Tyrosol is a quorum-sensing molecule in Candida albicans

Abstract

The human fungal pathogen Candida albicans shows a significant lag in growth when diluted into fresh minimal medium. This lag is abolished by the addition of conditioned medium from a high-density culture. The active component of conditioned medium is tyrosol, which is released into the medium continuously during growth. Under conditions permissive for germ-tube formation, tyrosol stimulates the formation of these filamentous protrusions. Because germ-tube formation is inhibited by farnesol, another quorum-sensing molecule, this process must be under complex positive and negative control by environmental conditions. The identification of tyrosol as an autoregulatory molecule has important implications on the dynamics of growth and morphogenesis in Candida.

Cells monitor their population density by releasing signaling molecules to which they respond. In bacteria, these autostimulatory compounds are usually small molecules, the concentration of which increases as the number of bacteria increases. After reaching a threshold concentration, these quorum-sensing molecules induce the population to cooperate in diverse behaviors such as bioluminescence, antibiotic production, virulence, biofilm formation, competence, sporulation, etc. (1). In Gram-negative bacteria, these compounds are often acyl homoserine lactones, and in Gram-positive bacteria, they are usually modified peptides (2).

One of the most striking quorum-sensing behaviors is the extremely long lag time shown by cells when they are diluted into fresh medium. The greater the dilution, the longer it takes for cells to initiate exponential growth. This “lag phase” is usually attributed to the requirement for autostimulatory compounds that are released into the medium as the cells proliferate but fall below some critical concentration when the cells are diluted into fresh medium. Compounds have been isolated from bacteria that shorten the lag time (3-5). Thus far, the phenomenon has not been investigated in fungi (6).

Candida albicans, a serious fungal pathogen (7-9), displays several density-dependent phenomena. After dilution, it has a long delay before resuming exponential growth. In addition, cell density controls the morphological switch between the cellular yeast form and the filamentous hyphal form: at high cell densities, Candida is in the yeast form; at low cell densities, the yeast-form cell develops a filamentous protrusion known as a germ tube (10, 11). Recently, farnesol and farnesoic acid have been identified as quorum-sensing molecules that block the morphological transition from yeast to the filament form at high cell densities in C. albicans (11, 12).

In this report, we show that the delay in growth after dilution is abolished by conditioned medium (CM). The active compound in the CM is tyrosol, which is released into the medium continuously during growth. In dilute cultures, tyrosol accelerates the formation of germ tubes. These results suggest that the morphogenesis of C. albicans is under complex positive and negative control by environmental conditions.

Materials and Methods

Growth Assay. C. albicans wild-type strain SC5314 was cultured in synthetic minimum medium (SD, pH 4.3) (13) at 30°C for 24 h. Cells were collected by centrifugation, washed twice with fresh SD (pH 4.3) (13), and reinoculated at indicated cell density in either SD or SD plus compounds. Cultures were allowed to grow at 30°C. Samples were taken at indicated time points, and cell densities were measured by either counting colony-forming units or measuring the OD absorption at 600 nm. For cultures at low cell densities (10⁵ and 10³ cells per ml), cells were concentrated by a filter unit (Millipore) before measuring OD absorption. The chemically synthesized tyrosol (Fluka) was purified by reverse-phase HPLC before being used in the assays.

Activity Purification. CM was prepared from overnight cultures (24 h at 30°C) by centrifugation and filtration to remove Candida cells. For analytical scale purification, 100 ml of CM was loaded to a C18 Sep-Pak cartridge (Waters). The column was washed with H₂O and 5% acetonitrile, respectively. The active component was eluted by 15% acetonitrile. The active fraction was lyophilized and purified further by C18 reverse-phase HPLC (column: Discovery C18, 25 cm × 4.6 mm, 5 mm; gradient: 6-15% acetonitrile in 60 min; flow rate: 1 ml/min). For preparatory scale purification, 10 liters of CM was loaded to 100 g of C18 resin (Waters) and washed extensively with H₂O and 5% acetonitrile. The active compound was eluted with 15% acetonitrile, lyophilized, and fractionated by a G25 resin (Bio-Rad). The final purification was achieved by C18 reverse-phase HPLC (column: Supelcosil LC-18, 25 cm × 10 mm, 5 mm; gradient: 9-15% acetonitrile in 60 min; flow rate: 2 ml/min).

NMR Measurements. NMR spectra were recorded on a Varian AS500 NMR spectrometer at 500 MHz for ¹H and 125 MHz for ¹³C. Data acquisition and processing were performed by using Varian vnmr 6.1 software according to the instructions. Chemical shifts of ¹H and ¹³C NMR spectra were referenced to the solvent peaks: 2.49 (δH) and 39.7 (δC) for DMSO-d6 and 4.80 (δH) for ²H₂O.

