The role of non-coding RNAs has long been underestimated. As they do not encode proteins, they did not appear to play an important biological role. However, it has now become clear that non-coding RNAs are the “puppeteers” in cells. If proteins can be said to represent the machinery that mediates molecular cell processes, then it is the tiny ribonucleic acids that control these machines. How can RNAs be detected in amongst the chaos of genomes? How can their structure be determined? How can their cellular targets be identified? Prof. Dr. Rolf Backofen’s team at the University of Freiburg is developing an indispensable tool for doing just this in the form of computer programmes that generate two-dimensional structural models of these tiny all-rounders.
Only around 1.5 per cent of the genome is translated into proteins. Nevertheless, around eighty to ninety per cent of the genome is transcribed from DNA into RNA. Why does the cell take the trouble to transcribe such a large percentage of the genome? What is the function of the large quantity of RNA generated? Researchers all over the world have found the answers to many, but not all, of these questions. Non-coding RNAs, which are rarely longer than a few hundred base pairs, can control the translation of protein precursors (so-called mRNAs, messenger RNAs) by binding to them, for example. They can also affect the activity of enzymes. RNAs are also known to regulate the transcription of genes by binding to the DNA. The absence of RNAs has also been shown to be associated with the development of cancer. “For many years we’ve just been looking at proteins and genes and we’ve asked ourselves why we were unable to understand how cells work,” said Prof. Dr. Rolf Backofen from the Institute of Computer Science (IIF) at the University of Freiburg. “It is now clear that non-coding RNAs are the molecules that pull the strings in a cell; they regulate a large proportion of the DNA and protein machineries.” Researchers are interested in finding out how the tiny molecules regulate this machinery. Biological experiments are not enough to elucidate these aspects because the function of RNAs does not depend exclusively on their base sequences. In any attempt to elucidate the function of RNAs, it is also important to take into account their two- and three-dimensional structures.
"Fortunately, nowadays we are able to predict the two-dimensional folding of RNA quite reliably," said Backofen. "This is very important because the two-dimensional folding is the basis for investigating the function of RNA." Scientists investigating the role of non-coding RNAs are really just starting out. In the majority of cases it is not even clear where in the genome the blueprints of RNAs are located. One way of possibly finding out more is to analyse the genome of different organisms. Similarities in the two-dimensional structure of the RNAs in groups of organisms of different evolutionary ages strongly suggest that these areas have barely changed during evolution. And evolutionary conservation suggests that these areas have an important function - they are assumed to be the blueprints of non-coding RNAs.
But how can researchers compare the two-dimensional structure of RNAs? "This requires intelligent computer algorithms," said Backofen. Backofen and his team are developing such algorithms. The computer programmes developed by the team enable the researchers to determine the most likely folding pattern of an RNA. The nucleotide sequence of RNAs can theoretically lead to many different folding patterns, quite in contrast to helical DNA. Different regions of the initial thread-like RNA molecule can bind to each other and form loops or so-called hairpin structures. Not all bindings are advantageous in energetic terms, and therefore break up again. Nevertheless, there are always several RNA variants in a cell. The energy required to establish a certain binding and form a two-dimensional structure can be determined by measuring the temperature required to dissolve the binding. The temperatures required to do this (so-called Turner parameters) are the basis of a computer programme being developed by Backofen and his team.
“In principle, we are able to predict the most energetically favourable folding structure of any area of an RNA molecule,” said Backofen. Adding up all the information leads to the most likely molecule structure, which the computer scientists can use to compare RNA pairs. The software LocARNA, which is available free of charge, has helped many biologists to discover non-coding RNAs in the genomes of a broad range of model organisms. In some cases, the programmes have also provided the researchers with decisive information on the RNAs’ potential biological function. One particular example is a project carried out under the “bacterial non-coding regulatory RNAs” priority programme in cooperation with Prof. Dr. Wolfgang Hess’ team from the Institute of Biology at the University of Freiburg. The objective of the project was to predict the cellular targets with which the RNAs would interact.
The programme not only calculates the likely interactions inside a molecule. It can also predict likely interactions between two molecules. Backofen and his colleagues succeeded in clarifying the two-dimensional structure of the molecules. The software was used to make this structure interact with different mRNAs and it calculated the most likely bindings in terms of energy state. The researchers found six mRNAs that were likely binding partners of non-coding RNA. The six possible target structures were subsequently investigated in closer detail by the team's biologists. It turned out that the activity of three out of the six mRNAs was regulated by non-coding RNAs. The remaining three RNAs were not fully transcribed, so it was not possible to obtain any information on their regulation. The researchers will now focus on finding out the effects of RNA regulation.
Backofen and his team are working with groups of researchers from Freiburg, Germany, Austria, New Zealand, Canada, the USA and many other countries. In future, the cooperation partners have plans to develop programmes that can make predictions on the temporary course of RNA interactions. Such programmes are necessary because cells are dynamic networks of molecular interactions. Complex mathematics and huge amounts of computing power will be necessary to be able to do this. However, one thing has already become clear: biologists will never be able to fully understand the processes in cells without such programmes. Therefore, empirical biology and computer sciences need to work together even closer than before.
Further information:Prof. Dr. Rolf Backofen Department of Bioinformatics Institute of Computer ScienceAlbert-Ludwigs-Universität Freiburg Georges-Köhler-Allee 106 D-79110 Freiburg Tel.: +49 (0)761/203-7461Fax: +49 (0)761/203-7462E-mail: backofen(at)informatik.uni-freiburg.de