ACCEPTED AUTHOR VERSION OF THE MANUSCRIPT: Evaluation of the tRNA-Leu (UUR) gene haplotype profile observed in canine mammary gland tumours based on comparative analysis with the MT-TL1 human gene

Evaluation of the tRNA-Leu (UUR) gene haplotype profile observed in canine mammary gland tumours based on comparative analysis with the MT-TL1 human gene, Abstract The aetiology and pathogenesis of many canine tumours are likely to be similar to cancers found in humans. This study aimed to evaluate a plausible link between changes in the tRNA-Leu (UUR) gene and the carcinogenesis process in dogs with mammary gland tumours. The whole mitochondrial DNA (mtDNA) isolated from blood and tumour tissues of 13 dogs with malignant mammary gland tumours was sequenced. The present work is the first report showing that some polymorphisms might occur at the corresponding positions in the human and canine mtDNA genome, which in turn may provoke similar deleterious effects. The homology between the human MT-TL1 and canine tRNA-Leu (UUR) genes was 84%. After resequencing of the whole mitochondrial DNA genome with the use of the NGS technology, two polymorphisms in two haplotypes were identified: m.2683G>A (observed in 18 out of 27 samples) and m.2678_2679insG (27 out of 27 samples). The m.2683G>A polymorphism corresponded to a deleterious change at m.3243A>G, which is linked with MELAS (Mitochondrial Encephalomyopathy, Lactic Acidosis, Stroke-like episodes) syndrome and with different types of cancers in humans as well. The comparative analysis of MT-TL1 and tRNA-Leu (UUR) led us to hypothesise that the m.2678_2679insG and m.2683G>A polymorphisms might influence the dog’s condition and might be linked with tumourigenesis, as observed in humans.

Although functional mitochondria are necessary for cancer cell growth and tumourigenesis, numerous mtDNA mutations and reduced mtDNA copy numbers altering OXPHOS (oxidative phosphorylation) have been reported to be common in cancers (Larman et al., 2012;Wallace, 2012;Singh et al., 2017 a, b). Different modifications of mitochondrial functions have been observed in human and canine cancer biology, including shifts in energy production, disruption of apoptosis signalling, an increase in mtDNA mutations, and altered antioxidant activity and reactive oxygen species production (Ślaska et al., 2013;Singh et al., 2017 c;Surdyka et al., 2017). There are several regions of mitochondrial DNA that have preserved its conservative character in many organisms, e.g. CYB, COI, and 12s rRNA, and are widely used as barcodes for identification purposes (Hebert et al., 2010;Kowal et al., 2020).
The information about functional regions in human mtDNA is stored in the Mitomap database (Lott et al., 2013); however, as far as we know, there is no similar database of canine mtDNA, except for the Mamit-tRNA database for comparison of tRNA genes (Helm et al., 2000). The specific information about the exact localisation of similar functional regions in canine mtDNA is limited and regards mainly the structure and changes in the D-loop (Pereira et al., 2004) for forensic purposes. This paper reports the polymorphisms in two haplotypes in the canine mtDNA tRNA-Leu (UUR) gene detected in a large-scale analysis of the mtDNA genome from dogs with mammary gland tumours.
The aim of this study was to present a plausible link between changes in two tRNA-Leu (UUR) gene haplotypes and the carcinogenesis process in dogs with mammary gland tumours, taking into account the comparative analysis of human MT-TL1 and canine tRNA-Leu(UUR) in order to find similarities.

