Molecular Basis of Inhibition of the Ribonuclease Activity in Colicin E5 by Its Cognate Immunity Protein

doi:10.1016/j.jmb.2006.02.014

J. Mol. Biol. (2006) 358, 571–579

Molecular Basis of Inhibition of the Ribonuclease Activity in Colicin E5 by Its Cognate Immunity Protein Ce´sar Luna-Cha´vez1†, Yi-Lun Lin2† and Raven H. Huang1,2,3* 1

Center for Biophysics and Computational Biology University of Illinois at Urbana-Champaign, Urbana IL 61801, USA 2

Department of Chemistry University of Illinois at Urbana-Champaign, Urbana IL 61801, USA 3

Department of Biochemistry University of Illinois at Urbana-Champaign, Urbana IL 61801, USA

Colicin E5 is a tRNA-specific ribonuclease that recognizes and cleaves four tRNAs in Escherichia coli that contain the hypermodified nucleoside queuosine (Q) at the wobble position. Cells that produce colicin E5 also synthesize the cognate immunity protein (Im5) that rapidly and tightly associates with colicin E5 to prevent it from cleaving its own tRNAs to avoid suicide. We report here the crystal structure of Im5 in a complex with ˚ resolution. The the activity domain of colicin E5 (E5-CRD) at 1.15 A structure reveals an extruded domain from Im5 that docks into the recessed RNA binding cleft in E5-CRD, resulting in extensive interactions between the two proteins. The interactions are primarily hydrophilic, with an interface that contains complementary surface charges between the two proteins. Detailed interactions in three separate regions of the interface account for specific recognition of colicin E5 by Im5. Furthermore, singlesite mutational studies of Im5 confirmed the important role of particular residues in recognition and binding of colicin E5. Structural comparison of the complex reported here with E5-CRD alone, as well as with a docking model of RNA–E5-CRD, indicates that Im5 achieves its inhibition by physically blocking the cleft in colicin E5 that engages the RNA substrate. q 2006 Elsevier Ltd. All rights reserved.

*Corresponding author

Keywords: structural biology; colicin; ribonuclease; protein–protein interaction; RNA cleavage

Introduction Colicins are defensive weapons produced by some Escherichia coli in order to survive under nutrition-deficient conditions.1 Two mechanisms are generally employed for the killing, formation of a pore in the cytoplasmic membrane of the target cell,2 or degradation of nucleic acids within the target cell.3–6 Because of the potential toxicity that would be caused by a colicin within the cell that produces it, it is necessary for the producing cell to express an immunity protein that can inhibit the activity of the colicin. Furthermore, in order to further minimize potential activity of colicin within the producing cell, the association of the immunity protein with its cognate colicin is rapid, specific, and high affinity. These properties make this system † C.L.-C. and Y.-L.L. contributed equally to this work. Abbreviations used: Im5, immunity protein 5; E5-CRD, E5 COOH-terminal region domain; Q, queuosine; MAD, multiple anomalous dispersion; rmsd, root-mean-square deviation. E-mail address of the corresponding author: [email protected]

ideal for the study of molecular recognition of a protein–protein interaction. Colicin classification is based on the type of the receptor that it binds for cell entry. The E group of colicins that bind to the vitamin B12 receptor (BtuB)7 are nucleases (with one exception, E1, which is a pore-forming toxin). Although colicin E5, the subject of our study, has also been shown to be a ribonuclease, it lacks sequence homology with other E group members of the colicin nucleases. Furthermore, the activity domain of colicin E5 does not contain a single histidine, the hallmark of the active site of other known ribonucleases. Masaki and his co-workers demonstrated that, unlike other E group colicin nucleases, which are either DNases or RNases for the ribosomal RNA, colicin E5 is a tRNA-specific ribonuclease.8 This is consistent with the fact that it has no sequence similarity with other colicin nucleases of the E group. Recently, we reported the crystal structure of the activity domain of colicin E5.9 The structure revealed that E5-CRD adopts a nuclease T1-like fold, but with totally different surface topology that presumably enables it to specifically recognize Q-containing tRNA anticodon region. Our structure, in combination

0022-2836/$ - see front matter q 2006 Elsevier Ltd. All rights reserved.

