Conserved Double Translation Initiation Site for Δ160p53 Protein Hints at Isoform’s Key Role in Mammalian Physiology

p53 is the most commonly mutated gene in human cancers. Two fundamental reasons for this are its long protein isoforms protect from cancer, while its shorter C-terminal isoforms can support cancer and metastasis. Previously, we have shown that the Δ160p53 protein isoform enhances survival and the invasive character of cancer cells. Here, we identified a translation initiation site nine codons downstream of codon 160—the known initiation codon for the translation of Δ160p53—that is recognized by the translation machinery. When translation failed to initiate from AUG160 due to mutation, it initiated from AUG169 instead, producing similar levels of a similar protein, Δ169p53, which promoted cell survival as efficiently as Δ160p53 following DNA damage. Interestingly, almost all mammalian species with an orthologue to human AUG160 also possess one for AUG169, while none of the non-mammalian species lacking AUG160 have AUG169, even if that region of the p53 gene is well conserved. In view of our findings, we do not believe that Δ169p53 acts as a different p53 protein isoform; instead, we propose that the double translation initiation site strengthens the translation of these products with a critical role in cell homeostasis. Future studies will help verify if this is a more general mechanism for the expression of essential proteins in mammals.


Introduction
p53 is a gene well conserved among species, from placozoans to humans [1]. It plays a central role in animals with true tissues (Histozoa) by maintaining the balance between cell growth/differentiation and transformation, and genetic variation and genomic catastrophe. Without p53, animals may develop cancer [2], and with too much, they may age prematurely [3]. The multiple and complex tasks of p53 are tissue-specific and shared between several RNA and protein products [4]. While longer products are well known for activating cell-cycle arrest, senescence and apoptosis, which counter tumor development [5][6][7], shorter C-terminal variants such as ∆160p53, ∆133p53 and ∆133p53β have been reported to induce cancer, invasion and metastasis [8,9]. Previously, we have shown that mutations enriched in cancer activate the translation of ∆160p53 from codon 160, which enhanced cell fitness and invasiveness [8,10]. Here, we identify a second active translation initiation site (TIS) a few nucleotides downstream of codon 160, and we investigate its effect on ∆160p53 expression and function. Understanding the mechanisms of ∆160p53 translation may help us devise a new strategy to target p53 in cancer.

Toeprinting Identifies Translation Initiation Site Downstream of Codon 160
Since we first identified ∆160p53 as a pro-oncogenic protein in 2013 [8,10], we have been interested in understanding the regulation of this molecule, especially at the translation level, which more clearly separates it from all the other p53 isoforms that can be produced from the same mRNA transcripts. If we could understand the unique regulation of this isoform then we might be able to target it specifically without affecting full-length (FL) p53 s function and that could finally make it possible to target p53 for cancer therapy. One of the common methods used to characterize the mechanisms for the regulation of translation initiation is toeprinting because it measures precisely the position of ribosomes (or ribosomal subunits) on mRNAs and can be used to identify (or confirm) TIS when it is performed in the presence of cycloheximide (CHX), an elongation inhibitor. Toeprinting was performed using in vitro transcribed wild-type (wt) FL p53 mRNA and two different primers: one well within the coding region of ∆160p53, but still within range of TIS160 (primer 1), and another in the close vicinity of codon 160 (c.160) that also covers ∆160p53 s 5 -untranslated region (5 -UTR) (primer 2). While toeprinting with primer 1 showed no fragments (no peaks in the graph), which means the reverse transcriptase could extend beyond the length limitation of the capillary (around 700 nucleotides) without being stopped by any strong RNA structure or RNA-binding protein ( Figure 1A), many were seen with primer 2 ( Figure 1B), suggesting that the 5 -UTR possesses strong secondary structures that prevent primer extension. This is a known feature of 5 -UTRs [11]. These structures are usually unwound by the ribosomes and were, with one exception, no longer observed in the presence of ribosomes from rabbit reticulocyte lysates (RRL) ( Figure 1C,D). Surprisingly, with RRL and CHX, the reverse transcriptase stopped at c.164 with either primer, which would correspond to the leading edge of 80S ribosomes initiating translation in c.169 and not c.160 [12]. Interestingly though, c.169 is also an AUG and thus a potential TIS. So, we next investigated if c.169 is used as a start site for translation in cells. Int. J. Mol. Sci. 2022, 23,

