abbreviations. HMG, high-mobility-group; Me2S0, dimethyl- sulphoxide.

Besides the canonical HMG proteins, other HMG-like proteins have been described. Drosophila melanogaster D1 protein [23] is a nuclear protein with many of the chemical properties of an HMG protein. The interaction of D1 with DNA has been studied both in situ and in vitro, showing that this protein seems to bind preferentially to well defined loci enriched in (A + T) sequences from satellite DNA nucleo- somes and free DNA [24 - 281. Some regulatory role in gene expression has been suggested for D1 protein [28]. Similar results have been reported recently for HMG2a from erythrocyte chromatin. This protein shows preferential affinity for (A + T)-rich region in prokaryotic and eukaryotic DNAs, as evidenced by nitrocellulose filter binding assays

We have previously reported the isolation of an HMG- like fraction from the chromatin of the fruit fly Ceratitis capitata [30]. We termed this fraction C1 because of its similarities to D. melanogaster D1 protein. C1 fraction can be resolved by CM-cellulose chromatography into different proteins, namely Clal , Cla2 and Clb, with M, of 25700, 28200 and 25700, respectively [31]. C1 proteins are aperiodical-coiled polypeptides in solution at all the ionic strengths tested, but they show some differences in amino acid composition. Thus, Cla l has one tryptophan residue per molecule while Cla2 and Clb lack this aromatic residue. Tryptic fingerprints also reveal some structural differences among them [31].

In this paper we present our results on the binding proper- ties of the different subfractions of C1 to C. capitata DNA. The thermodynamic parameters of the interactions have been determined by means of thermal denaturation curves of the complexes between C1 proteins and DNA. In the case of Clal , we have studied the fluorescence quenching of its unique tryptophan residue by the interaction with DNA. Also,

~291.

310

specific binding properties have been revealed by DNA reten- tion in nitrocellulose filters.

MATERIALS AND METHODS

Proteins

from crude histone HI. as described elsewhere [31]. C. capitata C1 and histone H1 proteins were obtained

Preparation of D N A

DNA from c'. capitata adulte pharates was prepared ac- cording to the method of Paul and Gilmour [32]. This DNA was labelled with 3H by the nick-translation procedure of Rigby et al. [ 3 3 ] , using [3H]dTTP.

Thermal denaturation studies

C. capitatu DNA at a concentration of 0.87mg/ml in 0.25 mM Na2EDTA, 1 mM phosphate, pH 7.5, was sonicated as described elsewhere [34]. Protein concentration was deter- mined as previously described [35]. Protein stock solutions in 0.25 mM EDTA. 1 mM phosphate, pH 7.5 were added in varying amounts to 57.5-pl aliquots of DNA solution, and the final volume was adjusted to 2 ml with NazEDTA/phosphate. Thermal denaturation profiles were obtained in a Beckman DU-8 spectrophotometer, at a rate of temperature increase of about 0.42"Cimin. The absorbance (average of five values) was recorded every 1 'C. To allow comparison with calculated melting curves, the data were normalized as a fraction coil versus temperature. Absorbance readings were first corrected for thermal expansion of the solvent and then hyperchro- micity at temperature t , H(t) , was calculated according to the following equation

H(t) = [ A ~ w d f ) - A260 (3O0C)I/A2a (30°C)

Fluorescence pro( edures

Clal stock solution in 0.25 mM Na2EDTA, 1 mM phos- phate, pH 7.5 was diluted to a concentration of 1 - 3 pM just before use. Fluorescence spectra were recorded in a Perkin- Elmer MPF-44E spectrofluorometer in the ratio mode, using 2.0-nm and 1 .O-nm slits in the excitation and emission mono- chromator, respectively. For titration, aliquots of DNA were added directly to the protein solution and the fluorescence changes at 340 nm were recorded. Excitation was at 275 nm. The data were corrected for solvent emission and for dilution. Protein inner-filter effects were avoided using samples whose 275-nm absorption values were lower than 0.02. DNA inner- filter effects were corrected using the following equation:

F(corrected) = F(observed x antilog [(A275 -A340)/2] .

