Tyr-94 Phosphorylation, PDP1, and the Warburg Effect

Figure 1. Crystal structure of Tyr-94 phosphorylated PDP1. This form of PDP1 is shown to have inhibited dephosphorylation activity, but the structure reveals that the phosphorylated Tyr-94 site is more likely to affect the lipoic acid binding site rather than the catalytic cleft, reflecting a more indirect course of action. (Shan et al., 2014)
Figure 1. Crystal structure of Tyr-94 phosphorylated PDP1. This form of PDP1 is shown to have inhibited dephosphorylation activity, but the structure reveals that the phosphorylated Tyr-94 site is more likely to affect the lipoic acid binding site rather than the catalytic cleft, reflecting a more indirect course of action. (Shan et al., 2014)

Tyr-94 Phosphorylation, PDP, and the Warburg Effect

Besher Tolaymat

ChemBio Spotlight Week 6

Paper: Tyr-94 Phosphorylation Inhibits Pyruvate Dehydrogenase Phosphatase 1 and Promotes Tumor Growth

Proper bodily functioning requires energy, and cellular energy is almost always in the form of ATP. Glycolysis is followed by either oxidative phosphorylation (OXPHOS) or lactic acid fermentation to form ATP, with the latter process being less efficient than the prior, but with the ability to be performed in hypoxic conditions. The Warburg Effect refers to the observation that cancer cells undergo rapid glycolysis followed by lactic acid fermentation rather than the more efficient oxidative phosphorylation pathway – regardless of the availability of oxygen. OXPHOS starts with acetyl-CoA production: pyruvate dehydrogenase (PDH) converts pyruvate into acetyl-CoA, pyruvate dehydrogenase kinase (PDK) and phosphatase (PDP) phosphorylate and dephosphorylate PDH, respectively, and all three enzymes are needed in a balance of sorts. PDH is inactivated via phosphorylation, so up-regulated PDK or inactivated PDP may be possible causes of the Warburg Effect by only allowing lactic acid fermentation to take place. In prior studies by the authors, certain kinases and acetyltransferases were correlated with cancer cell growth.1 In this study, the authors studied the effects of phosphorylation of PDP1 at Tyr-94 and the mechanism of action.

The authors performed several experiments to get at different levels of findings – from big-picture effects down to mechanistic details. Initially, PDP1 was found to be inhibited via Tyr-94 by FGFR1 (Fibroblast growth factor receptor 1), but not at other Tyr residues. Next, structural analysis showed that Tyr-94 is ~20Å from the catalytic cleft, which is likely too far to directly affect it, but it is closer to the location of lipoic acid binding to PDP1 (Figure 1). PDP1 activity is mediated by lipoic acid binding to E2 (an acetyltransferase). Thus, it was hypothesized that Tyr-94 phosphorylation inhibits PDP binding to lipoic acid and, consequently, to E2. To test this hypothesis, Shan et al. found that incubation of recombinant rPDP1 with rFGR1 resulted in decreased 3H-labeled lipoic acid binding to rPDP1 WT but not Y94F mutant. To test for the prevalence of this Tyr-94 phopshorylation among several cancer cell types, a specific phosphorylated-PDP1 antibody was used to find it was prevalent among several leukemia, breast, and lung cancer cells. Lastly and with great impact, it was shown that expression of a PDP1 Y94F mutant in cancerous cells increased oxidative phosphorylation under normoxic conditions, and decreased ATP production under hypoxic conditions, which was correlated to the idea that Y94F mutant relies more on OXPHOS as compared to WT cells.

The Warburg Effect has remained a mysterious phenomenon for many years, and only recently has it been shown that cancer cells undergo fermentation despite normoxic conditions. The mechanism by which cancerous cells switch to this metabolic fate has been mostly unknown. This paper fills in one of the gaps of this mystery with the finding that Tyr-94 phosphorylation of PDP1 inhibits its ability to de-phosphorylate and re-activate PDH, which may be integral to the shift towards lactic acid fermentation and the nature of the Warburg Effect in cancers. With a mechanism developed, more studies may explore PDP1 activators as possible anticancer agents.


  1. Kroemer, G. and Pouyssegur, J. (2008) Tumor cell metabolism: cancer’s Achilles heel. Cancer Cell 13, 472-482

16 Replies to “Tyr-94 Phosphorylation, PDP1, and the Warburg Effect”

  1. Hey Besher,

    I enjoyed the relevance of this article to our most recent class discussion of the Warburg effect and for sifting through some of its perplexities to increase our understanding for this Friday’s test.