Toeprinting Identifies Translation Initiation Site Downstream of Codon 160
Since we first identified Δ160p53 as a pro-oncogenic protein in 2013 [8,10], we have been interested in understanding the regulation of this molecule, especially at the translation level, which more clearly separates it from all the other p53 isoforms that can be produced from the same mRNA transcripts. If we could understand the unique regulation of this isoform then we might be able to target it specifically without affecting full-length (FL) p53′s function and that could finally make it possible to target p53 for cancer therapy. One of the common methods used to characterize the mechanisms for the regulation of translation initiation is toeprinting because it measures precisely the position of ribosomes (or ribosomal subunits) on mRNAs and can be used to identify (or confirm) TIS when it is performed in the presence of cycloheximide (CHX), an elongation inhibitor. Toeprinting was performed using in vitro transcribed wild-type (wt) FL p53 mRNA and two different primers: one well within the coding region of Δ160p53, but still within range of TIS160 (primer 1), and another in the close vicinity of codon 160 (c.160) that also covers Δ160p53′s 5′-untranslated region (5′-UTR) (primer 2). While toeprinting with primer 1 showed no fragments (no peaks in the graph), which means the reverse transcriptase could extend beyond the length limitation of the capillary (around 700 nucleotides) without being stopped by any strong RNA structure or RNA-binding protein ( Figure 1A), many were seen with primer 2 ( Figure 1B), suggesting that the 5′-UTR possesses strong secondary structures that prevent primer extension. This is a known feature of 5′-UTRs [11]. These structures are usually unwound by the ribosomes and were, with one exception, no longer observed in the presence of ribosomes from rabbit reticulocyte lysates (RRL) ( Figure  1C,D). Surprisingly, with RRL and CHX, the reverse transcriptase stopped at c.164 with either primer, which would correspond to the leading edge of 80S ribosomes initiating translation in c.169 and not c.160 [12]. Interestingly though, c.169 is also an AUG and thus a potential TIS. So, we next investigated if c.169 is used as a start site for translation in cells.

∆169. p53 Is a Newly Identified Translation Product of Full-Length p53 mRNA
We immediately noticed two close bands around 32 KDa (the size of ∆160p53) in several cell lines such as glioblastoma cell line LN229 and A549 lung adenocarcinoma cells (Figure 2A,B). By transfecting a ∆169p53 plasmid made with p53 codons 169 to 393 into p53-null H1299 cells, we saw that the band below ∆160p53 in A549 cells matched the size of ∆169p53 ( Figure 2B). H1299 cells expressing FL p53 constructs also exhibited the double band, which became much clearer with cancer-specific mutations R175H and R248Q ( Figure 2C). This goes in line with our previous finding showing that common gain-of-function mutations in cancer activate the translation of ∆160p53 and other short p53 isoforms [8]. Interestingly, the new band was induced by DNA damage (treatment with etoposide (Eto)) in LN299 cells ( Figure 2D) and endoplasmic reticulum (ER) stress (treatments with thapsigargin (Th) and tunicamycin (Tu)) in wt p53 breast adenocarcinoma cells MCF7 ( Figure 2E), which suggests this protein may play a role in stress response. Finally, we wanted to confirm that this band expressed from FL p53 transcripts, observed below ∆160p53 and with the size of ∆169p53, was really ∆169p53. So, we created a series of frameshift and TIS mutations. First, to verify that the band of interest was a translation product and not a product of protein degradation or cleavage, we deleted one nucleotide in codon 157 in order to create a frameshift between the sequences upstream and downstream of c.157. We also added an HA-tag to the C-terminus of FL p53 in frame with AUG160 and AUG169, but out of frame-because of the frameshift mutation in c.157-with all other TIS (i.e., TIS 1, 40 and 133). WB against HA revealed no bands corresponding to FL p53, ∆40p53 or ∆133p53, as expected, but presented the double band corresponding to ∆160p53 and ∆169p53, proving that these are products of translation from TIS160 and TIS169, respectively, and not cleavage products of FL p53 or other larger isoforms as those would lack the HA-tag and would not be recognized by the antibody ( Figure 3A). Furthermore, a mutation in c.160 shut down ∆160p53 s expression, while enhancing ∆169p53 s, which has its TIS coming next in line [11], and a mutation in c.169 erased ∆169p53 s band, leaving no doubts as to the identity of these proteins. The same TIS mutations removed the same bands in a mutant (R248Q) FL p53 WB, indicating that in this context also, these proteins are translated directly from TIS160 and TIS169 ( Figure 3B).

