Secreted aspartic proteases of Candida albicans, Candida tropicalis, Candida parapsilosis and Candida lusitaniae

Inhibition with peptidomimetic inhibitors

Authors


I. Pichová, Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, Flemingovo n. 2, 166 10 Prague 6, Czech Republic. Fax: + 4202 24310090, E-mail: iva.pichova@uochb.cas.cz

Abstract

The frequency of Candida infections has increased in recent years and it has been accompanied by a significant rise in morbidity and mortality. The secretion of aspartic proteases by Candida spp. was demonstrated to be one of the virulence determinants. Candida albicans is classified as the major human pathogen in the genus Candida. However, other species of this genus have been found to cause an increasing number of candidiases. We isolated secreted aspartic proteases (Saps) of C. albicans (Sap2p), C. tropicalis (Sapt1p), C. parapsilosis (Sapp1p), and C. lusitaniae (Saplp) from culture media. All the isolated proteases were N-terminally sequenced. Their specific proteolytic activities and sensitivity to series of peptidomimetic inhibitors modified in the type of scissile bond replacement as well as in the N- and C-termini were analyzed. The most divergent substrate specificity was observed for the Sap of C. tropicalis. The specificity of Sap of C. lusitaniae is most closely related to that of Sap of C. parapsilosis. We designed and prepared an inhibitor containing phenylstatine isoster that was equipotent towards all four proteases within the range of 10−10−10−9 m. The HIV-1 protease inhibitors ritonavir, saquinavir, indinavir, and nelfinavir were also tested for the inhibition of four Saps. Only ritonavir and saquinavir inhibited Sap2p, Sapt1p, Sapp1p, and Saplp in micromolar concentrations.

Abbreviations
SAP

gene encoding secreted aspartic protease

Sap

secreted aspartic protease

Sap2p

secreted aspartic protease 2 of Candida albicans

Sapt1p

secreted aspartic protease 1 of C. tropicalis

Sapp1p

secreted aspartic protease 1 of C. parapsilosis

Saplp

secreted aspartic protease of C. lusitaniae

Nph

4-nitrophenylalanine

Iva

isovaleryl

Sta

statine

Pst

phenylstatine (3S,4S)-4-amino-3-hydroxy-5-phenylpentanoic acid

PVDF

polyvinylidene fluoride.

Extracellular proteases of eukaryotic microbial pathogens have attracted the attention of many laboratories because of their potential role in pathogenesis. They facilitate the penetration into the host organism and counteract its defence system. Analysis of proteolytic enzymes of pathogenic microorganisms might lead to a design of inhibitors controlling these pathogens. A considerable amount of information is now available on the extracellular proteases of fungi, especially of the genera Aspergillus and Candida.

Candida albicans, the major human pathogen of the genus Candida, is a commensal organism in healthy individuals and seems to be almost universally present. The use of broad-spectrum antibiotics, steroids or other immunosuppressive agents, diabetes mellitus, local disorders of the gastrointestinal tract, AIDS, cancer chemotherapy or organ transplantation can enhance the risk of fatal consequences of candidal infection. Superficial infections, especially those of the mouth and vagina, are very common and often prove refractory to a treatment. However, other species of the genus Candida, such as C. tropicalis, C. parapsilosis, C. glabrata, C. krusei and C. lusitaniae, have been shown to cause an increasing incidence of mycoses. The role of secreted aspartic proteases of C. albicans in experimental and clinical candidiases has been demonstrated [1–4]. Secreted aspartic proteases (Saps) degrade a number of cellular substrates, including proteins related to immunological and structural defences, such as IgG heavy chains, α2-macroglobulin, C3 protein, β-lactoglobulin, lactoperoxidase, collagen and fibronectin [5–10].

The genome of C. albicans contains at least ten distinct genes encoding extracellular aspartic proteases and at least one gene encoding an intracellular aspartic protease [11–14]. The use of targeted gene disruption of C. albicans has shown that Saps 4–6 play an important role in the lethality of mice and guinea pigs [15]. The Saps 1–3 were shown to contribute significantly to the tissue damage in the reconstituted human epithelium used as a model of human oral candidiasis as well as during human oropharyngeal candidiasis in HIV-infected patients [3,16]. The genes encoding Sap1 and 2 (SAP1 and SAP2, respectively) are often expressed during vaginal candidiasis [2,17]. SAP4, SAP5 and SAP6 were induced in murine macrophages after phagocytosis of the Candida cells [4]. The obtained results show that individual members of the SAP gene family might be expressed at various stages of the infection process and specific Saps might have a special role in the infection.

