The effect of D380Y pathogenic mutation in human Yin Yang 1 on the protein’s structure and function*

Yang is a human transcription factor that controls a number of genes and takes part in the regulation of cell cycle, proliferation, differentiation, and neuronal de- velopment. Yin Yang 1 is composed of an N-terminal intrinsically disordered fragment and a C-terminal domain responsible for binding to DNA, composed of four zinc fingers. Recently, various alterations in the Yin Yang 1’s DNA binding domain were linked with an unexplained intellectual disability named Gabriele-de Vries syndrome. In this paper, a repetitively occurring substitution of as-partate-380 for tyrosine was analyzed to assess its impact on Yin Yang 1’s structure and DNA binding. The substitution was found to affect Yin Yang 1’s secondary and tertiary structure to a limited extent and to impair the specificity of its interaction with DNA.


INTRODUCTION
Yin Yang 1 (YY1) is a ubiquitously expressed human transcription factor that can act as a transcription initiator, activator or repressor. YY1 is crucial for cell cycle control, proliferation, differentiation, and neuronal development (Gordon et al., 2006;He & Casaccia-Bonnefil, 2008). Deletion of both yy1 alleles is lethal at the early stages of embryo development, while yy1 +/-heterozygotes exhibit neurological defects and growth retardation, which was shown in Mus musculus and Xenopus laevis (Donohoe et al., 1999;Morgan et al., 2004). Structurally, YY1 is composed of two parts, differing in both structure and function. The N-terminal fragment is intrinsically disordered (Gorecki et al., 2015) and fulfills a regulatory role through a number of protein-protein interactions. On the other hand, the C-terminal fragment of YY1 forms its DNA-binding domain (DBD) that includes four consecutive zinc fingers of C2H2 type (Houbaviy et al., 1996).
Pathogenic effects of substituting YY1 aspartate-380 for tyrosine was recently reported twice, in two different exome-sequencing studies on individuals with unexplained intellectual disabilities (Vissers et al., 2010;Gabriele et al., 2017). One of the studies was aimed at mutations in yy1 locus and in fact found several substitutions and deletions within the DBD (Gabriele et al., 2017). The finders hypothesized that the common effect of the discovered mutations could result from impaired DNA binding by the YY1 mutants. Aspartate-380 is located in the linker region that connects 3rd and 4th zinc fingers of YY1 (Fig. 1). No direct contact between aspartate-380 and DNA was observed in the crystal structure of YY1 nucleoprotein (Houbaviy et al., 1996). However, molecular simulations predict that D380Y substitution disrupts the network of hydrogen bonds on the surface of the YY1 molecule (Gabriele et al., 2017). Moreover, YY1 D380Y exhibited compromised performance in chromatin immunoprecipitation (Gabriele et al., 2017).
Our research group established a protocol for robust production of active, recombinant YY1 protein (Golebiowski et al., 2011) as well as a sensitive, quantitative equilibrium assay for YY1-DNA interaction (Golebiowski et al., 2012). These methods previously allowed us to precisely analyze the effect of another YY1 mutation, T372R, which is recurrent in insulinomas, and proved it to affect YY1 binding to DNA in a sequence-specific way (Nieborak & Gorecki, 2016). In this paper, we analyzed the impact of D380Y mutation on YY1 structure and its interaction with DNA.
Site-directed mutagenesis. pET21a/DBD vector was subjected to mutagenesis with the use of the QuikChange protocol (Stratagene). The sequences of the mutagenic forward and reverse primers were: GTGC-GAATCCATACCGGATACAGGCCCTATGTGTGC and GCACACATAGGGCCTGTATCCGGTATGGAT-TCGCAC, respectively. The obtained construct was sequenced to confirm the presence of the substitution.
