A pocket on the surface of the N-terminal BRCT domain of Mcph1 is required to prevent abnormal chromosome condensation

Abstract

Mcph1 is mutated in autosomal recessive primary microcephaly and premature chromosome condensation (PCC) syndrome. Increased chromosome condensation is a common feature of cells isolated from patients afflicted with either disease. Normal cells depleted of Mcph1 also exhibit a PCC phenotype. Human Mcph1 contains three BRCA1-Carboxyl Terminal (BRCT) domains, the first of which (Mcph1N) is necessary for the prevention of PCC. The only known disease-associated missense mutation in Mcph1 resides in this domain (T27R). We have determined the X-ray crystal structure of human Mcph1N to 1.6Å resolution. Compared with other BRCT domain structures, the most striking differences are an elongated, ordered β1-α1 loop and an adjacent hydrophobic pocket. This pocket is in the equivalent structural position to the phosphate binding site of BRCT domains that recognize phospho-proteins, although the phosphate-binding residues are absent in Mcph1N. Mutations in the pocket abrogate the ability of full-length Mcph1 to rescue the PCC phenotype of Mcph1^−/− mouse embryonic fibroblast cells, suggesting that it forms an essential part of a protein-protein interaction site necessary to prevent PCC.

The Microcephalin/Mcph1/BRIT1 gene is mutated in autosomal recessive primary microcephaly (MCPH) and premature chromosome condensation (PCC) syndrome ^{1; 2}. These diseases share characteristic symptoms of below average head diameter, reduced mental capabilities and a cellular phenotype of chromosome condensation in the G2 phase of the cell cycle. Disease-associated mutations in the Mcph1 gene have been identified that truncate the protein (S25X, 427insA) or that produce a missense mutation (T27R) ^{1; 3; 4}. Mcph1 expression is reduced in cell lines derived from breast cancers, leading to the suggesting that it may function as a tumour suppressor gene ⁵. Mcph1 has been linked with DNA damage regulation, repression of gene expression, cell-cycle progression and chromosome condensation, although the precise function of the protein in these processes is not yet clear ^{5; 6; 7; 8; 9}. Consistent with these functions, Mcph1 normally localises to the centrosome and relocates to sites of DNA repair after cells are exposed to ionizing radiation ⁵. Mcph1 contains three BRCA1-Carboxyl Terminal (BRCT) domains: a single domain at the N-terminus and a tandem pair at the C-terminus. BRCT domains function as protein-protein interaction modules, and tandem BRCT domains act as phospho-peptide binding modules ^{10; 11}. It is reasonable to predict that protein-protein interactions mediated by the BRCT domains of Mcph1 make important contributions to the functions of Mcph1. The interaction of the tandem BRCT domains of Mcph1 with γ-H2AX is crucial for localization to sites of DNA damage after ionizing radiation ¹². Mcph1 is recruited early in the DNA repair process, and its depletion disrupts the accumulation of downstream factors. The N-terminal BRCT domain is necessary for prevention of PCC, and contributes to the centrosomal localization of Mcph1, although binding partners for this domain have remained elusive ^{13; 14}.

The structures of several BRCT domains have been reported, including those present in the DNA repair proteins XRCC1 and the tandem domain pair of BRCA1 bound to a phospho-peptide ^{15; 16; 17}. Despite very low sequence homology, the secondary and tertiary structures of these domains are highly similar, consisting of a central four-stranded parallel β-sheet enclosed by three α-helices ¹⁵. In tandem domains, the phospho-peptide binding site lies at the interface of the two domains, with the residues that recognize the phosphate group on the first domain in the β1-α1 loop and on the N-terminus of α2 ^{16; 17; 18; 19; 20}.

To gain insights into the function of the N-terminal BRCT domain of Mcph1 we determined the X-ray crystal structure of this domain. The domain exhibits a hydrophobic pocket at the equivalent position to the typical phosphate binding site. We show that this pocket is crucial for the rescue of abnormal chromosome condensation in cells lacking wild-type microcephalin.

Human Mcph1 is an 835 amino acid protein that comprises three BRCT domains but no other recognizable domains (Figure 1a). We produced and crystallized a fragment of Mcph1 (Mcph1N) that encompasses the first BRCT domain comprising the first 95 amino acids of Mcph1. Molecular replacement failed with known BRCT domain structures, and as yet no novel individual BRCT domains have been solved using this method. The structure of Mcph1N was solved to 1.6Å by Single wavelength Anomalous Diffraction (SAD) phasing using protein crystals incorporating selenomethionine. The asymmetric unit comprises four copies of Mcph1N two of which extend the full 95 amino acids (chains A and B), and the other two lack the first three amino acids (chains C and D). The refinement and geometry statistics, and the final 2mFo-DFc electron density map are consistent with a high resolution, high quality structure (Table 1, Figure 1b).