Material and methods
We analysed 27 entire mitochondrial genomes. DNA was isolated from thirteen blood samples (n=13) and fourteen post-operative tumour tissues (n=14) obtained from thirteen dogs (one dog had two mammary gland tumours) ( Table 1). The dogs obtained neither chemotherapy nor radiotherapy. The tumour tissue sample was placed in a sterile container, and blood was sampled into sterile test-glasses with an anticoagulant. Histopathology analysis of the tumour tissue samples was performed. The malignancy degree of the mammary gland tumour was assessed using the 3-grade scale of malignancy, i.e. a sum of point values assigned to histomorphological traits according to Goldschmidt et al. (Goldschmidt et al., 2011). The DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) was used to extract DNA from tumour tissue and blood. DNA samples were assessed quantitatively and qualitatively by spectrophotometric measurement (DeNovix NanoDrop™ Alternative, DS-11) and electrophoretic separation in agarose gel. Selective amplification of mitochondrial DNA was held on the total genomic DNA using two pairs of primers obtained from the literature: F1418 and R11041, ~9,5kb PCR product, 9190F and R2382, ~9,8kb PCR product (Imes et al., 2012).
The library preparation and NGS sequencing on an Illumina MiSeq sequencer was performed according to the methodology proposed by Kowal et al. (2019). Full coverage of mitochondrial DNA was obtained after amplification of two long-range PCR products. PCR reactions were carried out using KAPA HiFi PCR Kit reagents (KAPA Biosystems, Wilmington, USA).
Mitochondrial DNA was sequenced in the DNA Sequencing and Synthesis Facility (oligo.pl) at the Institute of Biochemistry and Biophysics PAS. Approximately 1ng of the PCR DNA template mix was used as an input and an Illumina shotgun library was constructed using the Nextera XT Kit (Illumina, San Diego, CA) following manufacturer's instructions. The library sample was sequenced on an Illumina MiSeq sequencer (Illumina, San Diego, CA) using a 600cycle kit (v3) in a paired-end mode targeting at least 100x coverage. The data obtained after NGS sequencing was stored in NCBI BioProject number PRJNA679417. The sequence reads were filtered by quality and the remaining adaptors were removed using the fastp tool (Chen et al., 2018). Although the whole mitochondrial DNA sequences were obtained, we focused only on an analysis of the tRNA-Leu (UUR) gene in this article. The sequences of mtDNA from the blood and tumour were analysed with the use of bioinformatics programs (Unipro UGENE (v.34.0) (Okonechnikov et al., 2012). The differences in the nucleotide sequence were determined by comparison of the analysed sequences with the reference sequence (GenBank accession No. U96639) (Kim et al., 1998). The comparative analysis of the human and canine tRNA-Leu (UUR) genes was carried out in the UGENE program by building an alignment of human and dog reference sequence fragments (human Revised Cambridge Reference Sequence GenBank -rCRS -accession No. NC_012920.1). The mtDNA tRNA-Leu (UUR) genes obtained after sequencing were aligned with CLUSTAL O and MAFFT algorithms in order to obtain the highest similarity between the sequences. The secondary structure of the MT-TL1 gene was determined using the tRNA-scan program (Lowe et al., 2016). The HGVS (Human Genome Variation Society) (2016) nomenclature was used for the description of variants of sequences found in the mtDNA (den Dunnen et al., 2016).

Results
The overall comparative analysis of the human MT-TL1 gene and the canine tRNA-Leu  Figure 1).
Haplotype 1 was observed in nine out of the 13 dogs (69%), whereas haplotype 2 was observed in four out of the 13 dogs (31%). None of the analysed samples from the dogs with malignant mammary tumours had the same haplotype as the reference sequence from the GenBank database. All changes were homoplasmic. Haplotype 1 was present equally in dogs with all examined types of mammary gland tumours. We found no mutations in either the blood or the tumour tissue samples. MT-TER is a 28bp-long transcription terminator area spanning the end of RNR2 (16s rRNAcoding gene) and MT-TL1 in human mtDNA (Christianson et al., 1988).

Discussion
The dog has become a promising model for studying human genetic diseases, including cancers, hence the increased interest in canine genomics (Stacey et al., 2007;. In comparison to humans, dogs are affected by specific types of cancers more frequently: 35 times more often by skin cancer, four times more often by mammary gland cancer, eight times more often by bone cancer, and twice more often by leukaemia (Cullen et al., 2008). Mammary tumours are the second most common canine cancers with a high incidence in female dogs, accounting for 52% of all diagnosed tumours, and over 80% of them are malignant (Ślaska et al., 2013).