The Saps are synthesized as preproenzymes [18]. The Sap1p−3p and Sap6p were isolated and characterized [19,20]. The most detailed information is available for Sap2p, which is a protein composed of 342 residues, with pH optimum about 3. The crystal structures of Sap2p enzymes isolated from two different strains of C. albicans were determined [21,22]. The Sap2p structure most closely resembles the structure of rhizopuspepsin and pepsin. However, several regions of structural differences were detected between these enzymes. The substrate specificities of Sap1p, Sap2p, Sap3p and Sap6p were characterized using synthetic peptide mixtures [20] and the specificity of Sap2p was compared to that of secreted aspartic proteases of C. tropicalis and C. parapsilosis[23].

C. tropicalis is a fungal species frequently detected in cancer patients and is often associated with deep candidiasis [24,25]. C. tropicalis secretes dominantly Sapt1p in vitro in a medium containing BSA as the sole source of nitrogen. Sapt1p was isolated from the culture medium and its three-dimensional structure was determined [26]. Recently, a screening of C. tropicalisλEMBL3 genomic library was performed with the whole SAPT1 gene as a probe. Three new genes SAPT2, SAPT3 and SAPT4 were detected under low-stringency hybridization conditions. Recombinant Sapt1p, Sapt2p and Sapt3p were isolated and biochemically characterized [27].

Two genes encoding putative secreted aspartic proteases have been identified in C. parapsilosis by de Viragh et al. [28]. However, only one product has been identified as an extracellular aspartic protease. This protease has been partially characterized [29,30]. Fusek et al. [31] isolated two aspartic proteases secreted by C. parapsilosis. These proteases have similar molecular masses, though they vary in their N-terminal sequences and isoelectric points. The substrate specificities detected with a synthetic peptide also varied. Sono et al. [32] isolated Sapps of molecular masses 41 and 42 kDa with identical N-terminal sequences.

C. lusitaniae is an increasingly important nosocomial pathogen infecting immunocompromised patients, newborn infants, and the elderly. The majority of isolates have been recovered from the respiratory tract, stools, urine specimens, and blood [33–39]. Weak proteolytic activities of aspartic protease secreted by two C. lusitaniae strains have been detected in culture media using the fluorogenic substrate [40].

The options for treating candidiasis are limited and the number of Candida spp. that are resistant to polyenes and azoles have increased. Resistance to commonly used drugs has been largely observed among nonalbicans Candida species, especially C. glabrata, C. parapsilosis, C. lusitaniae and C. krusei. A significant dose-dependent protection against a subsequent lethal intranasal dose of C. albicans was observed by pretreatment of neutropenic mice with pepstatin A by intraperitoneal injection. The effect of pepstatin A was comparable to protection obtained with amphotericin B [41]. A reduced depth of invasion of parenchymal organs from the peritoneal cavity and the degree of tissue damage of peritoneal cavity by C. albicans was observed in mice pretreated with pepstatin A [42]. However, pepstatin A did not provide protection against C. albicans given intravenously to mice. Inhibitor was effective when injected before or concomitantly with the infecting agent to mice [43]. Infections of murine kidney with C. albicans were also reduced after injection of pepstatin A [44].

Recently, the effect of HIV protease inhibitors on the frequency of clinically apparent oral candidiasis in HIV-1-infected patients was observed. Experimental infections, adhesion assays with epithelial cells, and in vitro inhibition of proteolytic activities of Sap2p showed that these inhibitors directly inhibit the Sap of C. albicans[45–56].

We have isolated the Saps of four Candida species, C. albicans, C. parapsilosis, C. tropicalis, and C. lusitaniae. Here we describe their inhibition by a series of pepstatin A-based inhibitors and by clinically used inhibitors of HIV protease. Information on the secreted protease of C. lusitaniae is rather scarce and to our knowledge, this is the first study addressing its substrate specificity.

Materials and methods

Yeast strains and cultivation conditions

The strains of C. albicans F7-39/IDE99 and C. lusitaniae 408/IDE98, and C. parapsilosis Z1 50-119/IDE99 were obtained from P. Hamal (Faculty of Medicine, Palacky University, Olomouc, Czech Republic). C. albicans and C. lusitaniae were isolated from the throats of infected patients. C. parapsilosis was isolated from blood. C. tropicalis strain ATCC-75 was kindly provided by M. Monod (Centre Hospitalier Universitaire Vaulois, Switzerland). Candida strains isolated from patients with suspected nosocomial mycoses were identified using selective diagnostic media and biochemical tests including assimilation and fermentation of carbon and nitrogen sources. The results were confirmed using diagnostic kit AUXACOLOR (Sanofi Diagnostics Pasteur).