Figure 1. X-ray crystal structure of the first BRCT domain of Mcph1.

(a) The domain structure of Mcph1. The N-terminal BRCT domain (Mcph1N) is shown as a green box, the C-terminal tandem BRCT domains (Mcph1C) are shown as black boxes. (b) A stereo diagram showing a representative region of the refined structure and 2mFo-DFc electron density contoured at 1.0σ. All structure figures were made using PyMOL ²⁵. (c) The secondary structure assignment of Mcph1N is shown above a structure-based sequence alignment of its closest structural homologues. Abbreviated names of structural homologues are shown on the left: XRCC1C - XRCC1 C-terminal domain (1cdz chain A¹⁵); RNApII - RNA polymerase II (3ef1 chain A²⁶); 53BP1 - 53BP1 second domain of tandem (1gzh chain D²⁷, 1kzy chain D²⁸) ; MDC1 – MDC1 second domain of tandem (2ado chain A²⁹, 2azm chain A ¹⁹); BRCA1 - BRCA1 second domain of tandem (1jnx chain X ³⁰) ; ECT2 - ECT2 C-terminal domain (2cou); XRCC1N - XRCC1 N-terminal domain (2d8m); DNAIVN - DNA ligase IV N-terminal domain (2e2w), The residue numbering is taken from the coordinate files. (d) A cartoon representation of the superposition of Mcph1N (green) with XRCC1C (red). The secondary structure elements and the first and last residues of Mcph1N are labeled. (e) A magnified view of the β1-α1 loop region of Mcph1N. The putative H-bonds between sidechain residues and the mainchain are shown as dashed lines.

Table 1. Summary of crystallographic analysis.

His₆-tagged Mcph1N protein was expressed in B834 cells co-transformed with the CodonPlus RIL plasmid grown in SelenoMet Medium (Molecular Dimensions). The protein was purified using a 5ml chelating Sepharose column (GE Healthcare) that had been charged with nickel and equilibrated with 100 mM sodium phosphate pH 7.5, 0.5 M NaCl. The column was washed with 50 mM imidazole and eluted on a gradient 50–500 mM imidazole in this buffer. The eluted protein was cleaved using His-tagged TEV protease for 3 hours at 21°C. The cleaved protein was passed through a second nickel column to remove the tag and protease. The protein was dialysed into 20 mM sodium phosphate pH 6.0, 1 mM EDTA, 50 mM sodium chloride and applied to a SP Sepharose column (GE Healthcare). The protein was eluted on a gradient 50–750 mM sodium chloride in this buffer. The eluted protein was further purified by gel filtration on a Superdex 75 16/60 column equilibrated in 20 mM Tris-HCl pH 7.5, 80 mM NaCl. Crystals were grown at 18 °C using the hanging-drop vapour diffusion method using 100 mM MES pH 6.0, 18% PEG 6000, 200 mM ammonium chloride in the reservoir. 2 µl of 15 mg/ml protein stock was mixed with 2 µl of reservoir solution to form the drop and then microseeded. Crystals were flash-frozen in reservoir solution supplemented with 20% glycerol. X-ray diffraction data was processed using CCP4 software ²². The structure of Mcph1N was solved using Phenix ²³ to locate the selenium sites, automatically build the four protein chains and for refinement. Manual rebuilding was carried out using Coot ²⁴.

Crystals
Spacegroup		P21
Lattice constants	a (Å)	80.98
	b (Å)	34.21
	c (Å)	84.24
	β (°)	113.53
Data collection
X-ray source		Diamond I03
Wavelength (Å)		0.92
Resolution range (Å)		45.3-1.6
(Highest resolution shell)		(1.69-1.60)
Unique reflections		56453 (8180)
Completeness(%)		99.7 (99.9)
Multiplicity		4.3 (3.6)
Rmerge (%)		5.2 (33.6)
I/σ(I)		16.1 (3.2)
Refinement
Resolution range (Å)		45.3-1.6
Number of:	amino acids	377
	waters	469
Rfactor (%)		19.02
Rfreea (%)		22.21
Bond deviation (Å)		0.007
Angle deviation (°)		1.057
Molprobity output scores
All-atom Clashscore		8.15
Bad rotamers (%)		2.5
Ramachandran outliers (%)		0.0
Ramachandran favoured (%)		98.6

^aFree Rfactor was computed using 5% of the data assigned randomly ³⁰.