Filamentation Assay. C. albicans SC5314 was inoculated in SD (pH 4.3) at a density of 1 × 10⁶ cells per ml and incubated at 30°C for 24 h. Cells were collected by centrifugation and washed twice with distilled H₂O and reinoculated at 10⁴ cells per ml in prewarmed SD (pH 7.0) in the presence or absence of 20 μM tyrosol. Cultures were incubated at 37°C with shaking (200 rpm in a G10 Gyrotory Shaker, New Brunswick Scientific). Samples were taken at indicated time points and concentrated by 2,000-fold by filtration and centrifugation. The morphology of Candida cells was assessed directly by microscopy analysis.

Gene-Expression Analysis. C. albicans SC5314 was cultured in SD (pH 4.3) at 30°C for 24 h. Cells were collected by centrifugation, washed twice with fresh SD (pH 4.3), and reinoculated at 10⁸ cells per ml in SD (pH 4.3), 10⁵ cells per ml in SD (pH 4.3), and 10⁵ cells per ml in SD (pH 4.3) in the presence of 20 μM tyrosol. Samples were incubated for 60 min at 30°C. Cells were collected by filtration and stored at -80°C. Candida total RNA was prepared and used to amplify mRNA as described (14). The amplified mRNA was reverse-transcribed in the presence of amino-allyl dUTP and subsequently labeled with either Cy3 or Cy5 fluorescence dyes (Promega) (15). The Cy3- or Cy5-labeled cRNA was used to hybridize an oligo-based Candida genome array (Qiagen, Valencia, CA). Microarray raw data were extracted, normalized, and filtered to remove any genes with expression that did not change at least 2-fold in the experiment. Only genes showing a consistent pattern in two experimental duplicates were considered.

RT-PCR analysis was performed by using a Sensiscript RT kit according to the manufacturer instructions (Qiagen). The gene-specific probes for Northern analysis were amplified by PCR and labeled by random priming (Stratagene). The DNA sequences for these genes are available at http://genolist.pasteur.fr/CandidaDB/index.html. The primers used for PCR amplifications were ATGTTAGAAGGTAAATTTGAAGAAG and CTACTCATCATCATCGAATTTTGG (CaPol30), ATGCTGTCCAGATCAGCCAG and GTGGTTATTGACCTCCTTGT (CaPol1), ATGTCCGTACAAGAAACTCC and CCAACAATACCATCGGAAGC (CaRNR21), ATGTCCAATTTTGTATCATCACC and GTATGTAACAAGCTTGGTGC (CaMCM6), TGTTCATACTGGTGGAAGAG and GACTAAGCTTACACCAACTCTCAAAACGTTC (CaCDC46), ATGTCAGGTGGTAAAGGTAAAG and CTACAATTCTTGAGAAGCCTTAAC (CaHTA1), and CAAGCTTGATGGAAAGAAGACAATTTAACAC and CGGATCCGCATTTCCTCTATGGTGG (CaEDT1).

Results

To determine the effect of cell density on the growth of C. albicans, we diluted an overnight culture of the organism into fresh minimal medium (SD, 30°C, pH 4.3). When inoculated at low cell densities in SD, Candida went through a distinct lag phase before cells initiated exponential growth. The length of the lag phase increased dramatically with decreasing cell density (Fig. 1). In SD and at a density of ≥10⁷ cells per ml, C. albicans resumed exponential growth with no apparent lag (Fig. 1 A). However, at lower densities, the lag phase was significantly longer: at a density of 5 × 10⁵ cells per ml, they manifested a lag phase of 80 min (Fig. 1B); at 5 × 10³ cells per ml, the lag phase was ≈5 h (Fig. 1C).

Fig. 1.

Cell density controls the length of lag phase in C. albicans. An overnight culture of Candida strain SC5314 was diluted into SD (pH 4.3) at 5 × 10⁷ (A), 5 × 10⁵ (B), and 5 × 10³ (C) cells per ml in the absence (triangles) or presence (squares) of CM. The number of cells at each specific time point was acquired by measuring the OD absorption at 600 nm. The growth curve for each culture was prepared by plotting the logarithmic values of OD₆₀₀ vs. incubation time.

The addition of cell-free supernatant (CM) from high-density cultures (≥10⁸ cells per ml) to cells at low densities completely abolished their lag phase; in the presence of CM, even dilute cells (5 × 10³ cells per ml) underwent continuous exponential growth (Fig. 1 B and C). In contrast, the CM prepared from low-density cultures had no effect on the duration of the lag phase. These observations suggest that the CM of high-density cultures contains an activity that permits highly diluted cultures to resume exponential growth without a substantial lag. Furthermore, the activity was specific to the lag phase and had no effect on the exponential growth of C. albicans (Fig. 1 A).

