Alternative polyadenylation and cancer

This is a follow-up of sorts to a previous essay on the subject of alternative polyadenylation.  In the previous report, I discussed some bioinformatics studies that suggested that the 3′ UTRs of mRNAs change, in bulk, in the course of development in mammals.  The implication of these results is that poly(A) site choice in mammals is regulated, with important functional consequences.

A more recent study by Mayr and Bartel adds to this notion.  These authors studied 3′ UTR length in normal and cancer cells, and found a striking correlation between 3′ UTR length and the expression of oncogenes.  Specifically, higher expression (as is found in cancer cells) is correlated with shorter 3′ UTR.  As 3′ UTR length is determined by the position of the poly(A) site along a transcript, this implicates alternative polyadenylation as one mechanism by which oncogene expression is activated.

Putting the polyadenylation complex together

One of the reasons for the slow pace of entries here has been the intrusion of, um, life into my life.  (Yeah, figure that one out.)  Earlier this summer, a paper from my lab was published in BMC Cell Biology.  Seeing as it’s Open Access, and since it has a bit of relevance to a theme introduced in this essay, I thought I would point it out here.

The story in a nutshell – one of the subunits of the polyadenylation complex is the so-called CPSF30 protein, or its yeast relative Yth1.  (Yth1 looms large as one of the few subunits in the Giardia complex.)  What Drs. Suryadevara Rao and Randy Dinkins did was study the places within the cell where CPSF30 goes, and what happens when one co-expresses this protein with other polyadenylation complex subunits.  They did this by attaching the various proteins to fluorescent proteins and following the fusion proteins using microscopic techniques.

The results corroborated other studies that detailed interactions between various of these proteins.  However, a rudimentary deletion analysis showed that these interactions by and large involve parts of CPSF30 that are not found in the mammalian or yeast proteins.  Since the CPSF30 interacts with the other proteins of interest in this study (the 160, 100, and 73 kD subunits of the cleavage and polyadenylation specificity factor, or CPSF) in other eukaryotes, it stands to reason that the interactions themselves must have evolved independently.  This in turn suggests a somewhat different trajectory in the evolution of the complex in different eukaryotic lineages.  It also raises the possibility that the different complexes may process and polyadenylate RNAs in subtly different ways.

Some pretty pictures and a link to a fascinating movie may be found beneath the fold.  Enjoy.

Alternative polyadenylation in development

One of the things that is an open book is the true scope and physiological relevance of alternative polyadenylation.  A recent report in PNAS stirs this pot a bit (even if it leaves things still very much up in the air).  Briefly, this group has analyzed various large-scale gene expression repositories – ESTs, SAGE, and microarray – and found a tantalizing possible progression of 3′-UTR length during development.  Specifically, it seems as if global (or average) 3′-UTR length increases during the course of embryogenesis.  This change in the length of 3′-UTRs seems to be due to differential poly(A) site choice.  As I said, very tantalizing.

The abstract and brief commentary follows after the fold.

Where polyadenylation, siRNAs, and DNA methylation meet

It has become more apparent in recent years that the different aspects of gene expression – transcription initiation, transcription elongation, mRNA capping, splicing, and polyadenylation, transport of the mRNA to the cytoplasm, translation, and mRNA quality control – are rather extensively interconnected.  One corollary is that the polyadenylation complex, through various of its subunits, plays roles in various of these other processes.  This has been established for the most parts in mammalian and yeast models, but some recent work in plants is adding new and important variation to this theme.

A most recent of such studies has appeared online on PNAS.  This study, from the lab of Caroline Dean, reveals that the polyadenylation factor subunit FY (a homolog of the yeast protein Pfs2), acting in concert with the flowering regulator FCA, plays a crucial role in chromatin modifications that regulate the expression of the FLC gene.  Interestingly, this effect is not limited to just the FLC gene. Rather, other genes that are silenced by small RNA-mediated DNA methylation also require FY for this silencing.  This provocative finding seems to place FY in some sort of proximity to the small RNA-guided DNA methylation machinery, and may have some relevance to many aspects of transcription and mRNA quality control.

The abstract and citation follows. As always, enjoy.

Convergent transcription, polyadenylation, RNA editing, and regulation

One of the mechanisms by which polyadenylation may contribute to the regulation of gene expression (on paper, at least) involves gene pairs that are situated near each other and transcribed convergently.  In these instances, polyadenylation and transcription termination need to occur to prevent the production of RNAs that are anti-sense to the two members of the convergently-transcribed gene pair.  Overlapping transcripts could lead to the formation of double-stranded RNAs that could in turn trigger regulatory mechanisms, resulting in altered accumulation of the corresponding transcripts.

