Carotenogenesis in X. dendrorhous is a complex process with regulatory mechanisms that have not been fully elucidated. Several studies have reported that the amount and composition of carotenoids may be greatly modified depending on the carbon source used [12–14, 29, 30]. A common observation is that the synthesis of pigments is particularly low at glucose concentrations greater than 15 g/l [12, 13, 31]. However, until this study, there was no available data on how glucose exerts its repressive effect on carotenogenesis.
The results obtained in this work show that glucose has a regulatory effect on the expression of several genes in X. dendrorhous, as has been shown in other yeasts. The mRNA levels of the grg2 gene decreased dramatically when glucose was added to the culture. Moreover, the PDC gene was induced by glucose, as it is in the majority of phylogenetically related organisms [22–25]. In addition, we found that adding glucose to the media caused a decrease in the mRNA levels of all of the carotenogenesis genes involved in the synthesis of astaxanthin from GGPP. In the majority of these experiments, the effect of glucose reached its maximum between 2 and 4 h after addition. By 24 h after glucose addition, mRNA levels returned to baseline. No data were collected between 6 and 24 h after the addition of the sugar, but in most cases the recovery was estimated to occur completely within the first 8 h after the addition of glucose. Furthermore, the remaining glucose determinations showed that the kinetics of sugar consumption was slower than the return to basal gene expression levels. This finding suggests some type of adaptation mechanism, which over time diminishes the transcriptional response to the presence of glucose.
The global effect of glucose on the carotenogenesis pathway may be related to the presence of binding sites for the MIG1 general catabolic repressor in the promoter regions of the crtS , crtYB and crtI genes . Such sites are also present in the promoter region of the grg2 gene (unpublished data), suggesting that a homolog of the MIG1 regulator may mediate the glucose repression of these genes. However, further studies are needed to demonstrate the functionality and importance of these elements.
Interestingly, the repressive effect of glucose on crtYB and crtI is manifested in different ways on the alternative and mature transcripts of these genes. Considering that both transcripts of each gene come from a single transcriptional unit, their different expressions suggest the involvement of post-transcriptional regulatory mechanisms. Thus, the differential effect may be due to modification of the factors that control the alternative processing of these genes. This differential effect is in addition to previous observations that the amounts of the mature and alternative mRNAs for both genes vary during yeast growth, depending on the carbon source used, the age of the culture and the carotenoid content . The functions of the crtYB and crtI alternative transcripts are unclear [10, 15, 32], although it has been established that they are generated from anomalous splicing of the respective non-processed messenger. The alternative mRNA of the crtI gene conserves 80 bp of the first intron, while the alternative mRNA of the crtYB gene conserves 55 bp of the first intron and lacks 111 bp of the second exon. In both cases, the alternate splice results in mRNAs with several premature stop codons in their sequences , suggesting that the alternative transcripts may not encode functional proteins. Studies performed in our laboratory indicate that mutant strains that only express the alternative mRNA of the crtI gene are unable to synthesize astaxanthin and they accumulate phytoene , indicating that this mRNA does not encode a functional phytoene desaturase protein. Considering these observations, the biological significance of the glucose-mediated repression of the alternative crtYB and crtI mRNAs is not clear.
An important observation is that the glucose-mediated repression of the crtYB, crtI and crtS genes was seriously compromised in mutant strains incapable of synthesizing astaxanthin. This observation is consistent with previous reports that showed that a decrease in astaxanthin content causes an increase in the total amount of carotenoids, suggesting that astaxanthin may have a negative feedback effect on pigment synthesis . The results reported here indicate that an inability to synthesize astaxanthin would cause deregulation of a significant number of genes involved in the late stages of the pathway, thereby releasing it from repression by glucose and even increasing the availability of the messengers necessary for pigment synthesis.
By studying the effects of glucose on cell growth and early pigment production, we found that glucose promoted a high biomass production after 24 h, but completely inhibited carotenoid biosynthesis. Similar results were observed when other glucose-derived carbon sources were used, such as maltose and galactose (data not shown).
