What we’ve learned during the blogging hiatus

It’s been seven years or so since I updated this blog. In this post, I hope to summarize a direction my lab has taken in this time. If I am successful, you will see how interesting and exciting this research is, and the new directions we are exploring.

For the better part of 12 years, a main focus of research in my lab has been alternative polyadenylation in plants. I summarized a couple of seminal papers previously – here and here. One of the take-home points of this research was the scope of alternative polyadenylation – how many different poly(A) sites could be used, how usage shifts, and the impacts that shifts in poly(A) site choice have on gene expression. Since 2013, we have published many additional papers on alternative polyadenylation. The three summarized here help to develop a theme that guides current research in my lab.

Usually, the context for research on alternative polyadenylation has to do with alterations in the lengths of 3’ UTRs for a given set of transcripts. For good reason, as the lengthening and shortening of 3’ UTRs is an important mechanism by which alternative polyadenylation impacts gene expression. In plants, though, it does not seem that the regulation of 3’ UTR length is a general regulatory mechanism. (To be sure, it may apply in some specific instances. And it is true that not everyone who works in this area in plants would agree with me. But it is my opinion that the generality of this mechanism has not been shown in plants, and that it is probably not a significant player when it comes to gene regulation.)

Recall that the primary transcript may be processed and polyadenylated at many different positions – in 3’ UTRs, but also in introns, protein-coding regions, and even 5’-UTRs. (I have come to call these latter mRNA classes collectively as non-canonical.) Non-canonical mRNAs are curious, in that they would seem to have no useful protein-coding function. In addition, from what we know about RNA metabolism, one could readily hypothesize that these mRNAs might have different properties such as stability or translatability. In a collaboration with Dr. Julia Bailey-Serres’ lab, this is what Laura de Lorenzo, a postdoc in my lab, found. Laura showed that some classes of non-canonical mRNAs are decidedly less stable than canonical mRNAs encoded by the same gene. She also found that different classes are also translated with different efficiencies. Together with Reed Sorensen (in Julia’s lab), they further showed that hypoxia incited substantial shifts in the relative usages of canonical and non-canonical sites. The end result seemed to be that hypoxia induced a shift to usage of sites that are expected to produce non-functional mRNAs that have lower stabilities and altered translatabilities.

More recently, Manohar Chakrabarti (another postdoc in my lab) along with Laura published a study of stress-associated shifts in poly(A) site usage in sorghum. This was a collaboration with Dr. Ani Reddy’s lab at Colorado State University. In a nutshell, what Manohar and Laura found was that any of three different stresses – heat, drought, and salt – induced increased production of mRNAs derived from use of non-canonical poly(A) sites. In terms of individual genes, the scope of this shift usage was broad, and there was a sizable overlap in the sets of sites and genes impacted by each stress. It wasn’t really possible to tease out a coherent set of stress-specific alternative poly(A) sites. This was somewhat disappointing, since one goal in this project was (and is) to connect alternative polyadenylation with stress responses. It would have been nice to have a heat-specific motif that guides alternative polyadenylation during heat stress. Etc. and so on …

But the quest for connections between stress and polyadenylation hasn’t been entirely fruitless. My lab has a collaboration with Dr. Juan Carlos del Pozo in the Centro de Biotecnología y Genómica de Plantas (CBGP) in Madrid. Carlos is interested in many aspects of plant development and isolated a mutant with a deficit in lateral root development. (The first paper describing this mutant can be found by following this link.) When he mapped the mutation, be found that it affected the gene that encodes one of the Arabidopsis orthologs of Fip1, a core polyadenylation complex subunits. (I will discuss this mutant in another essay later this spring.) Laura conducted a gene expression/poly(A) site profiling study and found, not surprisingly, that poly(A) site usage differed dramatically in the mutant.

In addition to having a lateral root defect, the fip1-2 mutant turned out to be salt tolerant. This was encouraging, and places FIP1 somewhere in the salt stress response pathway in Arabidopsis. What was more remarkable, though, was that the fip1-2 mutant also had a fascinating polyadenylation phenotype. In wild-type Arabidopsis, salt induces a shift to increased production of mRNAs derived from non-canonical polyadenylation, much as it does in sorghum. In the mutant, though, no such shift was seen.

The ramifications of this last paper are many and very interesting. First, it suggests that the stress-induced changes in the levels of mRNAs derived from non-canonical polyadenylation are due to actual changes in the usage of these sites (as opposed to downstream differential effects of the stresses on the stabilities of these RNAs). Of course, it may be that FIP1 plays roles in the turnover of non-canonical mRNAs, but the simpler explanation is that it mediates stress-induced increases in usage of non-canonical poly(A) sites.

A second ramification is even more interesting. Usually, in the literature about stress-responsive transcription or gene expression, genes whose expression increase are associated with enhanced ability to counter the stress. The same tendency holds for stress-responsive alternative polyadenylation. In the case of non-canonical mRNAs, the idea is that their increased production is part of a global response that helps the plant respond to the stress. Since non-canonical mRNAs are typically non-functional, perhaps their production helps to re-model overall gene expression by shunting the transcriptional output for large numbers of genes into non- productive pathways, thereby promoting a more rapid remodeling of the proteome.

The result with the fip1-2 mutant suggests an interesting alternative. If the global poly(A) site remodeling was needed for improved survival upon exposure to stress, then one would expect a mutant that a unable to increase non-canonical poly(A) site usage to be more sensitive to the stress. But this is not what is seen. What the fip1-2 mutant seems to be telling us is that the shift in poly(A) site choice in wild-type plants may be responsible for some of the salt-associated phenotypes we seen in plants. Maybe one reason plants seem to be harmed by various stresses is because they deliberately shut down. This might be viewed as analogous to the hypersensitive response that is seen in many plants after challenge with pathogens.

This is an interesting idea and one that we will be exploring in the future. Regardless of how this research turns out, it seems as if FIP1 is probably a protein that connects stress signaling with poly(A) site choice. Figuring out the specifics will also occupy a great deal of time in my lab in the future.