The description of the protein encoded in this ORF: eukaryotic translation initiation factor 4 gamma
The translational frameshift (ribosome slippage) involved: 0
The ribosome read-through involved: no
The alternative forms of this protein occur by the alternative initiation of translation: not tested
The ORF absolute position (the base range includes START and STOP codons or their equivalents): 369-4559
This is the transcript with 5'-UTR of 368bp reported by Yan et al. (1992) and reported to GenBank accession
D12686 (GI:219612) and studied later by Gan and Rhoads (1998) with IRES reported between positions 1-357 (in
other words -368 to -12 relatively to the initiator AUG). The transcript is derived from the later called
gamma promoter (nomenclature used in Byrd et al., 2005 and Baranick et al., 2008).
This is the transcript studied by Baranick et al. (2008) as well. This splice variant is incompletely covered
by RefSeq record NM_004953.3 (it lacks part of the 5'-UTR, probably because of missing EST coverage) which
gives the hint where the promoter is located in the 5'-UTR cloned in 1992 and later reported as "IRES".
Anyway, as the mRNA is derived from the most downstream gamma promoter therefore it encodes N-terminally
shortened eIF4GI while the CDS spans over the shared exons 8-33 (as defined by Byrd et al., 2005 while not by
Byrd et al. (2002)).
Our own search for ESTs supporting existence of the region -368 to -219 failed. The sequence -218 to -1 is
covered by 12 EST records in dbEST to date from human. All these ESTs have its first nucleotide within this
region. Although EST mights not be full-length cDNA copies it seems highly probably the transcription start
sites are located only in this region. In other words, the region -368 to -219 is likely a promoter. Yan et al
(1992) determined the sequence of cDNA clone from human brain lambdaDR2 (Clontech) library and confirmed the
sequence except its very first base using placental mRNA prepared by guanidium isothiocyanate-based method.
However, there are no mentions that the sample was treated by DNase and we do not know about the lambdaDR12
library either. We further support under suspicion that the promoter is located in the region upstream of -218
because a 5'-terminus of a full-length cDNA clone (AL703840 alias DKFZp686M1927_r1) originates at position
-215. Another clone (AL704011 alias DKFZp686J1628_r1) at -213. It seems the D12686 sequence published by Yan
et al. (1992) on which is based this IRESite entry contains some sequencing errors as judged by alignment
to the RefSeq variants and the genome. However, the insertions/deletions/mutations are not in the "IRES" nor
presumed promoter region of the "5'-UTR" (-368 to -219).
Notably, Byrd et al. (2005) mentioned on page 18613 that they obtained from HeLa cells this gamma
promoter-based transcript which contained completely the 5'-UTR of the clone obtained by Yan et al. (1992).
Thus, this sometimes called "intronic IRES" containing a cryptic promoter might sometimes appear in a
Baranick et al. (2008) report that a 3'-splice acceptor site is present in the putative IRES region and
that this site is being not only used in their engineered bicistronic messages transcribed from CMV IE
promoter but also occurs in natural mRNAs as well (Supplementary Figure 12). Further, in Supplementary
Figure 7 they show also 5'-splice donor site located in the eIF4G IRES region.
Let us note that the 3'-splice acceptor site is optional as well as the two 5'-splice donor sites.
Byrd et al. (2002) mapped splice variants from alpha-, beta- and gamma-promoters. They investigated which
AUG codons act as initiator codons in translation and confirmed positions 275, 395, 536, 767, and 863
(AY082886). Mostly they studied the longest transcript bearing the eIF4GI CDS which is the only one derived
from the beta promoter (currently accessible in RefSeq under NM_182917.3). Nowadays the RefSeq record
NM_182917.3 is slightly longer than the original AY082886 which they termed SP4GI by 123 bp on the very
5'-end. In contrast, SP4GI is slightly longer than the clone called pAD-4GI (R. Rhoads), see Figure 3A in Byrd
et al. (2002). The corresponding IRESite entry of a beta promoter-derived mRNA is the longest of the three
just mentioned and accessible under IRESite_Id:576.
Byrd et al. (2005) studied mainly eIF4GI splice variants 5'-UTRs derived from alpha, beta promoter but
isolated cytoplasmic RNA without DNase treatment (probably not an issue as the amplicons spanned several exon
junctions confirming the cDNA was not contaminated). The first strand cDNA synthesis was carried using random
primers instead of oligo(dT) primers. eIF4GI specific primers were used for PCR amplification. However, they
focused on studies of 5'-UTRs of alpha-and beta- derived splice variants. They confirmed promoter activities
in both beta- and gamma-promoter regions immediately preceding the very first exon of the resp. splice variant
(Figure 1A, white thick bars).
Han and Zhang (2002) reported a promoter in the poly-pyrimidine tract (PPT) overlapping with the originally
reported "IRES" region using promoter-less based assay. The PPT region was found to bind C/EBPbeta
transcription factor. See Figures 3-5 in their article.
The IRES absolute position (the range includes START and STOP codons or their equivalents): 1-357
How IRES boundaries were determined: experimentally_determined
The sequence of IRES region aligned to its secondary structure (if available):
This region has been investigated recently by Baranick et al., 2008 and they postulate that most of the "IRES"
activity observed so far was due to the splice artifacts (low IRES activity in vivo when 5'-splice donor
depleted plasmid was used and low yields of reporter provirus sensitive to altered levels of splice products).
Please refer to Figures 1, 2, 3D in the article and Supplementary Figure 7.
Han and Zhang (2002) reported no IRES activity in direct RNA transfection.