Laboratory of Protein Structure: Nowotny Laboratory

Description of Current Research


   Our laboratory focuses on structural and biochemical studies of nucleic acid enzymes using protein crystallography as a primary method. The key results obtained recently in our laboratory concern three proteins: RNases H2, reverse transcriptases and UvrA.

1. Structural studies of bacterial RNases H2
   RNases H are small nucleases that specifically hydrolyze RNA in RNA/DNA hybrids. They are divided into two types—RNases H1 and RNases H2—that have a similar structure of the catalytic core but different domain organization and biochemical properties. The most important feature of RNases H2, differentiating them from type 1 enzymes, is their ability to cleave single ribonucleotides embedded in the DNA. Such single ribonucleotides occur quite frequently in genomic DNA and most often result from misincorporation by DNA polymerases. They must be removed to maintain genomic stability, and RNase H2 is the only enzyme that can initiate this process by cleaving the phosphate linkage on the 5’ side of the ribonucleotide. The removal is completed by the second cut on the 3’ side of the RNA by FEN-1 endonuclease.
   The mechanism of the specific cleavage of single ribonucleotides by RNase H2 was previously unknown. To elucidate this, we solved the crystal structures of Thermotoga maritima RNase H2 in complex with the nucleic acid substrate. The results showed that the nucleic acid is bound in a cleft between the N-terminal catalytic domain and C-terminal helical domain. The key element that ensures the substrate specificity of the enzyme is the recognition mechanism for a (5’)RNA-DNA(3’) junction. The RNA residue of the junction forms a network of interactions between its 2’-OH and the backbone of three protein residues—two glycines and an arginine—that form an element we call the “GRG motif.” The hydroxyl group of an absolutely conserved tyrosine residue from the C-terminal domain also forms a hydrogen bond with the 2’-OH group. This tyrosine also interacts with the second group of the junction, forming a stacking interaction with its ribose ring.
   This interaction can be efficient only if a 2’-OH group is absent from the ring and therefore is selective for deoxyribonucleotide in the second position of the junction. The stacking interaction leads to a deformation of the nucleic acid, changing the conformation of the phosphodiester backbone of the RNA-DNA junction. Because of this deformation, the phosphate group of the junction can participate in the coordination of a metal ion at the active site. This mechanism ensures very stringent substrate specificity. Only when a correct substrate is present (e.g., an RNA-DNA junction) that can be properly deformed can the metal ion be coordinated at the active site and the reaction proceed.
   The active site of RNase H2 is formed by four conserved carboxylate residues. In the wild-type structure solved in the presence of Ca2+ ions, we observed three ions at the active site. Two of the ions occupied positions very similar to the two catalytic metal ions in related enzymes, and we assume that RNase H2 uses a canonical two-metal ion mechanism. In this mechanism, one metal ion activates the attacking nucleophile, and the second ion stabilizes the transition state and reaction product.
   The studies of RNases H2 were performed in collaboration with Dr. Robert Crouch (National Institutes of Health, USA).

2. Structural studies of reverse transcriptases
   Reverse transcription involves the conversion of singlestranded RNA to double-stranded DNA and is essential for the proliferation of retrotransposons and retroviruses, such as human immunodeficiency virus (HIV-1). Reverse transcription is an intricate multi-step process catalyzed by very versatile enzymes called reverse transcriptases (RTs). These enzymes possess two activities: DNA polymerase synthesizes the new DNA, and RNase H degrades the RNA strand in the RNA/DNA intermediates of the reaction, thus removing the original genetic information.
   Retroviral RTs can be divided into two groups, based on their architecture. Dimeric enzymes, such as HIV-1 RT, are very well characterized structurally and biochemically, but the mechanism of monomeric RTs is less well understood. To characterize the mechanism, we solved the crystal structure of monomeric RT from xenotropic murine leukemia virus-related virus (XMRV) in complex with an RNA/DNA substrate. The structure comprised the polymerase domain, but the RNase H was disordered and hence not visible. The structure revealed that the active site and substrate contacts around it were well conserved between monomeric and dimeric RTs. Further toward the RNase H domain, however, substrate binding was mediated by a different set of residues. Our structure also revealed the role of a “pin” structure that guided the trajectory of the template strand and was important for DNA polymerase processivity. We also explained the ability of XMRV RT to perform so-called strand displacement DNA synthesis, during which a nucleic acid that is hybridized with the template ahead of the polymerase active site is removed concurrently with polymerization.
   In collaboration with Janusz Bujnicki and Dmitri Svergun we used small-angle X-ray diffraction data and combined them with the crystal structures to model the full-length enzyme.
These studies revealed that the RNase H domain was very mobile in the absence of nucleic acid and became organized on the substrate when it was present. Transient and infrequent interactions between the RNase H domain and substrate appear to be a universal feature of RTs. It likely allows the enzyme to regulate RNase H activity, which has to perform very precise cuts at several stages of reverse transcription.
However, the mechanism of the regulation of the RNase H interaction with the substrate is very different for the HIV-1 enzyme. Unlike in monomeric XMRV RT, the RNase H domain in dimeric RTs is rigidly placed on the non-catalytic subunit, and the substrate has to be deformed to reach its active site.
   Reverse transcriptases, therefore, exhibit an intriguing variety of mechanisms to perform their function.
   The studies of XMRV RT were performed in collaboration with Dr. Stuart Le Grice (National Institutes of Health, USA).