AUG160 and AUG169 May Have Co-Evolved in Mammals
We next questioned if ∆160p53 and ∆169p53, being so similar, would work together or separately, possess similar functions or not. To gain more insight into the co-existence and possible co-evolution of these two p53 forms, we aligned the p53 mRNA sequences of 37 different mammalian species ( Figure 4A). Of note, human c.160 shows complete conservation in all these species, which suggests ∆160p53 may play a key role in mammalian evolution and physiology. AUG169 orthologs (AUG169 orth ), in turn, are present in 32 (86%) of the species tested. This, when compared to the 81% amino acid conservation in the neighboring codons, also shows some evidence of selective pressure for a second TIS. More so when considering that none of the non-mammalian species lacking an orthologous of AUG160 have one for AUG169 either (0%), even if the neighboring amino acids show 62% conservation to the human peptide sequence ( Figure 4B). Altogether, these analyses suggest that it might have been advantageous for most mammals to develop a second TIS downstream of TIS160. On the other hand, species without a TIS160 orth never had any need to create a start codon downstream.
the neighboring codons, also shows some evidence of selective pressure for a second TIS. More so when considering that none of the non-mammalian species lacking an orthologous of AUG160 have one for AUG169 either (0%), even if the neighboring amino acids show 62% conservation to the human peptide sequence ( Figure 4B). Altogether, these analyses suggest that it might have been advantageous for most mammals to develop a second TIS downstream of TIS160. On the other hand, species without a TIS160 orth never had any need to create a start codon downstream.

TIS160 Is Redundant during Stress Response
Stress of the endoplasmic reticulum (ER) activated the expression of ∆160p53 and ∆169p53 ( Figures 2E and 5A). Since HDM2 is a strong regulator of p53 and is often inactivated during stress [7,13], we first tested if our results could be explained by an alleviation in HDM2-mediated degradation of ∆160p53 and ∆169p53. However, neither HDM2, nor an inhibitor of HDM2, Nutlin-3, had the expected effect on the levels of these proteins ( Figure 5B,C, respectively). Therefore, we further investigated the translation dynamics of the two isoforms. Next, we used a dual luciferase reporter system [14] for quick evaluation of translation from one, the other or both (or none) TIS under different cell conditions. Since the wild-type p53 sequence produced very low signals in this system, we added the cancer mutation R248Q, which enhances expression ( Figure 2C). Under normal cell growth, in the presence of a rich medium, removing a single TIS led to a large drop in translation ( Figure 5D). Conversely, under mild stress, each TIS alone was sufficient for high expression ( Figure 5E). These results are coherent in light of ∆160p53 s pro-survival capacity [8], which would not be crucial under normal growth conditions but would have to be ensured under stress to safeguard the life of the cell. Of course, this will only be true if both protein isoforms, ∆160p53 and ∆169p53, possess similar survival abilities, which we next verified. of translation from one, the other or both (or none) TIS under different cell conditions. Since the wild-type p53 sequence produced very low signals in this system, we added the cancer mutation R248Q, which enhances expression ( Figure 2C). Under normal cell growth, in the presence of a rich medium, removing a single TIS led to a large drop in translation ( Figure 5D). Conversely, under mild stress, each TIS alone was sufficient for high expression ( Figure 5E). These results are coherent in light of Δ160p53′s pro-survival capacity [8], which would not be crucial under normal growth conditions but would have to be ensured under stress to safeguard the life of the cell. Of course, this will only be true if both protein isoforms, Δ160p53 and Δ169p53, possess similar survival abilities, which we next verified.  Shown are averages ± s.d. of 4 experiments (# p > 0.05, * p < 0.05, ** p < 0.01 and *** p < 0.005 compared to "both TIS" samples).

