A new preprint from Brian Zid’s lab shows accumulation of P-body-like aggregates in old yeast cells with a suggested pathological role in aging. I think some key experiments are missing. A short review.
We do not yet understand the aging process at the cellular and molecular level in full detail. Many pathways contribute to the accumulating damage to molecular and cellular components. One highly studied “feature” of aging is the accumulation of protein aggregates, that in some cases leads to old-age associated neurodegenerative diseases.
The new perprint, which presents the research led by postdoc Joonhyuk Choi, looks at protein aggregates in aging yeast as a model system. They found the proteins that typically associate with P-bodies accumulate in huge aggregates in aging yeast, and they show some interesting data which suggests that these aggregates might be pathological and actually promote the aging process.
The model system:
They used an engineered yeast strain called the Mother Enrichment Program (MEP). The baker’s yeast (aka budding yeast) is asymmetrically dividing. This means that a mother cell produces a bud, which at the end of the cell cycle, detaches from the mother cell. The study of aging in yeast requires removing the daughter cells away from the mother, and counting how many times the mother buds (replicative life span). This can be done by physically removing the new bud every cycle (every 2 hours). It is performed with a fine needle under the microscope and is tedious. Early work from 1959 used this method to estimate a life-span of 24+/-8 budding cycles. The MEP strain is genetically engineered to delete two essential genes – but only in the daughter cells and only by the addition of an inducer. A very clever system.
Certain proteins were fluorescently tagged and single mother cells were followed under the microscope for several days using a microfluidics device. The proteins they tagged are known P-body related proteins: Dcp2, Lsm4, Xrn1, Edc3, Pat1, Dhh1. These proteins are the major enzymes involved in cytoplasmic 5′-3′ mRNA decay. They also tagged two other proteins – Pab1 (poly-A binding protein) as a marker for stress granules and Hsp104 as a marker for other protein aggregates.
So let’s talk a little bit about P-bodies. It was first described in 2003 by Roy Parker as an aggregate of mRNA and mRNA decay-processing proteins, and Parker suggested it was a hub of mRNA decay. This specific role was later disputed by Jeff Coller, Jeff Chao and Evelina Tuttucci, and their role is still unclear. P-bodies appear in yeast cells that experience many kinds of stress – but these are separate from stress granules. They also accumulate in cells defective in the mRNA decay pathway.
Results:
Zid’s paper shows very clearly the accumulation of aggregates which contain these decay factors as the yeast age – but this is separate from Pab1- or Hsp104 granules. The size and brightness of these aggregates increase with age and it seems that this aggregation accelerates at when the yeast reach 70% of their individual life-span. This also coincides with slowing of the cell division rate, which is an interesting correlation.
Amazingly, some mothers can dump this aggregate into their daughter. It seems that the mother gains 3-4 more cycles but this is toxic to the daughter, which in most cases stops dividing.
The last result shows that the cytosol acidifies as the yeast age (which they measure using pHluorin), and if this is counteracted by over-expressing a vacuolar H+-ATPase, the viability and replicative life span of the yeast is increased.
Overall they show very interesting results, but I am not convinced that this granule is pathological to yeast and/or promotes aging.
Major issues – key experiments that I think are missing or could be very informative:
- There are at least two known mutations that inhibit the formation of P-bodies in yeast (at least under starvation conditions) – dhh1/pat1 deletion and edc3/lsm4-deltaC. If indeed this is a P-body that becomes toxic to aged cells, can aging be delayed in these mutants? It may be that these P-bodies accumulate due to other aging processes, but do not directly cause aging. And why are they toxic to daughter cells? – well, maybe its not the aggregate itself which is toxic. Maybe it is accompanying mRNAs that affect the cell-cycle, or that this huge aggregate is transferred by mistake instead of another essential organelle/protein. I think it is too early to claim that it is pathologic.
- On that same note – treating cells with cycloheximide dissipates P-bodies, since mRNAs shuttle between P-bodies & polysomes, but with cycloheximide the mRNAs get stuck on polysomes. A short treatment might dissipate this age-related P-body and “rejuvenate” the mothers. If this aggregate dose not dissipate – it is probably not a classical P-body. If dissipates but the cells are not rejuvenated – then perhaps this aggregate is not toxic.