Nitrocellu1~)se.fi'ltc.r assa,vs These experiments were performed as described elsewhere

[36] with slight modifications. Before use, the Sartorius SM 113 06 filters were boiled in distilled water for 30 min and stored in a buffer containing 1 YO dimethyl sulphoxide (Me2SO), 0.1 M NaCl and 20 mM Tris/HCl, pH 7.5. Labelled DNA and proteins were incubated at room temperature during 15 min in the buffer described, without MezSO. The different mixtures contained 0.2 pg C . capitata DNA and increasing amounts of proteins in a final volume of 500 p1.

After incubation, the mixtures were filtered at 1 ml/min and the filters were then washed thoroughly with the Me,SO/ NaCl/Tris buffer, dried and counted for radioactivity in a Beckman LS-350 scintillation counter. Background radioac- tivity, defined as the amount of labelled DNA retained in the filters in the absence of proteins, was substracted for each experiment. These values varied over 0.2-0.8% of input radioactivity.

RESULTS

Thermal denaturation studies

In order to study the interactions among C1 proteins and DNA, we made protein . DNA complexes in 0.25 mM NazEDTA, 1 mM phosphate, pH 7.5, and observed how the C1 proteins altered the thermal denaturation of DNA.

Fig. 1 shows the experimental melting profiles obtained and their best-fitting theoretical curves, calculated according to the equations developed by McGhee [37], and the derivative melting curves from the experimental profiles. On the other hand, the concentration of both protein and DNA in the complexes, and the melting temperatures obtained in each case, are shown in Table 1 . In the conditions employed, the three C1 proteins stabilized DNA in a similar way. Therefore, these polypeptides seem to bind preferentially to the double- stranded DNA rather than to single-stranded DNA. It is worth noting that in spite of the apparent monophasicity of all the curves depicted in Fig. 1 they are at least biphasic as shown by their derivatives.

These interactions of C1 proteins with DNA seemed to be very similar for all of them. However, some differences were observed. The generation of the best-fitting theoretical melting curves from the experimental data, according to McGhee [37], allowed us to calculate the values of the intrinsic association constants (Ka), the number of base pairs occupied in the interaction (n) and the cooperativity factor (w). The three K, values were relatively low, ranging from 2.7 x lo4 M-' for Cla l to 2.0 x lo6 M-' for Cla2. while Clb has a K, value of 3.5 x lo5 M-' . The number of base pairs occupied in the interaction is low in the three cases: 1 , 2 and 3 for Clal , Cla2 and Clb, respectively. On the other hand, the interaction of the three C1 proteins with DNA showed a similar and relatively high cooperativity, as indicated by the values of the w factor: 30,20 and 35 for Clal , Cla2 and Clb, respectively.

Fluorescence studies

Taking advantage of the tryptophan residue present in Clal , we carried out some fluorescence experiments. First, it was observed that total saturation of Clal with DNA yielded an 85% quenching of the fluorescence of tryptophan. This percentage is clearly higher than the 30% reported for the chicken erythrocyte HMGl [18]. The reason for this difference is that while HMGl from vertebrates is able to fold 18, 91, Cla l is a random coiled polypeptide. Thus, the tryptophan residue of C la l is easily reached by any potential quencher (among them, the DNA), while the tryptophan residues of HMGl are protected inside a structural region of the protein [181.

Fig. 2 shows the titration curve of Cla l with C. capitata DNA in 0.25 mM NazEDTA, 1 mM phosphate, pH 7.5. Scatchard analysis of these data is shown in Fig. 3. Ex- perimental data were adjusted according to the model of McGhee and von Hippel [38], and the values of K,, n and ~i

