Sending Out a Hypoxia SOS

CAN'T STAND LOSING O₂

PHD enzymes destabilize HIF-1α by hydroxylating prolines on the oxygen-dependent degradation domain (ODD) of HIF-1α. HIF's second oxygen-dependent domain, CAD, interacts with the p300 transcriptional coactivator during hypoxia, initiating expression of genes involved in oxygen-response pathways. To delineate the mechanism behind this interaction, Whitelaw and colleagues separated the CAD from the ODD and analyzed the CAD by mass spectrometry under normoxic and hypoxic conditions.

They found that both the ODD and the CAD were hydroxylated during normoxic conditions. "When [Kaelin and Ratcliffe's] work was published, we thought that the C-terminal region we were studying might also be hydroxylated on a proline," Whitelaw says. But mass spectrometry analysis revealed that the modification occurred on an asparagine residue.

In collaboration with Bruick, who helped uncover the identities of the prolyl hydroxylase enzymes, Whitelaw and colleagues searched a library of Fe(II)-binding hydroxylase enzymes and found that one, FIH, had previously been shown to interact with the C-terminus of HIF. They then confirmed the identity of this molecule as asparagyl hydroxylase.

While discovery of this second hypoxic switch was important in terms of elucidating the mechanism of HIF activation, most hypoxia researchers view proline hydroxylation as the fundamental step. "Asparagyl hydroxylation is fine-tuning; it's a way of allowing further transcriptional regulation. But in most settings, the prolyl hydroxylation is going to cause the protein to very rapidly degrade," says Frank Bunn of Harvard University.

Hydroxylation of asparagines may also underlie a two-tiered oxygen-dependent HIF response. "It turns out that the asparaginyl hydroxylase can work reasonably well at very low oxygen levels. The other proline hydroxylase enzymes that control HIF seem to stop functioning at less stringent hypoxia," says Whitelaw. Thus, under moderately hypoxic conditions, HIF may be stable but inactive and remain that way until the oxygen levels decrease even further.

WHEN THE OXYGEN IS RUNNING DOWN

Hypoxia researchers have known for the past two decades that HIF proteins respond to changes in cellular oxygen levels, but the question of how HIF actually senses oxygen is still debated. Current models fit roughly into two categories. The scenario a majority of researchers support, states that oxygen availability controls the activities of both the prolyl and asparagyl hydroxylases; under hypoxic conditions, less oxygen is available as a substrate for these enzymes, and the rate of hydroxylation would decrease, causing stabilization of HIF and activation of gene expression.

A few favor another model in which the prolyl and asparagyl hydroxylases are regulated upstream by a proximal oxygen sensor. Paul Schumacker, at Northwestern University, describes that upstream sensor as a "master regulator." He and others argue that the HIF-mediated hypoxic response occurs when mitochondria release reactive oxygen species (ROS), initiating a signal transduction pathway that in turn inactivates the prolyl and asparagyl hydroxylases.

© 2002 AAAS

Hypoxia blocks prolyl hydroxylation which would otherwise trigger HIF-1α recognition by pVHL and subsequent degradation by the proteasome. Similarly, without the asparagyl hydroxylation modification, HIF-1α is allowed to freely interact with transcriptional coactivator p300 spurring expression of hypoxic response genes. (Reprinted from R.K. Bruick, S.L. McKnight, Science, 295:808–9, 2002.)

The latter model is attractive, says Schumacker, because it helps to explain data suggesting that overexpression of FIH and the prolyl hydroxylases is sufficient to shut off HIF activity. "You can have a profound effect on the transcriptional activation and the stability of HIF by manipulating the expression levels of the [enzymes]," Schumacker explains. If oxygen were the direct regulator of FIH and prolyl hydroxylase activities, tight controls would need to be in place to regulate the amount of these hydroxylases in the cell.

But, says Schumacker, "By having a single proximal regulator, an oxygen sensor, that controls the activity of the hydroxylases ... ultimately HIF stability and HIF transcriptional activity could be made more independent of the precise levels of the expression of the hydroxylases themselves." Further, the model could also explain other oxygen-dependent responses that do not require HIF; Schumacker's own research has shown, for example, that mitochondrial ROS mediate hypoxic pulmonary vasoconstriction, a HIF-independent process. "The mitochondrial model could act as a master regulator of hypoxic responses, conceivably responsible for regulating both the transcriptional responses mediated by HIF, and also explain the HIF-independent hypoxic response that we see in other cells or organs," notes Schumacker.

Others are not convinced. "It's an intriguing idea, but the evidence is rather limited. Unfortunately ROS levels are difficult to measure, which can make it hard to draw accurate conclusions. What's more, I think the HIF hydroxylases fit the bill perfectly as oxygen sensors," says Patrick Maxwell of Imperial College London. Likewise, Bunn points to recent work suggesting that the oxygen tension required to trigger PHD2 is in the physiologic range: "The Ratcliffe lab had reasonable data to suggest that the prolyl hydroxylase is the proximate, immediate oxygen sensor." (Ratcliffe was unavailable for comment.) Further experiments are needed to verify this conclusion. "There's lots of flux. And the oxygen tension is very high outside of the cell; there's a gradient. The mitochondria are active consumers of oxygen, so to really get a handle on what the oxygen tension is at the very intracellular site where prolyl hydroxylase is acting, is a very tough challenge," Bunn notes.

Bruick points out that while published data support a role for the PHD enzymes as oxygen sensors within the physiological range of cellular oxygen tension under hypoxic conditions, the role of FIH is less clear. "FIH appears to remain active at lower oxygen concentrations than the prolyl hydroxylase enzymes are, and may point to nonoverlapping roles of those two enzymes across a wider range of oxygen availability," he explains. Scientists may ultimately find that HIF activity depends on a combination of events. "I think all of this sort of points to an overall model that's developing, that the HIF transcription factor is highly regulated and there's a lot of fine-tuning going on," says Bruick.

EVERY LITTLE THING HIF DOES

Given the importance of HIF in development, ischemic diseases, and cancer, it's not surprising that scientists continue to study this protein. One area of interest is the possibility that other posttranslational modifications such as acetylation and phosphorylation play a role in HIF regulation. "We've only been focusing on very defined aspects of HIF, because genetics have been able to demonstrate the importance of these modifications. But there's potential that other modifications of HIF are also important to regulate stability and activity, and that they need to be addressed in the same rigorous manner as the proline and asparagine hydroxylations," notes Stanford University researcher Amato Giaccia.

Other groups are exploring the possibility that hydroxylase-dependent regulation is not limited to HIF. "I would regard it as astonishing if biology had come up with this mechanism of regulating protein-protein interactions purely for HIF. I think it will be regulating other protein-protein interactions. So the search is really on for that," says Maxwell.

Aileen Constans (aconstans@the-scientist.com)