- Does this granule actually contain mRNA? It is easy to test with an MS2-tagged mRNA, and is an important information as to the nature of this granule. Also – what happens if you over-express an MS2 version3-tagged mRNA? We know it forms large aggregates. Does this accelerates aging?
- I think figure 4 is over-interpreted in an attempt to tie acidification with the P-body formation. It appears that the normalized intensity of the P-bodies in wild-type and H+-ATPase overexpression in old age is similar. There are minute differences by how significant are they in affecting aging? These aggregates accumulate in either case. Furthermore, Figure 4F shows less data points compared to 4E, so any comparison is shaky. More data points are required, especially for the 48h time point. In my opinion, the acidification is a second, unrelated aging process.
- What is the % of mothers who transfer their P-body to their offspring? This is important data which is not clearly presented. Based on supplementary table, it looks like 15%. But the question is whether these were randomly identified, or whether these cases were cherry-picked, until you got more than 10 cases, for instance?
Minor points:
- It is unclear how the experiment with the toxicity of the granule in daughter cells was performed. Was the inducer of the MEP system removed at some point? Were daughters then separated from mothers and followed separately? The assay should be explained clearly.
- Why is Pat1 shown in S2 but not in 1A? also, all these other factors shown in A – do they show the same pattern of accumulation as Dcp2 or as Pat1?
- I know its small numbers, but is the granule transferred to daughters at a specific stage of the cell cycle (e.g. only small bud) or any stage before separation?
Questions that are interesting to me:
- Since these factors are important not only for mRNA decay but also for transcription and translation, I think it will be very interesting to determine if aging yeast show slowed or misregulated mRNA decay, or transcription or translation or all of the above.
- See question 1 but now in all kinds of mutant backgrounds related to aging, including the H+-ATPase overexpression strain.
- Why do the Hsp104 granule form, then disappear? shouldn’t it also accumulate over time? what happens in the H+-ATPase overexpression strain?
- What is the RNA content of this granule? If the granule forms due to accumulation of a (few) over-expressed mRNA(s) which fail to properly degrade during aging, perhaps these can be directly targeted for proper degradation by some way, thus delaying the granule formation.
- I wonder if there is an evolutionary reason for old mothers to effectively kill their offspring. Do they sense that this particular daughter cell is defective and use the granule-dump process to kill it before it makes more progeny?
I think that this work has the potential to have profound impact and I can’t wait to see this research continues.
Age-induced P-bodies become detrimental and shorten the lifespan of yeast
Joonhyuk Choi, Shuhao Wang, Yang Li, Nan Hao, Brian M. Zid
bioRxiv 2021.11.05.467477v1
greenfluorescentblog 
Hi Gal,
Thanks for reading through the manuscript and providing comments. One major point I would like to make is that while the first experiments were done using the MEP strain, from Figure 1C on, these experiments were all performed using a microfluidics device of non-MEP yeast (Li et al 2017 – http://haolab.ucsd.edu/li_PNAS_2017.pdf). These experiments are very laborious as the microfluidics devices can fail for a variety of reasons, but they provide really amazing data where we can track single mother cells every 15 minutes across their entire lifespan, and the daughter cells for some time until they are expelled from the device.
1. We did try some edc3 deletion experiments, and while we see changes in the trajectory of P-body formation, and the lifespan curves were distinct from wildtype, we did not see significant changes in average lifespan. This made interpretation complicated. One possibility is that P-bodies may have an antagonistic pleiotropic-like affect and be beneficial in early life (there is data that stress induced small P-bodies can be beneficial) and then these large age-induced P-bodies may become pathological in late life and by deleting a protein like Edc3 this may have detrimental and beneficial effects, making it difficult to interpret lifespan effects.
In terms of pathology. I would like to point out that at the same time we see strong detrimental effects in the daughter cells, we see lifespan extension in the mother cells. Its unclear how anything you describe would drive increased lifespan in the mother cell, with concomitant decrease growth in the daughter cell unless what is being transferred/removed is pathological?