31 1

Fig. 1. Thermal denaturation of complexes between C l proteins and DNA. Complexes between C. capitata C1 proteins and C. capitata DNA were formed in 0.25 mM Na2EDTA, 1 mM phosphate, pH 7.5. (A, B and C) Experimental melting profiles corresponding to the complexes formed with Cla l , Cla2 and Clb, respectively, at different r (w/w). (A, W , 0 ) Experimental points; (-, - - - -, - -) theoretical melting curves; (A) free DNA; (W) r = 1.0, r = 0.5 and r = 0.5 for Cla l . DNA, Cla2 . DNA and C l b . DNA complexes, respectively; ( 0 ) r = 4.0, r = 2.0 and r = 2.0, for Cla l . DNA, Cla2 . DNA and C l b . DNA complexes, respectively. Theoretical melting curves were calculated according to McGhee [37], using the following parameter values: protein Cla l , n = 1 bp, K, = 2.7 x lo4 M-' , w = 30; protein Cla2, n = 3 bp, K, = 2.0 x lo6 M-I, w = 20 and protein Clb, n = 2 bp, K, = 3.5 x lo5 M-', w = 35. In all cases, the theoretical melting curve for free DNA was generated with the following parameter values: A# = -34 kJ (mol bp)-', AX = - 103.7 kJ (mol bp)-' K- ' and o = 8.4 x (D, E and F) Derivative melting profiles of the melting curves shown in A, B and C, respectively

Table 1 . Melting characteristics of the complexes formed between C. capitata DNA and C1 proteins

Protein r(w/w) [DNA1 [Protein] Melting in the (as bP) temperature complex

1 2

CIM

Clal 1 .o 34.7 Cla l 4.0 24.4 Cla2 0.5 33.0 C1 a2 2.0 36.2 C l b 0.5 35.4 C l b 2.0 33.6

" C

1.95 55.4 67.1 3.89 - 74.0

0.443 55.1 67.6 1.77 - 72.0

0.486 55.7 70.1 1.95 - 79.0

were calculated. The values obtained were in good agreement with those calculated from the melting profiles. The K, value calculated by both techniques was identical, i.e. 2.7 x lo4 M-l . The n value was 0.8 base pair and w was 14.

Nitrocellulose filter assays Should a specific interaction between C1 proteins and C.

capitata DNA be present, the thermal denaturation studies and the spectrofluorimetry would have not detected it. These techniques are not useful to study interactions which involved a low percentage of the DNA present in the mixtures. To study the possibility of a specific binding of C1 proteins to C.

1.0 f

$ o . L / = , , , , , ,

U 0.2

'0 20 LO 60 80 100 120 IONAI (pM bp)

Fig. 2. Titration of C la l with DNA. The titration of Cla l with C. capitata DNA was carried out in 0.25 mM Na2EDTA, 1 mM phosphate, pH 7.5. Cla l concentration was 2.72 pM. The final quenching was 93.1 + 1.2% of initial intensity. Fluorescence intensity was measured at 340 nm after each addition of native C. capitata DNA. The data depicted were average of three independent titrations

capitata DNA we carried out nitrocellulose filter retention assays of complexes between C1 proteins and DNA, Under the conditions employed, proteins bind quantitatively to nitrocellulose filters while nucleic acids do not [39 - 411. The presence of specific interactions can thus be determined by making complexes between proteins and labelled DNA of different sizes and observing the resulting radioactivity reten- tion patterns [36]. We used [3H]DNAs of either about 40 kb or 1 .O f 0.3 kb from C. capitata (the latter obtained by sonica-

312

*O LO t

V

Fig. 3. Scatchard una1ysi.c of the C l a l jluorescence binding data obtained in 0.25 m M Na2EDTA, 1 mM phosphate, p H 7.5. The theo- retical binding curve was calculated according to McGhee and von Hippel [38], using the following parameter values: n = 0.8 bp, K , = 2 . 7 ~ lo4 M - ' and LL = 14

tion of the larger one). Sizes were determined by agarose gel elecrophoresis. using adequate molecular size markers (not shown). Also, the filtration assays were performed in 0.1 M NaCl to weaken the nonspecific interactions.

The interaction of Clal and Cla2 with the 1.0 + 0.3-kb [3H]DNA yielded the retention curves shown in Fig. 4A. The percentage of DNA bound by the two proteins was very similar (3%) and it was a function of the protein concentra- tion. However, in the range of concentration tested the curves did not reached a plateau. Instead, a straight line was present which grew steadily with the increasing amounts of protein. This result suggests that two kind of interactions were taking place in the binding of the Cla proteins to C. cupitutu DNA. One of these interactions was strong and it accounted for the 3% of the DNA retained in the filters. The other one, a weak interaction, should be responsible for the absence of a plateau. Its association constant has to be low because the DNA pres- ent in the assay was far from being saturated. This is in good agreement with the K, values ranging from lo4 to lo6 which were calculated by the thermal denaturation and spectro- fluorimetry studies, in which the ionic strength of the buffers was somewhat different.