The strains were maintained on Sabouraud dextrose agar. Single colonies from plates were inoculated into 1.2% yeast carbon base (HiMedia Laboratories Ltd). Medium was adjusted to pH 3 for the cultivation of C. parapsilosis and to pH 4 for the cultivation of C. albicans, C. lusitaniae and C. tropicalis. The medium was supplemented with filter-sterilized BSA (final concentration 0.2%). Cultivation was performed at 30 °C in a rotation shaker. The overnight culture was used for inoculation of larger volumes of the same medium (usually 0.5–1.5 L). The yeasts were cultivated until early stationary phase and harvested by centrifugation (10 000 g in a Beckman JA-14 rotor for 15 min). The supernatants were filtered through Durapore membrane filters (Millipore). The cell-free supernatants containing secreted proteases were used for further purification.

Purification of Saps of C. albicans and C. tropicalis

The protocol used for purification of proteases of C. albicans and C. tropicalis was similar. The cell-free supernatant was dialyzed against 15 mm sodium citrate buffer pH 5.6 and applied on the column of DEAE-Sephadex A-25 equilibrated in the same buffer. The protease was eluted with 100 mm sodium citrate buffer pH 5.6. The fractions containing an active protease were collected, analyzed, and if necessary, rechromatographed under the same conditions [57].

Purification of Saps of C. parapsilosis and C. lusitaniae

The protocol used for the purification of Sapp1p and Saplp was as follows. The cell-free supernatant was concentrated by ultrafiltration, dialyzed against 15 mm sodium citrate buffer, pH 3.5 and applied on the SE-Sephadex C-25 column equilibrated in 15 mm sodium citrate buffer pH 3.5. The protease was eluted using salt gradient (0–1 m NaCl) in 15 mm sodium citrate buffer, pH 3.5. The homogeneity of proteases was screened by SDS electrophoresis on 10% polyacrylamide gels.

Synthesis of inhibitors

Boc-protected amino acids used in this study were of commercial origin (Bachem, Buchs, Switzerland), with the exception of statine (Sta) analogs, which were prepared by stereoselective method described by Jouin and Castro [58]. All peptides were synthesized by an established solution method [59]. Crude inhibitors were purified by RP-HPLC on a Vydac 218T510 column with linear elution gradient from 0.05% trifluoroacetic acid in 50% aqueous MeOH to 0.05% trifluoroacetic acid in MeOH. Pooled fractions containing inhibitor were diluted with 10% acetic acid and lyophilized. The purity of the inhibitors, based on the isocratic RP-HPLC on a Vydac 218TP54 column, amino-acid analysis, and high-resolution fast-atom-bombardment mass spectrometry, was higher than 95%.

Proteolytic activity assays

Proteolytic activities of all the secreted aspartic proteases were determined at pH 3.3 using chromogenic pepsin substrate Lys-Pro-Ala-Glu-Phe-Nph-Ala-Leu (where Nph is 4-nitrophenylalanine) synthesized in our laboratory. The assays were performed in 100 mm sodium acetate buffer pH 3.3 (buffer A) at 37 °C. In all experiments, 1 mL of buffer A was mixed with 15 µL of substrate stock solution (5 mg·mL−1) and 1–5 µL of protease was added to the solution (i.e. a final concentration of the enzyme in the experiment was 1.5 nm). The cleavage of the substrate was followed by decrease of absorbance at 300 nm measured with an Aminco DW 2000 spectrophotometer.

The specific proteolytic activity of individual proteases was detected under the same conditions and was calculated according to an equation:

inline image

Where ΔA, is a change of the absorbance at 300 nm; d = 1 cm; ε, is a molar absorption coefficient determined at 300 nm; ce is a concentration of the protease; V is volume of the reaction solution; and v is volume of the enzyme in reaction. The concentration of Sap was determined by titration with pepstatin A.

Inhibition studies

Inhibition constants were determined by a spectrophotometric assay with the aforementioned substrate. Sap, 1.5 nmol, was added to 1 mL of buffer A containing 40 µm of the substrate and various concentrations of an inhibitor dissolved in Me2SO. The final concentration of Me2SO in each assay was below 2.5%. Ki values < 100 nm were estimated using the equation for competitive inhibition according to Williams and Morrison [60], Ki > 100 nm were assessed using the Dixon plot. All the values are the averages of at least two separate experiments that yielded the results within error limits of ± 10% for Ki values < 1 nm and ± 5% for Ki values > 1 nm.