To date, most research of alterations in canine mtDNA has been focused on the hyper- The knowledge of changes in the mitochondrial DNA in mammary gland tumours is insufficient, and no comparative analysis between the human and canine mtDNA sequences has been done before. Based on the comparative analysis of the human and canine mitochondrial DNA genome, it can be stated that the observed polymorphisms might affect tRNA-Leu (UUR) and cause pathological changes in the dog's organism or even lead to carcinogenesis. It is worth noticing that the tRNA-Leu (UUR) gene is responsible for transporting one of the most frequently used amino acids in the mitochondrial protein structure. Mitochondrially encoded proteins are mostly hydrophobic (Brandon et al., 2006). Therefore, it is not surprising that the tRNA transporting hydrophobic amino acids are the most evolutionarily conserved. The changes in the tRNA-Leu (UUR) structure were found between the D-arm and the acceptor stem and in the DHU loop. Determining whether these variants are pathogenic is critical, but confirmation of the effect of the variants on mitochondrial function may be challenging (Bulduk et al., 2020). Sonney et al. (2017) used available databases of the alignment of benign and pathogenic variants between diverse tRNAs, structural information, and comparative genomics to predict the impact of all possible single-base variants and deletions (Bulduk et al., 2020). The highest pathogenicity is mainly observed in the D-stem and Anticodon loops, but it should be emphasised that the disruptions in the D-loops might also affect the tertiary structure of tRNA.
The tRNA genes are crucial for the proper functioning of the OXPHOS chain and for the whole organism. One of the frequently observed mutations (m.3243A>G) in the MT-TL1 gene in humans causes mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS syndrome) (Bulduk et al., 2020). MELAS syndrome is a multi-organ disease with broad manifestations, including stroke-like episodes, dementia, epilepsy, lactic acidemia, myopathy, recurrent headaches, hearing impairment, diabetes, and short stature (El-Hattab et al., 2015). The most common mutation associated with MELAS syndrome is the m.3243A>G mutation in the MT-TL1 gene encoding the mitochondrial tRNA-Leu (UUR) gene. Over 80% of MELAS patients carry the m.3243A>G mutation in the mitochondrial genome (Rahman et al., 2009). The m.3243A>G mutation results in impaired mitochondrial translation and protein synthesis, including the mitochondrial electron transport chain complex subunits, leading to impaired mitochondrial energy production. The inability of dysfunctional mitochondria to generate sufficient energy to meet the needs of various organs results in the multi-organ dysfunction observed in MELAS syndrome (El-Hattab et al., 2015). Even though these symptoms are typical for patients diagnosed with MELAS syndrome, it cannot be confirmed that the same effect would be observed in dogs. The dogs taken to this study were diagnosed with malignant mammary gland tumours; moreover, the direction of the polymorphism is reversed (G>A instead of A>G) as in the MELAS case. However, it should be emphasised that lactic acidosis might appear as a significant biomarker provoked by a genetic alteration in the canine tRNA-Leu (UUR) gene. Reprogramming of biochemical pathways is a hallmark of cancer cells, and the generation of lactic acid from glucose/glutamine represents one of the consequences of such metabolic alterations (Brown et al., 2020). Cancer cells export lactic acid out to prevent intracellular acidification, thereby not only increasing lactate levels but also contributing to an acidic pH value in the extracellular milieu. Lactate and protons in the tumour microenvironment are not innocuous bystander metabolites, but have special roles in promoting tumour-cell proliferation and growth (Brown et al., 2020). Queen et al. (2017) suggest a plausible role of the m.3253 and m.3254 variants, which may stabilise the DHU stem and suppress the deleterious effect of the m.3243A>G change (Queen et al., 2017). It is yet unclear whether the same compensatory effect is observed in the reversed change in dogs with the m.2863G>A polymorphism. Although the polymorphism was not observed either in the tumour samples or even in the heteroplasmy state, it cannot be excluded that it provokes tRNA molecule instability.