Protein expression and purification. The plasmids encoding wild-type DBD and its D380Y variant were transformed into the E. coli strain BL21(DE3) (Novagen) which were grown on the Luria-Bertani (LB) agar plates with ampicillin (100 μg/mL) over night at 37°C. LBampicillin medium (50 mL) was inocu lated with a single colony from the plate and grown overnight at 180 rpm shaking at 37°C. The precultures were added to 1.5 L of LB-ampicillin and grown at 180 rpm shaking at 37°C. The expression was induced at an O.D. 600 of 0.6-1.0 with 1 mM isopropyl β-d-1-thiogalactopyranoside for three hours. At the time of induction, ZnCl 2 was added to the cultures to the final concentration of 0.1 mM. The cells were harvest ed by centrifugation at 7 000 × g for 15 min at 4°C, pel leted and stored frozen at -20°C.
The bacterial pellet from the 2 L culture was resuspended in 30 mL of buffer A (25 mM Tris-HCl, pH 7.9, 100 mM NaCl, 6 M guanidine hydrochloride, 1 mM phenylmethylsulfonyl fluoride, 2 mM β-mercaptoethanol). The sus pension was sonicated for six 45-s bursts on ice. The lysate was cleared by centrifugation (30 000 × g, 30 min, 4C), and the superna tant was loaded onto a 30-mL nickel-nitrilotriacetic acid gravity-flow column (Novagen) which was previously equilibrated in buffer A. The column was washed with 3 col umn volumes of buffer A. Proteins were then eluted in buffer A with addition of 300 mM imidazole. 1.0-mL fractions were harvest ed. The protein concentration was determined in triplicates with the Bradford assay (Bradford, 1976), calibrated with bovine serum albumin (BSA) (Sigma-Al drich). The purification process was monitored by so dium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, 1970) in 15% gel. Fractions containing DBD or its D380Y variant were pooled and reduced by a 30min incubation at 60°C with the addition of 10 mM ethylenediamine tetraacetic acid and 100 mM dithiothreitol (DTT). The proteins were refolded by a three-step dialysis at 4°C against buffer B (25 mM Tris-HCl, pH 7.9, 100 mM NaCl, 10 mM MgCl 2 , 0.1 mM ZnCl 2, 5 mM DTT) at 1:100 ratio. The final purification step was conducted using ÄKTA Explorer system (Amersham Pharmacia Biotech) on agarose resin with immobilized oligos of IgH promoter sequence, as described previously (Golebiowski et al., 2011). Briefly, the dialyzed samples were loaded onto the column and the specifically bound protein was eluted in the 10-100% gradient of buffers A1 (25 mM Tris-HCl, pH 7.9, 10 mM MgCl 2 , 0.1 mM ZnCl 2, 2 mM DTT) and A2 (25 mM Tris-HCl, pH 7.9, 1 M NaCl, 10 mM MgCl 2 , 0.1 mM ZnCl 2, 1 mM DTT). Protein content and purity in the collected fractions were analyzed with the use of SDS-PAGE. Selected fractions were pooled and dialyzed at 4°C against 500-fold excess of buffer C (25 mM Tris-HCl, pH 7.9, 100 mM NaCl, 10 mM MgCl 2 , 0.1 mM ZnCl 2 , 5 mM DTT), in which all the further analyses were performed.
Size-exclusion chromatography. Size exclusion experiments were performed using ÄKTA Explorer system (Amersham Pharmacia Biotech) on Superdex 200 Increase 10/300 column (GE). The separations were monitored by absorbance measurements at 220 nm and 280 nm to determine the elution volumes of DBD protein variants. The Stokes radii were calculated based on a calibration curve obtained for dextran blue and a set of standard proteins: thyroglobulin, apoferritin, BSA, myoglobin, cytochrome c.
Spectroscopic measurements. Circular dichroism (CD) measure ments were conducted using a J-710 spectropolarimeter (Jasco) equipped with an F25 temperature control unit (Julabo) in quartz cuvettes of 0.1 mm pathlength (Hell ma). The spectra were recorded in the range of 194-250 nm with 1 nm data pitch, 10 nm/min scanning speed, 8 s integrating time, and 2 nm bandwidth and averaged over three acquisitions. The protein concentration was 1.1 mg/ml and 1.3 mg/ml for wild-type and D380Y, respectively. All spectra were corrected for the effect from the buffer and all measurements were converted to molar residual ellipticity units. Secondary structure content was determined using SELCON3 program from CDPro package (Sreerama & Woody, 2000).

