Homo sapiens p53 protein mRNA, complete cds. Poly(A)-tail sequence was dropped.

The ORFs of the two possible p47 variants are 117/129bp downstream of the AUG codon of p53, so 39/43 aminoacid
residues shorter proteins are encoded. Existence of the p47 isoform was shown using different antibodies by
Yin et al. (2002).




We were in contact with Marco Candeias. In one of his answers we learned the following:

This is what we obtain using a forward primer inside CMV promoter for sequencing construct p53wt in the
oncogene paper:
5'-TAAAGACCCACTGCTTACTGGCTTATCGAAATTAATAC
GACTCACTATAGGGAGACCCAAGCTTACTGCCatgGAGGAGCCGCAGTCAGATCCTAGCG
TCGAGCCCCCTCTGAGTCAGGAAACATTTTCAGACCTATGGAAACTACTTCCTGAAAACA
ACGTTCTGTCCCCCTTGCCGTCCCAAGCAATGGATGATTTGATGCTGTCCCCGGACGATA
TTGAACAATGGTTCACTGAAGACCCAGGTCCAGATGAAGCTCCCAGAATGCCAGAGGCT
GCTCCCCGCGTGGCCCCTGCACCAGCAGCTCCTACACCGGCGGCCCCTGCACCAGCCCCC
TCCTGGCCCCTGTCATCTTCTGTCCCTTCCCAGAAAACCTACCAGGGCAGCTACGGTTTC
CGTCTGGGCTTCTTGCATTCTGGGACAGCCAAGTCTGTGACTTGCACGTACTCCCCTGCC
CTCAACAAGATGTTTTGCCA-3'

This is what we obtain using the reverse primer
5'-cactgaagacccaggtccag-3' inside p53 for sequencing construct p53 "C"
in the oncogene paper:
5'-AGAGGCAGGGGCAGTTCCGTCTTCCAGAGGACCACCCCCTTGGGGGCGGCCCGTGTTGTG
CCGGAAACCATACTTAGCACCATTCCGCCGGAGCAAAGACCCAAAGAGAAGCGCGATCAC
ATGGTCTTGCTGAGTTCGTGACCGCGCCGGGATCACTCTCGGCATGGACGAGCTGTACAA
GTGGAAGCTTGTACCAATGACGCGCGCGCGCGTGAGCGCGCGCGCGAAAGGATCCACTAG
TAACGGCCGCCAGTGTGCTGGAATTCACTGCCatgGAGGAGCCGCAGTCAGATCCTAGCG
TCGAGCCCCCTCTGAGTCAGGAAACATTTTCAGACCTATGGAAACTACTTCCTGAAAACA
ACGTTCTGTCCCCCTTGCCGTCCCAAGCAATGGATGATTTGATGCT-3'

The IRES of p53 was cloned into the pEGFP-N1 vector at restriction sites EcoRI(gaattc), BamH1(ggatcc). Like
this: gaattc actgccATGGAGGAG...CCCAAGCAATGGA ggatcc accggtcgccaccATGGTG
so a 130 nt sequence of p53 was inserted





The sequence obtained through the reverse primers shown above aligns best in the GFP and stem-loop region to
entry EU716638 and the alignment indicates the number of sequencing errors in this single-read:

>gb|EU716638.1|  Synthetic construct N-EGFP/centrosomal protein 97kDa fusion protein
gene, complete cds
Length=3336

 Score =  165 bits (182),  Expect = 1e-37
 Identities = 174/212 (82%), Gaps = 13/212 (6%)
 Strand=Plus/Plus

Query  21   CTTCCAGAGGACCACCCCCTTGGGGG-CGGCCC-GTGTTG-TGCCGGA-AACCA-TACTT  75
            || ||||  || ||||||| | || | |||||| ||| || |||| || ||||| ||| |
Sbjct  552  CTACCAGCAGAACACCCCCATCGGCGACGGCCCCGTGCTGCTGCCCGACAACCACTACCT  611

Query  76   -AGCACC-ATTCCGCCG-GAGCAAAGACCCAAA-GAGAAGCGCGATCACATGGTCTTGCT  131
             |||||| | ||||||  |||||||||||| || ||||||||||||||||||||| ||||
Sbjct  612  GAGCACCCAGTCCGCCCTGAGCAAAGACCCCAACGAGAAGCGCGATCACATGGTCCTGCT  671

Query  132  G-AGTTCGTGACCGC-GCCGGGATCACTCTCGGCATGGACGAGCTGTACAAGTGGAAGC-  188
            | ||||||||||||| |||||||||||||||||||||||||||||||||||| | ||||
Sbjct  672  GGAGTTCGTGACCGCCGCCGGGATCACTCTCGGCATGGACGAGCTGTACAAG-GAAAGCG  730

