Candida albicans strains with a deletion of the mitogen-activated protein kinase CEK1 gene are defective in the yeast to hyphal transition on solid surfaces in vitro. The virulence of a cek1Δ/cek1Δ null mutant strain was compared with its wild-type parent strain (WT) in a novel model of localized candidiasis. The mammary glands of lactating mice (at day 5 postpartum) were infected for 2, 4 and 6 days with 50 μl suspension containing 1×105, 1×106 and 1×107 blastospores before death. Infected and non-infected control glands were evaluated pathologically. All animals infected with cek1Δ/cek1Δ null mutant strains showed no lesions while 65% of animals infected with the WT strain had severe lesions characterized by widespread heterophilic infiltration, necrosis, and abscess formation. As an additional control, animals infected with the disrupted strain complemented with the WT CEK1, on a replicating plasmid, also showed severe pathological changes similar to the WT strain. These results clearly demonstrate that the CEK1 gene codes for a virulence determinant of C. albicans and that the mouse mastitis model is well suited for the discriminative study of the pathogenicity of different C. albicans strains.
Candida albicans is an opportunistic infectious agent associated with morbidity and mortality in immunocompromised patients, such as those suffering from AIDS and those treated with prolonged chemotherapy and radiation. Recently, some advances have been made in defining new virulence determinants through the isolation of genes and the analysis of the phenotypes and pathogenicity of genetically engineered null mutants of these genes.
Strains containing disruptions of C. albicans genes involved in chitin synthesis (CH3)  multidrug resistance (CaMDRI)  and apical cell growth and morphogenesis (PHR1)  are less virulent than their isogenic parental strains, revealing the importance of the processes controlled by these genes in pathogenicity. In addition, auxotrophic respiratory mutants also show reduced virulence . Mitogen-activated protein (MAP) kinase signal transduction pathways transmit extracellular signals from the cell surface to the nucleus for expression of a variety of eukaryotic responses. Recently, C. albicans genes encoding components of MAP kinase cascades, the MAP kinase MKC1 gene, the MAP kinase kinase kinase kinase CST20 gene, and a related gene, CaCLA4, have been found to be involved in virulence in models of murine systemic candidiasis [6, 11, 12]. The MAP kinase gene MKC1, a homolog of the baker's yeast Saccharomyces cerevisiae SLT2 (MPK1) gene, is required for cell integrity and its product is thought to act as part of the PKC1-mediated MAP kinase pathway , whereas the CST20 gene is required for carbon source-dependent yeast to hyphal switching on agar surfaces. The other components of the Cst20p (Cst20 protein)-driven kinase cascade in C. albicans include the Hst7p MAP kinase kinase homolog [4, 10, 11], the Cek1p MAP kinase homolog ) and the Cph1p transcription factor , all of which are structural or functional homologs of components of the S. cerevisiae pheromone response MAP kinase cascade. A double disruption of the CPH1 gene and a gene encoding another transcription factor, EFG1, together, blocks both hyphal differentiation and virulence . However, unlike the cst20 null mutant , null mutants of the HST or CPH1 gene are still virulent [11, 14]. We therefore decided to explore whether the Cek1p MAP kinase was involved in virulence using a localized infection model.
We have recently developed a murine model of localized candidiasis  in which intact immunocompetent lactating mice were infected with C. krusei by inoculation through the lactiferous duct. No sign of systemic spread was observed in all of the infected animals when a postmortem examination was performed. Therefore, this approach constituted a relevant model of naturally occurring fungal infection of a single organ . In the present study, we used the localized murine mastitis model to investigate the virulence of a C. albicans strain with a deletion of the CEK1 MAP kinase gene and the parental pathogenic clinical isolate SC5314 [6, 8].
2Materials and methods
Male and female BALB/c mice, 6 weeks old (Bomholtgaard, Ry, Denmark) were allowed to acclimatize for 2 weeks before mating. The mice were housed in groups of four females and one male in size III Macrolone cages (Scanbur, K?ge, Denmark) and were transferred to size II single cages in a laminar flow cabinet (Scanbur) after confirmation of pregnancy (vaginal plug). The cabinets had negative pressure and the temperature was maintained at 20±1°C. The mice were given pellet diet (R36, Lactamin, Stockholm, Sweden) and tap water ad libitum. Light/dark ratio and humidity were maintained at 12:12 h and at 50–60%, respectively.
2.2Candida strains and gene disruption procedures
The C. albicans strains used were the parental strain SC5314 , hereafter referred to as wild-type strain (WT), and strain 43b-16 (cek1Δ::hisG-URA3-hisG/cek1Δ::hisG, ura3Δ::imm434/ ura3Δ::imm434) containing a deletion of both alleles of the CEK1 gene, hereafter referred to as null-mutant strain (NM), or as the cek1Δ/cek1Δ null mutant. Disruption of both alleles of the CEK1 gene was performed using a sequential disruption technique . As an additional control, we infected animals with the cek1Δ/cek1Δ null mutant that has been transformed with the wild-type CEK1 gene on a replicating plasmid (pVECCEK1 or pCCa4, ), hereafter referred to as reconstituted strain (RS).