The early glucose-mediated inhibition of carotenoid synthesis can be explained, at least partially, by the decrease in the mRNA levels of the carotenogenesis genes. A previous study showed that overexpression of crtYB causes an increase in the amount of pigments produced and that overexpression of crtYB and crtI cause a change in the relative composition of the carotenoids synthesized . These results indicate that changes in the mRNA levels of the carotenogenesis genes have a direct effect on pigment biosynthesis, supporting the importance of gene expression in the regulation of the pathway. However, it was not possible to establish a priori whether the inhibition of pigment synthesis caused by the addition of glucose was due only to a decrease in the expression of the carotenogenesis genes. Previous work has shown that high glucose concentrations (between 20 and 80 g/l) cause a reduction in the X. dendrorhous growth rate; the low carotenoid production may be associated with that inhibition . However, our results indicate that, at least under the conditions tested, glucose can induce an important increase in biomass. Thus, the inhibition of carotenoid biosynthesis reported here cannot be explained by a reduction in growth rate. Another possibility is that the inhibition of pigment production is a consequence of the cell growth promoted by glucose, in contrast with the lack of growth observed in the control culture. However, the experiments were designed to evaluate the effect of glucose and ethanol over a short period of time after the addition of the carbon source (only during the first six hours). In most cases, the maximum effect on carotenogenic gene expression was observed between 2 and 4 h after the treatment; during this time, biomass variations were very low. In addition, X. dendrorhous exhibited very poor growth in other carbon sources, such as galactose, sorbose and succinate, registering a growth level equivalent to the control condition (data not shown), preventing these carbon sources from being used as a "growing" control.
It is well known that glucose has a global effect on cell metabolism, causing induction of genes related to glycolysis and fermentative metabolism and thereby repressing many of the genes involved in secondary metabolism and the use of alternative carbon sources . The induction of the PDC gene and repression of the invertase-coding gene INV (data not shown) in response to glucose addition suggests that this phenomenon may also occur in X. dendrorhous. Therefore, the inhibition of pigment synthesis in response to glucose may also be a consequence of inhibition of the components of respiratory metabolism, which control the availability of substrates for the carotenogenesis pathway.
However, in contrast to glucose, non-fermentable carbon sources generally cause an increase in the synthesis of pigments in X. dendrorhous [12, 13]. Accordingly, our results indicate that the addition of ethanol causes an increase in the total amount of carotenoids 24 h after treatment. Strikingly, even when there was an increase in the biomass, ethanol induced de novo synthesis of pigments, as evidenced by an increase in the relative amounts of intermediate carotenoids in the pathway. These results agree with a previous report by Gu and coworkers, in which the addition of ethanol at different stages of growth caused an increase in the total amount of carotenoids . The authors proposed two mechanisms by which ethanol induced the biosynthesis of pigments. Initially, ethanol and its subsequent conversion into acetate by the aldehyde oxidase enzyme may have generated superoxide radicals, resulting in the induction of carotenoid synthesis. This outcome is consistent with previous work examining the induction effects of some reactive oxygen species on carotenogenesis [27, 36, 37]. Alternatively, acetate may have continued along its metabolic pathway towards the generation of acetyl coenzyme A, with the latter becoming the substrate for the synthesis of isoprenoids by the mevalonate pathway. This outcome is in agreement with previous reports demonstrating that the addition of mevalonate  and several other non-fermentable carbon sources [12, 29] causes an increase in pigmentation production in X. dendrorhous, probably because of their direct conversion into isoprenoid precursors. Our results suggest that there is a possible third mechanism underlying increased pigmentation production, which is mediated by the increase in expression of the crtYB and crtS genes caused by the addition of alcohol. The increase in pigment synthesis mediated by ethanol is likely due to a combination of these proposed mechanisms as well as other factors not yet elucidated.