J. MOZ. BiOl. (I 986) 188, 115-l 18 Crystallization of a Complex of cro Repressor with a 17 Base-pair Operator Crystals of the 2 cro repressor complexed to a 17 base-pair synthetic binding site related to the 0,3 operator have been obtained. The complex crystallizes in the hexagonal space group P6, (or P64 with unit cell dimensions a = b = 154.8 8, c = 85.6 A. Preliminary photography reveals that the crystals are stable to X-rays and display measurable reflections to a resolution of about 3.7 A. The diffraction patterns suggest that the cro- DNA complexes are arranged in an open hexagonal network with the DNA fragments stacked end-to-end. The DNA is in the B-form but appears to be bent or curved int,o an approximate superhelix. Tnteractions between proteins and nucleic acids are ubiquitous and are of fundamental importance in molecular biology. Progress towards a molecular understanding of such interactions has been made rerently with the successful structure determina- tions of several DNA binding proteins, including the ,J cro repressor (Anderson et al., 1981), the catabolite gene activator protein (McKay & Steitz, 1981), the amino-terminal domain of the A repressor (Pabo & Lewis, 1982), the restriction enzyme EcoRI complexed to a 12 base-pair fragment (Frederick et al., 1984), the DNA binding protein II (Tanaka et al., 1984), the Klenow fragment of DNA polymerase I from Escherichia coli (Ollis et al., 1985) and the phage 434 repressor bound to a 14 base-pair fragment (Anderson et al., 1985). Structural comparisons between the first three of these proteins have revealed a common helix-turn- helix substructure (Steitz et al., 1982; Ohlendorf ft al.. 1983a) which, on the basis of model building, has been proposed to be responsibk for the formation of specific complexes between these proteins and their DNA binding sites (Anderson et al.. 1981; Ohlendorf et al., 1982; Pabo & Lewis, 1982; Weber & Steitz, 1984). Furthermore, amino acid sequence homology suggests that the helix- t,urn-helix unit, occurs in a number of other DNA binding prot,eins (Anderson et al., 1982; Matthews et al., 1982: Sauer et al., 1982; Ohlendorf et al., 1983a,b). Model building suggests that the recognition of its DNA binding sites by cro is due, for the most part,, to the formation of a net,work of complementary hydrogen bonds and hydrophobic interactions between specific amino acid side-chains and the exposed bases and the thymine methyl groups within the DNA major groove (Ohlendorf et al., 1982). Independent, evidence from electrostatics calculations (Ohlendorf et al., 1983c; Matthew & Ohlendorf, 1985), nuclear magnetic resonance studies (Arndt et al., 1983; Kirpichnikov et al., 1984; Metzler et al.. 1985), chemical modification of cro (Takeda et al., 1985), cro mutants (A. Pakula & R. Sauer, reported in Pabo & Sauer, 1984) and site- directed mutagenesis (Eisenbeis et al., 1985) strongly supports the general feat,ures of the proposed cro-DNA complex. To ascertain the detailed basis of sequence- specific DNA recognition, it is essential to obtain cryst,als of cro complexed to suitable DNA fragments. Complexes of cro with six- and nine- base-pair DNA duplexes have been described (Anderson et al., 1983). Here, we report the crystallization of the 1 cro repressor complexed to a 17 base-pair synthetic operator. The complementary strands of the 17 base-pair fragment were synthesized by solid phase phospho- triester methods (Miyoshi et al., 1980). The sequence of one strand is 5’-d(TpApTpCpApCpCpGpCpGp- GpGpT>GpApTpA)-S’, where the underscored bases are found as A, A and G, respectively, in the native tight-binding 0,3 operator. The main objective in the choice of this sequence was to conserve all the expected sequence-specific inter- actions present in the cro-0,3 operator complex, but to break the exact S-fold symmetry of the DNA so as to prevent self-annealing of individual DNA strands. Tn solution, cro forms a more stable complex with this synthetic operator than with the 0,3 operator site (Y. Takeda et al., unpublished results). Crystals were initially grown by the vapor diffusion technique from solutions where the initial droplet concentrations were 2.5 mg cro monomer/ ml, 2.5 mg duplex 17 base-pair fragment/ml, both of which were buffered in 20 mw-sodium cacodylate (pH 6.9), 100 miv-NaCl. The reservoir concentration ranged from 3.5 M to 4.0 M-NaCl in different experiments. These crystals grew very rapidly (overnight) but were accompanied by a sticky crystalline precipitate, which made mounting technically difficult. Furthermore, t’hese crystalliza- tion conditions led to a large number of nucleations.