Biological context

Single stranded DNA-binding proteins (SSBs) are ubiquitously present in all known cellular organisms (Mushegian and Koonin 1996) including viruses (Newport et al. 1981; Sun and Shamoo 2003). This gives credence to their vital role in maintaining the integrity of the genome by coordinating many DNA metabolic events which include DNA replication, recombination, and repair (Richard and Khanna 2009). SSB proteins bind single-stranded DNA (ssDNA) via a highly conserved oligonucleotide-binding (OB) domain and function to temporarily bind and protect exposed ssDNA generated during these events (Murzin 1993; Suck 1997). They also typically possess additional structured or unstructured regions that are involved in protein–protein interactions that orchestrate downstream processing of DNA.

The ssDNA-binding OB domain is made up of a five-stranded bent antiparallel β-sheet that forms a closed β-barrel, which is conserved across organisms from all three domains of life. Crenarchaeal and bacterial SSBs display a ‘simple’ domain organisation with only one OB fold followed by a flexible C-terminal extension, whereas the extensively studied human replication protein A (hRPA) has multiple OB domains distributed across three subunits (Iftode et al. 1999). Over the last decade, we have characterized the biological role and molecular details of two new human SSBs, hSSB1 (NABP2/OBFC2B/SOSSB1) and hSSB2 (NABP1/OBFC2A/SOSSB2), both of which have a simple domain organisation (Ashton et al. 2013, 2016, 2017; Bolderson et al. 2014; Croft et al. 2017; Kariawasam et al. 2016; Paquet et al. 2015, 2016; Richard et al. 2008, 2011a, b; Richard and Khanna 2009; Touma et al. 2016, 2017).

Mouse models and mammalian cell line experiments have shown that SSB1 and SSB2 display some overlapping functions (Huang et al. 2009; Li et al. 2009). For example, both human SSBs (hSSB) have been shown to be important in the repair of DNA double strand breaks via HR (Paquet et al. 2015; Richard et al. 2008, 2011a) and the UV induced DNA damage response (Richard et al. 2008). A more recent study has shown that both mouse SSB1 and SSB2 (mSSB1 and mSSB2, respectively) function cooperatively to alleviate replication induced stress in hematopoietic and progenitor stem cells (Shi et al. 2017). However, there are some important differences in the protein sequences of hSSB1 and hSSB2, both in the OB fold, and to a larger extent in the disordered C-terminal extension. These differences most likely result in significant structural and functional divergences as well as cell- and tissue-specific roles for hSSB1 and hSSB2. Indeed, higher expression levels of SSB2 compared to SSB1 in the testis, spleen and thymus in mice point to a specific role of SSB2 in repair and recombination in these tissues (Boucher et al. 2015; Kang et al. 2006). More recently, Vernin et al., have reported that hSSB2 can be downregulated by tumor suppressor miRNAs, miR17 and miR21, which results in increased genomic instability and abnormal cell proliferation (Vernin et al. 2014). Furthermore, in contrast to mSSB1, depletion of mSSB2 in wild-type MEFs revealed rapid cell death demonstrating a critical role in cell viability at early embryonic stages (Feldhahn et al. 2012; Gu et al. 2013). A hSSB2/RARA fusion gene, where the C-terminal extension of hSSB2 is deleted, has also been reported to result in variant acute promyelocytic leukaemia, showing for the first time an involvement of hSSB2 in human disease (De Braekeleer et al. 2014; Won et al. 2013).

While we have recently described a data-driven model of the solution structure of the OB domain of hSSB1 bound to ssDNA (Kariawasam et al. 2016; Touma et al. 2016), comparatively little is known about the molecular details of ssDNA recognition by hSSB2. In this study, we present the solution-state backbone resonance assignments of hSSB2 at 298 K as determined by NMR. We have also mapped the DNA-binding interface of hSSB2 which reveals similarities and differences in the binding mode between hSSB1 and hSSB2. These data will further our understanding of the important functional differences that exist between these two essential human SSBs.

Methods and experiments

Protein expression and purification

An Escherichia coli codon-optimised construct of hSSB2 OB (1-125) (GeneArt) was directionally cloned into a pGEX6p vector (with a GST expression tag) using BamHI and EcoRI restriction sites. Protein expression of this construct was achieved in BL21(DE3) E. coli cells, and induced with 0.2 mM IPTG at 25 °C for 16 h. The harvested cells were lysed via sonication in lysis buffer (10 mM MES, pH 6.0, 50 mM NaCl, 3 mM TCEP, 0.5 mM PMSF, 0.1% Triton X-100). The soluble fraction extracted using centrifugation was purified via GSH affinity chromatography, followed by HRV-3C protease cleavage at 4 °C overnight (leaving five additional residues GPLGS, on the N-terminus of the OB construct). The cleaved protein was loaded onto a HiTrap HP Heparin (GE) column equilibrated with NMR buffer (10 mM MES, pH 6.0, 50 mM NaCl, 3 mM TCEP). A 500-mL linear gradient consisting of 50–1000 mM NaCl was used to elute hSSB2 protein. SDS-PAGE was utilised to analyse the fractions which corresponded to a distinctive absorbance peak from the cationic gradient elution, and fractions identified as hSSB2 were pooled and concentrated. The concentration of hSSB2 was determined using the theoretical extinction coefficient for hSSB2 and the absorbance value at 280 nm.