3. Structural and biochemical studies of UvrA DNA repair protein
   DNA constantly undergoes detrimental chemical modifications (also called DNA damage) that occur spontaneously or are caused by physical and chemical factors.
   To maintain the genetic stability of the cell and protect the organism, these modifications need to be corrected. One of the primary pathways to achieve this is nucleotide excision repair (NER). The most important feature of NER is its ability to recognize a wide variety of DNA lesions of unrelated chemical structures. Different proteins are involved in NER in bacteria and eukaryotes, but the principle is the same. The site of damage is located, its presence is verified, and the DNA is incised on both sites of the lesion. The DNA fragment that contains the lesion is removed by a helicase, and the gap is filled by DNA polymerase.
   n bacteria, the first component of the pathway, which locates the lesion, is UvrA protein. It is a dimeric adenosine triphosphatase (ATPase) from the ATP-binding cassette (ABC) family. After the damage is found, the DNA is handed over to UvrB, which possesses weak helicase activity and verifies the presence of the lesion. UvrC nuclease executes the two cuts on the two sides of the modification.
   The key unanswered question in NER is how its remarkably wide specificity is achieved. To elucidate this, we sought to solve the crystal structure(s) of a UvrA protein in complex with modified DNA. In our extensive crystallization trials, we used UvrA proteins from two bacterial species and DNA oligonucleotides that contained a single thymine residue with a fluorescein moiety attached through a flexible tether. We used DNA duplexes with a modified thymine residue in one of the DNA strands and duplexes that consisted of palindromic oligonucleotides that contain symmetrically placed modified thymines in both strands. We reasoned that the symmetry of such DNAs would reflect the two-fold symmetry of the UvrA dimer and hence promote crystallization. Indeed, we only obtained crystals with the palindromic oligonucleotides.
   We then used biochemical assays to verify that each of the strands of the palindromic substrates can be independently processed by the NER machinery that consisted of UvrA, UvrB, and UvrC. The crystals diffracted X-rays up to 2.9 Å resolution, and the structure was solved using molecular replacement. In the structure, the DNA is bound in a cleft that runs across the UvrA dimer. The interactions between the protein and nucleic acid are formed almost exclusively with the terminal regions of the DNA duplex. We identified a conserved, positively charged patch on the surface of the protein that forms extensive contacts with the DNA backbone.


Model of full-length monomeric XMRV RT prepared based on crystal structures and small-angle X-ray scattering data. The DNA polymerase domain is shown in pink, cyan, and yellow. Connection and RNase H domains is green and orange. The RNA/DNA substrate is shown red and blue.


   The key to DNA damage recognition by UvrA is the conformation of the DNA. The duplex is bent by approximately 15 degrees, stretched in the middle, and unwound. Only this deformed conformation is complementary with the protein surface. The DNA deformations we observe are also often seen in various modified DNAs in free, unbound form. Unwinding is a particularly common feature of many damaged DNAs.
   Therefore, we proposed that UvrA uses an indirect readout mechanism to detect the presence of the damage. The protein senses the deformations of the DNA caused by the lesion. At the same time, it may also adjust those deformations so that the duplex fits to the protein surface. Modified DNA duplexes are known to be more flexible and easier to deform.
   UvrA probes the conformation of the DNA symmetrically on both sides of the lesion without directly interacting with the site of modification itself. Its dimeric structure is ideally suited for this purpose, but the symmetrical damage detection does not provide information about which strand is damaged and needs to be incised. This is most likely the role of the UvrB protein, which is recruited to the DNA after UvrA finds the damage site.
The mechanism of indirect readout we described is unique.
   Eukaryotic NER proteins, for which crystal structures are available, such as UV-DDB and XPC/HR25 complexes, form specific contacts with the site of lesion and use base flipping to probe the strength of the base pair hydrogen bonds to detect the damage.