A Novel cis Element Achieves the Same Solution as an Ancestral cis Element During Thiamine Starvation in Candida glabrata

Abstract

Regulatory networks often converge on very similar cis sequences to drive transcriptional programs due to constraints on what transcription factors are present. To determine the role of constraint loss on cis element evolution, we examined the recent appearance of a thiamine starvation regulated promoter in Candida glabrata This species lacks the ancestral transcription factor Thi2, but still has the transcription factor Pdc2, which regulates thiamine starvation genes, allowing us to determine the effect of constraint change on a new promoter. We identified two different cis elements in C. glabrata - one present in the evolutionarily recent gene called CgPMU3, and the other element present in the other thiamine (THI) regulated genes. Reciprocal swaps of the cis elements and incorporation of the S. cerevisiae Thi2 transcription factor-binding site into these promoters demonstrate that the two elements are functionally different from one another. Thus, this loss of an imposed constraint on promoter function has generated a novel cis sequence, suggesting that loss of trans constraints can generate a non-convergent pathway with the same output.

Keywords: Candida glabrata; PDC2; THI2; cis evolution; thiamine.

Figure 1

Phylogenetic relationships and presence or absence of thiamine signal transduction pathway transcription factors and thiamine pyrophosphatases (TPPases). Using a phylogeny of yeast (Gabaldón and Carreté 2016; He et al. 2017), presence or absence of genes was determined previously (Wapinski et al. 2007; Huerta-Cepas et al. 2014). C. glabrata lacks THI2 as do the other “glabrata group” yeast (not presented in figure), but only C. glabrata contains the PMU array of genes (Gabaldón Estevan et al. 2013). We believe that the S. cerevisiae THI pathway behaves similar to the ancestral pathway, and C. glabrata has lost Thi2, gained PMU3, and is unable to synthesize thiamine de novo. WGD (and the star) refers to the whole genome duplication event, and CTG clade refers to the altered codon usage of C. albicans.

Figure 2

The CgPMU3 promoter contains an 11 bp UAS necessary for thiamine starvation dependent expression. A) After truncation analysis of the CgPMU3 promoter, a MEME analysis identified a region that appeared conserved in THI promoters (1000 bp of each THI promoter and 270 bp of the CgPMU3 promoter). Searching the C. glabrata genome for a consensus GACRNANNACG using a pattern match algorithm (Skrzypek et al. 2017), yielded 116 genes with this element in the 1 kb upstream of the start codon, including CgPMU3, but no other known THI regulated genes. The gray shading indicates nucleotides in common with CgPMU3. The number after the promoter name indicates the nucleotide (under the arrow) upstream from the start codon. B) Characterization of the 11 bp CgPMU3 UAS. The first four samples show truncation analysis and the next five samples have mutations introduced into the full-length (1000 bp) wild-type promoter. Promoter induction was assayed during thiamine starvation by measuring the fluorescence of cells containing plasmids with these promoters driving YFP. C) Scanning mutagenesis of the 11 bp CgPMU3 UAS. Single mutations were introduced into the full-length promoter, replacing the native nucleotide with either a T or a G, except when the native nucleotide was a T/G, in which case the T/G was replaced with an A/C. For this and the following figures, the data presented is the mean and standard deviation of at least three independently grown samples.

Figure 3

Deletion of the 11 bp CgPMU3 UAS in THI promoters does not eliminate thiamine starvation dependent expression. The putative CgPMU3 UAS (Figure 2A) was precisely deleted in the full-length promoters of CgPMU3 (and replaced with a PacI restriction site), CgPET18, and CgTHI10 and assayed for YFP expression in high and no thiamine conditions. While necessary for CgPMU3, this UAS is not important for induction of other THI promoters.

