PEA is a highly bioactive lipid that activates intracellular and membrane-bound receptors distinct from those activated by eCBs. The molecular machinery controlling PEA production and inactivation in mammalian cells is not fully understood 
. Our results show that S. cerevisiae
produce and inactivate PEA, and that this process is significantly altered by single gene deletions or heterologous expression of mammalian proteins involved in neurodegenerative diseases. Thus, our findings show that forward and reverse genetics in yeast can be used to identify the molecular components that control PEA production and inactivation.
Using CI-GC-MS and isotope-dilution, we demonstrated that S. cerevisiae
produce significant amounts of PEA, a lipid that contains a C16:0 acyl chain, while not producing detectable amounts of anandamide, 2-AG, OEA, SEA, HEA, and DEA, all of which contain acyl chains with more than 16 carbon atoms. Accordingly, classic studies showed that lower eukaryotes cannot produce fatty-acid chains containing more than 16 carbon atoms despite the presence of two enzymes, fatty acid elongase and enoyl reductase, that are known to lengthen carbon acyl chains 
The amount of PEA we detected in yeast (650 pmol/g) is comparable to that in mouse brain (100–500 pmol/g) and cultured neural cells (1800–2200 pmol/g of protein) (for review see 
). However, reliable PEA quantification required only 3 mg of yeast homogenate sample but 30 mg of mouse brain homogenate 
. This 10-fold difference suggests that the polysaccharide wall in yeast is less invasive when analyzing PEA by CI-GC-MS than the rich lipidic environment of the brain.
PEA is also produced and hydrolyzed by plants, where it is thought to be a response mechanism to stressors and regulator of plant growth 
. PEA was also quantified in Ciona intestinalis
and in leech 
. Together with our results, these studies show that PEA is produced by a wide variety of cell types and organisms. Whether PEA constitutes a building block of yeast membranes or carries the function of signaling lipid in these lower eukaryotes remains to be determined. Note that while we do not suggest that the function of PEA as a signaling lipid in mammalian cells is conserved in yeast, our data indicate that the basic machinery involved in its production and inactivation might be conserved, and therefore yeast constitute an incredibly efficient genetic tool to dissect the molecular mechanisms involved in PEA metabolism.
Genetic evidence demonstrating the involvement of specific enzymes in PEA production and inactivation is sparse. To our knowledge, only two examples have been reported: Nape-Pld, which affects PEA production, and Faah, which affects PEA inactivation 
. By screening 10 yeast gene deletion strains, we identified two additional candidates: Spo14 (coding for a Pld-like activity) and Yju3 (coding for a Mgl-like serine hydrolase activity). These genes have mammalian orthologs: Spo14 is related to Pld1 (coding for a phosphatidylcholine-hydrolyzing Pld1), and Yju3 is related to Mgl (coding for a monoglyceride lipase) and to some extent to Abhd12 (coding for a serine hydrolase recently implicated in eCB signaling 
). The implication of a yeast homologue of Mgl and Abhd12 in PEA hydrolysis is remarkable for two reasons. First, it is known that mammalian Mgl does not hydrolyze PEA 
, suggesting that the mammalian Mgl evolved so as to preferentially hydrolyze acylesters rather than acylamide. Second, the substrate specificity of Abhd12 is still unknown, and our data suggest that this newly discovered serine hydrolase might use PEA as substrate and thus be involved in its inactivation in mammalian cells.
Approximately 50% of the PEA hydrolysis expressed by yeast is resistant to heat and microwave inactivation, but is fully blocked by enzymatic digestion. This result agrees with the fact that enzymes resistant to heat-induced denaturation are expressed by some lower organisms 
. While the identity of this heat-resistant enzyme expressed by yeast remains to be determined, our results already suggest that it is unlikely to be a serine hydrolyze, since it is insensitive to the broad serine-hydrolase inhibitor MAFP.
We found that heterologous expression of huntingtin fragments 25Q and 103Q in yeast increased PEA level by ~2 fold, showing that the ability of huntingtin to control PEA levels does not depend on the length of the polyglutamine repeat that determines the age of onset of Huntington's disease. Considering the proposed role of huntingtin in the regulation of gene expression 
, this result suggests that increased PEA levels measured here might not be due to a direct involvement of huntingtin in PEA metabolism, but rather due to huntingtin-mediated impairment of transcription of key molecular steps controlling PEA levels (e.g.
increased expression of enzymes producing PEA or decreased expression of enzymes inactivating PEA) 
. We also found that α-synuclein expression increased PEA levels by ~2 fold, suggesting that the metabolism of this signaling lipid might also be impaired in patients with Parkinson's disease carrying this mutation, as well as mice models carrying this mutation. To our knowledge, no such measurement has been reported, even though α-synuclein is known to affect lipid metabolism 
The molecular steps occurring between the presence of huntingtin and α-synuclein in yeast, and a change in PEA metabolism remain to be determined, since performing these experiments in yeast strains lacking Spo14 and Yju3 did not prevent the effects of these pathogenic proteins, suggesting that they regulate the expression of other proteins. Here, two points can be made. First, it is unlikely that the increase in PEA levels induced by Htt25Q, Htt103Q and α-synuclein results from the down-regulation of a PEA inactivating enzymes, since we found that the genetic mutation of YJU3
leads to a 49% reduction in PEA hydrolysis without increasing PEA levels (compare and Figure S1
). Second, the molecular mechanism involved in these changes in PEA metabolism could be dissected out by using combinations of the forward and reverse yeast genetic approaches described here.
In summary, identifying the molecular steps that control the production and inactivation of signaling lipids is essential for understanding their biological functions and for developing new therapeutic approaches. We found that S. cerevisiae produce and inactivate PEA, and identified several yeast genes involved in its production and inactivation. PEA metabolism in S. cerevisiae was affected by heterologous expression of mammalian proteins involved in Huntington's disease and Parkinson's disease, suggesting that PEA signaling is affected in those conditions. Thus, our results provide proof-of-concept for the use of reverse and forward yeast genetics to gain a genome-wide understanding of the molecular steps in PEA metabolism.