RESULTS
To study the prospective changes in YY1's structure and function upon D380Y substitution, we produced a recombinant protein carrying the mutation. The DNAbinding domain of YY1 (DBD) was used, as the fragment is solely responsible for DNA binding (Figiel & Gorecki, 2017). Standard protocols were used for the expression and purification, that previously had been optimized for the wild-type DBD (Golebiowski et al., 2011). Notably, the final purification step employed an affinity column with immobilized DNA of sequence recognized by wild-type YY1. If the D380Y variant did not specifically bind DNA, this could reduce the purification yield, though the mutant protein could still non-specifically bind to the column through electrostatic interactions. No such reduction was in fact observed; the procedure was repeated twice and no significant difference in the production yield was observed between the D380Y mutant and the wild-type DBD1. The effect of D380Y mutation in human YY1

Structure analysis
The analyzed substitution completely changed the character of the side chain, replacing a charged aspartate moiety with the aromatic ring of tyrosine. Changing the protein charge, hydrophobicity and possibly the network of hydrogen bonds on its surface could severely disturb the spatial structure. To check this, we compared the secondary and tertiary structures of the wild-type DBD and its D380Y mutant.
Secondary structure content was assessed with the use of CD measurements in the far-UV range. The spectra obtained for the two protein variants had roughly similar shapes but differed regarding the intensity of a broad negative band centered at about 215 nm (Fig. 2). However, deconvolution of the spectra yielded a similar content of particular secondary structure types in the wild-type protein and the D380Y mutant (Table 1), suggesting that the propensity to fold into zinc finger motifs was retained in the mutant protein. D380 residue is located in the linker region between the third and fourth zinc fingers. Although the region had no regular arrangement in the crystal structure (Houbaviy et al., 1996), the mutation could possibly impair the proper folding of both zinc fingers. Yet the amount of particular secondary structures differed only to a small extent between the wild-type and D380Y DBD. Possibly, the mutation affected the distribution of conformations available for the protein.
Next, we analyzed the impact of D380Y mutation on the overall tertiary structure of DBD. Size exclusion chromatography separations allowed us to compare the compactness of conformation of wild-type and D380Y DBD variants. Substituting tyrosine for aspartate changed the protein's charge and hydropathy, the two properties that determine the propensity of the ordered protein structure (Campen et al., 2008). Wild-type DBD has a compact, ordered structure (Gorecki et al., 2015). The compactness of the D380Y mutant was loosened compared to the wild-type protein, as demonstrated by its increased Stokes shift and a higher ratio of apparent and formula molecular masses, presented in Table 2.

DBD1 D380Y binding to DNA
Protein-DNA interaction can be affected by mutations in many different ways. The number of contacts can be decreased by substitutions of either positively charged residues that electrostatically interact with the phosphate backbone or residues that make base-specific contacts in the grooves of DNA helix. The preferred consensus sequence may also be changed. Such an effect was observed for T372R mutation of YY1, which is recurrent in insulinomas. The 1st and 4th zinc fingers of YY1 participate in DNA binding to a different extent in the wild type protein and in the T372R mutant (Cromer et al., 2015). Similarly, the D380Y mutation might also affect the concerted DNA binding by YY1's 3rd and 4th zinc finger, as it changes the sequence of their linker into a non-optimal one. In order to capture such subtle differences, we measured the affinity of DBD1 D380Y for three different sequences, derived from promoters actually bound by YY1, but characterized by a varying similarity to the YY1 consensus motif: the immunoglobulin heavy chain enhancer, the adeno-associated virus P5 promoter and the adeno-associated virus P5 initiator.
Fluorescence anisotropy measurements demonstrated that DBD1 D380Y bound DNA weaker than the wildtype protein (Fig. 3). The affinity for IgH sequence was decreased the most. As a result, the binding constants (presented in Table 3) were similar for all three sequences, even though the values differed for the wild type protein. Thus, we observed no differences in the affinity of the mutated protein for the tested sequences, although such a difference was clearly visible in the wildtype protein (Golebiowski et al., 2012).