The ability of the CM to attenuate the lag phase provided the assay for an activity-directed purification of the molecule. In the presence of CM a dilute culture (10³ cells per ml) became turbid in 24 h, whereas in its absence it did not. The purification involved size fractionation and reverse-phase HPLC chromatography (Fig. 2A). The structure of the purified molecule was solved by measuring one-dimensional NMR spectra (¹H and ¹³C NMR) and two-dimensional NMR spectra (¹H ¹H COSY and ¹H-¹³C HMBC, heteronuclear multiple bond correlation spectra) (Fig. 2B). These spectral data indicate that the purified signaling molecule is 2-(4-hydroxyphenyl)ethanol (tyrosol), a compound derived from amino acid tyrosine (16, 17). To confirm that the purified molecule is tyrosol, we compared its chemical properties to those of tyrosol synthesized chemically. The compound from the CM and tyrosol had identical UV-absorption spectra, HPLC elution profiles, and NMR spectra. Moreover, the chemically synthesized tyrosol shortened the lag phase of dilute cultures of C. albicans (Fig. 2C). All these results indicate that tyrosol is responsible for the growth-stimulatory activity of CM. Tyrosol has been isolated previously from the culture medium of a number of fungi including C. albicans, but its autostimulatory effects had not been investigated (16-18).

Fig. 2.

Tyrosol is responsible for the growth-promoting activity of CM. (A) A typical analytical scale reverse-phase HPLC elution profile of the activity. The active peak is labeled as QS (quorum signaling). The values on the x axis represent the elution time, and the values on the y axis represent the UV absorption at 215 nm. (B) NMR spectral data for the purified compound. HMBC, heteronuclear multiple bond correlation. (C) The chemically synthesized tyrosol shortens the lag phase of Candida after dilution. An overnight culture of Candida was diluted to 5 × 10³ cells per ml in SD (pH 4.3) in the absence (triangles) or presence (squares) of tyrosol (20 μM). The number of cells at each specific time point was acquired by plating and subsequently counting colony-forming units on yeast extract/peptone/dextrose agar plates.

The accumulation of tyrosol in the medium increases with increasing cell density (Fig. 3). The concentrations of tyrosol in CM prepared from Candida cultures at different densities were measured by analytical scale reverse-phase HPLC. Tyrosol could be detected in the medium when the culture reaches a density of ≥10⁷ cells per ml and accumulated to a level of ≈3 μM when the culture is ≥10⁸ cells per ml. This result indicates that tyrosol is released into the medium continuously during the growth.

Fig. 3.

The accumulation of tyrosol in the medium increases with increasing cell density. An overnight Candida culture was diluted to 1 × 10⁵ cells per ml in SD (pH 4.3), and its growth (squares) was measured by counting colony-forming units. The yield of tyrosol at specific time points (triangles) was obtained by measuring its concentration in the medium by analytical scale reverse-phase HPLC.

Externally added tyrosol shortens the lag phase. The concentration of externally added tyrosol required to shorten the lag phase (≥10 μM) is greater than that present in CM (≈3 μM). The observation suggests that CM may contain other compounds that also play direct or indirect roles in abolishing lag phase. Tyrosol is unlikely to be an essential growth factor because it cannot serve as a carbon source and does not affect the rate of exponential growth or the total yield of cells. Moreover, neither tyrosine, the presumptive precursor of tyrosol, nor other amino acids (histidine, tryptophan) affect the extent of the lag phase.

Tyrosol also accelerates the morphological conversion of Candida yeast-form cells to filaments. After dilution into permissive medium, the yeast-form cell develops a filamentous protrusion known as a germ tube (10). Standard minimal medium can be made permissive for germ-tube formation if the pH is raised from 4.3 to 7 and the cells are grown at 37°C. When diluted in SD (pH 7, ≤10⁴ cells per ml), ≈5% and ≈15% of cells formed a germ tube after 1 and 2 h, respectively, whereas in the presence of tyrosol, ≈55% and ≈80% of cells formed a germ tube (Fig. 4A). After 6 h, the majority of the cells formed a germ tube even in the absence of tyrosol (Fig. 4B). At higher cell densities (≥10⁵ cells per ml) in this medium, most of the cells form germ tubes with or without tyrosol. Moreover, tyrosol does not induce germ tubes in cells grown in noninducing conditions. These observations suggest that under conditions favorable for germ-tube formation, dilute cells grown without tyrosol take longer to initiate growth and therefore the formation of a germ tube.

Fig. 4.

Tyrosol promotes the germ-tube formation in C. albicans.(A) Microscopy analysis of the morphogenesis of Candida in the presence or absence of tyrosol. Diluted Candida cultures (1 × 10⁴ cells per ml) were incubated in SD (pH 7) at 37°C in the presence or absence of tyrosol (20 μM) and photographed at a magnification of ×400. At 1 h in SD (pH 7) without tyrosol, ≈5% of cells formed germ tubes, whereas in the presence of tyrosol, ≈55% of cells formed germ tubes. At 2 h, only ≈15% of cells formed germ tubes when tyrosol was absent, whereas ≈80% of cells formed germ tubes when tyrosol was present. (Scale bar, 25 μm.) (B) Quantitation of the germ-tube formation in Candida when diluted to 10⁴ cells per ml (percentage).