572 with the mutational studies, allowed us to construct a RNA docking model and propose a likely mechanism of tRNA cleavage by colicin E5 without the involvement of a single histidine. Our previous study, however, did not provide information on how the RNA cleavage activity of colicin E5 is inhibited by its cognate immunity protein. Accordingly, we report here the crystal structure of E5-CRD in complex with its immunity protein. In addition, we have carried out a series of Im5 mutations, in which the amino acid residues implicated by the structure to be important in the interaction with E5-CRD are mutated to alanine. These mutants, and the wild-type protein, were assayed for their ability to inhibit RNA cleavage by colicin E5. The results shed light on the molecular basis of colicin E5 inhibition by Im5. Furthermore, together with previously published structures of other colicin-immunity protein complexes (colicin E3, colicin E7, colicin E9 and colicin D), our structure provides additional insight into the molecular mechanisms that are involved in specific recognition processes and protein–protein interactions.

Results Structure of Im5–E5-CRD complex The crystal structure of the Im5–E5-CRD complex was determined using multiple wavelength anomalous dispersion (MAD) phasing, and the structure was refined with multiple rounds of model building and simulated annealing. The asymmetric unit contained a single Im5–E5-CRD complex. The electron density map of the final round of refinement lacked density for the first three residues of Im5, and the first 11 and last four residues of E5CRD. Therefore, the structure of Im5 contains residues 4–109, and the structure of E5-CRD

Crystal Structure of Im5–E5-CRD

contains residues 12–104. The same N-terminal 11 residues are disordered in the structure of the free E5-CRD,9 indicating that they likely belong to the flexible linker region (between the catalytic domain and the receptor binding domain) of colicin E5. As shown in Figure 1(a), the folding of Im5 begins with a loop at the N terminus, which has been shown to interact with E5-CRD through Lys4 (described in detail later). It is followed by two consecutive a helices (a1 and a2) and three b strands (b1, b2 and b3). A third a helix followed by the fourth b strand at the C terminus complete the structure. Three-dimensional arrangements of these secondary structures result in an overall folding of Im5 with a four-stranded b sheet on one side and three a helices packed on the other (Figure 1(a)). The overall structure of E5-CRD in the Im5–E5CRD complex is essentially the same as the one in ˚ upon superfree E5-CRD,9 with the rmsd of 1.0 A imposition of the Ca atoms of the entire protein (residues 11–104) between these two structures. The only noticeable difference is a small conformational change at the tip of the internal loop formed between b1 and b2 (Figure 1(b)), indicating likely conformational flexibility of this region upon association with the RNA substrate. As described previously, E5-CRD adopts folding of a four stranded b-sheet (b1-b4) packed across by an a helix (a2), as commonly seen in other microbial ribonucleases such as ribonuclease T1 (Figure 1(b)). An additional helix from the N terminus (a1), an internal loop formed between b1 and b2, and a C-terminal loop form a unique RNA-binding cleft, which is occupied by part of Im5 in the complex (Figure 1). Interactions between Im5 and E5-CRD Association of Im5 with E5-CRD results in the formation of a compact protein–protein complex,

Figure 1. Ribbon representation of the overall structure of Im5–E5-CRD. (a) Top view of the structure of the complex. Im5 is in magenta and E5-CRD is in blue. (b) Side view of the structure of the complex. This Figure and Figures 2, 3(a), 5 and 6 were made with PyMOL.25

Crystal Structure of Im5–E5-CRD

573

Figure 2. Surface representation of the structure of Im5–E5-CRD. (a) The overall structure of the complex oriented the same as in Figure 1(b). (b) Structures of Im5 and E5-CRD after brought apart from each other from the complex to illustrate the shape complementary in the interface. (c) Front view of the surfaces of the interface to demonstrate the charge complementary in the interface of the complex.