    My question in regard to this article is about the phosphorylation of Y94. What enzyme is causing this phosphorylation? Is the enzyme causing this phosphorylation genetically linked.. that is, are somatic mutations the cause of misregulation? If I missed this information, please point me to it. If not, would you consider these questions viable avenues for future work?

    In regard to the suggested treatment of ‘PDP1 activators as novel anticancer agents,’ I feel as though the authors offer controversial evidence that offsets the above quote. Even if PDP1 could be activated by dephosphorylation of Y94, there is still the issue of lysine acetylation and misregulation at Y381. Furthermore, these two residues are stated as ‘distinct and independent molecular mechanism[s],’ meaning that dephosphorylation at Y94 does not effect any activity at Y381. To me, this says that there is not a single treatment to the inhibition of PDP1 and that while activation of Y94 may be effective, PDP1 may still be inhibitied and PDK could be upregulated and inhibiting PDH at serine residues. How do you feel about these ideas? Do you think that an activator of PDP1 could act on both tyrosine residues by virtue of their chemicals identities despite their mechanistic differences?

    1. All good questions Ryan. They don’t seem to give the enzyme a specific name, so I guess we’re going to call it Tyrosine 94 kinase? I’m not entirely sure if this enzyme would truly only phosphorylate PDP1 at Y94 or if it could phosphorylate other tyrosine kinases in similar environments, but this might be a question they haven’t answered yet. They don’t say anything specifically about it being a somatic mutation so I’m not sure of the answer to that. We know that family histories of cancer are something physicians keep in mind, but I wonder how genetically linked these kinds of mutations are. You also bring up some good points in your next questions. I think it’s important to keep in mind that when they attempted to “treat” the cells, the tumors shrunk in size but were not completely resolved, which aligns with what you’re mentioning. They know there is so much going on in these cells, and so far they’ve found a few acetylations and phosphorylations that are important, and furthermore targeting these mutations does affect the tumor, but they’re not at the level of treatment quite yet. In addition, I think there needs to be more done in order to determine if all of these mutations they’re finding are linked and more importantly if there’s causation – is it a cascade of one mutation causing another? In the bigger picture, this begs the question: is cancer itself a giant cascade, and what is that initial ripple?

  2. Nice job, Besher! I very much appreciate how closely this paper relates to our recent class discussion and offers insight into the mechanisms underlying the Warburg effect. I have a few thoughts and questions surrounding this topic. We know that cancer is a disease of mass destruction and the molecular mechanisms of a cell’s process of “switching identities” from normal to cancerous are complex. I wonder whether preferentially performing glycolysis provides an advantage for the cancer cell’s survival. Perhaps the cancer cells are not investing as much energy such that they perform simple reduction chemistry in lactate production (instead of the complex chemistry that is OXPHOS), which gives them some type of survival advantage? I also think that it is very interesting that there are so many different phosphorylation events that produce the Warburg effect in concert. Do you know if ALL of them are required to produce the full Warburg effect or if some are more important than others? Thanks! ☺

    1. I remember during our discussion about an article in class there was a slide that proposed two hypotheses as to why cancer cells might prefer glycolysis & reduction rather than OXPHOS – (1) they have plenty of access to ATP and don’t need to make it, and/or (2) cancer cells aren’t interested in the energy, but may have other reasons to avoid OXPHOS. We discussed the second hypothesis in terms of retaining biomass – that cancer cells can have more biomass if they avoid converting the carbon atoms of carbohydrates into CO2 and instead keep them as lactate. I think both hypotheses could have some truth to them. A responsible cell would be one that takes global metabolism into account and should either perform glycolysis or OXPHOS depending on the body’s needs – but a cancer cell (if we are going to anthropomorphize) only cares about itself, so ATP production might not be a goal anymore. Likewise, if the cancer cell wants to proliferate, it needs biomass, which may validate the second hypothesis. Furthermore, I wonder if the mitochondria is a key player here – OXPHOS needs the mitochondria, but glycolysis & fermentation can take place anywhere. I believe we learned in bio 3 that p53 (“the genome’s guardian”) is associated with the mitochondria, and we know that the mitochondrial potential is affected significantly in cancerous cells – maybe this change is mainly there to deter p53 and other tumor suppressors, and as a side-effect, conversion to pyruvate and OXPHOS are less favorable? Here’s one paper I found after a quick search:
      it’s titled “SIRT3 and SIRT4 are mitochondrial tumor suppressor proteins that connect mitochondrial metabolism and carcinogenesis”, personally I’d love to see where this paper takes cancer studies

  3. Hi Besher! Great job reviewing an interesting article! Forgive me if I missed the obvious here, but can you clarify some things about oncogenic tyrosine kinase for me? So, the authors keep referring to the existence of oncogenic tyrosine kinases in cancer cells and their role in the phosphorylation of the Tyr-94 residue they are interested in, but what really makes these tyrosine kinase residues oncogenic? It seems that tyrosine kinases could serve many key roles in healthy cells and cancer cells. Would cancer cells have a mix of normal and oncogenic tyrosine kinases or would they all be oncogenic? If they were all oncogenic, couldn’t this have larger implications and a wider impact than just the phosphorylation of PDP1 at Tyr-94? Thanks!

