Altering the DNA-binding specificity of Mu transposase in vitro

Deciphering the Roles of Multicomponent Recognition Signals by the AAA+ Unfoldase ClpX

ATP-dependent protein remodeling and unfolding enzymes are key participants in protein metabolism in all cells. How these often-destructive enzymes specifically recognize target protein complexes is poorly understood. Here, we use the well-studied AAA+ unfoldase-substrate pair, Escherichia coli ClpX and MuA transposase, to address how these powerful enzymes recognize target protein complexes. We demonstrate that the final transposition product, which is a DNA-bound tetramer of MuA, is preferentially recognized over the monomeric apo-protein through its multivalent display of ClpX recognition tags. The important peptide tags include one at the C-terminus ("C-tag") that binds the ClpX pore and a second one (enhancement or "E-tag") that binds the ClpX N-terminal domain. We construct a chimeric protein to interrogate subunit-specific contributions of these tags. Efficient remodeling of MuA tetramers requires ClpX to contact a minimum of three tags (one C-tag and two or more E-tags), and that these tags are contributed by different subunits within the tetramer. The individual recognition peptides bind ClpX weakly (KD>70 μM) but impart a high-affinity interaction (KD~1.0 μM) when combined in the MuA tetramer. When the weak C-tag signal is replaced with a stronger recognition tag, the E-tags become unnecessary and ClpX's preference for the complex over MuA monomers is eliminated. Additionally, because the spatial orientation of the tags is predicted to change during the final step of transposition, this recognition strategy suggests how AAA+ unfoldases specifically distinguish the completed "end-stage" form of a particular complex for the ideal biological outcome.

The μ transpososome structure sheds light on DDE recombinase evolution

Studies of bacteriophage Mu transposition paved the way for understanding retroviral integration and V(D)J recombination as well as many other DNA transposition reactions. Here we report the structure of the Mu transpososome--Mu transposase (MuA) in complex with bacteriophage DNA ends and target DNA--determined from data that extend anisotropically to 5.2 Å, 5.2 Å and 3.7 Å resolution, in conjunction with previously determined structures of individual domains. The highly intertwined structure illustrates why chemical activity depends on formation of the synaptic complex, and reveals that individual domains have different roles when bound to different sites. The structure also provides explanations for the increased stability of the final product complex and for its preferential recognition by the ATP-dependent unfoldase ClpX. Although MuA and many other recombinases share a structurally conserved 'DDE' catalytic domain, comparisons among the limited set of available complex structures indicate that some conserved features, such as catalysis in trans and target DNA bending, arose through convergent evolution because they are important for function.

The AAA+ ClpX machine unfolds a keystone subunit to remodel the Mu transpososome

A hyperstable complex of the tetrameric MuA transposase with recombined DNA must be remodeled to allow subsequent DNA replication. ClpX, a AAA+ enzyme, fulfills this function by unfolding one transpososome subunit. Which MuA subunit is extracted, and how complex destabilization relates to establishment of the correct directionality (left to right) of Mu replication, is not known. Here, using altered-specificity MuA proteins/DNA sites, we demonstrate that transpososome destabilization requires preferential ClpX unfolding of either the catalytic-left or catalytic-right subunits, which make extensive intersubunit contacts in the tetramer. In contrast, ClpX recognizes the other two subunits in the tetramer much less efficiently, and their extraction does not substantially destabilize the complex. Thus, ClpX targets the most stable structural components of the complex. Left-end biased Mu replication is not, however, determined by ClpX's intrinsic subunit preference. The specific targeting of a stabilizing "keystone subunit" within a complex for unfolding is an attractive general mechanism for remodeling by AAA+ enzymes.

Changing the recognition site of a conjugative relaxase by rational design

TrwC is a relaxase protein, which starts and finishes DNA processing during bacterial conjugation in plasmid R388. TrwC recognizes a specific sequence of DNA (25 nucleotides) in the donor cell: the nic-site. As a model example, a single transversion C24G in nic avoids DNA processing by TrwC. Using this simple model, our objective was to obtain a proof of principle that TrwC specificity can be changed. Several structures of DNA-TrwC complexes were used as reference to design a focused saturation mutagenesis library (NNK) randomizing amino acid Lys262, since its side chain seems to sterically hinder the recognition of the C24G nic mutation by wild-type TrwC. Using bacterial conjugation as an in vivo selection system, several TrwC variants were found that show changes in substrate specificity. These variants were also tested in a competitive assay to evaluate their conjugation efficiency.