DALI was used to identify the most similar BRCT domain structures and to produce the structure-based sequence alignment (Figure 1c ²¹). The sequence identity between Mcph1N and the other BRCT domains is in the range 9 – 20%, and no residues are identical in all nine proteins. These domains are more highly conserved at the secondary and tertiary structure level (DALI RMSD, root mean square deviation, 1.7 – 3.4 Å). The only structural feature of Mcph1N for which a match was not found among other BRCT domains is the extended loop between β1 and α1, which protrudes an additional ~8Å in Mcph1 when compared to XRCC1 (Figure 1d). This loop is held in an ordered conformation by H-bonds between the mainchain and the sidechains of Ser17 and Asn19 (Figure 1e).

To identify potential surface features that might play a role in the function of this domain, we analyzed the conservation of Mcph1N homologues. Mcph1N is well conserved among vertebrates and many of the conserved residues are in the four β-strands that are buried in the interior of the protein (Figure 2a). The surfaces of the protein, including the outsides of the four helices, are mostly not conserved, although there is a striking surface patch of conserved residues formed by β1, the β1-α1 loop and α1 (Figure 2b). This patch abuts the site of the microcephaly mutation Thr27, which suggested to us that it might be of functional relevance. We therefore included residues from this region in our functional analysis, as described below.

Figure 2. Identification of potential functional sites on the surface of Mcph1N.

(a) The sequence conservation among vertebrate Mcph1 homologues is shown using the color scheme: red for identical; magenta for conserved, grouped by (ILVM), (YF), (DE), (RK), (ST); white for non-conserved. The sequence alignment was made using ClustalW ³¹. (b) Sequence conservation as defined in (a) mapped onto the surface of Mcph1N. The microcephaly mutation T27R and residues that form a conserved surface patch are marked. (c) The two different crystal contacts that involve the surface pocket formed by Trp16 and Tyr56 are shown using a cartoon representation of the protein mainchain and a stick representation of the interacting sidechains. The colour scheme used is: chain A green, chain B yellow, chain C magenta, chain D cyan. (d) The left panel shows the binding of phospho-peptide (grey carbon atoms) to the first domain of BRCA1 BRCT tandem pair (orange cartoon and carbon atoms) ¹⁹. The central panel shows the equivalent region of Mcph1N in the same view. The right panel shows the electrostatic potential (colour ramped from red=negative to blue=positive, calculated using PyMOL) of the Mcph1N surface in a view rotated by ~90°.

Inspection of the crystal packing contacts revealed a surface pocket composed of Trp16 and Tyr56 that was involved with two different contacts (Figure 2c). In one type of crystal contact, the sidechain of Lys45 of chain C or chain D packs into the pocket of chain A or chain B respectively. In the other type of crystal contact, the sidechain of Met1 of chain A or chain B packs against the pocket of chain C or chain D respectively. In the first domain of tandem BRCT domains, the equivalent region forms the phosphate-binding pocket in the phospho-peptide binding site, as shown for BRCA1 in the left panel of Figure 2d ¹⁹. In Mcph1N this region is incompatible with phosphate binding because the crucial phosphate-binding residues (i.e. Lys1702, Ser1655 using BRCA1 numbering) are replaced by Thr59 and Glu14 (Figure 2d, centre panel). In addition, the surface of Mcph1N in and around this pocket is acidic rather than basic (Figure 2d, right panel). The tryptophan sidechain is conserved in vertebrates and the tyrosine is less well conserved (Figure 2a), although the substitutions could conserve the hydrophobic nature of the pocket. The fact that this unusual surface feature of Mcph1N occupies the position of the phospho-binding pocket in other BRCT domains suggested to us that it might be involved in protein-protein interactions, and that this site should be included in our functional analysis.

