Brewer’s/baker’s yeasts (Saccharomyces cerevisiae) are extremely convenient object for the study of mitochondrial processes. In this organism, mitochondria have considerable functional similarity with human mitochondria. Moreover, yeast can perform vital functions in the absence of functionally full-fledged mitochondria, allowing for deletion of the genes of certain mitochondrial proteins and study of the phenotypic manifestations of such mutations. The mitochondrial genome of S. cerevisiae is a structure (nucleoid) organized in space. The nucleoid contains up to 10 mtDNA copies that interact with different proteins. Mitochondrial DNA located in the nucleoid is composed of three-dimensional organization, closely connected with the mechanisms of replication, transcription, and inheritance. Laying in the nucleoid also protects the DNA from the attack of reactive oxygen species, appearing during oxidative phosphorylation [7].

Abf2p is one of the most widely represented proteins interacting with mtDNA in the mitochondrial nucleoids in yeast. It was extracted for the first time in its pure form in 1979 [8], and the name ‘Abf2p’ was given to it in 1991 [9]. This protein contains two domains — HMG1 (High-Mobility Group 1) and HMG2 (HighMobility Group 2). Abf2p is presently the most studied among other proteins of yeast mitochondrial nucleoid.

Abf2p is a unique protein with the highest basicity among nucleoid proteins [10]. It generates negative DNA supercoiling when in contact with a circular plasmid in the presence of topoisomerase 1 [10]. Abf2 mutant yeast can support mtDNA growth in YPG medium containing glycerol as a carbon source. However, when culturing in media with fermentable carbon sources such as glucose, there is gradual loss of mtDNA [10]. It is also proved that in Abf2p gene mutation, the number of mitochondrial recombination events when paired with wild-type strain significantly reduces [11]. Besides, Abf2p stabilizes Holliday recombination junction intermediates, which also points to the importance of this protein in recombination [12].

Despite the large number of studies on Abf2p, the molecular mechanisms of its involvement in mtDNA recombination processes remain unexplored. The presence of two HMG domains in mitochondrial DNA-binding protein Abf2p is typical of most HMGB proteins: they are known to bind DNA minor groove with limited specificity or completely non-specific [13]. Besides, it is assumed that HMG domains that make up these proteins bind DNA independently of each other with similar efficiency [13]. However, mitochondrial HMG proteins apparently have much broader functionality than their nuclear homologues. For example, human mitochondrial DNA-binding protein, TFAm, apart from compaction of nucleoids, is a transcription factor and is probably involved in the processes of mtDNA recombination and repair [14]. TFAm is similar to bacterial HU proteins when it comes to its multi-functionality and DNA binding characteristics [14]. It is assumed that Abf2p also possesses multifunctional properties, as it is a structural analog of TFAm in yeast mitochondria.

In this paper, recombinant proteins corresponding to two Abf2p domains — HMG1 and HMG2 — were obtained. The DNA-binding ability of these domains with respect to linear DNA duplex and structure imitating late homologous recombination intermediates (Holliday junction) was studied.

METHODS

Gene cloning and expression

ABF2 gene sites corresponding to domains HMG1 (amino acid residues 27–115) and HMG2 (amino acid residues 112–183) were amplified using primer pairs abfhmg1F (GATACATATGGGTCCTAAAAGGCCC ACATC) / abfhmg1R (CGTCCTCGAGAGGAAGTTTTTCGTCA AACTCC) and abfhmg2F (GGCGCATATGGAGTTTGACGAAAAA CTTCC) / abfhmg2R (GAGGCTCGAGAGCATTATATTCTTGG ATAGC) respectively. Yeast genomic DNA strain BW303 was used as a template. The obtained amplification products were treated with restriction endonucleases NdeI and XhoI (Thermo Fisher Scientific, USA) and cloned into expression vector pET32a (Novagen, USA). In this way, we obtained the pET32a_HMG1 and pET32a_HMG2 vectors. Conformity of cloned sequences with the reference sequence was verified through Sanger sequencing in a post-genomic lab belonging to the Institute of Physico-Chemical Medicine of Russia, a Federal Medical & Biological Agency. The plasmids obtained were transformed into expression strain Escherichia coli B834 (DE3), single colonies of transformants were transferred to a 2xYT liquid medium with 100 μg/ml of carbenicillin, cultured at 37 °C under vigorous agitation to OD600 ~ 0.6–0.8. After that, expression of cloned genes was induced by addition of isopropyl β-D-1-thiogalactopyranoside (IPTG) to a concentration of 0.25 mM. Culturing was continued at 30 °C for 3 hours. The cells were then pooled by centrifugation (3000 g, 10 min) and re-suspended in the starting buffer (25 mM sodium phosphate buffer with pH 7.4; NaCl 500 mM; imidazole 20 mM). The re-suspended cells were sonicated in 4 pulses by 15 seconds each at 20 % amplitude. The cell lysates obtained were centrifuged at 17,000 g for 20 min. Supernatants containing recombinant proteins were pooled and further purification of the target product was carried out by metal-chelate affinity chromatography in 1 ml HisTrap column filled with Ni-NTA sepharose (GE Healthcare, USA), using high-performance protein chromatography system AKTA Purifier (GE Healthcare System, US) according to manufacturer’s recommendations. Elution of target proteins was monitored by absorbance at 280 nm wavelength. A fraction with the target protein was immediately transferred into the storage buffer (25 mM Tris-HCl, pH 7.4; 150 mM NaCl; 5 mM EDTA) using a 5 ml HiTrap Desalting column that contained Sephadex G-25. Efficient protein transfer to the buffer for storing was monitored through the absorption peak ratio at 280 nm wavelength to the conductivity of the solution.

Protein concentrations were determined by absorbance at 280 nm wavelength on NanoPhotometr spectrophotometer (Implen, Germany) according to known molar extinction coefficient (12090 M^–1 × cm^–1 for both proteins) and taking into account exact molecular weights: 12632.39 12 Da for HMG1 and 10177.58 Da for HMG2.

Assembling DNA structures

The structures used (DNA duplex and Holliday structure) were collected from the following oligonucleotides — x-FAM (FAM-AGTCTAGACTGCAGTTGAGTCCTTGCTAGGAC GGATCCCT), x-com (AGGGATCCGTCCTAGCAAGGACTCAA CTGCAGTCTAGACT), b (AGGGATCCGTCCTAGCAAGGGGC TGCTACCGGAAGCTTCT), r (AGGAATTCAACCACCGCTCAA CTCAACTGCAGTCTAGACT), h (AGAAGCTTCCGGTAGCAGC CTGAGCGGTGGTTGAATTCCT) – which are similar to those in the study by Duckett & Lilley [15]. Linear double-stranded DNA was pooled from x-FAM and x-com oligonucleotide, while Holliday structure was collected from x-FAM, b, r and h oligonucleotides. The pooling was performed as follows. 5l of 1 μm of solution of labeled oligonucleotide (x-FAM) and 5l 2 mM of solutions of other oligonucleotides were added to 20 mcl of double buffer (Tris-HCl 25 mM, pH 7.5; NaCl 150 mM), and if necessary, adjusting the volume of the reaction mixture to 40 μm of deionized water. The reaction mixture was heated in a water bath to a temperature of 95 °C, and then passively cooled over 3 hours to 25 °C.