˚ 2 (Figure 2(a)). with a total surface area of 9050 A The interaction between Im5 and E5-CRD is ˚ 2 buried in extensive, with a surface area of 2425 A the interface. This constitutes more than 20% of the entire surface area of each individual protein (5828 ˚ 2 for E5-CRD, respectively). This for Im5 and 5647 A extensive interaction is achieved in part because the shapes of the two molecules at the interface are quite complementary (Figure 2(b)). In addition to shape complementation, the surface charges between these two proteins are also complementary to each other (Figure 2(c)). Specifically, the top and

middle regions of E5-CRD are positively charged (Figure 2(c), right panel, presumably the binding cleft for the negatively charged RNA substrate), while their binding partners in Im5 are negatively charged (Figure 2(c), left panel). On the other hand, the bottom region of E5-CRD is negatively charged while its binding partner in Im5 is positively charged. As discussed below, the detailed interactions in these three regions of the interface undoubtedly confer specificity of Im5 for E5-CRD. The N-terminal residues in Im5 interact primarily with residues in E5-CRD near the C terminus

574

Crystal Structure of Im5–E5-CRD

Figure 3. Molecular recognition of E5-CRD by Im5. (a) Surface representation of Im5 and E5-CRD as shown in Figure 2(c) with the additions of boxes to show the regions of specific interactions. (b), (c) and (d) Detailed interactions between Im5 and E5-CRD in the regions indicated in (a). The main-chains are colored magenta and light blue, and the side-chains are colored orange and cyan for Im5 and E5-CRD, respectively. The hetero-atoms are individually colored, with nitrogen in blue and oxygen in red. The Figures in (b), (c) and (d) were made with Ribbons.26

(Figure 3(b)). Four residues from each protein (K4I, H8I, S26I and K28I from Im5, and E24C, D88C, T90C, and D91C from E5-CRD) take part in the interaction. Specifically, the side-chain of K4I forms two hydrogen bonds with the side-chain and the main-chain carbonyl groups of E24C. H8I forms a hydrogen bond with the side-chain of T90C, which also forms another hydrogen bond with the side-chain of D88C from the same molecule. The side-chain of S26I forms a hydrogen bond with the side-chain of D88C, which also interacts with K28I through a water molecule. In addition to its interaction with D88C, K28I also forms a hydrogen bond with the side-chain of D91C. Residues in the middle region as well as the very C terminus of Im5, located in the three-stranded b sheet, interact with the N-terminal residues of E5CRD (Figure 3(c)). Three residues from Im5, D52I, T70I and T106I, and three residues from E5-CRD, Q16C, K17C and R25C, are involved in specific interactions. The most extensive interaction is from D52I. Its side-chain forms hydrogen bonds with the side-chains of all three residues from E5-CRD. It also forms a fourth hydrogen bond with either a well-positioned water molecule or the side-chain of S68I from the same molecule. The side-chain of T70I forms a hydrogen bond with the side-chain of K17C, and the side-chain of T106I forms a hydrogen bond with the side-chain of Q16C.

The C-terminal residues of Im5 interact with residues at the tip of the RNA-binding loop (Figure 3(d)). Three residues from Im5, N91I, D95I, and A67I, are involved in recognition of a single residue K52C from E5-CRD. Each of the sidechains of N91I and D95I, as well as the main-chain carbonyl group of A67I, forms a hydrogen bond with the side-chain of K52C. The interaction is further enhanced by a well-positioned water molecule, which forms hydrogen bonds with all the three components from Im5 for the recognition of K52C. Mutational study of Im5 Based on the structure, we generated several Im5 single-site alanine mutants of residues implicated by the structure to be important for the Im5–E5CRD interaction. These Im5 mutants were overexpressed in E. coli and purified. The relative activity assay, a modification of a method employed for the study of colicin E9,10 was carried out (Figure 4). As expected, wild-type Im5 is able to inhibit RNA cleavage by E5-CRD at concentrations of E5-CRD that are up to 100% relative to that of Im5 (Figure 4, top panel). On the other hand, the K4A and D95A mutations resulted in w85% loss of inhibitory ability (i.e. only w15% of E5-CRD is required to cleave half of RNA substrate in the

Crystal Structure of Im5–E5-CRD

575 Molecular basis of inhibition of E5-CRD by Im5

Figure 4. Inhibition of RNA cleavage of E5-CRD by wild-type and mutated Im5. In the presence of constant concentration (100 mM) of wild-type or mutated Im5 (marked on the left side), the stem–loop RNA was incubated with increasing concentration of E5-CRD (marked on the top as percentile relative to the concentration of Im5) for 10 min at 25 8C. The resulting samples were analyzed by denaturing PAGE. The top band represents the intact stem–loop RNA substrate and the bottom band indicates the cleaved RNA product. Lesser percentile of E5-CRD required for the RNA cleavage indicates more severe of the mutation in Im5 to its inhibitory ability of E5-CRD.

