ERK1 and ERK2 present functional redundancy in tetrapods despite higher evolution rate of ERK1

Abstract

Background: The Ras/Raf/MEK/ERK signaling pathway is involved in essential cell processes and it is abnormally activated in ~30 % of cancers and cognitive disorders. Two ERK isoforms have been described, ERK1 and ERK2; ERK2 being regarded by many as essential due to the embryonic lethality of ERK2 knock-out mice, whereas mice lacking ERK1 are viable and fertile. The controversial question of why we have two ERKs and whether they have differential functions or display functional redundancy has not yet been resolved.

Results: To investigate this question we used a novel approach based on comparing the evolution of ERK isoforms' sequences and protein expression across vertebrates. We gathered and cloned erk1 and erk2 coding sequences and we examined protein expression of isoforms in brain extracts in all major clades of vertebrate evolution. For the first time, we measured each isoforms' relative protein level in phylogenetically distant animals using anti-phospho antibodies targeting active ERKs. We demonstrate that squamates (lizards, snakes and geckos), despite having both genes, do not express ERK2 protein whereas other tetrapods either do not express ERK1 protein or have lost the erk1 gene. To demonstrate the unexpected squamates' lack of ERK2 expression, we targeted each ERK isoform in lizard primary fibroblasts by specific siRNA-mediated knockdown. We also found that undetectable expression of ERK2 in lizard is compensated by a greater strength of lizard's erk1 promoter. Finally, phylogenetic analysis revealed that ERK1 amino acids sequences evolve faster than ERK2's likely due to genomic factors, including a large difference in gene size, rather than from functional differences since amino acids essential for function are kept invariant.

Conclusions: ERK isoforms appeared by a single gene duplication at the onset of vertebrate evolution at least 400 Mya. Our results demonstrate that tetrapods can live by expressing either one or both ERK isoforms, supporting the notion that ERK1/2 act interchangeably. Substrate recognition sites and catalytic cleft are nearly invariant in all vertebrate ERKs further suggesting functional redundancy. We suggest that future ERK research should shift towards understanding the role and regulation of total ERK quantity, especially in light of newly described erk2 gene amplification identified in tumors.

Fig. 1

ERK1 and ERK2 protein appear first in bony vertebrates (Bichir). Western-blot analysis of samples from brain, spinal cord or cultured cells: immunoblots were incubated with polyclonal anti-ERK antibody (ERK1 and ERK2; panels a, b, c) or anti-phosphoERK antibody (pERK1 and pERK2; panels d, e, f). Taxonomic names are listed in Additional file 7. Predicted MW sizes of vertebrate ERKs range from 41.7 kDa for lamprey’s ERK to 44.1 for zebrafish’s ERK1. SC: spinal cord; CE: cerebellum; OP: optic lobe. Positions of ERK1, ERK2, phospho-ERK1 (pERK1) and phospho-ERK2 (pERK2) are indicated on the sides

Fig. 2

Evolution of vertebrate ERK sequences. a Maximium-likelihood phylogenetic analysis of erk nucleotide and (b) derived ERK amino-acid sequences, at key nodes of vertebrate evolution. erk1/ERK1 and erk2/ERK2 phylogenies are indicated by red and blue branches respectively. Key taxonomic clades for this manuscript (tetrapods and Chondrichthyes) are indicated by brackets. Accession numbers and common names are listed in Additional file 7. Branch length is indicative of degree of sequence divergences. The order of species in the derived trees is noted to vary slightly between different topological methods and whether it is calculated using protein or nucleotide sequences. Given that all vertebrate ERK sequences are highly conserved, minor variations in tree placement can be expected

Fig. 3

ERK2 is not detected in snakes and lizards. a, b Western-blot analysis performed as in Fig. 1 incubated with a polyclonal anti-ERK antibody (upper panels) or anti-phosphoERK antibody (lower panels). a brain extracts from snakes, lizards, crocodile and turtles (full taxonomic names in Additional file 7). Control brain extracts from cattle. CO: cortex, CE: cerebellum, ME: medulla, OL: olfactory bulb, PI: pineal gland, W: whole brain. Note that degraded bands are observed in turtle E. orbicularis likely due postmortem sampling. b extracts from lizard embryo fibroblasts were stimulated or not (NS) one hour in presence of FCS alone (FCS) or FCS + phosphatase inhibitor NaVO4^—(FCS + NaVO4^—) to activate ERKs

