Limits...
Malleable nature of mRNA-protein compositional complementarity and its functional significance.

Hlevnjak M, Zagrovic B - Nucleic Acids Res. (2015)

Bottom Line: We show that for most proteins there exist cognate mRNAs that improve, but also significantly worsen the level of native matching (e.g. by 1.8 viz. 7.6 standard deviations on average for H. sapiens, respectively), with the least malleable proteins in this sense being strongly enriched in nuclear localization and DNA-binding functions.Even so, we show that the majority of recodings for most proteins result in pronounced complementarity.Our results suggest that the genetic code was designed for favorable, yet tunable compositional complementarity between mRNAs and their cognate proteins, supporting the hypothesis that the interactions between the two were an important defining element behind the code's origin.

View Article: PubMed Central - PubMed

Affiliation: Department of Structural and Computational Biology, Max F. Perutz Laboratories, University of Vienna, Campus Vienna Biocenter 5, 1030 Vienna, Austria.

Show MeSH

Related in: MedlinePlus

A schematic of the recoding procedure. We recode native mRNA sequences by varying the pyrimidine content of the codons corresponding to Leu, Ile, Val, Thr, Ala, Pro, Gly, Ser or Arg, while protein sequences do not change. Amino acids with codons whose pyrimidine content can vary are shown in blue, and their corresponding codons in green when native, or red when recoded. In steered recoding (left), we randomly change one codon in each out of 10 000 cycles, evaluate the complementarity by using the Pearson correlation coefficient (Rnew) and accept the change if it goes in the desired direction or reject it if not, finally optimizing mRNA-protein matching (Rbest) or mismatching (Rworst). In non-steered recoding (right), we recode native mRNAs independently 10 000 times without optimizing the native level of matching (Rnative), resulting in a range of Rs (R1, R2, …, R10000) for each native mRNA.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4381073&req=5

Figure 1: A schematic of the recoding procedure. We recode native mRNA sequences by varying the pyrimidine content of the codons corresponding to Leu, Ile, Val, Thr, Ala, Pro, Gly, Ser or Arg, while protein sequences do not change. Amino acids with codons whose pyrimidine content can vary are shown in blue, and their corresponding codons in green when native, or red when recoded. In steered recoding (left), we randomly change one codon in each out of 10 000 cycles, evaluate the complementarity by using the Pearson correlation coefficient (Rnew) and accept the change if it goes in the desired direction or reject it if not, finally optimizing mRNA-protein matching (Rbest) or mismatching (Rworst). In non-steered recoding (right), we recode native mRNAs independently 10 000 times without optimizing the native level of matching (Rnative), resulting in a range of Rs (R1, R2, …, R10000) for each native mRNA.

Mentions: Two types of recoding procedures were used for sampling the mRNA sequence space: steered and non-steered. In steered recoding, each mRNA sequence went through 10 000 steps of a Monte Carlo-type procedure in which at each step a single, randomly chosen codon was reassigned to another synonymous codon. Reassignments were carried out only at those positions at which choosing a synonymous codon could lead to a change in pyrimidine content, and were attempted following the frequency of a given codon type in the standard genetic code. In this way, each step of the Monte Carlo procedure resulted in a newly recoded mRNA, which then served as the input for the next cycle of recoding. Since in steered recoding the aim was to either increase or decrease the level of matching between a protein's PR profile and its cognate mRNA PYR profile as compared to the native mRNA, after each codon reassignment step the two profiles were compared by calculating the Pearson R between them. If the goal was to optimize matching and a given codon change resulted in improved matching, the change was accepted, and if not, the codon selection was repeated. Conversely, if the goal was to optimize mismatching, changes which increased the level of mismatching were selected and others rejected. In this way, steered recoding progressively either increased or decreased the level of profile-matching between an mRNA and its cognate protein, and resulted in a single best- or worst-matched mRNA for each protein in a given proteome (Figure 1, left). In non-steered recoding, on the other hand, each mRNA was recoded independently 10 000 times by each time randomly changing all degenerate codons that could lead to a change in pyrimidine content. In this type of recoding, the same native mRNA served as input for a new recoding cycle, finally resulting in a set of 10 000 independent, recoded mRNA variants per each native mRNA in the transcriptome (Figure 1, right).