Although the structure that we have previously reported was unliganded E5-CRD, crystal packing indicated that two E5-CRD molecules interacted with each other, with the C-terminal loop of each molecule inserted into the RNA-binding cleft of the other.9 Furthermore, with additional functional analyses of amino acid mutants in the RNA-binding cleft to guide us, we created an RNA docking model.9 This information has now been incorporated with the present results of the Im5–E5-CRD complex to illustrate the proposed positions of the different ligands for the E5-CRD molecule (Figure 5). Accordingly, Im5, the second molecule of E5-CRD in our previously reported structure, as well as the docked RNA, all appear to occupy the same cleft in E5-CRD. Importantly, the active site of the enzyme, identified by our previous mutational studies, is located within the cleft. Therefore, the molecular basis of inhibition of E5-CRD by Im5 is the direct steric block of the RNA-binding cleft as well as the active site in E5-CRD by Im5. Presumably, steric blocking is sufficient only because the high affinity of Im5 for E5-CRD insures complete inhibition of E5-CRD. This affinity is achieved by extensive interaction between Im5

presence of these two mutants; Figure 4, second and bottom panels). On the other hand, the H8A and K28A mutations exhibited more minimal effects, reducing the inhibitory ability by only 10–15% of wild-type Im5 (i.e. cleavage of half the amount of RNA substrate required w90% of E5-CRD with H8A mutant and w85% with K28A mutant; Figure 4, third and fourth panels). The D52A and N91A mutations have an intermediate effect, reducing 35–45% of inhibitory ability of wild-type Im5. These results thus demonstrate the relative importance of these amino acid residues in the specific interaction with colicin E5.

Discussion The crystal structure of Im5–E5-CRD reported here, in combination with our previous structural and biochemical studies of E5-CRD, provide insight into the likely mechanism of inhibition of colicin E5 by its cognate immunity protein. Furthermore, comparison of the structure of the complex reported here with four previously reported structures of colicins in complex with their cognate immunity proteins furthers our understanding of the strategies that have evolved to accomplish the inhibition of nuclease toxins.

Figure 5. Occupation of a common cleft in E5-CRD by various molecules. The structure of E5-CRD is represented as surface and colored based on the surface electrostatic potential, with negative charge in red and positive charge in blue. Part of other molecules occupying the RNA binding cleft is depicted as a rod, with the peptide from Im5 (residues 62–71) in magenta, the peptide from the second E5-CRD (PDB accession code 2A8K, residues 90–97) in yellow, and the docked anticodon region of tRNA (residues 31–37) in green.

576 and E5-CRD through their hydrophilic surfaces, including the specific interactions in the three separate regions of the interface, as discussed previously. Mechanistic comparisons to the inhibitions of other colicins by their cognate immunity proteins The first structural information available, in terms of inhibition of a particular colicin by its cognate immunity protein, was from the crystal structures of the catalytic domains of colicin E7 and colicin E9 in complex with their cognate immunity proteins.4,11 Both colicin E7 and colicin E9 are DNases. In fact, they may be evolutionarily related because they are highly homologous in amino acid sequences, both in colicins (65% sequence identity) and the immunity proteins (50% sequence identity).4 Therefore, it is not surprising that they are structurally homologous and hence we consider them as one class here. These structural studies indicated that the immunity proteins block the DNA-binding cleft of colicins, but the active site remains open.4,11,12 The structure of the catalytic domain of colicin E3 in complex with its cognate immunity protein was reported not long after the structures of colicin E7 and E9.13 Colicin E3 is an RNase, cleaving E. coli 16 S ribosomal RNA at the site between nucleotides 1493 and 1494. Colicin E3 and its cognate immunity protein have no sequence similarity to their corresponding partners of colicin E7 and colicin E9. Therefore, it was reasonable to predict that they might be structurally different. However, structural studies revealed that the E3 immunity protein does not bind the active site but rather binds at an adjacent site.13 Therefore, despite structural differences, the mechanism of inhibition appears to be similar to that of colicin E7 and