N-terminal sequence analysis

To determine N-terminal sequences of Saps, proteins were separated by SDS electrophoresis on 10% polyacrylamide gels, transferred to the poly(vinylidene difluoride) (PVDF) membrane and stained with Coomassie Blue G-250. Automated Edman degradation of proteins was carried out in an Applied Biosystems Procise sequencer. The phenylthiohydantoins were identified by HPLC on an Ultrasphere ODS column.

Results

Cultivation, purification and characterization of Saps

The level of expression of the Sap of C. lusitaniae was significantly lower in comparison with those of C. albicans, C. tropicalis, and C. parapsilosis. The yield of the secreted protease of C. lusitaniae isolated from the culture medium represented about 10% of the production of other studied proteases. Purity and homogeneity of the enzyme preparations were controlled using SDS/PAGE and N-terminal sequencing of isolated proteins transferred to the PVDF membrane. The preparation of C. albicans protease yielded a pure protein. Its N-terminal sequence, QAVPVT-, corresponded to Sap2p [61]. Sapt active fractions revealed also a single band in SDS/PAGE with the N-terminal sequence SDVPTTI-, corresponding to the N-terminal sequence of Sapt1p [62]. Similarly, the purification of the cell-free supernatant resulting from the cultivation of C. parapsilosis yielded a homogeneous preparation with the N-terminal sequence DSISLS-. As the isoforms of aspartic proteases secreted by C. parapsilosis are known to differ in isoelectric points [31], we have performed an isoelectric focusing of purified Sapp. We have detected a single spot corresponding to the pI 5.3. Both pI value and the N-terminal sequence of our Sapp isolate corresponded to Sapp1p defined by Fusek et al. [31] as the CPAP 1 isoform. Secreted aspartic protease of C. lusitaniae has not been characterized yet. The N-terminal sequencing of the Saplp transferred to the PVDF membrane has revealed the sequence of 25 amino acid residues of the following sequence: GSYPETLKDVDDVSYVVDIYKGSDK-. A comparison of N-terminal sequence of Sap1p with the corresponding regions of Sap2p, Sapt1p, and Sapp1p is shown in Fig. 1.

Figure 1.

Comparison of 25 N-terminal residues of secreted aspartic protease of C. albicans (Sap2p), C. tropicalis (Sapt1p), C. parapsilosis (Sap1p), and C. lusitaniae (Saplp). Identical amino acids residues are shaded and similar residues are underlined.

The comparison of the specific proteolytic activities of these Saps using chromogenic peptide substrate Lys-Pro-Ala-Glu-Phe-Nph-Ala-Leu showed that the protease secreted by C. lusitaniae had the highest specific activity (i.e. 3.42 µkat·mg−1). The values detected for Sap2p, Sapt1p, and Sapp1p were 2.65 µkat·mg−1, 1.39 µkat·mg−1, and 1.33 µkat·mg−1, respectively.

Inhibition of Saps of C. albicans and C. tropicalis

The series of 13 inhibitors derived from the structure of pepstatin A was tested for the inhibition of Saps of two main pathogens of genus Candida, i.e. C. albicans and C. tropicalis.

The three-dimensional structure of Sap2p of C. albicans complexed with pepstatin A [21] shows that this general aspartic protease inhibitor binds into the active cleft in an open conformation. This extended hexapeptide contains a central Sta residue with the nonscissile bond -(CHOH-CH2)- which forms the hydrogen bonds between its hydroxyl group and two catalytic aspartates Asp32 and Asp218. The insertion of two additional carbon atoms in the backbone of the inhibitor results in the displacement of the P′1 side chain of alanine into the S′2 subsite of the active site. Thus the Sta residue and its analogs are considered to be bound in the -S1-S′1- subsites of Sap2p [21] and other aspartic proteases.

The first set of inhibitors tested with Sap2p and Sapt1p (Table 1) consists of the pepstatin A analogs containing C-terminal Sta residues terminated by a carboxyl, a carboxamide, and a methoxycarbonyl (Inh1Inh3). In Inh4Inh6 the central Sta was replaced by the more rigid phenylstatine residue (Pst). The Inh7 contains a Boc instead of an isovaleryl residue (Iva) at its N-terminus. The results show that none of the modifications introduced into the pepstatin A had significantly changed their binding to the active site of Sap2p. The decrease of the Ki values of pepstatin A analogs with terminal methoxycarbonyl group (Inh3, Inh6) by one order of magnitude was detected for Sapt1p. The crystal structure of Sapt1p revealed several secondary-structure differences compared to the structure of pepsin that could influence the substrate specificity of Sapt1p [26].