On the other hand, when the [3H]DNA of the complexes was the larger one ( ~ 4 0 kb), the results obtained were those depicted in Fig. 4B. Again, the percentage of DNA bound to the filters was only a fraction of total input DNA. It is worth noting that while the shape of the retention curves was similar to those shown in Fig. 4A, the amount of DNA bound to the filters was clearly higher. Thus, only 3% of the 1.0 f 0.3-kb DNA was retained while about 25% of the 40-kb DNA was bound to nitrocellulose, indicating a marked size-dependence of the binding. The possibility of a preferential binding of C1 proteins to DNA nicks can be ruled out because thermal denaturation studies suggest a binding to double-stranded DNA. It has been reported that some proteins bind preferen- tially to DNA ends [42, 431, but this does not seem to be the case of C1 proteins because an interaction with DNA ends would have yielded curves with a 100% retention of DNA, independently of its size.

Therefore, the results obtained suggest that the strong interaction has to be sequence-specific and that the number of binding sites along the DNA molecule is limited. Clb, a protein distinct from the two Cla proteins, although they are closely related [31], was able to bind twice as much DNA

as the Cla proteins in the range of concentration assayed. Furthermore, histone H1, a protein known to bind nonspecifically to DNA, retained almost 100% of the input DNA. On the other hand, all the curves shown in Fig. 4 have a sigmoidal shape (the only exception is the one of histone Hl). Thus the specific interaction of C1 proteins with C. capitata DNA was cooperative, thus suggesting that the nucleotide sequences involved in this binding have to be near one another. As it has been pointed out above, the nonspecific interactions also proved to be cooperative.

DISCUSSION The results reported here show the existence of both

nonspecific and specific interactions between C1 proteins and C. capitatu DNA. Data from the melting studies and from spectrofluorimetry can be fitted to the models of McGhee [37] and McGhee and von Hippel (381, thus suggesting that the interactions are nonspecific. Also, some differences among the different subfractions can be appreciated.

The nonspecific interaction has a relatively low K, value, although a difference of two orders of magnitude exists between K, values for Cla l and Cla2. being K,(Clal) < K,(Clb) < K,(Cla2). Recently, the interaction of chicken erythrocyte HMGl with DNA has been studied [18]. The K, values for C1 proteins are similar to that reported for this HMG1, which varies from 3.0 x lo4 M- ' to 3.0 x 10' M- ' depending on the ionic strength.

Perhaps the most striking difference between chicken erythrocyte HMGl and C1 proteins with respect to their nonspecific interaction with DNA is the number of base pairs occupied as a result of the binding. Thus, Butler et al. [18] reported an n value of about 14 bp for HMG1, while our results with C1 proteins suggest an n value lower than that, i.e. 1 - 3 bp, as calculated by both thermal denaturation and, in the case of Clal , fluorescence studies. It is worth noting that in spite of an amino acid composition homology among HMGl and C1 proteins given their HMG condition, the amino acid distribution may be very different. HMGl from vertebrates is a protein with an asymmetric distribution of charged amino acid residues, with domains enriched in some particular amino acids [44, 451. This does not seem to be the case with C1 proteins as suggested by their tryptic fingerprints and by the fact that they are not able to fold to a significant degree [31]. Thus, it may be possible that the nonspecific interaction of C1 proteins with DNA was a 'territorial' bind- ing in the way suggested by Manning for some ligands [46] rather than being entirely local in nature.

The existence of a specific interaction in the binding of C1 proteins to DNA seems to be more interesting. Some conclusions can be drawn from our results. (a) The K, values seems to be clearly higher than those of the nonspecific inter- actions; thus, the nonspecific binding should be practically irrelevant in the nitrocellulose filter assays, where their pres- ence is observed for lack of a plateau. (b) The specific binding is cooperative, so that the sequences involved in the interac- tion should not be statistically distributed along the DNA molecule, but in relatively close proximity. A conformational change in the DNA could also explain the cooperativity ob- served. However, circular dichroic spectra of soluble complexes formed between DNA and Cla l and Cla2 showed no significant differences to that of free DNA [31], thus making this an unlikely hypothesis. (c) Slight differences among the C1 proteins concerning their specific binding to DNA can be detected.