Crystal Structure of Im5–E5-CRD

colicin E9, blocking the RNA-binding cleft, but not the active site. Perhaps the most interesting comparison, as far as the structure of Im5–E5-CRD is concerned, is with the structure of the catalytic domain of colicin D in complex with its cognate immunity protein, recently reported by two groups.14,15 Like colicin E5, colicin D is a tRNA-specific ribonuclease, cleaving four isoaccepting tRNAs for Arg between nucleotides 38 and 39.16 In fact, colicin E5, colicin D and a suicide ribonuclease toxin, PrrC, are the only known tRNA-specific ribonucleases to date.17 Furthermore, a Dali search18 revealed that the catalytic domain of colicin E5 has the same structural folding as that of colicin D and they are modestly structurally homologous (Z scoreZ3.5, ˚ ; Figure 6). Both of them belong to a rmsdZ3.2 A well-known fold for a superfamily of microbial ribonucleases, represented by the well-studied ribonuclease T1. Although they have similar folding, the topologies of RNA binding clefts of colicin E5 and colicin D are different, resulting in different locations of the active sites of the enzymes (Figure 6). The immunity protein for colicin D has also been shown to block both RNA binding and the active site, and thus the mechanism of inhibition of colicin E5 and colicin D by their cognate immunity proteins appears to be the same. Interestingly, however, these two immunity proteins differ in structure, as discussed below. Folds of known immunity proteins With four available examples (as indicated above, colicin E7 and colicin E9 are treated as one example because of their high structural homology), it is of some use to discuss the folding of the immunity proteins, in order to provide possible hints of their evolutionary origin. As discussed previously, Im5 adopts a fold of mixed a/b protein (Figure 6(a)). Another known immunity protein that is also a a/b

Figure 6. Structural comparison of Im5–E5-CRD complex to that of ImmD-colicin D. (a) Ribbons representation of the side view of the structure of the Im5–E5-CRD. The representation is the same as in Figure 1(b) with the exception that the side-chains of two catalytic residues (D46 and R48) are represented in sticks. (b) The structure of ImmD-colicin D (PDB entry 1V74). ImmD is colored red and colicin D is in yellow. The side-chains of the catalytic residue H611 and its nearby K608 and K610 are represented in sticks. The secondary structures of the common folding between colicin E5 and colicin ˚ based on Dali search. D are labeled (a1-a2, b1-b4) and the rmsd between these two molecules is 3.2 A

577

Crystal Structure of Im5–E5-CRD

protein is the immunity protein for colicin E3.13,19 However, their topologies are different from each other. The immunity proteins for colicins E7/E9 and colicin D are all four-helix bundles (Figure 6(b)).4,11,14,15 Since colicin E5 and colicin D are structurally and functionally similar, but their immunity proteins are quite different, it is possible that the acquisition of cognate immunity proteins by colicin E5 and colicin D occurred by distinct evolutionary pathways. Surface charges in the interface of colicins and their immunity proteins Whether the immunity protein blocks the binding of the DNA/RNA substrate alone, as in the case of colicin E7/E9 and colicin E3, or blocks the RNA substrate as well as the active site of the enzyme, as in the case of colicin E5 and colicin D, the interaction between a particular colicin and its cognate immunity protein is hydrophilic in nature because of hydrophilic property of DNA or RNA substrate. However, The interaction between Im5 and E5-CRD appears to differ from other known colicin-immunity protein complexes in terms of charge distribution in the interface. The interactions in the regions of (c) and (d) shown in Figure 3(a) result from contribution of positively charged surface from E5-CRD and negatively charged surfaces from Im5, which is no different from other known colicin-immunity protein structures. On the other hand, the interaction in the region of (b) is the opposite, with the positively charged surface from the Im5, and the negatively charged surface from the colicin E5. In other cases of colicinimmunity protein interaction (colicin E7/E9, colicin E3, and colicin D), the surface charges are more or less homogenous, with the colicin providing positively charged surface and the immunity protein providing the negatively charged one. In this sense, the interaction of Im5 with colicin E5 is unique among known examples of colicinimmunity protein pairs.