Fig. 4

Only ERK1 protein and erk1 mRNA are expressed significantly in lizards. a siRNA sequences targeting anole lizard (A. carolinensis) erk1 mRNA (ERK1-A, ERK1-B and ERK1-C) and anole lizard erk2 mRNA (ERK2-A, ERK2-B and ERK2-C). In red, the nucleotides that differ when both erk1 and erk2 isoforms are aligned. b, c siRNA pools targeting erk1 or erk2 or control-siRNA were transfected three days prior to mRNA harvesting. b Absolute quantification of erk1 mRNA (left panel) or erk2 mRNA (right panel) after siRNA transfection, in A. sagrei and A. carolinensis lizard embryo fibroblasts (LEFs). erk1 or erk2 mRNA quantities were determined relatively to quantities of linearized plasmids harboring either A. carolinensis erk1 or erk2 cDNAs as described in materials and methods. To ease comparison between erk1 and erk2, data are expressed as percentage of siControl quantities. Bars represent mean +/− s.d. n = 4. c Western-blot analysis of samples from two independent transfections loaded on the same gel (Experiment 1 and 2) (insert) Fluorescence quantification of western blot. d Number of erk1 and erk2 mRNA molecules per 25 ng of total RNA from reptile samples: lizard (A. carolinensis), Crocodile (C. niloticus), turtle (T. scripta elegans) and mouse NIH3T3 cells. Bars represent means +/− s.d. n = 3 (e) Firefly luciferase activity driven by mouse and anole lizard ERK promoters (1 kb upstream from start ATG). Normalization by co-transfection of a plasmid expressing Renilla luciferase driven by the thymidine kinase promoter. Transfection was performed in A. carolinensis fibroblasts (left panel) or mouse NIH3T3 fibroblasts (right panel). Bars represent means +/− s.d. n = 4

Fig. 5

Only amino-acid neutral for function diverge among mammalian ERKs. a Ratio of non-synonymous rate (Ka) over synonymous rate (Ks) for all available erk1 and erk2 pairs of mammalian gene sequences (37 pairs). Multiple sequence alignment was generated and Ka/Ks values were then computed at each codon. Abscissa: distance from start ATG codon. The 5′ end of the sequences (first exon) could not be analyzed because some sequences were incomplete. b Still images were generated for 3D PDB structure 4QTB (Human ERK1) displaying the following highlighted positions: amino-acids interacting with substrates forming the DEJL(KIM) motif are colored in blue, light blue for docking groove and darker blue for acidic patch; amino-acids interacting with substrates via the DEF domain (FXFP) are colored in green; the threonine and tyrosine phosphorylated upon activation by MEK are colored in white and the “face of the kinase” is circled by a dashed white line. Amino-acids that diverge among at least 5 % of 49 mammalian ERK1s are colored in yellow (17 amino-acids). i) front-side of the kinase where ATP-transfer occurs in the catalytic cleft between the two lobes ii) left-side of the kinase iii) back-side of the kinase iv) right-side of the kinase

Fig. 6

In vertebrates, erk2 genes are larger than erk1 genes. Genomes from vertebrates were screened for the presence of the full length sequences of erk genes, and then the size from ATG to stop codon was calculated in kilobases. In teleosts and mammals, animals for which both erk1 and erk2 genes sequences are available were preferred, only one pair of mammalian erks is not from the same animal. Full erk1 gene sequence of coelacanth (L. chalumnae) is not available. For teleosts, the erk genes of T. fubripes and T. nigroviridis were not taken into account due to their known extreme genome compactness that would skew the size distribution (smallests vertebrate genomes) their erk1 gene size being 5 and 3.9 kb respectively and their erk2 gene size being 9.7 and 10 kb respectively). For teleosts and mammals, the average sizes of the genes are written on the upper left side of the graphical bars, and the p-value that evaluates the statistical significance of the difference between erk1 and erk2 gene sizes is indicated. For reptiles, the fold difference between the sizes of erk1 genes and erk2 genes is indicated on top of the bars with an arrow. Silhouettes are from http://phylopic.org/