1Structural Genomics Consortium, University of Toronto, 101 College Street, Toronto, Ontario M5G 1L7, Canada.2Department of Physiology, University of Toronto, Toronto, Ontario M5S 1A8, Canada.
Correspondence should be addressed to Jinrong Min: email@example.com
The PWWP domain is a small structural unit that can bind both DNA and histone, hence involved in chromatin targeting. Here we summarize the structural features of the human PWWP domains and highlight their ligand binding ability from a structural view. Their association with biological functions is also discussed.
The PWWP domain is a structural unit of 100-150 amino acids and named after the conserved Pro-Trp-Trp-Pro motif (Stec et al., 1998; Stec et al., 2000). It is found throughout eukaryotes, ranging from unicellular organisms to human. There are more than 20 PWWP domain containing proteins in the human genome, most of which are chromatin associated. Moreover, the PWWP domain is often found to coexist with other chromatin-associated domains, such as the PHD domain and Bromodomain, in a single polypeptide. More and more studies are revealing the involvement of the PWWP domain in chromatin biology. Here, we will first summarize the structural features of the PWWP domains, then move to their ligand binding ability, and finally discuss their association with biological functions.
The PWWP domain belongs to the Royal superfamily, which also includes chromodomain, Tudor domain, and MBT domain (Maurer-Stroh et al., 2003). The Royal superfamily shares a common structural feature, i.e., an antiparallel β-barrel-like structure formed by 3-5 β-strands. For PWWP domains, the β-barrel is consisted with 5 antiparallel β-strands (β1-β5). The linker between β2 and β3 may also contain additional secondary elements. The unique feature of the PWWP domain is the presence of a α-bundle of 1-6 α-helixes following the β-barrel (Wu et al., 2011). This α-bundle region is very variable and diverse at both the sequence and structural levels. Therefore, only the β-barrel part of the PWWP domain could be reliably predicted in protein domain databases, such as SMART and Human Protein Reference Database. The conserved Pro-Trp-Trp-Pro motif is located at the beginning of the β2 strand and it is packed against the α-bundle, indicating its critical roles in protein folding and stability. Generally speaking, the PWWP domain can fold as an independent functional unit. However, recent studies reveal that the PWWP domain of ZMYND11 (also known as BS69) functions together with the preceding Bromodomain and zinc finger as an integral functional module (Wang et al., 2014; Wen et al., 2014).
DNA binding ability of the PWWP domain
The first three-dimensional structure of a PWWP domain was determined for the mouse DNA methyltransferase Dnmt3b (Qiu et al., 2002). Structural analysis of this PWWP domain revealed a prominent positively charged surface, suggesting a potential role in DNA binding (Qiu et al., 2002). Moreover, it is a common feature that the PWWP domain is rich in lysine and arginine residues and has a predicted pI of more than 9. Indeed, other PWWP domains have been reported to be able to bind DNA, such as HDGF (Lukasik et al., 2006; Yang and Everett, 2007), MSH6 (Laguri et al., 2008), PSIP1 (also known as LEDGF and p75) (Eidahl et al., 2013; van Nuland et al., 2013), and ZMYND11 (Wang et al., 2014). In vitro binding assays revealed that the PWWP domain binds DNA in a nonspecific manner (Qiu et al., 2002; Lukasik et al., 2006; Laguri et al., 2008).
To date, no structure of a PWWP-DNA complex is available in protein structure database. Facilitated by NMR chemical shift perturbation experiments, several groups tried to map the DNA binding site on different PWWP domains (HDGF (Lukasik et al., 2006), MSH6 (Laguri et al., 2008), and PSIP1 (Eidahl et al., 2013; van Nuland et al., 2013)). Notably, the perturbed residues are consistently localized on one side of the protein, centering on the β1-β2 arch region and the PWWP motif, which also overlaps with the patch of highly positively charged surface area.
Histone binding ability of the PWWP domain
The structural similarity of the PWWP domain with other Royal superfamily members, which can recognize methylated lysine and arginine (Adams-Cioaba and Min, 2009), prompted people to propose the PWWP domain as a potential histone "reader" in 2005 (Nameki et al., 2005). Later on, it was demonstrated that the zebrafish Brpf1-PWWP domain can bind histones directly (Laue et al., 2008) and the fission yeast protein Pdp1 can recognize H4K20me specifically (Wang et al., 2009; Qiu et al., 2012). The crystal structure of BRPF1-H3K36me3 complex fully established the notion that the PWWP domain can recognize methylated histone (Vezzoli et al., 2010; Wu et al., 2011). After that, many other PWWP domains were reported to recognize methylated histone peptides, e.g., DNMT3A-H3K36me3 (Dhayalan et al., 2010), PSIP1-H3K36me3 (Pradeepa et al., 2012; Eidahl et al., 2013; van Nuland et al., 2013), MSH6-H3K36me3 (Li et al., 2013), HDGF2-H3K79me3 (Wu et al., 2011), HDGF2-H4K20me3 (Wu et al., 2011), and ZMYND11-H3.3K36me3 (Wen et al., 2014).