Materials and Methods Cloning, overexpression and purification of proteins Preparations of native Im5–E5-CRD complex and the individual wild-type Im5 and E5-CRD proteins have been described.9 For the preparation of selenomethionine derivatives needed for structural determination through MAD phasing, 5 ml of overnight culture was used to inoculate 3 l of minimal media supplemented with ampicillin and thiamine at final concentrations of 100 mg/l and 50 mg/l, respectively. The cell culture was grown at 37 8C to A600 nm of 0.6 and then 40 ml of methionine suppression buffer (comprised of 200 mg of L-Lys-HCl, L-threonine and L-phenylalanine, and 100 mg of L-leucine, L-isolutine and L-valine in water) was added. After another 30 min of incubation at 37 8C, the cell culture was induced with 1 mM isopropyl-b-D-galactopyranoside (IPTG) and addition of 50 mg/l of selenomethionine. Incubation at

37 8C was continued for another 18 h and the cells were then harvested. For Im5 mutants, it was necessary to create a new vector that expressed only Im5 protein because expression of some Im5 mutants together with E5-CRD was toxic to E. coli cells. Therefore, the gene encoding Im5 was PCR amplified from the plasmid encoding Im5–E5CRD. The resulting PCR product was purified using the Qiagen gel extraction kit and digested with the restriction enzymes SacI and HindIII (New England Biolab). The digested PCR product was inserted into pLM1 vector, resulting in pLM1-Im5 expression vector. The plasmids containing Im5 mutants were created by the QuickChange method using the recombinant pLM1-ImE5 plasmid as the template. Recombinant pLM1-Im5 vectors containing the genes of wild-type Im5 or Im5 mutants were transformed into BL21(DE3) E. coli expression cells. A 5 ml overnight preculture was used to inoculate 2 l of LB supplemented with 50 mg/ml ampicillin. Cells were cultured at 37 8C until an A600 nm of 0.6 and then induced with 1 mM IPTG for another 4 h. Purification using an FPLC system was as follows: cell pellets were suspended in Q Sepharose buffer A (20 mM Mes (pH 6.5), 10 mM NaCl, 1 mM DTT) and cells were lysed using a French Press. Cell lysates were centrifuged and supernatants were passed through a preparatory Q Sepharose Fast Flow column. The protein was eluted with Q Sepharose buffer B (same as Q Sepharose buffer A except 1 M NaCl). Fractions containing Im5 were pooled, saturated (NH4)2SO4 was added to bring the final concentration of (NH4)2SO4 to 1 M. The sample was then injected into a HiLoad hydrophobic column equilibrated with HiLoad buffer A containing 20 mM Mes (pH 6.5), 1.0 M (NH4)2SO4, and 1 mM DTT. Proteins were eluted with HiLoad buffer B (same as HiLoad A except the concentration of (NH4)2SO4 was reduced to 0.1 M). Fractions containing Im5 were pooled, concentrated and washed with Q Sepharose buffer A. The sample was then applied to the MonoQ anion exchange column and Q Sepharose buffers were used to elute proteins. Finally, fractions containing Im5 from MonoQ were purified by Superdex 200 size-exclusion using a buffer containing 20 mM Mes (pH 6.5), 10 mM NaCl and 1 mM DTT. Crystallization, data collection, and structural determination All crystals were obtained by the hanging-drop vapor diffusion method at 4 8C. Both native and selenomethionine derivative crystals were grown against a well solution of 50% saturated (NH4)2SO4, 100 mM Tris–HCl (pH 8.5), and 100 mM NaCl. Orthorhombic crystals that ˚ resolution were obtained in one week. diffracted to 1.15 A To carry out data collection at a low temperature, the crystals were first soaked stepwise in cryo-protecting solutions containing all the components of the well solution and 5-to-25% (v/v) of glycerol. The soaked crystals were then mounted in a nylon loop, and flashfrozen in liquid nitrogen. The MAD data were collected with the help of Dr H. H. Robinson at beamline X12C at Brookhaven and the native data were collected at beamline 19-ID at Advanced Photon Source (APS). Data were reduced with Denzo and Scalepack.20 For phase ˚ was determination, the resolution range from 25.0 to 1.9 A chosen. Three of four expected sites were found using Solve.21 Phases were improved, and 90% of the model was automatically built using Resolve.22 Several rounds of manual building using O,23 followed by refinement

578

Crystal Structure of Im5–E5-CRD

Table 1. Statistics of data collection and refinement Inflection Crystals Space group Unit cell ˚) Resolution (A ˚) Wavelength (A Unique reflections Redundancy Completeness (%) Average I/s (I)a Rsymb (%) Refinement ˚) Resolution (A Reflections (free) Rcrystalc (Rfreed) (%) ˚) rmsd bonds (A rmsd angles (deg.) Ramachandran statistics Favored/allowed/generous Disallowed (%)