Table 1. The inhibition of Sap of C. albicans (Sap2p) and C. tropicalis (Sapt1p) by pepstatin A-derived inhibitors.
Inhibitor Ki (nm)
number Sap2p Sapt1p
1IvaValValStaAlaStaOH1.4 5.2
2IvaValValStaAlaStaNH21.4 3.0
3IvaValValStaAlaStaOMe1.4 0.4
4IvaValValPstAlaPstOH4.2 6.1
5IvaValValPstAlaPst NH28.0 7.7
6IvaValValPstAlaPstOMe5.0 0.2
7BocValValStaAlaStaOMe1.6 1.6

The inhibitors with the identical sequence -VVPstAPstOMe but different N-terminal residues were tested for the inhibition of Sap2p and Sapt1p. To produce these inhibitors, the N-terminal Iva of Inh6 was replaced by Boc (Inh8), tert-butylacetyl (Tba; Inh9) (CH3)2CHOCO residue (Inh10), acetyl (Inh11), and phenylacetyl (Inh12), respectively (Table 2). The results confirmed that the S4 subsite of Sap2p prefers bulky residues. Introduction of a small acetyl group to the N-terminus (Inh11) resulted in one order of magnitude decrease of inhibition of the Sap2p compared to the inhibition activities of corresponding compounds containing bulky N-terminal residues. Similar fitting of inhibitors into the Sap2p was detected for Inh8 and Inh10, terminated with Boc and (CH3)2CHOCO residue, respectively. A comparison of inhibition constants of these inhibitors for Sap2p and Sapt1p showed that the shapes of the S4–S5 subsites of both enzymes are different. The compounds with Iva and Boc (Inh6 and Inh8) at the N-terminus represent very potent inhibitors of Sapt1p. They inhibit the protease in subnanomolar concentrations. Interestingly, the acetyl group of the Inh11 fits into the S4 subsite of Sapt1p in a comparable manner with more bulky (CH3)2CHOCO residue of the Inh10. Sapt1p lacks a short helix B (loop 2) between Ser118 and Val119 and therefore, the binding cleft of Sapt1p is wider, particularly in the nonprimed side between subsites S3–S5 [26]. This may explain different substrate specificity of Sapt1p compared to other aspartic proteases.

Table 2. The inhibition of Sap of C. albicans (Sap2p) and C. tropicalis (Sapt1p) by pepstatin A analogs modified at the N-terminus.
Inhibitor Ki (nm)
number Sap2p Sapt1p
 8(CH3)3COCOValValPstAlaPstOMe3.9 0.5
 9(CH3)3CCH2COValValPstAlaPstOMe6.2 4.8
10(CH3)2CHOCOValValPstAlaPstOMe3.7 2.2
11CH3COValValPstAlaPstOMe14.4 2.8
12C6H5CH2COValValPstAlaPstOMe5.0 6.1
13(CH3)3OCOChaValValPstAlaPstOMe1.6 14.1

The crystal structure of a pepstatin A–Sap2p complex [21] and a hexapeptide with hydroxyethylene isostere (A-70450)–Sap2p complex [21,22] suggest that an extension of the inhibitor at the P4 could result in a better fit to the hole around the Ser118 of Sap2p. To confirm a requirement for an extension of the inhibitor, we have prepared a compound N-terminally extended by a cyclohexylalanyl (Inh13). Its Ki values measured for Sap2p and Sapt1p were 1.6 and 14.1, respectively. Thus, such extension resulted in better fitting of the Inh13 in the Sap2p in comparison to shorter inhibitors listed in Table 2. Interestingly, the Inh13 displayed the lowest inhibition activity for Sapt1p from all 13 compounds of this series.

Comparison of inhibition of Saps of C. albicans, C. tropicalis, C. parapsilosis and C. lusitaniae

Most of the efforts to characterize Saps of different Candida spp. have been focused on proteases of C. albicans, C. tropicalis and C. parapsilosis. Information concerning the production of the Sap of C. lusitaniae is less frequent [40] and data concerning characterization of substrate specificity are missing in the literature. To compare the specificity of Saplp with that of Sap2p, Sapt1p and Sapp1p, we have performed the inhibition study using rationally designed analogs of pepstatin A. The effects of ritonavir, nelfinavir, indinavir, and saquinavir, clinically used inhibitors of aspartic HIV-1 protease, were also studied for these proteases.

