The long-awaited demonstration of protein pyrophosphorylation by IP7 in vivo?

Being wanted, he may be more wonder’d at… .

William Shakespeare, Henry IV, Part 1

The “high-energy” of inositol pyrophosphates such as IP7 has long been considered as their raison d’être (see ref. 1). Unsurprisingly, therefore, the excitement was palpable when these pyrophosphates were shown to transphosphorylate proteins in vitro (1). This observation was hailed as “a new mechanism in second messenger biology” (2). However, there is also a more circumspect viewpoint which requests evidence that this transphosphorylation actually occurs in vivo (2).

Thus, the demonstration by Azevedo et al. (3) that IP7 hinders viral exit from cells by perturbing the interaction between a molecular motor, KIF3A, and AP3B1, the β-subunit of the AP3 adaptor complex, may initially attract less attention than the authors’ interpretation of their back-phosphorylation data. This analysis leads the authors to what even they believe is their “more important” conclusion: that IP7 pyrophosphorylates AP3B1 in vivo. But there is an alternative explanation for these data.

To understand my reservation, both the nature of the back-(pyro)phosphorylation assay and the transphosphorylation event need to be appreciated fully. Azevedo et al. (3) note that an already pyrophosphorylated protein would be poorly pyrophosphorylated by [³²P]IP7 in vitro. So, AP3B1 was exogenously expressed in a strain of S. cerevisiae (vip1Δ) with elevated IP7 levels. When this AP3B1 was extracted and incubated with [³²P]IP7, the adaptor was not pyrophosphorylated. Thus, the authors argue, AP3B1 must already have been pyrophosphorylated by IP7 in vivo. Conversely, there was considerable back-pyrophosphorylation of AP3B1 by [³²P]IP7 in vitro after it was isolated from kcs1Δ yeast, in which inositol pyrophosphates are virtually eliminated.

However, in vitro, inositol pyrophosphates can phosphorylate only an appropriate Ser residue that is first primed by phosphorylation by casein-kinase II (CK2) (4). Then, inositol pyrophosphates convert the phospho-Ser to diphospho-Ser. Thus, I propose that the back-phosphorylation assay actually could be recording the CK2-mediated monophosphorylation of the appropriate Ser in AP3B1 in vivo. In this alternate scenario, this Ser is monophosphorylated by CK2 in kcs1Δ yeast, rendering AP3B1 susceptible to back-pyrophosphorylation by IP7 in vitro. Similarly, it can be argued that high levels of IP7 in vivo prevent AP3B1 from being back-pyrophosphorylated by IP7 in vitro because the appropriate Ser is not first monophosphorylated. This different interpretation of how changes in IP7 levels regulate AP3B1 phosphorylation and function in vivo can take some support from a scattered—albeit not entirely consistent—literature that demonstrates that CK2 activity is modulated by inositol polyphosphates (see ref. 5).

Neither of these competing interpretations of the back-phosphorylation data is perfect. My proposal does not explain the “downward gel shift” AP3B1 experiences in kcs1Δ cells (3). Equally, although Azevedo et al. (3) believe this gel shift reflects lack of AP3B1 pyrophosphorylation, no shift is observed following its pyrophosphorylation in vitro (3). Thus, the authors and I agree that there are additional, more complex posttranslational modifications of AP3B1 to be resolved. Nevertheless, the fact that there is an alternate explanation for the back-phosphorylation data leaves us still waiting for a clear demonstration that inositol pyrophosphates can transphosphorylate proteins in vivo.