Malleable nature of mRNA-protein compositional complementarity and its functional significance.

Hlevnjak M, Zagrovic B - Nucleic Acids Res. (2015)

A schematic of the recoding procedure. We recode native mRNA sequences by varying the pyrimidine content of the codons corresponding to Leu, Ile, Val, Thr, Ala, Pro, Gly, Ser or Arg, while protein sequences do not change. Amino acids with codons whose pyrimidine content can vary are shown in blue, and their corresponding codons in green when native, or red when recoded. In steered recoding (left), we randomly change one codon in each out of 10 000 cycles, evaluate the complementarity by using the Pearson correlation coefficient (Rnew) and accept the change if it goes in the desired direction or reject it if not, finally optimizing mRNA-protein matching (Rbest) or mismatching (Rworst). In non-steered recoding (right), we recode native mRNAs independently 10 000 times without optimizing the native level of matching (Rnative), resulting in a range of Rs (R1, R2, …, R10000) for each native mRNA.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4381073&req=5

Figure 1: A schematic of the recoding procedure. We recode native mRNA sequences by varying the pyrimidine content of the codons corresponding to Leu, Ile, Val, Thr, Ala, Pro, Gly, Ser or Arg, while protein sequences do not change. Amino acids with codons whose pyrimidine content can vary are shown in blue, and their corresponding codons in green when native, or red when recoded. In steered recoding (left), we randomly change one codon in each out of 10 000 cycles, evaluate the complementarity by using the Pearson correlation coefficient (Rnew) and accept the change if it goes in the desired direction or reject it if not, finally optimizing mRNA-protein matching (Rbest) or mismatching (Rworst). In non-steered recoding (right), we recode native mRNAs independently 10 000 times without optimizing the native level of matching (Rnative), resulting in a range of Rs (R1, R2, …, R10000) for each native mRNA.
Mentions: Two types of recoding procedures were used for sampling the mRNA sequence space: steered and non-steered. In steered recoding, each mRNA sequence went through 10 000 steps of a Monte Carlo-type procedure in which at each step a single, randomly chosen codon was reassigned to another synonymous codon. Reassignments were carried out only at those positions at which choosing a synonymous codon could lead to a change in pyrimidine content, and were attempted following the frequency of a given codon type in the standard genetic code. In this way, each step of the Monte Carlo procedure resulted in a newly recoded mRNA, which then served as the input for the next cycle of recoding. Since in steered recoding the aim was to either increase or decrease the level of matching between a protein's PR profile and its cognate mRNA PYR profile as compared to the native mRNA, after each codon reassignment step the two profiles were compared by calculating the Pearson R between them. If the goal was to optimize matching and a given codon change resulted in improved matching, the change was accepted, and if not, the codon selection was repeated. Conversely, if the goal was to optimize mismatching, changes which increased the level of mismatching were selected and others rejected. In this way, steered recoding progressively either increased or decreased the level of profile-matching between an mRNA and its cognate protein, and resulted in a single best- or worst-matched mRNA for each protein in a given proteome (Figure 1, left). In non-steered recoding, on the other hand, each mRNA was recoded independently 10 000 times by each time randomly changing all degenerate codons that could lead to a change in pyrimidine content. In this type of recoding, the same native mRNA served as input for a new recoding cycle, finally resulting in a set of 10 000 independent, recoded mRNA variants per each native mRNA in the transcriptome (Figure 1, right).

Bottom Line: We show that for most proteins there exist cognate mRNAs that improve, but also significantly worsen the level of native matching (e.g. by 1.8 viz. 7.6 standard deviations on average for H. sapiens, respectively), with the least malleable proteins in this sense being strongly enriched in nuclear localization and DNA-binding functions.Even so, we show that the majority of recodings for most proteins result in pronounced complementarity.Our results suggest that the genetic code was designed for favorable, yet tunable compositional complementarity between mRNAs and their cognate proteins, supporting the hypothesis that the interactions between the two were an important defining element behind the code's origin.

View Article: PubMed Central - PubMed

Affiliation: Department of Structural and Computational Biology, Max F. Perutz Laboratories, University of Vienna, Campus Vienna Biocenter 5, 1030 Vienna, Austria.

Show MeSH
Related in: MedlinePlus