The region surrounding m.3243 is an etiologic hotspot for mutations in humans (Moraes et al., 1993). Lorenc et al. (2003) analysed mtDNA variations in various cancer samples by comparing them with normal tissue controls and identified the m.3243A>G mutation in a colon cancer sample and m.3244G>A in lung cancer. A similar mutation was observed by Mayr et al. (2008) and Meierhofer et al. (2006) in studies of renal cell oncocytoma (Meierhofer et al., 2006;Mayr et al., 2008). The latest findings presented by Xu et al. (2019) revealed the presence of m.3243A>G in a non-invasive encapsulated follicular variant of papillary thyroid carcinoma.
Thus, it cannot be excluded that m.2683A>G plays a significant role in cancer progression or proliferation in canine malignant mammary gland tumours. However, this hypothesis should be supported by further research conducted on a larger cohort.
The functional MT-TER region localised in 3229-3256 bp spans the MT-RNR1 and MT-TL1 genes in human mtDNA (Christianson et al., 1986(Christianson et al., , 1988Kruse et al., 1989;Hyvärinen et al., 2007). The MT-TER region was recognised as a plausible termination transcription region in mitochondrial DNA. Transcription and replication of mtDNA have been regarded as interlinked processes. The primer for initiation of DNA replication has been assumed to be a product of transcription by the mitochondrial RNA polymerase. However, there is no consensus concerning the mechanism by which 3' ends are generated for extension by DNA polymerase, variously proposed to be RNA processing by endonuclease MRP or protein-independent termination at one of the conserved sequence blocks of the non-coding region (NCR) (Hyvärinen et al., 2007). Mitochondrial transcription termination factors (MTERFs) regulate mitochondrial gene transcription and metabolism in numerous types of cells (Sun et al., 2019). Termination at the 3' end of rDNA is brought about by a transcription termination factor mTERF (Christianson et al., 1988;Fernández-Silva et al., 2003), which has also been proposed to interact with the RNA polymerase in initiation site selection (Martin et al., 2005). However, the mechanism of termination of heavy strand promoter (HSP) transcription is still unclear. It was previously suggested that mitochondrial termination factor 1 (MTERF1) bends the mtDNA connecting the HSP1 promoter site and its apparent tRNA-Leu (UUR) termination site. MTERF1 would then induce transcription termination through base flipping and DNA unwinding. This model was originally proposed to explain the 50-fold higher abundance of mitochondrial rRNAs. However, more recent evidence contradicts this hypothesis (D'Souza et al., 2018). Yet, it should be emphasised that the MT-TER region in humans has a high homology rate compared with the corresponding canine mtDNA region (unpublished data -79% comparing human and canine reference sequences and 86% comparing the human reference sequence and the sequence obtained from dogs with mammary gland tumours with two tRNA-Leu (UUR) polymorphisms). It is worth underlining that changes in this region might alter mitochondrial DNA transcription and replication in dogs, leading to disruptions that may provoke cancer progression and proliferation.

Conclusions
The comparative analysis of the human and canine mtDNA genome revealed a high homology rate (84-87%) in the MT-TL1 gene. The presence of the canine polymorphism m.2683G>A corresponded to the human m.3243A>G deleterious alteration, which is frequently diagnosed in MELAS syndrome patients. Moreover, such an alteration in humans has frequently been linked with different types of cancers, i.e. colon cancer, renal cell oncocytoma, and papillary thyroid carcinoma. Thus, it cannot be excluded that the presence of the polymorphisms in canine tRNA-Leu (UUR) might be linked with mammary gland tumours. It is also worth mentioning that the polymorphisms are located in the plausible MT-TER region in dogs responsible for transcription termination. However, the association of these two polymorphisms with the carcinogenesis process cannot be clearly confirmed. The authors emphasise that this research is based on bioinformatics analysis, and a further experiment based on additional canine cancer samples should be carried on for full confirmation of the hypothesis and evaluate the rate of association of the identified mitochondrial mutation with the occurrence of mammary gland tumours or other cancers in dogs.