We have synthesized a series of pepstatin A analogs (Table 3) varying in the type of scissile bond replacement, in the C-terminal modifications as well as in the length of the nonprimed sites of compounds. Pepstatin A (Inh1) inhibited the Saplp with the Ki value 0.5 nm. Only the Sap of C. parapsilosis was inhibited with pepstatin A more effectively with the Ki value 0.3 nm. The replacement of isovaleryl group with Boc at the N-terminus of pepstatin A (Inh14) resulted in the increase of inhibition activities for Sap2p and Sapt1p. The Ki values of Inh14 for Sapp1p and Saplp were not changed. The replacement of both, central and C-terminal Sta residues with phenylstatine (Inh15) caused the decrease of inhibition activities for Sap2p, Sap1p, and Saplp. The modifications of C-terminal carboxyl by carboxamide (Inh16) and methoxycarbonyl (Inh8) improved the effectiveness of the inhibition of Sapt1p, Sapp1p, and Saplp. The configuration of both phenylstatine residues in these inhibitors is 3S, 4S. A dramatic decrease of the inhibition activities for all the proteases was achieved with the replacement of central 3S, 4S Pst by the 3R, 4R Pst (Inh17). The inhibition activity of the Inh17 dropped by 3–4 orders of magnitude for all proteases compared to the Inh8 containing both Pst with 3S, 4S configuration. Interestingly, the change of the 3S, 4S configuration of terminal Pst into the 3R, 4R configuration made the Inh18 the best compound from our series with comparable effectiveness for all four proteases tested here. This is very unique situation considering differences in the substrate specificities of Sap2p, Sapt1p, Sapp1p, and Saplp as documented by both our results and the results published by Fusek et al. [23].

Table 3. The inhibition of Sap of C. albicans (Sap2p), C. tropicalis (Sapt1p), C. parapsilosis (Sapp1p), and C. lusitaniae (Saplp) by pepstatin A-based inhibitors.
  Ki (nm)
Inhibitor number Sap2pSapt1pSapp1pSaplp
 1IvaValValStaAlaStaOH1.45.10.30.5
14BocValValStaAlaStaOH0.61.60.30.5
15BocValValPstAlaPstOH2.91.86.70.3
16BocValValPstAlaPstNH21.11.00.10.3
 8BocValValPstAlaPstOMe3.90.50.30.4
17BocValVal(3R,4R) PstAlaPstOMe61025617447
18BocValValPstAla(3R,4R)PstOMe1.30.250.40.1
19BocValValPstAlaOMe8347914.61.2
20BocValPstAlaPstOMe591804411.7
21BocValVal(3R,4R)PstAlaOMe> 100000> 100000> 100000> 100000
22Boc(CH2)10ChstAlaPstOMe8.49341.11.4

To study the effect of the chain length of inhibitors on their activity, we have prepared truncated analogs. The Inh19 contained a deletion of terminal phenylstatine residue in the P′3 position and Inh20 lacked valine in the P3 position. The removal of these residues caused a comparable decrease of Ki values by two to three orders of magnitude for Sap2p, Sapt1p, and Sapp1p. The inhibition of Saplp was not significantly altered by Pst deletion in the P′3 position of the Inh19. No inhibition of any of the tested proteases was detected for Inh21 having the similar structure to the Inh19 except the 3R, 4R configuration of the central phenylstatine residue. This result confirmed that the S configuration of hydroxyl group of central Pst is crucial for the formation of the hydrogen bond with catalytically active aspartates in all proteases tested here.

The last Inh22 in Table 3 represented a different structure compared to rationally designed inhibitors of series discussed here. It contained a cyclohexylstatine in the P1-P′1 positions and the chain of 10 CH2 groups in the nonprimed site. This inhibitor was efficiently bound into the active site of Sapp1p with the Ki value 1.1 nm, in the Saplp with the Ki value 1.4 nm, and in the Sap2p with Ki value 8.4. However, for Sapt1p, it represented the least efficient inhibitor with the Ki value 934 nm. We prepared an analogous inhibitor where ten CH2 residues were replaced by two valines. The Ki values of this compound with the structure BocVVChstAlaPstOMe were 2.9 and 12.4 nm for Sap2p and Sapt1p, respectively. This result is in a good agreement with the Sapt1p structure showing a wide nonprimed site that can influence the substrate specificity in a manner different from other Saps [26].

In a separate set of experiments we have compared the inhibition of Saps of C. albicans, C. tropicalis, C. parapsilosis, C. lusitaniae with HIV-1 protease inhibitors ritonavir, saquinavir, nelfinavir, indinavir. Table 4 summarizes the Ki values obtained for these inhibitors. The best inhibitor from this set was ritonavir that inhibited all four proteases in micromolar concentrations. The Ki values of saquinavir for Sap2p, Sapp1p, and Saplp were one order of magnitude higher. Sapt1p was inhibited with similar potency by both saquinavir and ritonavir. Indinavir inhibited only the Sapt1p and nelfinavir did not inhibit any of the proteases.

