DNA looping generated by DNA bending protein IHF and the two domains of lambda integrase

Abstract

The multiprotein-DNA complexes that participate in bacteriophage lambda site-specific recombination were used to study the combined effect of protein-induced bending and protein-mediated looping of DNA. The protein integrase (Int) is a monomer with two autonomous DNA binding domains of different sequence specificity. Stimulation of Int binding and cleavage at the low affinity core-type DNA sites required interactions with the high affinity arm-type sites and depended on simultaneous binding of the sequence-specific DNA bending protein IHF (integration host factor). The bivalent DNA binding protein is positioned at high affinity sites and directed, by a DNA bending protein, to interactions with distant lower affinity sites. Assembly of this complex is independent of protein-protein interactions.

Fig. 1

Protein binding sites in attP and attL and their relation to relevant plasmids. The coordinates for attP and attL show 0 within the 7-bp overlap, with positive numbers to the right (P′ arm) and negative numbers to the left (P arm or B arm) (14). The relative orientations of Int binding sites are indicated by arrows for both arm-type (P1, P2, P′1, P′2, P′3) and core-type (C, C′, B) sites. Also indicated are the binding sites for IHF (H1, H2, H′), Xis (X1, X2), and FIS (F). The attP sequences contained in pMJB11 (16) and pJTT58 (7) are noted. The attL region in pSN55 (38) deviates from the canonical sequences in pPH201 (34) by the introducdon of an Eco RI site at +12 and is the parent of the 4-bp (pLV5) and 10-bp (pLV4) insertion mutants. PLV5 was constructed by cleavage of pSN55 with Eco RI, filling in the 3′ ends with the Klenow fragment of DNA polymerase and ligation. The pLV4 plasrnid was constructed by cleavage of pSN55 with Eco RI and ligation of a single unphosphorylated Eco RI—Sma I adaptor (New England Biolabs).

Fig. 2

Int and IHF promoted nuclease protection of att DNA’s containing different combinations of protein binding sites. Nuclease protection assays were performed with NCS (neocarzinostatin) in the presence of the indicated amounts of Int and IHF (given as recombination units per 20 μl) (24). The plasrnid pJT58 does not contain core sites and pMJB11 does not contain arm sites (Fig. 1). The protein binding sites (Fig. 1) protected by Int and IHF are shown on the left. The attP DNA is an Nco I–Aat II fragment from plasmid pWR1 (34) that was 5′ end–labeled on the bottom strand at the Nco I site. DNA from pJT58 is a Nco I–Ssp I fragment 5′ end–labeled on the bottom strand at the Nco I site. DNA from pMJB11 is an Eco RI–Sal I fragment 5′ end–labeled on the top strand at the Eco RI site. The reaction mixture (100 (μl volume) consisted of 120 mM Nacl, 10 mM MgCl₂, 50 mM tris-Hcl (pH 7.4), 10 mM 2-mercaptoethanol, 10 percent (volume by volume) glycerol, and bovine serum albumin (BSA) at 2 mg/ml as described (17). The ³²P-labeled fragments (1 × 10⁻¹¹ to 5 × 10⁻¹¹M) and proteins were purified as described (17, 29).

Fig. 3

Nuclease protection of attL and attL spacing mutants. Nuclease protection assays in the presence of the indicated concentrations of Int and IHF were performed as in Fig. 2. Protein binding sites are to the right of the protected regions for attL and +10 spacing mutant and to the left of protected regions for +4 mutant. Sequences of att site DNA’s are shown in Fig. 1. The attL DNA is a Bam HI–Ban II fragment from pSN55 that was 5′ end–labeled on the bottom strand at the Bam HI site. The +4 attL spacing mutant is a Bam HI–Ban II fragment from pLV5 that was 5′ end–labeled on the bottom strand at the Bam HI site. The +10 attL spacing mutant is a Bam HI–Dra III fragment from pLV4 that was 5′ end–labeled on the bottom strand at the Bam HI site.

Fig. 4

Int cleavage of DNA substrates with and without P′ arm Int sites. (Top) Representation of Int cleavage (open arrow) of substrates that contain (left) or lack (right) the P′ arm sites and are labeled with ³²P at the 5′ end (*). (Bottom) Gel electrophoresis of reactions with the indicated substrates. The amount of Int and IHF in each reaction is given as the number of recombination units per 20 μl. The left margin shows the relative mobility of covalent complexes (att-Int) (29), att DNA with the P′ binding sites [att (+P′)], and att DNA lacking the P′ binding sites [att (−P′)]. Reaction mixtures (20 μl) consisted of 10 mM tris-HCl (pH 7.4), 80 mM NaCl, 5 mM EDTA, BSA at 2 mg/ml, 4 × 10⁻¹⁰M DNA; the reactions were stopped by the addition of SDS to a concentration of 0.1 percent. Samples were analyzed by electrophoresis on a 5 percent acrylamide, 0.1 percent SDS gel that was subsequently dried and autoradiographed. The substrate containing P′ arm Int sites was an Xba I–Nco I fragment from pSK3 5′ end–labeled on the top strand at the Xba I site. The substrate lacking P′ arm Int sites is an Xba I–Bst BI fragment from pSK3 5′ end–labeled on the top strand at the Xba I site. Plasmid pSK3 contains attL and has an Xba I site at +1 from pSN84 (30) and a Bst B1 site at position +48 from pLV1 (39), and a Hind III site that was introduced at +23. Restriction fragments used to generate these substrates were prepared by cleavage of pSK3 with Xba I, ³²P-labeling of the 5′ ends (17), filling in the 3′ ends with unlabeled nucleotides and the Klenow fragment of DNA polymerase, and cleaving with either Nco I or Bst BI. The ³²P-labeled fragments were purified by electrophoresis on a 5 percent polyacrylamide gel with subsequent elution on NA-45 DEAE membranes (Schleicher & Schuell, Inc.) as described (30) and incubated with the appropriate proteins at room temperature for 2 hours.

Fig. 5

Patterns of IHF-mediated enhancement of Int binding (filled symbols). IHF binding at the H′ site (◆) mediates enhanced binding (filled symbols) of Int at core-type (●) and arm-type (■) sites, att DNA (—), non-att DNA ( ) and the end of a DNA fragment ( ) are indicated. The pattern of enhanced binding for both the P′ arm and the core Int binding sites depends upon the presence or absence of adjacent DNA and other Int binding sites.

Fig. 6

Representation of potential interactions in the attL loop structure. The IHF-induced bending of attL DNA (double solid line) is shown promoting interactions between an arm-type binding site and a core-type binding site via the two autonomous DNA binding domains of one Int monomer.