Biochemistry zyxwvut 1983, 22, 5 zyxwvuts 169-5 176 5169 Direct Evidence for the Preferential Binding of Escherichia zyxw coli RNA Polymerase Holoenzyme to the Ends of Deoxyribonucleic Acid Restriction Fragment st Paul MelanGon,*Richard R. Burgess, and M. Thomas Record. Jr.* ABSTRACT: Escherichia coli RNA polymerase holoenzyme has been observed to form a variety of nonpromoter complexes with DNA restriction fragments in experiments performed with the nitrocellulose filter assay [Melangon, P., Burgess, R. R., zyxwvut & Record, M. T., Jr. (1982) Biochemistry 21,431843311. Here we report the use of this assay to investigate aspects of the weak (heparin-sensitive) interactions of RNA polymerase core and holoenzyme with a 1600 base pair (bp) fragment ofaT7 DNA which contains no promoters or TB (tight binding; heparin-resistant) sites. Under the ionic conditions investigated [50 mM NaCl/10 mM MgC12/10 mM sodium N-(2- hydroxyethy1)piperazine-N’-ethanesulfonic acid (pH 7.7)], both core and holoenzyme bind to the linear DNA fragment and cause comparable levels of filter retention. When the DNA fragment is self-ligated into a circular molecule (non- supercoiled), the extent of binding of holoenzyme (but not that of core) is dramatically reduced. This directly proves our previous hypotheses that holoenzyme recognizes and prefer- entially binds to the ends of DNA fragments and that this mode of binding is responsible for most of the heparin-sensitive filter retention of nonpromoter fragments. The residual mode Escherichia coli RNA polymerase (RNAP),’ the multi- subunit enzyme responsible for the synthesis of RNA in E. coli, exists in two major forms: core (subunit structure zyxwvu a2&3’) and holoenzyme zyxwvutsrq (cuz&3’u). The additional presence of the u subunit allows the holoenzyme to recognize specific (promoter) regions on the DNA, from which RNA synthesis is correctly and efficiently initiated. Both forms of the enzyme exhibit general affinities for DNA. Some nonpromoter interactions of holoenzyme may play a role in the promoter search mechanism [see von Hippel et al. (1982)]. In addition, they reduce significantly the solution concentration of holoenzyme under most in vitro conditions used to investigate binding to promoters. When the nitrocellulose filter binding assay is used, the potential for retention of DNA by RNAP bound at non- promoter sites must be considered in addition to the reduction of free enzyme concentration. A wide range of techniques has been used to study such nonspecific interactions. A survey of the results is provided by Shaner et al. (1983). We have recently studied the interactions between RNAP holoenzyme and an unfractionated HaeIII digest of T7 DNA by using the nitrocellulose filter binding assay (Melangon et From the Department of Chemistry, College of Letters and Science, and the Department of Biochemistry, College of Agricultural and Life Sciences (P.M. and M.T.R.), and the Department of Oncology, School of Medicine (R.R.B.), University of Wisconsin-Madison, Madison, Wisconsin 53706. Receiued April 7, 1983. This work was supported by National Institutes of Health Grants GM 23467 (to M.T.R.) and CA-23076 and GM-28575 (to R.R.B.). *Address correspondence to this author at the Department of Chem- istry, University of Wisconsin-Madison, *Recipient of a Bourse d’excellence from the Fonds F.C.A.C. (Que- bec). 0006-2960/83/0422-5169$01.50/0 of binding of holoenzyme detected with the circular DNAs was considered in determining the amount of protein bound at ends only. To calculate end-binding constants (KE), the amount of protein bound nonspecifically (which does not ap- pear to cause efficient filter retention) was also taken into consideration. At 0 OC, we obtain a value for KE of (2.1 f 0.5) zyxwv X lo8 M-I, in good agreement with that determined earlier. This value of KE is relatively constant over the tem- perature range 0-37 OC. The magnitude of KE indicates that ends can effectively compete with some promoters for RNA polymerase. Therefore, for in vitro promoter binding studies where enzyme is not in excess, end binding (like nonspecific binding) must be considered in the analysis of the promoter binding data, as discussed earlier [Shaner, S. L., MelanGon, P., Lee, K. S., Burgess, R. R., & Record, M. T., Jr. (1983) Cold Spring Harbor Symp. Quant. Biol. 47, 463-4721. The apparent greater specificity for DNA ends of holoenzyme relative to core polymerase is discussed in terms of a steric model in which the u subunit helps to reduce the affinity of holoenzyme for interior DNA sites through unfavorable steric contacts that are absent in an end complex. al., 1982). Two major classes of nonpromoter interactions were detected by this assay. [Other assays can detect and quantify a third weaker class of nonpromoter complexes [cf. Shaner et al. (1983) and Kadesch et al. (1981a, 1982)], but that class is not likely to give rise to efficient filter retention.] The two classes detected by filter binding consist of (i) complexes which form instantaneously on the time scale of mixing and are sensitive to a 10-s challenge with the polyanion heparin and (ii) complexes which form more slowly (tllz = 2-3 min), are resistant to a heparin challenge, and are designated TB (for tight binding; Kadesch et al., 1981b). At 0 “C, only the first class is detected. Whereas the first class of complexes was observed on all HaeIII fragments investigated, TB complexes appear to be located on only a subset of the fragments. This last property allowed the isolation of DNA restriction frag- ments that did not appear to carry either promoter or sig- nificant TB sites and that were therefore suitable to study the heparin-sensitive binding detected by the filter assay. The comparison of binding constants determined by using DNA molecules of different lengths (800 and 2000 bp) provided strong but indirect evidence that the ends of DNA fragments show a markedly higher affinity for RNAP holoenzyme than do random interior sites and that end binding could be re- sponsible for most of the filter retention of promoter-free DNA fragments (MelanGon et al., 1982). Using the same end- binding hypothesis but taking into consideration the reduction ‘ Abbreviations: RNAP, Escherichia coli RNA polymerase; TB, tight binding; BSA, bovine serum albumin; Tris, 2-amino-2-(hydroxy- methyl)-1,3-propanediol; Hepes, N-(2-hydroxyethyl)piperazine-N’- ethanesulfonic acid; EDTA, ethylenediaminetetraacetic acid; bp, base pair(s); L, linear; CC, closed circular; NC, nicked circular. 0 1983 American Chemical Society