Table 4. The inhibition of Sap of C. albicans (Sap2p), C. tropicalis (Sapt1p), C. parapsilosis (Sapp1p), and C. lusitaniae (Saplp) by HIV-1 protease inhibitors.
 Kim)  
 Sap2pSapt1pSapp1pSaplp 
Ritonavir0.31.711.9 0.8 
Saquinavir6.83.716.915.9 
Indinavir> 10001.0> 1000> 1000 
Nelfinavir> 1000> 1000> 1000> 1000 

Discussion

The goal of this work was to compare the sensitivity of Saps of different pathogenic Candida spp. that cause an increasing incidence of mycoses to rationally designed inhibitors. These proteases represent an important virulence factor and they are becoming a significant therapeutic target for combating the candidal infections. The amount of information available on inhibitors of Saps is limited. Novel pepstatin-like inhibitors isolated from Streptomyces sp. [63,64] were not as efficient as pepstatin A. Secretion of Sap of C. albicans was efficiently inhibited by Lys-Nva-FMDP [65]. A series of hexapeptides with hydroxyethylene isostere (A-70450 analogs) was tested for the inhibition of Sap2p [22]. The results of the inhibition study as well as the results obtained from the analysis of the crystal structure of inhibitor A-70450 with Sap2p have revealed that the S3 large subsite of the Sap is critical for designing inhibitors of the Sap of C. albicans. Most information is today available on the inhibition of Saps by pepstatin A, which was used in several in vivo and in vitro inhibition studies of Saps of Candida albicans. Pepstatin A exerted a strong curative effect on experimental vaginitis in rat infected by the high-level Sap strain of C. albicans isolated from HIV-positive women [2]. The pepstatin A inhibition of high vaginopathic potential of vaginal isolates from HIV-positive women producing high levels of Sap was also detected [66]. The addition of the pepstatin A reduced the number of Candida cells adhering to epithelia in experimental oral infection [67] and the extent of lesions of reconstituted human epithelium caused by C. albicans[16]. Pepstatin A also exhibited a protective role during infection of the epidermis by C. albicans in an in vitro model of cutaneous candidiasis [1] and inhibited cavitation on the surface of skin caused by a clinical strain of C. albicans[68]. A pretreatment of mice with pepstatin A reduced the depth of invasion of parenchymal organs by C. albicans from the peritoneal cavity and the degree of tissue damage of the cavity [42]. However, an incomplete reduction of the morphologic alterations in the presence of pepstatin A was demonstrated in some experimental models of candidiasis [1,69,70]. Most probably, pepstatin A inhibition of Saps is insufficient and other virulence factors could contribute to tissue damage during Candida infection. Therefore further improvements of the inhibitor structure might contribute to better inhibition of Saps.

Here we present a comparison of activity of a series of peptidomimetic inhibitors, derived from the structure of pepstatin A, towards Saps of four Candida spp. The Ki values of pepstatin A for Sap2p, Sapt1p, Sapp1p, and Saplp were 1.4 nm, 5.1 nm, 0.3 nm, and 0.5 nm, respectively. The best inhibitor of Sap2p with the structure BocValValStaAlaStaOH (Inh14) inhibited the enzyme with the Ki value 0.6 nm. Other modifications have not improved its activity for Sap2p. Three inhibitors with the structure IvaValValPstAlaPstOMe (Inh6), BocValValPstAlaPstOMe (Inh8), and BocValValPstAla(3R,4R)PstOMe (Inh18) had subnanomolar Ki values for Sapt1p. The inhibitors BocValValPstAlaPstOMe (Inh8), BocValValStaAlaStaOH (Inh14), BocValValPstAlaPstNH2 (Inh16), and BocValValPstAla(3R,4R)PstOMe (Inh18) inhibited Sapp1p with subnanomolar Ki values. The Saplp was inhibited with the same analogs of pepstatin A with the subnanomolar Ki values and inhibitor BocValValPstAlaPstOH (Inh15) inhibited in the same concentration only protease secreted by C. lusitaniae. These results show that aspartic proteases secreted by different Candida spp. display diverse substrate specificities. However, several common subsite requirements of the enzymes from different Candida species have been found. The results have confirmed that S configuration of the hydroxyl group of the inhibitor isosteric group, which is bound into the S1-S′1 enzyme subsites and interacts with the catalytically active aspartates, is crucial for the inhibition. The P3 and P′3 residues of the inhibitors play also significant role in the inhibition of Candida Saps. These residues stabilize the proper binding of the inhibitor and influence the interactions of other side chains of inhibitor residues with the protease active subsites. The deletion of valine from the P3 position of the Inh20 had the comparable effect on the efficiency of the inhibition of Saps of C. albicans, C. tropicalis, and C. parapsilosis as the deletion of phenylstatine from the P′3 position of the Inh19. Although the specificities of individual proteases are different, we have succeeded in designing the inhibitor with the structure BocVVPstA(3R4R)PstOMe (Inh18) efficient in nanomolar and subnanomolar concentrations towards all the proteases tested here. Interestingly, this inhibitor contains the C-terminal phenylstatine residue in the 3R,4R configuration. To test the inhibitory effect of the pepstatin A and its analogs on Candida spp., the yeasts were cultivated in the presence of different concentrations of selected inhibitors in medium with BSA as the sole nitrogen source. The results confirmed that the production of Saps was inhibited and the growth of Candida spp. was significantly decelerated in the presence of 1 µm pepstatin A (the data are not shown). Moreover, results from testing the effect of the inhibitors tested in this study with four Saps on Candida growth on a solid medium corresponded to the in vitro inhibitory effects of isolated proteases (unpublished data).