25.0–1.9 0.9796 20,134 10.4 98.8(97.1) 43.5(30.9) 3.7(5.5)

Peak

25.0–1.9 0.9791 20,151 10.6 99.2(98.2) 42.0(33.0) 3.8(5.7)

Hard remote

Native

25.0–1.9 0.9680 20,139 10.7 98.9(96.0) 41.4(31.3) 3.6(5.9)

I222 61.79, 73.64, 110.11 90.00, 90.00, 90.00 25.0–1.15 1.0000 87,669 6.2 98.2(82.8) 20.1(8.8) 3.4(19.1) 25.0–1.15 75,886 (6733) 19.2 (19.8) 0.0044 1.30 92.8/7.2/0.0 0.0

Mean figure-of-merit (FOM) for phasingZ0.65 from SOLVE and 0.79 after RESOLVE. a I/s(I) is the mean reflection intensity/estimated error. b Rsym Z SjIKhIij=SI, where I is the intensity of an individual reflection and hIi is the average intensity over symmetry equivalents. c Rcrystal Z SjjFo jKjFc jj=SjFo j, where Fo and Fc are the observed and calculated structure factor amplitudes. d Rfree is equivalent to Rcrystal but calculated for a randomly chosen set of reflections that were omitted from the refinement process.

with CNS,24 resulted in a final model with an R-factor of 19.2% (Rfree 19.8%). Final refinement statistics are given in Table 1. Assay of inhibition of RNase activity of E5-CRD by Im5 and its mutants A stem–loop RNA with the sequence of 5 0 -CACGGCUGUAAACCGUG-3 0 , in which the first and last five nucleotides form the stem and the middle seven nucleotides form the loop, was synthesized chemically and purchased from Dharmacon Research (Boulder, CO). The RNA was radiolabeled as reported.9 A mixture of 100 mM of Im5, 2 mM of 32P radiolabeled stem–loop RNA, and 0–100 mM of E5-CRD in 10 ml of buffer (100 mM NaCl, 20 mM Tris–HCl (pH 7.6), 5 mM MgCl2) was incubated at room temperature for 10 min. The reaction was stopped by addition of 10 ml of gel loading buffer and heating at 95 8C for 5 min. The sample was then loaded onto a 20% (w/v) denaturing polyacrylamide gel. Electrophoresis was carried out in a 1!TBE buffer at 400 V for 3 h. The gel was dried and exposed to a phosphor screen (Fisher Scientific). Gel bands were visualized with a phosphorimager (Amersham Biosciences) and analyzed with ImageQuant (Molecular Dynamics, Sunnyvale, CA). Protein Data Bank accession numbers The coordinates of the structure of Im5–E5-CRD complex have been deposited in the RSCB Protein Data Bank with accession code 2FHZ.

Acknowledgements This research was supported by NIH grant CA90954 and by start-up funds from the University

of Illinois at Urbana-Champaign. We thank Howard H. Robinson for collecting MAD data at Brookhaven and staffs of beamline 19ID at APS (S. Ginell, A. Joachimiak and Y. Kim) for their support during native data collection. We also thank Dr D. Brunner for the ColE5-099 plasmid and D. Kranz for helpful discussions and critical reading of the manuscript.

References 1. James, R., Kleanthous, C. & Moore, G. R. (1996). The biology of E colicins: paradigms and paradoxes. Microbiology, 142, 1569–1580. 2. Cramer, W. A., Cohen, F. S., Merrill, A. R. & Song, H. Y. (1990). Structure and dynamics of the colicin E1 channel. Mol. Microbiol. 4, 519–526. 3. Schaller, K. & Nomura, M. (1976). Colicin E2 is DNA endonuclease. Proc. Natl Acad. Sci. USA, 73, 3989–3993. 4. Kleanthous, C., Kuhlmann, U. C., Pommer, A. J., Ferguson, N., Radford, S. E., Moore, G. R. et al. (1999). Structural and mechanistic basis of immunity toward endonuclease colicins. Nature Struct. Biol. 6, 243–252. 5. Boon, T. (1971). Inactivation of ribosomes in vitro by colicin E 3. Proc. Natl Acad. Sci. USA, 68, 2421–2425. 6. Senior, B. W. & Holland, I. B. (1971). Effect of colicin E3 upon the 30 S ribosomal subunit of Escherichia coli. Proc. Natl Acad. Sci. USA, 68, 959–963. 7. Di Masi, D. R., White, J. C., Schnaitman, C. A. & Bradbeer, C. (1973). Transport of vitamin B12 in Escherichia coli: common receptor sites for vitamin B12 and the E colicins on the outer membrane of the cell envelope. J. Bacteriol. 115, 506–513. 8. Ogawa, T., Tomita, K., Ueda, T., Watanabe, K., Uozumi, T. & Masaki, H. (1999). A cytotoxic ribonuclease targeting specific transfer RNA anticodons. Science, 283, 2097–2100.