Figure 1. Complex structure of BRPF1-PWWP and H3K36me3 (PDB ID: 2X4X). Green: H3K36me3 peptide; magenta: β barrel; cyan: α-bundle; yellow: insertion elements between β2 and β3. Residues of the conserved PWWP motif, aromatic cage and methylated K36 are shown in stick mode.
Structural analysis of these PWWP-histone complex structures identified a conserved cage formed by three aromatic residues for methyl-lysine binding. The third residue (W/Y) of the PWWP motif and the residue (F/Y/W) immediately preceding this motif are involved in forming this cage. Another aromatic residue (F/Y/W) comes from the end of β3 strand (Wu et al., 2011). Sequence alignment reveals that most PWWP domains have this conserved cage for potential methylated-histone binding (Wu et al., 2011). However, several exceptions exist. The PWWP domains of RBBP1, RBBP1L1, MBD5, and NSD1 (N-terminal) have incomplete aromatic cage. RBBP1-PWWP indeed did not show any binding to methylated histone peptides (Gong et al., 2012).
The histone variant H3.3 possesses a signature motif 'S31...A87AIG90' that is distinct from the 'A31...S87AVM90' sequence signature of the canonical histone H3.1/2. The Bromo-Zinc-PWWP cassette of ZMYND11 specifically recognizes H3.3K36me3 (Wen et al., 2014). In addition to the K36me3 recognition by the conserved aromatic cage of PWWP domain, S31 of H3.3 is specifically recognized by the Bromodomain and Zinc finger domains synergistically. Glu251 from the Bromodomain and Asn266 from the Zinc finger form hydrogen bonds with the hydroxyl group of S31.
Nucleosome binding ability of the PWWP domain
The mentioned PWWP-binding sites H3K36, H3K79 and H4K20 are at a close proximity to DNA in nucleosome context. The fact that PWWP domain can bind both DNA and methylated histone suggests a synergy mechanism among these interactions. The binding affinities of PWWP domain towards histone peptide and DNA oligos are relatively too weak. But the affinity of PSIP1-PWWP for methylated nucleosomes (Kd ~1.5 mM) is four orders of magnitude higher than for a methylated peptide (Kd ~17 mM) and two orders higher than for isolated DNA (Kd ~150 mM) (van Nuland et al., 2013). In another study, similar trend was observed for PSIP1: H3KC36me3 nucleosome (Kd ~48 nM), H3K36me3 peptide (Kd >6.5 mM), and DNA (Kd ~1.5 mM) (Eidahl et al., 2013). Studies on ZMYND11 (Wen et al., 2014) and fission yeast Pdp1 (Qiu et al., 2012) show that a synergy was not detected in vitro assays using histone peptides and DNA oligos, so it is likely that the synergy between histone and DNA may contribute only at the nucleosomal level.
Strikingly, in all the available histone-PWWP complex structures, the histone peptides bind to PWWP domains in a similar orientation, that is, the peptides is almost perpendicular to the β4 strand. Further comparing the DNA binding surface identified by NMR and the histone binding mode suggests that PWWP domains may bind to nucleosome through a conserved mode, i.e., the PWWP domains adopt distinct and conserved interfaces to engage histone and DNA, respectively.
Functional roles of PWWP domain containing proteins
Although the PWWP domains mainly function as chromatin interactors and recognize H3K36me3 mark, they are involved in diverse biological processes, depending on the protein context they settled.