Interactions of phage Mu enhancer and termini that specify the assembly of a topologically unique interwrapped transpososome

The higher-order DNA-protein complex that carries out the chemical steps of phage Mu transposition is organized by bridging interactions among three DNA sites, the left (L) and right (R) ends of Mu, and an enhancer element (E), mediated by the transposase protein MuA. A subset of the six subunits of MuA associated with their cognate sub-sites at L and R communicate with the enhancer to trigger the stepwise assembly of the functional transpososome. The DNA follows a well-defined path within the transpososome, trapping five supercoil nodes comprising two E-R crossings, one E-L crossing and two L-R crossings. The enhancer is a critical DNA element in specifying the unique interwrapped topology of the three-site LER synapse. In this study, we used multiple strategies to characterize Mu end-enhancer interactions to extend, modify and refine those inferred from earlier analyses. Directed placement of transposase subunits at their cognate sub-sites at L and R, analysis of the protein composition of transpososomes thus obtained, and their characterization using topological methods define the following interactions. R1-E interaction is essential to promote transpososome assembly, R3-E interaction contributes to the native topology of the transpososome, and L1-E and R2-E interactions are not required for assembly. The data on L2-E and L3-E interactions are not unequivocal. If they do occur, either one is sufficient to support the assembly process. Our results are consistent with two R-E and perhaps one L-E, being responsible for the three DNA crossings between the enhancer and the left and right ends of Mu. A 3D representation of the interwrapped complex (IW) obtained by modeling is consistent with these results. The model reveals straightforward geometric and topological relationships between the IW complex and a more relaxed enhancer-independent V-form of the transpososome assembled under altered reaction conditions.

Directed evolution of a recombinase for improved genomic integration at a native human sequence

We previously established that a unidirectional site-specific recombinase, the phage phiC31 integrase, can mediate integration into mammalian chromosomes. The enzyme directs integration of plasmids bearing the phage attB recognition site into pseudo attP sites, a set of native sequences related to the phage attP recognition site. Here we use two cycles of DNA shuffling and screening in Escherichia coli to obtain evolved integrases that possess significant improvements in integration frequency and sequence specificity at a pseudo attP sequence located on human chromosome 8, when measured in the native genomic environment of living human cells. Such integrases represent custom integration tools that will be useful for modifying the genomes of higher eukaryotic cells.

Domain III function of Mu transposase analysed by directed placement of subunits within the transpososome

Assembly of the functional tetrameric form of Mu transposase (MuA protein) at the two att ends of Mu depends on interaction of MuA with multiple att and enhancer sites on supercoiled DNA, and is stimulated by MuB protein. The N-terminal domain I of MuA harbours distinct regions for interaction with the att ends and enhancer; the C-terminal domain III contains separate regions essential for tetramer assembly and interaction with MuB protein (IIIalpha and IIIbeta, respectively). Although the central domain II (the 'DDE' domain) of MuA harbours the known catalytic DDE residues, a 26 amino acid peptide within IIIalpha also has a non-specific DNA binding and nuclease activity which has been implicated in catalysis. One model proposes that active sites for Mu transposition are assembled by sharing structural/catalytic residues between domains II and III present on separate MuA monomers within the MuA tetramer. We have used substrates with altered att sites and mixtures of MuA proteins with either wild-type or altered att DNA binding specificities, to create tetrameric arrangements wherein specific MuA subunits are nonfunctional in II, IIIalpha or IIIbeta domains. From the ability of these oriented tetramers to carry out DNA cleavage and strand transfer we conclude that domain IIIalpha or IIIbeta function is not unique to a specific subunit within the tetramer, indicative of a structural rather than a catalytic function for domain III in Mu transposition.

Examination of the transcription factor NtcA-binding motif by in vitro selection of DNA sequences from a random library

A recursive in vitro selection among random DNA sequences was used for analysis of the cyanobacterial transcription factor NtcA-binding motifs. An eight-base palindromic sequence, TGTA-(N(8))-TACA, was found to be the optimal NtcA-binding sequence. The more divergent the binding sequences, compared to this consensus sequence, the lower the NtcA affinity. The second and third bases in each four-nucleotide half of the consensus sequence were crucial for NtcA binding, and they were in general highly conserved. The most frequently occurring sequence in the middle weakly conserved region was similar to that of the NtcA-binding motif of the Anabaena sp. strain PCC 7120 glnA gene, previously known to have high affinity for NtcA. This indicates that the middle sequences were selected for high NtcA affinity. Analysis of natural NtcA-binding motifs showed that these could be classified into two groups based on differences in recognition consensus sequences. It is suggested that NtcA naturally recognizes different DNA-binding motifs, or has differential affinities to these sequences under different physiological conditions.