In a population of mouse embryonic fibroblasts (MEFs) prepared from Mcph1^−/− knockout mice, approximately 30% of the cells exhibit PCC, which recapitulates the defect observed in cells derived from microcephaly or PCC syndrome patients ¹⁴. The introduction of wild-type Mcph1 in these Mcph1^−/− MEFs reduces the proportion of cells that exhibit PCC to 2–3%, which is similar to that observed in wild-type MEFs. 40–50% of the Mcph1^−/− MEFs also have abnormally condensed chromosomes during metaphase, which can be rescued by the expression of wild-type Mcph1, but not by a mutant lacking the Mcph1N domain ¹⁴. The Mcph1N domain is therefore crucial for the function of Mcph1 in prevention of abnormal chromosome condensation, and we decided to identify essential residues for this function within Mcph1N. Based on the above structural analysis, we designed several mutations in full-length Mcph1 and tested their abilities to rescue abnormally condensed metaphase chromosomes in Mcph1^−/− MEFs. We made Y24A and K53A to disrupt the conserved surface patch (Figure 2b), W16A and Y56S to disrupt the pocket (Figure 2c) and the T27R mutation that was found in a microcephaly patient ⁴. The position of these mutations within the Mcph1N structure is shown in Figure 3a. Metaphase spreads were prepared from Mcph1^−/− MEFs expressing Mcph1 and the proportion of cells exhibiting abnormally condensed chromosomes were counted (Figure 3b, c). As previously observed, fewer than 10% of wild-type MEFs displayed abnormally condensed chromosomes, whereas almost 50% of the Mcph1−/− MEFs had abnormally condensed chromosomes. Mcph1 containing mutations in the conserved surface patch was able to rescue this metaphase chromosome condensation defect to an extent similar to wild-type Mcph1, ruling out a critical role for this patch in this chromosome condensation function of Mcph1. The T27R mutant of Mcph1 was only slightly impaired in this function, consistent with the previous report that this patient’s cells only shows mild PCC phenotype. In contrast, the pocket mutations were significantly reduced in their abilities to rescue this chromosome condensation defect, with almost 30% of cells expressing the W16A mutant of Mcph1 exhibiting abnormally condensed chromosomes. This is very similar to the proportion previously observed with a fragment of Mcph1 lacking the entire Mcph1N domain ¹⁴, and therefore the W16A mutation recapitulates the effect of the loss of the domain with respect to chromosome condensation. None of the mutations compromised the ability of Mcph1 to relocate to γ-H2AX foci upon irradiation (Figure 3d), consistent with the previous report that the C-terminal BRCT domains are required for Mcph1 localization to sites of DNA breaks ¹². We used circular dichroism spectroscopy and a thermal shift assay to show that neither pocket mutations nor the T27R microcephaly mutation disrupt the fold of the Mcph1N domain or reduce its thermal stability (Figure 3e,f). Based on these observations, we conclude that the pocket is a crucial feature required for Mcph1 function in the regulation of chromosome condensation.

Figure 3. Pocket mutations do not rescue the abnormal chromosome condensation phenotype of Mcph1^−/− cells.

(a) The position of residues mutated in this study on the structure of Mcph1N. (b,c) Rescue of premature chromosome condensation (PCC) phenotype of Mcph1^−/− MEFs by wild-type (WT) or mutant Mcph1. (b) Shows the quantification of metaphase cells with condensed chromosomes observed in each indicated cell line. (c) Shows examples of metaphase spreads seen in these cell lines. (d) Proper localisation of wild-type and mutant Mcph1 to γ-H2AX foci upon ionizing radiation (IR). Cell culture, transfection, chromosome spreads, irradiation and immunofluorescence were carried out as described in ¹⁴. (e) Circular dichroism spectroscopy shows that the WT, T27R, W16A and Y56S Mcph1N domain proteins are folded. Spectra were collected using 1 mg/ml samples in 10mM sodium phosphate pH 7.5, 20 mM sodium chloride in a 0.01 cm pathlength quartz cell at 18 °C on an Aviv 202SF instrument (Aviv Biomedical). Data were processed using CDTool software ³². (f) A thermal shift assay shows that the T27R, W16A and Y56S Mcph1N domain proteins are at least as thermally stable as the WT domain protein. The thermal shift assay is based on the increased fluorescence of a dye upon binding to the exposed hydrophobic core of proteins unfolded by a temperature gradient ³³. 25 ml samples containing 0.3 mg/ml protein and 5x Sypro Orange dye (Sigma) in 10 mM sodium phosphate pH 7.5, 20 mM sodium chloride were heated from 4–95 °C at 2 °C/min in a BioRad IQ5 instrument and monitored at excitation and emission wavelengths of 485 nm and 585 nm. The mean of eight denaturation curves was calculated for each protein and the buffer alone; the baseline was then subtracted. Data were fitted to a Boltzmann sigmoid function to yield temperatures of 50% denaturation (apparent T_m) of 52.6 °C (WT), 52.7 °C (T27R), 55.1°C (W16A), 59.5 °C (Y56S).

The phosphate-binding pocket is a conserved feature of the first domain of tandem BRCT repeats and the equivalent site was postulated to be a site of protein-protein interaction on the C-terminal BRCT domain of XRCC1 ^{15; 16; 17; 18; 19; 20}. We have now shown that the equivalent surface on the N-terminal BRCT domain of Mcph1, on the β1-α1 loop and the N-terminus of α2, is a functionally important site for Mcph1 function in prevention of abnormal metaphase chromosome condensation. This region of the BRCT domain is the most divergent structurally from other BRCT domains, due to the more elongated β1-α1 loop and the presence of a pair of surface aromatic sidechains. This pocket is presumably a crucial element of the interface of a protein-protein interaction. Future work will identify the binding partners of Mcph1 that mediate its function in prevention of PCC, and define the structural basis of their interactions.