Previous work has shown that germ-tube formation is suppressed by farnesol, which accumulates in the medium as the cells proliferate (11, 12). Therefore, germ-tube formation in Candida is under the control of at least two quorum-sensing molecules: farnesol, which inhibits formation, and tyrosol, which promotes it. Presumably, at high cell densities, at which the majority of cells are in the yeast form, the suppressive effects of farnesol predominate.

Although the annotation and assembly of the Candida genome is incomplete, it was possible to identify genes with transcript levels that were regulated by tyrosol using whole-genome arrays (15, 19, 20). Genes with expression that is regulated by tyrosol were identified by comparing gene-expression profiles of Candida cells grown to high density (10⁸ cells per ml) with those at low density (10⁵ cells per ml) with or without tyrosol, respectively. The data were analyzed to identify those genes with expression that increased or decreased in proportion to cell density and levels that were affected by tyrosol at low cell density. The known genes with transcripts that showed the greatest reduction after dilution but no reduction when tyrosol was present are those encoding proteins involved in DNA synthesis and cell-cycle regulation (Fig. 5A). This group includes the genes encoding DNA-dependent DNA polymerase, DNA-replication factors, ribonucleotide reductases, cell-cycle regulators, DNA-repair enzymes, and chromosome-segregation factors. The expression analysis also identified a key regulator of filamentation in Candida, CaEDT1 (Q.F., H.C., B. Guo, and G.R.F., unpublished data). These data were supported by RT-PCR and Northern analysis of some of these genes (Fig. 5B). Moreover, the transcript of one of the genes, CaEDT1, the level of which was restored by tyrosol, also showed a cell-density-dependent increase (Fig. 5C).

Fig. 5.

Genes regulated by tyrosol. (A) Candida genome array analysis identified a cluster of genes regulated by cell density and tyrosol (≥2.0-fold changes). The level of these transcripts are reduced after dilution but not reduced when tyrosol is present. (B) Semiquantitative RT-PCR confirmed the genome array data. Cells from a saturated Candida culture were washed and reinoculated at either high (10⁸ cells per ml, lane 1) or low (10⁵ cells per ml, lanes 2-5) density in SD (pH 4.3), and total RNA was extracted for RT-PCR analysis. Lane 1, 10⁸ cells per ml; lane 2, 10⁵ cells per ml (30 min); lane 3, 10⁵ cells per ml (60 min); lane 4, 10⁵ cells per ml plus 20 μM tyrosol (30 min); lane 5, 10⁵ cells per ml plus 20 μM tyrosol (60 min). C. albicans rRNA was used as a loading control. (C) The level of CaEDT1 mRNA increases with cell density relative to rRNA controls. Cells from an overnight culture were washed and reinoculated in SD at different cell densities for 30 min, and mRNA was extracted for Northern analysis. Lane 1, 10⁵ cells per ml; lane 2, 10⁶ cells per ml; lane 3, 10⁷ cells per ml; lane 4, 10⁸ cells per ml; lane 5, a saturated overnight culture.

Discussion

The identification of tyrosol as a density-sensing molecule has important implications for the dynamics of both growth and morphogenesis in Candida. Dilute cultures do not have sufficient external tyrosol to stimulate their growth and must synthesize it to initiate cell division. Presumably, the internal concentration must reach some threshold level that permits the cells to initiate exponential growth. These requirements for growth can be satisfied by the addition of exogenous tyrosol.

The transcriptional profiling data have identified a number of transcripts with abundance that is reduced rapidly after dilution. Among them are transcripts encoding the DNA-replication machinery and cell-cycle-control proteins. One explanation for the delay in growth is that these transcripts, which encode functions critical for cell division, are destabilized in dilute cultures. If dilution causes rapid degradation of the transcripts for these functions, then tyrosol must stabilize them. Alternatively, tyrosol may be involved in the transcriptional regulation of these genes.

The stimulation of germ-tube formation by tyrosol probably is a reflection of the abolition of the lag phase. In the absence of tyrosol, the diluted Candida cells experience a delay before they initiate the growth of germ tubes. In its presence, the cells initiate growth and therefore the formation of germ tubes immediately after dilution. Previous study has identified another quorum-sensing molecule, farnesol, that inhibits the formation of germ tubes when the cultures reach high density (11). The reciprocal control of germ-tube formation by farnesol and tyrosol must reflect a complex structure of metabolic regulation in response to environmental cues that stimulates the synthesis of these regulatory molecules.