NMR Spectroscopy & data processing

NMR experiments were performed using ~ 0.1–0.5 mM hSSB2 OB construct in NMR buffer containing 10% D2O. Proton chemical shifts were referenced to 4,4-dimethyl-4-silapentanesulfonic acid (DSS) at 0 ppm. 13C and 15N chemical shifts were referenced indirectly to the same signal. 1H15N-HSQC, 1H 13C-HSQC, 1H15N13C CBCA(CO)NH, 1H15N13C HNCACB, 1H15N13C HNCO, 1H15N13C HN(CA)CO and 1H15N-NOESY NMR spectra were recorded on either a 800 or 600 MHz spectrometer (Bruker Avance III) equipped with 5-mm TCI cryoprobes at 298 K. The chemical shifts of DNA-bound 15N13C hSSB2 were unambiguously identified by gradual additions of 0.2, 0.4, 0.6, 0.8 and 1 equimolar volumes of 6T (6 × Thymines) ssDNA (HPLC purified; Sigma-Aldrich). The data collected were processed using Topspin (Bruker Biospin) and assignments were made using Sparky (Goddard and Kneller, University of California at San Francisco).

Assignments and data deposition

The hSSB2 construct utilized throughout this study contained the complete OB fold domain (1–125) and an additional five residues as a result of the HRV-3C cleavage. The 1H15N HSQC spectrum of folded hSSB2 is shown in Fig. 1. Apart from residue N2 and all proline residues, all backbone chain atoms were assigned (99.6% completeness). Notably, in analogy to the closely related hSSB1 protein (Kariawasam et al. 2016; Touma et al. 2016), spectra recorded to obtain side-chain chemical shift assignments were not of sufficient quality, most likely due to signals experiencing intermediate exchange. Further attempts to improve the quality of the spectra by increasing the temperature above 298 K [a strategy that proved successful for the related SSB from Sulfolobus solfataricus (Gamsjaeger et al. 2013, 2014)] resulted in substantial protein degradation, as expected for a human protein. All chemical shifts (including unusual ones such as K15) have been verified (using the assignment program and software provided by the BioMagResBank BMRB) and deposited into the BMRB (http://www.bmrb.wisc.edu) under the Accession Number 27184.

Fig. 1
figure 1

1H 15N-HSQC spectrum of hSSB2 OB domain (1–125 with five additional residues from HRV-3C cleavage, ~ 0.5 mM) showing backbone amide resonances. Note that for clarity, residues 41, 42, 48, 55, 89 and 103 located to the left of the centre of the spectrum and residues 5, 9, 25, 96, 97 and 123 to the right of the centre of the spectrum are not labelled. The spectrum was recorded at a proton resonance frequency of 800 MHz at 298 K in 10 mM MES, pH 6.0, 50 mM NaCl, 3 mM TCEP

Full size image

Additionally, the DNA-binding interface of hSSB2 in solution has been mapped and compared to the previously published DNA-binding interface of hSSB1. Figure 2a depicts a portion of a 1H 15N HSQC spectrum of hSSB2 in the absence and presence of 6T ssDNA. Further analysis of the weighted chemical shift perturbations (Ayed et al. 2001) from 1H 15N-HSQC spectra of hSSB2 compared to hSSB1 reveals the important role that aromatic residues (shown in blue) play in the recognition of ssDNA (Fig. 2b). More specifically, hSSB2 residues W59, F78, W82 and Y89, which are conserved in hSSB1 (W55, Y74, F78 and Y85), exhibit significant chemical shift changes upon the addition of ssDNA. The similarity between the chemical shift profiles of both SSBs indicates that hSSB2 utilises mostly the same set of residues for ssDNA-binding as hSSB1. For instance, based on the chemical shift change of H40 in hSSB2, the hydrogen bond formation between this residue and ssDNA in hSSB1 is most probably conserved in hSSB2. However, there are likely small but important differences in ssDNA recognition between the SSBs. For example, closer inspection of the chemical shifts changes of lysines 35 and 83 in hSSB2 (Fig. 2) reveals that the role of these residues in ssDNA-binding could be the exact opposite to hSSB1 (corresponding residues K31 and K79, respectively). While K35 may be in involved in ssDNA-binding (significant chemical shift change), K83 is likely not important (minor shift change), as opposed to hSSB1 where K31 displays no contacts to any DNA bases and K79 forms a hydrogen-bond with the ssDNA.

Fig. 2
figure 2

a Section of the 1H 15N HSQC spectrum of ~ 0.3 mM hSSB2 alone (black) and a 1:1 mixture of hSSB2 and ssDNA (6T, grey). Assignments and directions of movement are indicated. b Weighted backbone chemical shift changes of HN and N atoms (Ayed et al. 2001) for hSSB2 (top) and hSSB1 (bottom) (Touma et al. 2016) upon binding to ssDNA. Note that residue numbers of hSSB2 differ by four compared to hSSB1 based on the sequential alignment of the SSBs [see Fig. 1 in Touma et al. (2016)]. Residues exhibiting changes larger than the average are coloured in red or blue. Aromatic residues (conserved between hSSB1 and hSSB2) that are essential for ssDNA-binding are coloured in blue

Full size image

Overall, our data show that the ssDNA-binding interface is conserved between hSSB1 and hSSB2, with minor but significant differences, potentially contributing to changes in binding. Future studies using NMR and other biophysical experiments will help distinguish the molecular details of ssDNA recognition and thus help to gain a more comprehensive understanding of the functional differences that these novel hSSBs display in the framework of genome maintenance.