∆160p53 and ∆169p53 Have Similar Survival Capacities
To gain a better idea of the role of TIS169 in cell physiology we compared ∆160p53 and ∆169p53 in their capabilities (or lack of) to affect cell survival under conditions of mild stress (low concentration of etoposide, a DNA damaging agent, or exposure to hydrogen peroxide (H 2 O 2 ), which causes oxidative damage) since ∆160p53 was previously shown to promote resistance to mild stress though it failed to rescue cells with severe damage [8]. Etoposide induced cell death over a period of 21 h (27%), but this was greatly minimized (down to 11% and 14%) by either isoform, as both displayed very similar pro-survival qualities ( Figure 6A). H 2 O 2 treatment on the other hand was equally toxic after 24 h, but its effect could not be counteracted by the isoforms ( Figure 6B). Importantly, however, in both instances, DNA damage or oxidative damage, ∆160p53 and ∆169p53 proteins led to exactly the same phenotype, supporting our model in which both protein forms play the same role in the cell and function together as the result of a robust bipartite translation initiation process. drogen peroxide (H2O2), which causes oxidative damage) since Δ160p53 was previou shown to promote resistance to mild stress though it failed to rescue cells with sev damage [8]. Etoposide induced cell death over a period of 21 h (27%), but this was grea minimized (down to 11% and 14%) by either isoform, as both displayed very similar p survival qualities ( Figure 6A). H2O2 treatment on the other hand was equally toxic af 24 h, but its effect could not be counteracted by the isoforms (Figure 6B). Importan however, in both instances, DNA damage or oxidative damage, Δ160p53 and Δ169p proteins led to exactly the same phenotype, supporting our model in which both prot forms play the same role in the cell and function together as the result of a robust bipar translation initiation process.  Shown are averages ± s.d. of 2-4 experiments (# p > 0.05, * p < 0.05 and *** p < 0.005 compared to "Etoposide/− (no p53 expression)" sample (A) or to "− (no drug and no p53 expression)" (B) sample or as indicated). (C) Proposed model: Under normal cell growth conditions TIS160 and TIS169 show low activity and there is low expression of the corresponding protein isoforms, but under stress conditions or with mutation each of these individual TIS becomes more active and initiate translation of the two protein forms, ∆160p53 and ∆169p53, which have equal capacity to induce cell growth and survival, and which could lead to recovery from stress or, in the case of continued activation, cancer formation or cancer development.

Discussion
Based on our results we propose a model in which, during stress conditions, either AUG, 160 or 169, is sufficient for translation and phenotype manifestation via ∆160p53 and ∆169p53 s functions; though under normal conditions these TIS may be too weak to provide additional survival advantages to the cell, which is desirable ( Figure 6C). This double-TIS system seems to have appeared early in mammals, together with ∆160p53, which suggests it is a key regulatory feature of ∆160p53; i.e., for most species, ∆160p53 might not be effective enough or advantageous enough without it. During stress or mutation, translation from TIS169 can be almost as efficient as from TIS160 (Figures 2 and 5) and may play an important role in the cell's adaptation to stress ( Figure 6A) or in promoting tumor formation/development, if permanently active. It will be interesting to investigate if other stress-response genes with important roles in cell-fate decisions also possess comparable mechanisms for the regulation of translation, especially in mammals. HDM2, for example, a known proto-oncogene and p53 regulator with a strong influence on cell fitness also harbors a second downstream AUG in codon 7. In fact, Bazykin et al. observed that downstream AUGs among the first 30 codons are under selection as alternative TIS in eukaryotes, and are particularly common in transcription factors [15]. The authors proposed that these might contribute to increased efficiency of translation and/or for translation of different N-terminal protein variants. Here, our research validates the first hypothesis for the case of TIS169, though we have previously shown that it is the second hypothesis that applies to TIS40 in p53 [16]. Future research might be able to determine if the exact distance and/or sequence from upstream to downstream AUG is sufficient to predict the role of alternative AUGs in different genes. Our findings also help elucidate the mechanisms of regulation of an essential p53 isoform with possible roles in carcinogenesis and may contribute to a new therapeutic strategy in the future.

Toeprinting
Toeprinting was adapted from a previously described protocol [19]. Briefly, 500 ng of in vitro-transcribed (mMESSAGE mMACHINE T7 Ultra kit; Ambion, Austin, TX, USA) full-length p53 mRNA was combined with 20 pmol of 6-FAM 5 -labelled primer 1 (5 -ATGTAGTTGTAGTGGATGGTGG-3 ) or primer 2  in 50 mM Tris-HCl, pH 7.5, and heated to 68 • C for 2 min and cooled to 37 • C for 8 min, and immediately added to the translational mixture, comprising 20 µM amino acid mixture minus methionine, 20 µM amino acid mixture minus leucine (Promega, Madison, WI, USA), 1.5 U/µL ribonuclease inhibitor (RNaseOUT) plus or minus rabbit reticulocyte lysate (50%), and 500 µg/mL CHX (Sigma). The reactions were incubated at 30 • C for 20 min. Primer extension was performed using SuperScript III Reverse Transcriptase as described by the manufacturer (Invitrogen). Primer extension products were then purified using QIAquick PCR Purification Kit (Qiagen, Hilden, Germany) and run on an Avant 3100 DNA sequencer (ABI) capillary electrophoresis along with a size standard (GS600Liz; Applied Biosystems, Waltham, MA, USA) or a sequencing ladder generated using the same primers, but unlabelled, and analyzed with Peak Scanner 2 software.