313

Protein/DNA lw/w)

Fig. 4. Binding of C1 proteins to C. capitata DNA. A constant amount of 0.2 pg of 3H-labelled C. capitata DNA was incubatd with increasing amounts of C1 proteins. After a 15-min incubation at room temperature in a buffer composed of 0.1 M NaCl and 20 mM Tris/HCl, pH 7.5, the mixtures were filtered. The filters were then washed, dried and counted for radioactivity. The background radioactivity, defined as described under Materials and Methods, was substracted. The data depicted are from four independent experiments. (A) DNA size was about 400 bp; (-) C l a l , (----) Cla2. (B) DNA size was about 47kbp; (-.-.-) Clal, ( . . . . . . ) Cla2, (----) C l b and (-) C. capitata histone H1

D. melunogaster D1 and C. cupitatu C1 proteins are closely related [31], and they could be insect-specific HMG-like pro- teins. The preferential binding loci for D1 have been deter- mined to be (A + T)-rich satellite sequences [28]. A canonical HMG protein such as erythrocyte HMG2a also exhibits a marked preference for this kind of reduced stability regions, at physiological ionic strength and temperature [29]. The cooperativity shown by the C1 proteins in their specific in- teraction with C. cupitutu DNA could be explained in terms of a binding to repeated sequences such as satellite sequences are. However, we actually do not know the precise location of the nucleotides to which C1 proteins specifically bind. Also, we do not know whether the diversity of C1 and the different behaviour of its subfractions are or are not connected with a variety of functions.

Nevertheless, we have recently reported our preliminary results on the location of C1 proteins in different fractions of C. capitata chromatin showing that these polypeptides are present, although not exclusively, in a subfraction of active chromatin [47]. Further work will have to be performed to ascertain the location and the number of base pairs involved in the specific interaction of these proteins with DNA.

Comision Asesora de Investigacion Cientifica y Tecnica (Spain). This work has been supported in part by grant 1472/82 from the

REFERENCES

1.

2.

3.

4.

5.

6. 7.

Goodwin, G. M., Sanders, C. & Johns, E. W. (1973) Eur. J . Biochem. 38, 14-49.

Kuehl, L., Salmond, B. & Tran, L. (1984) J. Cell Biol. 99, 648 - 654.

Goodwin, G. H., Walker, J. M. & Johns, E. W. (1978) in Cell nucleus (Busch, H., ed.) vol. VI, pp. 181 -219, Academic Press, New York.

Johns, E. W. (1982) in The HMG chromosomalproteins, pp. 1 - 7 (Johns, E. W., ed.) Academic Press, New York.

Spiker, S., Mardian, J. K. W. & Isenberg, I. (1978) Biochem. Biophys. Res. Commun. 82, 129- 135.

Weber, S. & Isenberg, I. (1980) Biochemistry 19,2236-2240. Mayes, E. L. (1982) in The HMG chromosomalproteins (Johns,

E. W., ed.) pp. 9-40, Academic Press, New York.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25. 26.

27. 28. 29.

30.

Cary, P. D., Crane-Robinson, C., Bradbury, E. M., Javaherian, K., Goodwin, G. H. & Johns, E. W. (1976) Eur. J. Biochem.

Baker, C., Isenberg, I., Goodwin, G. H. &Johns, E. W. (1976)

Carballo, M., Puigdomenech, P. & Palau, J. (1983) EMBO J . 2,

Levy, W. B. & Dixon, G. H. (1978) Can. J . Biochem. 56,480-

Stoute, J. A. & Marzluff, W. F. (1982) Biochem. Biophys. Res.

Bonne-Andrea, C., Harper, G., Sobczak, J. & De Recondo, A.

Javaherian, K., Liu, L. F. & Wang, J. C. (1878) Science (Wash.

Javaherian, K. & Sadeghi, M. (1979) Nucleic Acids Res. 6,3569- 3580.

