Werner helicase wings DNA binding

Summary

In this issue of Structure, Kitano et al. describe the structure of the DNA-bound winged-helix domain from the Werner helicase. This structure of a RecQ/DNA complex offers insights into the DNA unwinding mechanisms of RecQ family helicases.

RecQ DNA helicases are central genome maintenance enzymes found in nearly all organisms. Humans have five different RecQ proteins: RecQ1, BLM, WRN, RecQ4, and RecQ5. The study of RecQ proteins has gained focus due to three rare genetic diseases, Bloom’s, Werner’s, and Rothmund-Thomson syndromes, which arise, respectively, from mutations in the genes that encode BLM, WRN, and RecQ4 proteins. Although they are distinct in their clinical manifestations, these diseases share common characteristics of increased genomic instability and cancer predisposition, consistent with roles for each RecQ helicase in cellular genome maintenance (Vindigni and Hickson, 2009). An important, but thus far ill-defined, feature of RecQ proteins is the mechanism underlying DNA binding and unwinding. In this issue of Structure, Kitano et al describe the first molecular image of any RecQ protein domain bound to DNA, offering new insights into RecQ function (Kitano et al., 2010).

RecQ proteins catalyze ATP-dependent unwinding of DNA, preferentially acting on DNA structures that resemble replication and recombination intermediates such as replication forks, Holliday junctions, D-loops, and G-quadruplexes. Most RecQ proteins share three conserved elements: helicase, RecQ-C-terminal (RQC, which includes Zn²⁺-binding and winged-helix (WH) subdomains), and “helicase-and-RNaseD-like-C-terminal” (HRDC) domains. The helicase and RQC domains generally comprise a minimal “catalytic core” domain, whereas the HRDC domain is often dispensable for helicase activity. Each of the RecQ domains has been projected to play a role in DNA binding, although structural models that define how such binding occurs and is coordinated during DNA unwinding have been hampered by the lack of direct structural studies of RecQ/DNA complexes. In addition to these core domains found in most RecQ family members, some RecQ proteins also contain domains that are important for proper subcellular localization or that encode elements facilitating protein interactions, oligomerization, or extra enzymatic activities, such as the exonuclease domain of WRN.

To understand how RecQ proteins function, several labs have used structural and biochemical approaches to examine model RecQ enzymes. These efforts have revealed the structures of each of the major RecQ domains, including isolated domains from human WRN (Hu et al., 2005; Kitano et al., 2007; Perry et al., 2006), and RecQ1 (Pike et al., 2009), two bacterial RecQ proteins (Escherichia coli (Bernstein and Keck, 2005; Bernstein et al., 2003) and Deinococcus radiodurans (Killoran and Keck, 2008)), and Saccharomyces cerevisiae Sgs1 (Liu et al., 1999). With this wealth of available structural information from the past decade of studies on RecQ proteins, the new X-ray crystal structure of the DNA-bound WRN WH domain described in this month’s Structure provides represents an important advance for understanding how full-length RecQ proteins might bind and unwind DNA.

The WRN WH/DNA structure reveals two unexpected features. The first is that the WH domain binds duplex DNA in an unconventional way relative to other characterized WH-containing proteins. In most cases, a prominent helix of the WH fold protrudes into the major groove of duplex DNA to form a complex. This arrangement facilitates sequence-specific DNA binding that can induce a bend in the DNA (Gajiwala and Burley, 2000). In contrast to this binding mode, the WRN recognition helix does not appear to be directly involved in DNA binding. Instead, an interhelical loop (the α2–α3 loop) serves as the prominent DNA binding site by interfacing with the major groove of the DNA through a series of hydrogen bonds and salt bridge contacts to backbone phosphates. Interestingly, the recognition helix in the human RecQ1 catalytic core structure is buried and not surface accessible (Pike et al., 2009), however the analogous α2–α3 loop of RecQ1 is surface exposed, making it available for a similar DNA binding arrangement (Figure 1A, left). In contrast, the WH domain arrangement in the E. coli RecQ catalytic core structure differs from that of RecQ1, leaving both the recognition helix and the analogous α2-α3 loop surface exposed in the E. coli enzyme (Bernstein et al., 2003; Pike et al., 2009) (Figure 1A, right). Kitano and coworkers suggest that use of the unusual DNA binding surface used in the WRN WH domain may be important for the broad, non-sequence-specific activities of RecQ proteins.

Figure 1.

Comparison of RecQ helicases. A. The WRN WH/DNA complex (blue) is shown overlaid with human RecQ1 catalytic core (orange, left), and with E. coli RecQ catalytic core (tan, right). B. Close-up showing the β-wing and important residues.

The second unexpected feature is that the structure captured a partially unwound form of the DNA duplex in which the β-wing of the WH fold appears to have wedged apart the terminal basepair on the DNA duplex (Figure 1B). Subsequent biochemical analyses indicate that three hydrophobic residues in the β-wing (Figure 1B) are important for WRN WH DNA binding affinity. These observations dovetail with earlier studies involving human RecQ1 that delineated an important role for the β-wing in DNA unwinding; mutating a phenylalanine at the β-wing tip abrogates DNA unwinding (Pike et al., 2009). Again in contrast with the human RecQ proteins, the E. coli RecQ WH β-wing is significantly smaller and lacks analogous hydrophobic residues when compared to either WRN or RecQ1 (Figure 1B) and mutation of a histidine at the tip of the E. coli RecQ β-wing does not alter helicase activity (Pike et al., 2009). These observations leave open the question of whether all RecQ proteins use a common mechanism for DNA binding and unwinding. Nonetheless, the observation that the WH β-wing is important for DNA binding and unwinding in WRN and RecQ1 supports a model in which the WH domain splits the DNA duplex using the β-wing as a wedge. One can imagine that the helicase motor would drive this reaction by pulling on the single-stranded DNA that would be projecting from the partially unwound duplex. Consistent with this model, superimposition of the WRN WH/DNA structure onto the RecQ1 structure indicates that the 3′ end of one the strands could be oriented toward the motor (Figure 1A).

Although this new study reveals a major role for the WH domain in DNA unwinding, many questions remain. One major question is whether the model posited above, in which RecQ helicase domain translocation on single-stranded DNA drives unwinding by the WH domain, is complete or if there are additional undiscovered components. Additional structural and biochemical experiments with catalytically active RecQ proteins will be essential for fleshing out the validity of this or other models. A second question is whether use of the WH β-wing as a DNA-unwinding wedge is conserved in all RecQ family members. As described above, significant structural differences between E. coli RecQ and the human RecQ proteins studied to date indicate that the β-wing is not as important in the E. coli RecQ enzyme. Moreover, the orientation of the E. coli RecQ WH domain relative to the rest of the protein is different than that observed in human RecQ1 (Pike et al., 2009), which may point to a functional difference in how the domains act during DNA binding and unwinding. Alternatively, it may be that DNA binding reorients the E. coli RecQ WH domain to be more similar to that of RecQ1. A final major question arises from the fact that all RecQ structures produced to date are missing one or more domains. This leaves the question of how any full-length RecQ protein coordinates multiple domains to bind and unwind DNA unresolved. This issue is further complicated by the fact that different RecQ proteins have distinct substrate preferences and the domains of these diverse family members play different roles in substrate binding. While this new structure takes us one important step closer toward understanding how RecQ proteins function, there is still much to learn about the structure and mechanism of RecQ DNA helicases.