    1. Phosphorylation is one of the most common ways of turning things on/off, and tyrosine kinases, I believe, should really not be seen as different from any other kinase – they’re phosphorylating a substrate and turning on/off that substrate’s activity. Tyrosine kinases can have many normal functions and are actually involved in some signal transduction sequences and cell-to-cell signaling. I feel pretty comfortable in saying that not all tyrosine kinases would have to be malfunctioning for a cell to be cancerous – it looks like there are more than 100 known 3D structures of tyrosine kinases on the PDB – and I think if all the tyrosine kinases in a cell were malfunctioning, cancer may very well be the least of its problems. Specifically in the context of this paper, I think a Tyr-94 kinase is a normal occurence in cells. PDP (when phosphorylated) just removes a phosphate group off of PDH and turns it off, and you could imagine several reasons as to why that would be helpful in a normal cell (turning off OXPHOS to undergo fermentation in a working muscle cell?). I think like in many other cases, a constitutively phosphorylated PDP1 is the problem.

  4. Hi Besher,

    Great job! In the results section, I noticed that the authors mentioned that Tyr-94 phosphorylation was observed in cell lines that were tumorigenic or leukemogenic cell lines, but that this modification was not seen in any of the normal (non-cancerous) cells from human foreskin or keratinocytes. Thus, do you think that targeting this particular phosphorylation event could serve as a useful treatment method for cancerous cells in humans? The authors seem to hint at this possibility in the last sentence of the paper, where they mention that PDP1 activators may serve as “novel anticancer agents,” which would be extremely exciting because non-cancerous cells would not be harmed in this treatment, thus minimizing side-effects of the medicine. The authors did show, though, that the reduction in Tyr-94 phosphorylation decreased the size of the tumor, but did not eliminate it, so will this treatment be enough to “cure” the cancer? Could the efficacy of this type of treatment perhaps be improved by combining it with an agent that inhibits the acetylation of Lys-202, which is also known to inhibit PDP1?

    1. I think the last thing you suggested – combining several effects at once – would be the best bet. I think it makes sense to anyone with a basic understanding of cancer (we’ll assume it’s not a gain-of-function cancer mutation here) that there is a lot going on to make a cell cancerous, and thus, there needs to be many things targeted at once for a therapy to be successful. I think it is almost expected that the therapy they tried only decreased the size of the tumor without destroying it. After all, with our understanding that aerobic glycolysis is there to promote tumor growth and not tumor sustainability, cancer cells may not need this process to stay alive. I also think that their recent publications are reflective of what’s going on. They’re finding several discrete & independent (or not?) things going on – acetylations here and phosphorylations there – in cancer cells, which we would expect in such a complex disease. Again, I think there needs to be so many things going wrong at once for a cancer to develop, and we’re only starting to see that with these kinds of papers.

  5. Besher – Nicely done! In reading this paper, I realized that they cited themselves…a lot. Two of the papers they cited (references 11 and 12) were referred to almost 10 times! I looked those papers up, and they were published in Molecular Cell in 2011 and 2014 by a team that included a majority of the authors in this paper (http://ac.els-cdn.com/S1097276511008914/1-s2.0-S1097276511008914-main.pdf?_tid=0eda0d7c-be11-11e4-8b41-00000aacb35d&acdnat=1424994300_fa7af91075ef74758168dab99b5e9de4 and http://ac.els-cdn.com/S109727651400032X/1-s2.0-S109727651400032X-main.pdf?_tid=1152652c-be11-11e4-a383-00000aacb35d&acdnat=1424994304_85f1322396a6b4ed80ee02eeffc0eabb). According to wikipedia, JBC has an impact factor of 4.6 and Mol. Cell. has one of 14.464. Now, I’m not one to focus on such an odd statistic, but my question here is about contextualizing this lab’s work in the biochemical landscape. I think all three of their papers are a bit vague in purpose – the one from 2011 is titled in part “… is important for cancer cells,” and they all passively refer to the findings as significant because they point toward the Warburg Effect. Now, I know my question is rather vague, so only tackle the parts you find relevant, but in general, what exactly is this lab doing? Are they randomly finding Tyr kinases and ascribing context and relevance to them later, or are they systematically clarifying the cellular process that promotes tumor proliferation? Do you see a trajectory for the lab historically, or are they just sort of stabbing in the dark? Is there any reason that this paper is in JBC even though it is similar to the other two, which were in Molecular Cell? I’m just trying to see the impact here – for some reason, it feels unfocused to me.