2. I like this idea of trying to temporally change the presence of large P-bodies for the aforementioned antagonistic pleiotropy reasons. We haven’t tried this yet, but it could be worth exploring.
3. We did some preliminary experiments where we see mRNAs in the young granules and then changes in mRNA localization over time, but it was very unclear if this was because of changes in reporter, or actual changes at the level of the granule localization.
4. We don’t think that the large P-bodies are major drivers of lifespan, as we only see a small increase in lifespan when cells transfer them to daughters. Similar to if you cure something like sporadic Alzheimer’s disease in an 80 year old. Will they live longer? Potentially yes. Will they live 40 years longer? No. We do find it interesting though that if you slow the acidification of the cytosol, this also leads to a significant, >5 generation delay in the formation of the very intense P-bodies.
5. The % of mothers that transfer is less than 10% (we had to use a selection of the non-transfer cases to make sure cells were correctly aged-matched), and the percentage of daughters that receive an intense P-body is much less than that since a mother with a large P-body undergo ~6 divisions, most of those daughters will not have a large P-body transferred to them.
Minor
1. As mentioned earlier this was using microfluidics devices not the MEP strain.
2. Similarly, Pat1 was done in the microfluidics device but not using the MEP strain. We see high amounts of overlap between Pat1 and Dcp2 in our microfluidics device.
3. Though the device allows us reasonably high time resolution, we are still only taking images every 15 minutes. With the low numbers as you mentioned, and this time window we don’t have the resolution to answer that question.
I agree these are interesting questions, but as mentioned these microfluidic experiments are both laborious to perform but also take a large amount of time to analyze the data. The segmentation of the budding events is very difficult to automate since daughters can arise from the mother or from other daughters in the device. This makes doing large numbers of experiments/mutants difficult to perform.
Its unclear why you think Hsp104 foci disappear? Generally, they appear relatively early, and sustain throughout the lifespan, but they don’t progressively increase like Dcp2.
We are also very interested in what the composition (RNA and protein) of the large granules are. This will have to be performed in a MEP strain because we cannot isolate large amounts of cells from the microfluidics.
Its unclear if the passage of the pathological granule to the daughter is an active process or instead just a random event. Overall since this would happen in late life, and most yeast cells in the wild would never make it to such a time, there may not be much selective pressure to regulate this event.
LikeLiked by 1 person
Hi Brian,
Thanks for taking the time to reply.
” Figure 1C on, these experiments were all performed using a microfluidics device of non-MEP yeast ” – this was definitely not clear from the text. I really thought that all experiments were done on MEP in microfluidics – and i wasn’t the only one. I didnt mention in the post, but we had a journal club on this paper. We all missed this.
1. edc3 – interesting and complicated, i agree. But I suggested edc3/lsm4deltaC strain which is completely devoid of PB. Maybe it will simplify things?
regarding the other point – yes, that is the simplest explanation. but I think other possibilities should be explored. e.g. perhaps the PB is transferred at the expenses of an essential organelle? and now this extra organelle in the mother is beneficial.
3. hmmm… well, it would have been nice if you show at least in the early PB, but i agree that this is not very informative, except to show you tried 🙂
4 & 5. If you explain like that also in the text, the results and their meaning will be better understood.
minor 3 – too bad… would have been interesting if it only happens at a certain stage of the cell cycle.
Hsp104 – ok, i see my mistake. it doesnt disappear, but it does reache a peak, then declines in intensity and size. That’s an odd dynamic for a protein aggregate.
your last point – I haven’t thought of it this way.
Again, thanks for your long and thoughtful reply.
LikeLike
“this was definitely not clear from the text” This is the text from the first line of the 2nd paragraph of the results when we start to use the microfluidics device – “To track the age-induced formation of P-bodies across time,we tagged the P-body marker Dcp2 with mRuby2 in wildtype cells and followed the formation of foci in a microfluics device under normal growth conditions over their entire lifespans (Fig. 1C and 1D).”
It is an important point that we hope others will not miss. I guess I was under the impression that the field doesn’t think of the MEP strain as anywhere close to a wildtype, but you believe this should be clarified more?
LikeLike
Yes, definitely clarify this point. I read the paper 3 times and didnt get it until you explained in your reply.
LikeLike