Query  189  TTGTACCAATGACGCGCGCGCGCGTGAGCGCG  220
            ||  | | ||| ||   |||||||||  ||||
Sbjct  731  TTCGAACCATGGCGGTGGCGCGCGTGGACGCG  762


The putative IRES (a 120bp fragment between Met1 and Met40 of p53) was cloned by Candeias et al., (2006) into
two monocistronic constructs 'A' and 'B' after a presumed stem loop while the latter construct had mutated
Met1 of p53. Construct 'C' contained GFP ORF followed similarly by the presumed stem loop and by the 120bp
long region between Met1 and Met40. Construct 'D' contained in addition to construct 'C' also the 5'-UTR of
p53 mRNA. All constructs utilized CMV promoter.

The mono- and bicistronic constructs 'A-C' were shown in Figure 3a. In monocistronic experiments a presumed
hairpin was placed upstream of the p53/47 AUG codon or in addition the AUG codon of the full-length p53 the
was mutated. Placement of just the hairpin increased amounts of p53/47 isoform 6.1x while mutation of the
AUG codon resulted in 11.8x higher yields of the p53/47 protein. This could be interpreted with the ribosome
scanning theory which scanned for the AUG codon. While the one of p53 was mutated ribosomes initiated
synthesis at the AUG codon of p53/47 isoform. One would not expect that mutation of an upstream AUG codon or
even of a AUG codon probably encompassed in the putative IRES to affect IRES-mediated initiation from a
downstream codon.

The only bicistronic construct "C" shown in Figure 3a merely shows that with the introduction of an upstream
ORF encoding the GFP protein synthesis of both p53 and p53/47 dropped seriously. Again, this looks like a lot
less ribosome leaked through the GFP coding region and therefore much less of p53 isoforms were obtained. The
yields of full-length p53 dropped in comparison to the monocistronic construct down to 0.4 while p53/47
amounts increased 2.6 times. Notably, yield of the shorter p53/47 isoform is higher and one might be tempting
to interpret this as IRES-mediated increase. We believe this difference is due to the fact that the p53 AUG
codon is rather "soon" behind the presumed stem-loop and therefore the ribosome does not re-initiate the
synthesis so effectively after it has finished GFP synthesis. Although efficiency of re-initiation decreases
with the length of the intercistronic region the ribosome probably needs some "spacer" to gain its speed after
reading-through stronger secondary structure. Therefore in this scenario after bypassing the "extra" 3*40bp
between the two AUG codon it can at higher efficiency initiate synthesis of p53/47 isoform. This explanation
could also be applied to the construct 'D' shown in Figure 4a where the 5'-UTR 135bp long was inserted in
between the hairpin and the p53 AUG codon.


The scenario does not change much with introduction of a STOP codon causing the full-length p53 isoform to be
degraded via non-sense mediated decay (Figure 3b) and to re-initiate 3bp downstream at the p53/47 AUG codon.

Possible presence of a cryptic promoter was ruled out using promoter-less plasmid (Figure 3c). Maybe to some
extent by Northern-blot of the 'C' construct but it is not clear to which part of the p53 CDS the 171bp long
probe hybridized (Figure 3c).

In vitro translation using rabbit reticulocyte lysates with increasing presence of 7m-GTP showed that p53
synthesis is more sensitive to presence of the 7m-GTP than p53/47 isoform (Figure 3d). Probably more
interesting results could have been obtained with 7m-GpppG cap analogue. Unfortunately it is not clear from
the text how authors separated the two isoforms (note the CPM units in the graph). Similar note applies to
Figure 3e with the addition that it is not clear why tweaking the context of AUG codon of full-length p53
could result in altered p53/47 yields from a downstream AUG if mediated by IRES. Notably, decreasing yields of
the p53 isoform were followed by slight increase of p53/47 yields, again resembling that ribosomal leaky
scanning model could be applied well.




Yang et al. (2006) also reported IRES in the p53 transcript. They cloned region -131 to -1 (obtained from
from pRST/hUTR-luc-plus vector) into pRF plasmid from A. Willis (IRESiteID:184) using SpeI/NcoI sites (called
pR5UTRF). Maybe they used PCR step to introduce the sites into the DNA insert (no primers sequences mentioned
in the article).  Similarly, it is not clear how they deleted the 70bp region from the 5'-end (pRDNF plasmid).
In Figure 3a they mention RT-PCT without revealing primer sequences. Regarding the Figure 3b and 3c, similarly
one cannot figure out the sequence of pGL3/5UTR and pGL3E/5UTR plasmids without more thorough description.
The attempts to prove that there was no cryptic promoter or aberrant splicing cannot be judged without the
detailed information.