It is in this vein that a recent study from Gordon Carmichael’s lab at the University of Connecticut is of interest.  Briefly, these authors report that the early-to-late transition in gene expression in cells infected with the mouse polyoma virus is accomplished (at least in part) by a reduction in polyadenylation efficiency of the primary transcript encoding the so-called late genes.  Interestingly enough, this reduction in polyadenylation efficiency seems to be due to A-to-I editing of the region around the polyadenylation signal.  This editing in turn may be traced to an overlap of the early and late transcripts, such that double-stranded RNAs (the substrate for the A-to-I editing complex) that include the late polyadenylation signal are produced and edited before pre-mRNA processing occurs.

What we’re talking about

A recent paper from James Manley’s lab details a proteomic analysis of the polyadenylation complex.

In this study, the polyadenylation complex was purified using an affinity technique, attaching a functional polyadenylation signal to a series of MS2 coat protein binding sites (of course, all within the context of an RNA), incubating this RNA with Hela cell extracts, and purifying the RNA using a maltose binding protein-MS2 coat protein fusion. The protein components of the complex were characterized using the so-called Multidimensional Protein Identification Technology (MudPIT). In addition to the usual players* (CPSF, CstF, CFIm) were found proteins suggestive of links with transcription, splicing, and DNA repair. Curiously, the sole poly(A) polymerase found was not the canonical enzyme but an isoform first identified as an enzyme that adenylates the RNA present in the so-called Signal Recognition Particle (SRP). Moreover, readily detectable was a testis-specific CstF64 variant (CstF64-tau) as well as the canonical 64 kD subunit of CstF. CFIIm was largely absent.

Interesting stuff

While my yard is recovering from the ice, and I from today’s UK game, I thought I would toss out a few interesting abstracts that touch on important and contentious issues.  Peek beneath the fold and, as always, enjoy.

An interesting take on poly(A) signals

One interesting facet of RNA biology is the matter of the occurrence and function of structural RNA units within RNAs inside of living cells.  Excellent examples of these are the many so-called riboswitches, motifs that bind metabolites and alter the functionalities of RNAs in which they reside.  As more and more examples of RNAs with catalytic activities become known, questions naturally arise as to whether such activities might impact RNA function in vivo.  A recent study poses just such a question for what is among the simplest of known catalytic RNAs, namely the manganese-dependent ribozyme.  This enzyme consists of little more that a GAAA-UUU complex.  These two motifs, that need not be adjacent in the RNA, are expected to occur at a high frequency in natural RNAs.  A recent report indeed finds that RNAs that possess such motifs indeed may be Mn-dependent ribozymes.  The physiological signifiance of this finding is, in my opinion, a bit of an open issue.  But the possibilities are fascinating.  The last sentence of the abstract (that follows) is especially provocative.

Structure? Who needs structure?

One of the more interesting aspects of the polyadenylation complex is the nature of the protein-protein interactions that help shape the machinery. A recent study has revealed a fascinating side of this story. It involves the interaction between a subunit known as Fip1 (Fip = Factor Interacting with Poly(A) polymerase) and poly(A) polymerase (affectionately abbreviated as PAP). This study is remarkable for two reasons. For one, it dispels some previously-held notions about the functional significance of this interaction. In addition, it reveals that Fip1 is a member of a large class of proteins that share a common feature – namely, they do not have a rigidly-specified 3-dimensional structure in and of themselves.

If I could only remember ….

A recent essay described the process of cytoplasmic polyadenylation, and the functioning of the process in animal oocyte maturation and early development.  Cytoplasmic control of translation is important, not only in the oocyte, but in other settings as well.  One of these is in the neuron, where gene expression in repsonse to synaptic stimulation can be controlled by the activation of stored mRNAs in the cytoplasm.  While details of this control have remained poorly understood, a suggestive requirement (well, it’s suggestive when it comes to the general subject of this blog) of CPEB for memory in Drosophila lends itself to the hypothesis that the activation of stored mRNAs in the neuron involves cytoplamsic polyadenylation.

It is thus of interest to see that Drosophila homologue of Gld2 (the cytoplasmic poly(A) polymerase mentioned here) is required for long-term memory.  The abstract of the recent report follows.  As always, enjoy.