    1. These are some great questions Zach, and I’m not sure I can answer them all sufficiently, but I’ll give them a shot. To me, after reading the introduction of this paper especially, it seems like this paper is more of a continuation of the earlier 2014 paper (initially submitted in 2013). In my opinion, that last paragraph of the introduction boiled down to “here’s what we’ve found before (acetylations), and oh we just found one more”. Maybe the first two papers were making discoveries new enough that they deserved a higher impact journal, whereas this paper was simply adding on to a system previously investigated in some capacity. In terms of your “stabbing in the dark” comment, I was thinking that when I started to read your question, and I’m not sure I’m thinking any differently right now. I think the fact that they’re studying cancer cells has everything to do with it though. In such a complex system, it may be hard/impossible to tell what is causing what and what the source of the cancerous phenotype is – “so why don’t we focus in on one aspect (the PDC and its role in the Warburg Effect) and see what we can find?” is what I think may be going on. I think this approach makes studying cancer a little more manageable, although it may make for less interesting publications.

  6. Besher,

    I was honestly dreading reading this and coming up with a question the night before the exam, given the relatively mushy state of my brain. The fact that this essay is epically relevant is just brilliant. Somebody give this man a cookie.

    Anyways, at the end of the paper, the authors propose that the phosphorylation of tyrosines they under covered in their paper might be relevant to other enzymes and determine how these important metabolic enzymes function in cancerous cells. They also say it would be interesting to see how these phosphorylation interactions could coordinate and effect each other. I was wondering if you could give a brief way about which you could accomplish this experimentally?

    The next part of my question is more general and just something I was kind of thinking about. It seems that cancer cells generally express different isoforms of enzymes (embryonic form of pyruvate kinase) or modify enzymes (PDC) in such a way that all of the modifications tend towards the same goal, which is to promote glycolysis, and presumably cell replication. To me, this seems unlikely that all of these just happen coincidentally, and this is what gives you a tumorogenic cell. I feel like it would make more sense if there were a more general, overarching switch that was responsible for at least a large portion of normal metabolic processes and that mucking with this switch results in many of the cancerous changes that happen at a cellular level. I’m not sure how much sense this makes, I was just curious if you had any thoughts on this matter.

    1. Kind words up front, Tommy! I think I’m going to use one word to try answer both of your questions: correlation. For your first question, let’s say they have ways of probing for the different phosphorylated residues & enzymes on several cells. I’m sure there’s some statistical operation they could do to test if statistically these phosphorylation events are coincidental or correlated. In a simple example, if they found that 90% of cells that have one phosphorylated enzyme also have a second one, then there’s a better chance that one is affecting the other – or that maybe the machinery that controls tyrosine kinases itself is malfunctioning (in other words, is one tyrosine kinase affecting the activity of another, or is there a third party that is in charge of tyrosine kinases that is acting on both?).
      For your second question, there may very well be one big switch that is malfunctioning and causing the cancerous phenotype. However, I think that “coincidental” view is also valid – maybe cancer is just the perfect storm? I’m not exactly sure if I’m justified in pointing out this flaw, but think about their methodology. They looked at cancer cells and found that many/all of them had phosphorylations at Tyr-94 (and in previous papers, acetylations at the respective residues). What I am trying to point out is that they started out with a cell with problems and looked for the problems – but what about healthy cells? Are there healthy cells out there that have these phosphorlations & acetylations that are still undergoing OXPHOS because they don’t have other problems that the cancer cells have? Maybe some of these findings are more common than we think, but they don’t cause cancer because there are a dozen other things that have to go wrong as well. To give a possible scenario, let’s just focus on two aspects of cancer cells as we understand them: (1) uncontrolled growth and (2) aerobic glycolysis. We think that both are necessary for a cell to be cancerous, but what if a cell has just the one? What if the cell can successfully divide very rapidly (because it knocks out the checkpoints of mitosis), but is limited by OXPHOS, and so cannot create the biomass, and so doesn’t actually end up dividing? And what if another cell can undergo aerobic glycolysis all it wants, but if it tries to divide rapidly, the tumor-suppressor cells knock it down? Maybe we have several of these types of cells in our body all the time, and once in a while a cell has both properties and then a cancer develops.