Selectivity represents a problem of inhibitors structurally analogous to pepstatin A. It is known that pepstatin A and its analogs efficiently inhibit human pepsin and cathepsin D. We have determined Ki values of selected pepstatin analogs synthesized in this study for human pepsin and cathepsin D. All compounds inhibiting Saps in subnanomolar or nanomolar concentrations inhibited very efficiently also both aforementioned human enzymes (data are not shown).

Comparison of sensitivity of Sap2p, Sapt1p, Sapp1p, and Saplp to tested inhibitors leads to a conclusion that Sap of C. tropicalis is divergent from the rest of tested proteases. The substrate specificity of Sap of C. lusitaniae is most closely related to that of Sap of C. parapsilosis. However, more detailed analysis of the substrate specificity and the crystal structure of Saplp are necessary to confirm this suggestion.

Saps of Candida spp. display also different specific proteolytic activities. The highest specific proteolytic activity was detected for the Sap of C. lusitaniae. This result, together with its low cultivation yield, allows us to speculate that the lower level of secretion of Saplp can be compensated with its higher proteolytic activity necessary to hydrolyze particular amount of the substrate.

Recently, the effect of HIV protease inhibitors ritonavir, saquinavir, indinavir, and nelfinavir on Sap of C. albicans was demonstrated [52]. Ritonavir and saquinavir were shown to inhibit the Candida adherence to epithelial cells [71]. Candida viability assays (colony formation, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide assay, propidium iodide staining) confirmed that C. albicans is inhibited with retroviral inhibitors in the order ritonavir > indinavir > saquinavir in concentrations higher than 0.1 mg·mL−1[53,54]. Our results are in good agreement with these data and show also that the proteases secreted by C. tropicalis, C. parapsilosis, and C. lusitaniae are inhibited by some HIV-1 protease inhibitors. The most effective inhibition of Sap2p was observed with ritonavir (Ki = 0.34 µm). The Ki value of ritonavir for Sapt1p, Sapp1p, and Saplp are one order of magnitude higher. Saquinavir inhibited all four proteases with Ki values in micromolar concentrations. Interestingly, indinavir inhibited only the Sapt1p with the Ki value 1 µm. In our experiments, nelfinavir did not inhibit any of the Saps tested here. This result corresponds to the testing of indinavir for an antifungal effect on C. albicans in vitro in a broth dilution test [50].

We have used pepstatin A analogs for the mapping of substrate specificities of Saps of four pathogenic Candida spp. We have succeeded in designing and synthesizing subnanomolar inhibitors that inhibited all Saps tested here better than pepstatin A. The protective role of pepstatin A in experimental candidiasis and in vivo reduction of infection in mice shows that derivatives or analogs of pepstatin A might be useful for the treatment of vaginitis, and oral or skin infection with Candida spp. Testing of our subnanomolar inhibitors with all isoenzymes secreted by C. albicans and C. tropicalis might confirm their possible use during infections.

Acknowledgements

This work was supported by the Grant Agency of the Czech Republic under Contract numbers: 303/98/1612, 303/98/P252, 303/01/0831 and research project Z 4055905.

Footnotes

  1. Enzymes: secreted aspartic protease of Candida albicans or candidapepsin (EC 3.4.23.24); secreted aspartic protease of C. tropicalis or canditropsin (EC 3.4.23.24); secreted aspartic protease of C. parapsilosis or candiparapsin (EC 3.4.23.24); secreted aspartic protease C. lusitaniae (EC 3.4.23.24).

  2. Note: the sequences reported in this paper have been deposited in the SwissProt databank under the accession numbers P28871(candidapepsin), Q00663(canditropsin), and P32951(candiparapsin).

Ancillary