The role of non-AUG codons in translation initiation and localisation of mitochondrial proteins

The DNA sequence in every cell of the body, termed the genome, stores the instructions to make all the proteins that are the essential building blocks needed for living. The code for each protein is stored in shorter stretches of DNA called genes. Converting the DNA sequence of a gene (which has only four "letters") to that of its corresponding protein (made up of twenty different kinds of amino acids) requires two major processes: transcription and translation.

Transcription is the copying of the sequence information in the DNA into a similar molecule called messenger RNA (mRNA). The mRNA messages are then decoded into the amino acid sequence using the process of translation (so called because it goes from the language of DNA/RNA, to the language of amino acids). The "Genetic Code", which is used to translate nucleotides (RNA) into amino acids (protein), is well established and it is easy to predict what amino acids are encoded by a given stretch of DNA. The process is complicated, however, by the presence of "untranslated regions" at the ends of the mRNA, which do not encode any protein sequence. As a consequence, in order to correctly translate a protein, it is important to know where translation begins.

The complex molecular machine that translates the mRNA sequence is called the ribosome, which starts making a protein when it finds a particular sequence in the mRNA called a translation initiation codon, which usually has the sequence "AUG". This project will advance our knowledge regarding the nature of these translation initiation codons. Many of the rules of translation initiation remain unclear therefore the more we know, the more we can understand from existing data. This is particularly true as improvements in sequencing technology means that an ever-increasing proportion of our knowledge about the protein universe is derived purely from applying these rules computationally to DNA sequence data.

It is well established that multiple proteins can be produced from a single gene by generating different mRNA sequences during transcription. A much more recent finding is that translation can similarly produce different proteins from the same mRNA by the ribosome beginning to translate the protein at different positions, making longer or shorter versions of the protein. This project is concerned with how and why different translation initiation codons are used and how widespread this phenomenon is. So far, there are only a few examples where this has been discovered but those that are known are very important. In fact alternative initiation codons can be used to make new forms of proteins which have completely different functions or go to different places within the cell. Furthermore, it is now becoming clear that the initiation codon itself does not have to be the AUG triplet and the use of what we describe as non-canonical initiation codons is the focus of our proposed work.

We have successfully identified translation initiation from non-AUG codons, and in this project we will particularly focus on genes that make proteins with roles in the "batteries" of the cell, the mitochondria. We believe that important signals within the proteins which help target them to this part of the cell have been ignored because they are made by starting from non-AUG codons. This means that computational methods using the wrong rules will have missed them. We have already proven this phenomenon in one gene, and once we have successfully identified novel initiation codons in further candidates, we will then examine what the consequences are for the proteins that are produced. We will certainly identify the signals that are involved in moving proteins to the mitochondria, but may also find new roles for the newly identified protein sequence.

Eukaryotic translation initiation is a key point in gene expression. Central to translation initiation is the selection of the initiation codon, usually an AUG. However, other non-AUG codons may be used and we will investigate the usage of alternative initiation codons (AICs) and how these result in synthesis of N-terminally extended proteins.

We have identified instances where the open reading frames of a number of proteins can be extended by initiation at non-AUG AICs. In particular, RPP25 has novel N-termini, translated from GUG and CUG codons, containing sequences that target the extended proteins to the mitochondria and nucleolus. As RPP25 is a subunit of the nuclear RNaseP and the mitochondrial equivalent MRP, our work has therefore identified the signals necessary for RPP25 to localise to these subcellular compartments. We have further bioinformatic evidence for this phenomenon in other genes, suggesting that mitochondrial targeting peptides (MTPs) have been hidden within genomic data because non-AUG codons do not figure in the rules that databases apply when annotating new genes.

We will examine further candidate genes using the assays we have developed, and the presence of N-terminal MTPs will be confirmed by immunofluorescence. As targeting of proteins to mitochondria and other intracellular destinations is a dynamic event, we aim to visualise both protein movement and initiation codon selection in real time.

We will also study how novel N-termini alter other functions of the proteins, beyond their targeting role. One finding from our work is that non-AUG codons are essential in forming a competent MTP. We will determine whether novel N-termini have new sites of post-translational modification, which only occur when an amino acid other than methionine is used for initiation.

This work will thus examine the importance of alternative TIC selection in the generation of protein isoform diversity, a previously neglected aspect of gene expression.

This proposal works alongside several long term projects we are undertaking to examine the expansion of the proteome by use of alternative translation initiation codons, which will have major implications in the areas of biotechnology (in particular proteomics studies) and human health. This work will make us reconsider mechanisms of gene expression and I believe that alternative initiation codons may represent as important a method for generating diversity in the proteome as alternative splicing.

Our research will benefit those working in the proteomics community as the use of alternative translation initiation codons is not often considered in such studies, meaning a considerable portion of proteomic data may be missing. Information regarding the use of alternate initiation codons, will therefore be useful in performing such analyses, so will be made available in public gene expression databases.
This project aims to identify further candidates which are likely to be expressed as different isoforms from alternative translation initiation codons. As this work expands, the physiological roles of such different isoforms will be explored. This work will therefore be of interest to workers in many fields, not just gene expression or mitochondrial biology, where this work is focused.

A long-term objective of our research is to determine whether the use of alternative initiation codons can be used as a biomarker to determine the status of gene expression in diseased tissues. The groundwork for this approach has already been proven for the chaperone BAG-1, and we are studying the translational regulation of the most 5' non-canonical CUG initiation codon of this protein as part of another BBSRC funded project. Increased nuclear staining of BAG-1 in breast cancer (correlating with increased usage of the most 5' non-canonical CUG initiation codon) is associated with improved long-term prognosis and MJC is a co-investigator on a grant awarded by the Breast Cancer Campaign to Dr Ramsay Cutress (Faculty of Medicine, University of Southampton) to investigate BAG-1 as a therapeutic target in HER2 positive breast cancer. Being able to assay a patient sample for different isoforms of a protein that arise from alternative initiation codons, where the isoforms have defined properties, may enable diagnoses to be made. One simple way in which this could be performed would be a simple ELISA test, if suitable antibodies to different isoforms could be developed. Furthermore, it is likely that at least some of the N-terminal extensions identified during the course of this project will represent possible therapeutic targets.

There is already a large and successful infrastructure in place at the University of Southampton to exploit novel biological research in the clinical environment. I also benefit from the extensive networking opportunities afforded to me in the newly built Institute for Life Sciences and as an active member of the South Coast RNA Network as well as the wider Translation Control community. This will ensure that I am able to disseminate this work to the largest audience possible and create strong collaborations in the future. The selection of alternative initiation codons may be amenable to therapeutic interventions and prospective collaborations with industrial partners will be sought at the earliest stages of this project, in order to create the biggest potential impact in improving quality of life and generating economic success.