579

Crystal Structure of Im5–E5-CRD

9. Lin, Y. L., Elias, Y. & Huang, R. H. (2005). Structural and mutational studies of the catalytic domain of colicin E5: a tRNA-specific ribonuclease. Biochemistry, 44, 10494–10500. 10. Wallis, R., Leung, K. Y., Pommer, A. J., Videler, H., Moore, G. R., James, R. & Kleanthous, C. (1995). Protein–protein interactions in colicin E9 DNaseimmunity protein complexes. 2. Cognate and noncognate interactions that span the millimolar to femtomolar affinity range. Biochemistry, 34, 13751–13759. 11. Ko, T. P., Liao, C. C., Ku, W. Y., Chak, K. F. & Yuan, H. S. (1999). The crystal structure of the DNase domain of colicin E7 in complex with its inhibitor Im7 protein. Struct. Fold. Des. 7, 91–102. 12. Hsia, K. C., Chak, K. F., Liang, P. H., Cheng, Y. S., Ku, W. Y. & Yuan, H. S. (2004). DNA binding and degradation by the HNH protein ColE7. Structure (Camb.), 12, 205–214. 13. Carr, S., Walker, D., James, R., Kleanthous, C. & Hemmings, A. M. (2000). Inhibition of a ribosomeinactivating ribonuclease: the crystal structure of the cytotoxic domain of colicin E3 in complex with its immunity protein. Struct. Fold. Des. 8, 949–960. 14. Graille, M., Mora, L., Buckingham, R. H., Van Tilbeurgh, H. & De Zamaroczy, M. (2004). Structural inhibition of the colicin D tRNase by the tRNAmimicking immunity protein. EMBO J. 23, 1474–1482. 15. Yajima, S., Nakanishi, K., Takahashi, K., Ogawa, T., Hidaka, M., Kezuka, Y. et al. (2004). Relation between tRNase activity and the structure of colicin D according to X-ray crystallography. Biochem. Biophys. Res. Commun. 322, 966–973. 16. Tomita, K., Ogawa, T., Uozumi, T., Watanabe, K. & Masaki, H. (2000). A cytotoxic ribonuclease which

17. 18. 19.

20.

21. 22. 23.

24.

25. 26.

specifically cleaves four isoaccepting arginine tRNAs at their anticodon loops. Proc. Natl Acad. Sci. USA, 97, 8278–8283. Masaki, H. & Ogawa, T. (2002). The modes of action of colicins E5 and D, and related cytotoxic tRNases. Biochimie, 84, 433–438. Holm, L. & Sander, C. (1993). Protein structure comparison by allignment of distance matrices. J. Mol. Biol. 233, 123–138. Soelaiman, S., Jakes, K., Wu, N., Li, C. & Shoham, M. (2001). Crystal structure of colicin E3: implications for cell entry and ribosome inactivation. Mol. Cell, 8, 1053–1062. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode Methods in Enzymology, vol. 277, pp. 307–326. Academic Press, San Diego, CA. Terwilliger, T. C. & Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallog. sect. D, 55, 849–861. Terwilliger, T. C. (1999). Reciprocal-space solvent flattening. Acta Crystallog. sect. D, 55, 1863–1871. Jones, T. A., Zou, J.-Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110–119. Bru¨nger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921. DeLano, W. L. (2002). The PyMOL Molecular Graphic System 2002, DeLano Scientific, San Carlos, CA. Carson, M. (1987). Ribbon models of macromolecules. J. Mol. Graph. 5, 103–106.

Edited by I. Wilson (Received 3 January 2006; received in revised form 31 January 2006; accepted 6 February 2006) Available online 24 February 2006