  7. Hey Besher,

    Very cool paper! Its neat to look at some of the more specific characteristics integral to the Warburg effect, especially given that we’ve seen some inter-related effects thereof in class. I was curious as to whether there were other possible sources of the oncogenic effects besides those specifically defined by the researchers: specifically regarding the acetylation of Lys-202. and Lys-321.

    While they cited a 2014 paper http://www.sciencedirect.com/science/article/pii/S109727651400032X
    as they laid claims to the effects of acetylation of the aforementioned Lysine residues, the paper they cited appears to list the Lys-321 and Lys-202 residues within different activity spheres. Given that Lys 321 appears to have a more focused effect on PDHA1 than PDP1 specifically, what other mechanisms do you think might exist in order to mediate the warburg effect?
    Do you feel there’s any likelihood of one pathway (PDP vs PDHA) having a more concerted effect?

    1. I certainly do think there are other sources of the oncogenic nature than those found so far. Personally, when I looked at this group’s work, I thought of it more as “Here’s what we’ve found so far in these cancer cells” – their findings aren’t an end-all list of cancer cell properties, but simply what they’ve found. I expect them to continue what they’re doing and to keep finding other modifications that are involved in these cells – and maybe, when there are enough dots, we can connect them with some overarching idea. In terms of other mechanisms that may promote the Warburg effect – I think until the discoveries are made, there are as many as you can imagine. Conversion of pyruvate to acetyl-CoA occurs in the mitochondrial matrix, so any way of stopping pyruvate from crossing the mitochondrial membrane (maybe it is phosphorylated?) would deter that pathway. Maybe there is some effect that recruits the enzymes of lactic acid fermentation to pyruvate once it is formed after glycolysis so that the pyruvate is consumed before it has a chance to enter the citric acid cycle. Lastly, I would not expect one (PDP vs PDHA1) as being more influential than the other – if one is functional while the other isn’t, then there should not be a major issue and pyruvate should be converted to acetyl-CoA successfully.

  8. Hi Besher- thanks for your work here. It’s always good to see a spotlight on a topic which we have discussed in class. I am still a little unclear as to the relationship between PDH, PDK, and PDP. So PDP de-phosphorylates pyruvate dehydrogenase complex, therefore opening up the path to OXPHOS- so a mutant PDP should decrease OXPHOS, right? Later in your spotlight, you state:

    “Lastly and with great impact, it was shown that expression of a PDP1 Y94F mutant in cancerous cells increased oxidative phosphorylation under normoxic conditions, and decreased ATP production under hypoxic conditions, which was correlated to the idea that Y94F mutant relies more on OXPHOS as compared to WT cells.”

    I would think that expression of that tyrosine mutant would decrease OXPHOS under all conditions, not only under hypoxic conditions, so I find the correlation that the authors make to be a little confusing – this is probably due to my own misunderstanding, though. Decreased ATP production means lactic acid fermentation, not OXPHOS, right? I suppose that the increased level of OXPHOS in normoxic tissue could indicate an increased dependence on the process. I’m just surprised that they can perform OXPHOS, given the nature of the mutation…

    I suppose that the conclusion is that this residue could be important in cancer cells because cancer cells perform lactic acid fermentation despite the oxygen supply in their environment, but this mutant “choose” to perform OXPHOS. That is an interesting correlation- I’m just not sure how all of the enzymatic players come together.

    I apologize for the long comment; I was working through my thoughts!

    1. What you said at the beginning is right – phosphorylated PDH cannot function properly (it can’t convert pyruvate to acetyl-CoA). PDP that is phosphorylated at Y94 cannot function properly either (it can’t de-phoshporylate PDH), which yields the same effect. As for your next question, it made sense to me while I was writing this up, but right now I’m completely drawing a blank. I found the quote in the last part of the results: “such a low oxygen condition resulted in
      decreased ATP levels in these cells as compared
      with control PDP1 WT rescue cells. This is consistent with our
      hypothesis that cells expressing PDP1 Y94F mutant rely more
      on oxidative phosphorylation for ATP production as compared
      with control WT cells, so under hypoxic condition where oxygen is insufficient to sustain oxidative phosphorylation level, PDP1Y94F cells showed decreased ATP levels and subsequently reduced cell proliferation.” It looks like the PDP1 Y94F mutant cells can’t undergo lactic acid fermentation as well as the rescue cells, so when they were both put in a hypoxic state, the rescue cells could make more ATP?

Comments are closed.