Isackson, P. J., Fishback, J. L., Bidney, D. L. & Reeck, G. R. (1979) J . Biol. Chem. 254, 5569-5572.

Javaherian, K. & Amini, S. (1978) Biochem. Biophys. Res. Commun. 85, 1385-1391.

Butler, A. P., Mardian, J. K. W. & Olins, D. E. (1985) J. Biol. Chem. 260, 10613-10620.

Abercrombie, B. D., Kneale, G. G., Crane-Robinson, C., Bradbury, E. M., Goodwin, G. H., Walker, J. M. &Johns, E. W. (1978) Eur. J . Biochem. 84, 173- 177.

Weisbrod, S. & Weintraub, H. (1979) Proc. Natl Acad. Sci. USA

Weisbrod, S., Groudine, M. & Weintraub, H. (1980) Cell 19,

G a d , B., Panet, A. & Ceder, H. (1980) Proc. Natl Acad. Sci.

Rodriguez-Alfageme, C., Rudkin, G. T. & Cohen, L. H. (1976)

Rodriguez-Alfageme, C., Rudkin, G. T. & Cohen, L. H. (1980)

Levinger, L. & Varshavsky, A. (1982) Cell 28,375-385. Levinger, L. & Varshavsky, A. (1982) Proc. Natl Acad. Sci. USA

Levinger, L. (1985) J. Biol. Chem. 260, 11 799 - 11 802. Levinger, L. (1985) J. Biol. Chem. 260, 14311-14318. Brown, J. W. &Anderson, J. A. (1986) J . Biol. Chem. 261,1349-

Franco, L., Montero, F. & Rodriguez-Molina, J. J. (1977) FEBS

62, 583-590.

Biochemistry 15, 1645 - 1649.

1759-1764.

491.

Commun. 107, 1279- 1284.

M. (1984) EMBO J . 3,1193-1199.

DC) 199, 1345-1346.

76, 630 - 634.

289-301.

USA 77, 1787 - 1790.

Proc. Natl Acad. Sci. USA 73,2038 - 2042.

Chromosoma (Berl.) 78, 1-31.

79,7152-7156.

1354.

Lett. 78, 317-320.

314

31. Marquez, G., Moran, F., Franco, L. & Montero, F. (1982) Eur.

32. Paul, J . & Gilmour, R. S. (1968) J . Mol. Biol. 34, 305-317. 33. Rigby, P. W. J . . Dieckmann, M., Rhodes, C. & Berg, P. (1977)

34. Britten, R. J., Graham, D. E. & Neufeld, B. R. (1974) Methods

35. Barbero, J . L., Franco, L., Montero, F. & Moran, F. (1980)

36. Dastugue, B. & Crepin, M. (1979) Eur. J . Biochem. 99,491 -498. 37. McGhee, J. D. (1976) Biopolymers 15, 1345-1375. 38. McGhee, J. D. & von Hippel, P. H. (1974) J. Mol. Biol. 86,469-

39. Jones, D. W. & Berg, P. (1966) J . Mol. Biol. 22, 199-209.

J . Biochem. 123, 165-170.

J . Mol. Biol. 113, 237-251.

Enzymol. 24, 363 -41 8.

Biochemistrj 19,4080-4087.

489.

40. Riggs, A. D. & Bourgeois, S. (1968) J . Mol. Biol. 34, 361 -364. 41. Riggs, A. D., Suzuki, H. & Bourgeois, S. (1970) J . Mol. Biol. 48,

42. Melacon, P., Burgess, R. R. & Record, T. Jr (1982) Biochemistry

43. Melacon, P., Burgess, R. R. & Record, T. Jr (1983) Biochemistry

44. Walker, J. M., Goodwin, G. H. & Johns. E. W. (1976) Eur. J .

45. Walker, J. M., Gooderham, K . &Johns, E. W. (1979) Biochem.

46. Manning, G. S. (1979) Biopolymers 18, 2929-2942. 47. Caballero, R., Marquez, G. & Montero, F. (1986) hi. J . Biochem.

67 - 84.

21,4318-4331,

22, 5169-5176.

Biochem. 62,461 -469.

J. 179,253-255.

18,901 -908.