DISCUSSION
The aim of this work was to evaluate how YY1 protein's structure and its interaction with DNA is affected  Table 1. Secondary structure content of the wild-type and D380Y DBD protein inferred from their CD spectra (Fig. 2). Values denote the number of amino acid residues forming the given type of secondary structure.   Golebiowski and others (Golebiowski et al., 2012) by D380Y substitution that was discovered in two cases of unexplained mental retardation. Aspartate-380 locates within the linker that connects the third and fourth zinc fingers of YY1 (Houbaviy et al., 1996). Linkers between the consecutive zinc fingers are important structural elements for protein-DNA interactions. They exhibit a consensus sequence GEK and a similar spatial arrangement within the molecules of various proteins (Wolfe et al., 2000). Deviations from the consensus impede the concerted binding of the linked zinc fingers to DNA. For instance, the linker between the first and second zinc fingers of YY1 is shorter than the consensus, impeding the interaction between the first zinc finger and DNA (Figiel & Gorecki, 2017;Chen et al., 2019). Aspartate-380 is located at the 3rd position of the linker. Although it is not the optimal glutamate residue, its side chain has a very similar chemical character. On the contrary, D380Y substitution changes the properties completely, influencing both the hydropathy and the charge, and possibly also introducing steric hindrance. Based on the molecular simulations, D380Y substitution is expected to disrupt the network of hydrogen bonds and salt bridges on the YY1-DNA interface (Gabriele et al., 2017). Our observations confirmed that the substitution affects YY1's structure and its interaction with DNA.

Wild
The CD spectra registered in the far-UV range differed between the two variants but still allowed us to infer that the native secondary structure was retained in the D380Y mutant. Size exclusion chromatography separations indicated that the overall compactness of the protein molecule was slightly affected by D380Y substitution, as it affected both the charge and the hydropathy of the protein. Moreover, the mutation compromised YY1's affinity to DNA and reduced its specificity. D380 locates in a region that is important for the interaction, as it links the two zinc fingers that make the majority of specific contacts with the nucleobases (Houbaviy et al., 1996) (Fig. 1) and contribute the most to YY1-DNA binding (Chen et al., 2019).
Interestingly, the changes in fluorescence anisotropy of the DNA-bound ROX were more pronounced upon binding of DBD1 D380Y than of the wild-type protein. This could be explained by several mechanisms; firstly, the mutant protein could be unstable and more prone to aggregation than the wild-type protein, though the hydrodynamic parameters did not confirm this. Alternatively, the mutant protein binding could engage the first zinc finger to a greater extent, decreasing the mobility of the dye, which is attached to the 5' end of the reverse DNA strand. Such an effect was previously ascribed to YY1 T372R mutant (Cromer et al., 2015). Finally, YY1 could electrostatically interact with the fluorescent dye attached to DNA duplex within the nucleoprotein, altering its fluorescence parameters. Such interaction could be abolished by D380Y substitution, as described elsewhere for TraI36 mutants, E187A and R150A (Anderson et al., 2008).
One of the premises for investigation of YY1 D380Y variant was its impaired chromatin immunoprecipitation, previously reported by Gabriele and others (Gabriele et al., 2017). Apart from a compromised affinity for DNA, this could also be caused by abnormal interaction with trans regulators. This is in line with the finding of Gabriele and others (Gabriele et al., 2017) that the peaks retained in YY1 D380Y immunoprecipitates were enriched in COREST, p300, PAX5 and HDAC1 motifs, compared to those present only in wild-type YY1 immunoprecipitates. In fact, D380 is located in a region involved in many protein-protein interactions, i.e. with E1A (Lewis et al., 1995), ATF2a (Zhou et al., 1995), SAP30 (Huang et al., 2003), or p53 (Sui et al., 2004).
The presented study was limited to an in vitro approach. An in vivo analysis would give a fuller insight into the malfunctions caused by the D380Y substitution. However, it would not allow for separate analyses of particular structural aspects and interactions. The simplified model enabled a detailed, quantitative analysis and allowed us to draw unequivocal conclusions regarding the character of changes in YY1 interaction with DNA. Functional studies on cells should be the next step to comprehend a full picture of the effects of D380Y mutation and fill the gap between the case-control study of Gabriele and others (Gabriele et al., 2017) and the presented in vitro study of the underlying mechanisms.
To conclude, D380Y substitution was independently suggested as the cause of unexplained mental retardation twice (Vissers et al., 2010;Gabriele et al., 2017). We found that the substitution affected YY1's secondary and tertiary structure only to a limited extent, yet it impaired DNA binding and decreased its specificity.