The first class of PWWP domain containing proteins is directly involved in modifying chromatin. DNMT3A/B are the de novo DNA methyltransferases responsible for the establishment of DNA methylation patterns during development (Chédin, 2011). The interaction of the PWWP domain with H3K36me3 is involved in targeting of DNMT3A to chromatin and guides DNA methylation (Dhayalan et al., 2010). Disruption of the PWWP domain abolishes the ability of DNMT3A and DNMT3B to methylate the major satellite repeats at pericentric heterochromatin (Chen et al., 2004; Ge et al., 2004). NSD1/2/3 are histone mono- and di-methylases for H3K36 (Wagner and Carpenter, 2012). Histone-modifying enzymes often contain "Reader" domains to bind their product to propagate the resultant mark along the chromatin. Indeed, the N-terminal PWWP domains of NSD2/NSD3 are able to recognize H3K36me2/3 in vitro (Wu et al., 2011). GLYR1 (also known as N-PAC), was found to be a cofactor of H3K4 demethylase LSD2, and GLYR1 positively regulates the H3K4 demethylase activity (Fang et al., 2013). BRPF1/2/3 are scaffold proteins of the MOZ/MORF histone acetyltransferase complexes (Ullah et al., 2008). Brpf1 is required for histone acetylation, the maintenance of cranial Hox gene expression, and the proper determination of pharyngeal segmental identities during development of zebrafish (Laue et al., 2008) and medaka fish (Hibiya et al., 2009). PWWP domains of BRPF1/2 are able to recognize H3K36me3 (Vezzoli et al., 2010; Wu et al., 2011), a mark present in transcribed Hox genes (Wang et al., 2007), consistent with BRPF1's crucial role in promoting the expression of these genes.
The second class of PWWP domain containing proteins is involved in DNA repair. MSH6 binds to H3K36me3 mark in a PWWP-dependent manner and this interaction mediates MutSa (one of the two DNA-mispair recognizing complexes) association with chromatin in cells (Li et al., 2013), hence contributing to DNA mismatch repair. MUM1 (also known as EXPAND1) contributes to the chromatin architecture and, via its direct interaction with the DNA damage mediator 53BP1, plays an accessory role to facilitate damage-induced chromatin changes and is important for efficient DNA repair and cell survival following DNA damage (Huen et al., 2010). In vivo chromatin association of MUM1 relies on its PWWP domain-mediated binding to nucleosomes, and in vitro assays revealed its preference of H3K36me3 (Wu et al., 2011). PSIP1 promotes the repair of double strand breaks (DSBs) by the homologous recombination repair pathway (Daugaard et al., 2012). PSIP1 is constitutively associated with chromatin through its PWWP domain that binds preferentially to methyl-lysine histone markers characteristic of active transcription units (probably H3K36me3). It also binds CtIP in a DNA damage-dependent manner, thereby enhancing its tethering to the active chromatin and facilitating its access to DNA DSBs.
ZMYND11 was originally identified as an adenovirus E1A-binding protein that inhibits the transactivation function of E1A (Hateboer et al., 1995). It also functions as a co-repressor of cellular transcription factors including the c-Myb (Ladendorff et al., 2001) and interacts with a set of chromatin remodeling factors, including ATP-dependent helicases, histone deacetylases, and histone methyltransferases (Velasco et al., 2006). Most recent study reveals that its Bromo-Zinc finger-PWWP cassette specific recognizes the histone variant H3.3K36me3 and regulates transcription elongation (Venkatesh and Workman, 2014; Wen et al., 2014). Although ZMYND11 is associated with highly expressed genes, it functions as an unconventional transcription co-repressor by modulating RNA polymerase II at the elongation stage. Overexpression of ZMYND11 suppresses cancer cell growth in vitro and tumour formation in mice (Wen et al., 2014).
PSIP1 was first isolated as an transcriptional co-activator (Ge et al., 1998a) and it associates with transcriptional activators and components of the basal transcriptional machinery including RNA pol II subunits (Ge et al., 1998b). It is also an essential subunit of the MLL complex in MLL oncogenic transformations via HOX gene regulation (Yokoyama and Cleary, 2008). PSIP1 short isoform (p52) can interact with splicing factors and contributes to the regulation of alternative splicing. PSIP1 can also interacts with the HIV integrase and directs viral cDNA integration into transcribed genes. Although there are multiple chromatin (DNA) binding domains in PSIP1 protein, the PWWP domain is essential for its in vivo functions, especially due to its recognition of H3K36me3.
It is becoming clear that the PWWP domain functions through its interactions with chromatin with methylated histone mark. However, the structural basis for these interactions in a nucleosome context remains unclear. Given the PWWP domains are often coexist with other chromatin interactors, further studies on their crosstalk are required to understand their biological functions. It is also valuable to consider these interactions in context of the protein/complex they settled. Overall, through binding chromatin, the PWWP domains are engaged in many essential cellular processes.