Figure 4

Fine scale truncation analysis of THI promoters uncovers a 13 bp UAS that is not present in CgPMU3 A-D) We truncated THI promoters in 100 bp intervals and then further narrowed down to regions where we observed a >90% decrease in thiamine starvation induction. E) With 50 bp regions of the THI promoters around the site of truncation, we performed a MEME analysis and identified a 13 bp region which was not present in the 1000 bp CgPMU3 promoter. The TTCCCTBTAAWTG consensus is only found in 4 promoters in the C. glabrata genome, and those genes do not appear to be regulated by thiamine starvation based on previous RNA-seq data (Nahas et al. 2018). Each promoter element has at least one mismatch from the consensus, suggesting some permissiveness in the element. The arrow indicates the nucleotide number upstream from the start codon and the gray shaded regions are conserved nucleotides. F) A schematic of the location of the two elements in the five most upregulated THI pathway promoters with the arrows indicating where a truncation reduced expression. The blue boxes correspond to the CgPMU3 UAS and the orange boxes correspond to the THI UAS.

Figure 5

Deletion of the THI UAS eliminates thiamine starvation inducible expression of THI promoters. The putative THI UAS (Figure 4E) was precisely deleted, and replaced with a PacI restriction enzyme site, in the full-length promoters of CgPET18 and CgTHI10 and assayed for YFP expression in high and no thiamine conditions.

Figure 6

The THI UAS is able to substitute for the CgPMU3 UAS, but the CgPMU3 UAS cannot replace the THI UAS. A) Deletion of the CgPMU3 UAS in the context of the full-length CgPMU3 promoter results in a severe defect in thiamine starvation inducible expression of YFP; however, replacement of the CgPMU3 UAS with either the CgPET18 or CgTHI10 UAS restores upregulation of the CgPMU3 promoter. B) Deletion of the THI UAS in the context of the full-length promoter results in a severe defect in thiamine starvation inducible expression of YFP; however, replacement of the CgPET18 UAS or the CgTHI10 UAS with the CgPMU3 UAS does not restore upregulation of the promoters.

Figure 7

Deletion of regions most similar to the THI UAS in S. cerevisiae promoters does not abrogate thiamine starvation regulation, but deletion of regions near the UAS reduces expression. A) Mutation of the region most similar to the THI UAS has little effect on upregulation in two S. cerevisiae promoters (see Figure S5 for details on the sequence). B) A scanning deletion of the promoter region of ScTHI5 and C) ScTHI20 uncovers 20 bp that appear important for expression. These regions span the Thi2 binding site and are near the putative CgTHI UAS.

Figure 8

CgTHI10 and CgPMU3 respond differently to the introduction of a ScThi2 binding site. A) A ScThi2 binding site (forward and reverse orientation) was introduced into the CgTHI10 promoter with and without the THI UAS deleted. These plasmids were transformed into S. cerevisiae strains and assayed for fluorescence in thiamine starvation conditions. For there to be increased expression of CgTHI10 in S. cerevisiae, the Thi2 binding site must be incorporated in the forward orientation and expression requires the THI UAS. B) The CgPMU3 promoter tolerates the ScThi2 binding site in either orientation and does not require the CgPMU3 UAS to function in S. cerevisiae, but expression is still Thi2 and Pdc2 dependent. CgPMU3 with a Thi2 binding site leads to higher level expression of the promoter in S. cerevisiae relative to CgTHI10. It is unclear why the two promoters have such different expression levels.

Figure 9

Model of transcription factor binding sites in thiamine starvation regulated promoters in S. cerevisiae and C. glabrata. A) ScThi2 binding may be the “anchoring” step, and because Pdc2 is in a complex with Thi2, Pdc2 is then able to bind to degenerate sequences nearby, leading to the recruitment of the transcriptional machinery. B) THI promoters in C. glabrata (other than CgPMU3) behave similar to S. cerevisiae promoters but only require Pdc2 and Thi3. C) For the CgPMU3 promoter, it seems likely that a novel transcription factor has been co-opted into the THI pathway to act as a functional analog to Thi2, and it may bind both the CgPMU3 UAS and CgPdc2. Regardless of where Pdc2 binds, it is still required for the recruitment of the transcriptional machinery.