2.3Preparation of C. albicans inocula
The strains were grown on complete culture media, yeast extract/peptone/dextrose (YPD) and were incubated at 30°C for 18 h. The cultures were harvested and stock suspensions made in sterile water (Pharmacia, Uppsala, Sweden). Decimal dilutions of the suspensions were made by adding 0.1 ml of the stock suspension and 0.9 ml of sterile water. Fifty microliter suspensions from three different dilutions were incubated again on YPD agar plates for 24 h at 30°C in order to establish the number of colony forming units (cfu/ml, hereafter called cfu) of the stock suspension. The number of colonies were counted from the three dilutions and the average was calculated. By adding sterile water, standard suspensions of 20×105, 106, and 107) cfu were made from the stock suspension and stored at 4°C for up to 24 h until inoculation.
The lactating females were housed individually in Macrolone size II cages. At day 5 postpartum the pups were removed and the mother was inoculated with 50 μl suspension containing between 1×105 and 1×107 blastospores during intraperitoneal anesthesia with 0.3 ml propanidid (Sombrevin; Lenau, Copenhagen, Denmark) as described previously . Fifteen and 17 lactating female mice were used for each of the two C. albicans strains WT and NM, respectively. The infected animals were killed at 2, 4 or 6 days after inoculation. As a control for successful genetic manipulation, six lactating females were inoculated with the RS at a dose of 1×107 blastospores and killed after 4 days.
2.5Blood and tissue sampling
The details of these procedures were as described previously . Briefly, blood sampling was done before and after infection, stored at 4°C for 24 h and centrifuged. Serum was stored at −20°C until analysis. Animals were killed by injection with 10% pentobarbitone sodium at a dose of 200 mg/kg b.wt. (Apoteksbolaget, Malmö, Sweden). At death, the left (L) and right (R) mammary glands were examined macroscopically. The four most posterior glands (L4 and L5 as well as R4 and R5) were removed and fixed in 10% neutral buffered formalin together with the brain, liver, spleen, kidneys, uterus, heart, lungs, gastrointestinal tract and pancreas. Tissues were embedded in paraffin wax, sectioned at 5 μm and stained with hematoxylin and eosin (HE), periodic acid-Schiff (PAS), Van Gieson stain and by Grocott's methenamine-silver (GMS) method for histology.
3Results and discussion
No gross or microscopic lesions were observed in extramammary organs. All the 15 animals infected with the NM strain showed virtually no lesions in the mammary gland. In contrast, 11 out 17 animals (65%) infected with the WT strain showed severe lesions characterized by widespread heterophilic infiltration, presence of blastospores and hyphae within mammary gland alveoli, and necrosis and abscess formation. Fig. 1 shows these lesions in a mouse infected with 105 cells 2 days after inoculation. All the animals infected with RS had similar lesions as those characterized by animals infected with the WT strain. Fig. 2 demonstrates these lesions in a mouse infected with 107 cells 4 days after inoculation.
The present study focused on the effect of CEK1 gene deletion on the virulence of C. albicans. To confirm that the reduced virulence was due to the genetic deletion, animals were infected with the cek1 null mutant strain complemented with wild-type CEK1 gene (pVECCEK1 or pCCa4 (RS)). It was observed that the C. albicans strain containing a deletion of the CEK1 gene (cek1Δ/cek1Δ (NM)) resulted in virtually no lesions while glands infected with the wild-type parental strain (SC5314 (WT)) and RS strains showed signs of host damage and contained abscesses with blastospores and hyphae. The cek1Δ/cek1Δ strain was completely unable to invade or colonize mammary glands even when very high doses of cek1Δ/cek1Δ cells were used as inocula. Hence, the C. albicans CEK1 gene, and thus its MAP kinase product, Cek1p, is required for virulence in this model of localized candidiasis.
Why are cek1Δ/cek1Δ null mutants unable to infect mammary glands? Although both C. albicans virulence factors and host responses help in determining the outcome of infection, a look at the phenotype of cek1Δ/cek1Δ null mutants in vitro, reveals a possible mechanism. MAP kinases, such as Cek1p, transmit extracellular signals to the transcriptional machinery in order to coordinate the expression of appropriate genes. In C. albicans, several components of a MAP kinase cascade play a role in invasive hyphal growth on agar surfaces, probably in response to nutritional cues. Cek1p is involved in this process. In vitro, cek1Δ/cek1Δ null mutants, like mutants in the CST20[10, 11], HST7[10, 11], and CPH1 genes, are defective in the yeast to hyphal switch on agar-surfaces in response to some nutritional starvation and as a result are less able to penetrate the agar [5, 10, 11, 13]. In addition, cek1 null mutants have a defect in mycelial growth on serum in vitro. Mastitic glands may provide an environment where Cek1p-activating MAP kinase cascade is required for tissue penetration. However, one cannot exclude the possibility that cek1Δ/cek1Δ null mutants are more susceptible to host defense mechanisms or are less virulent for other reasons. Taken together, these studies demonstrate the importance of enzymatic MAP kinase signal transduction cascades in virulence.
In conclusion, we have established a role of the C. albicans MAP kinase Cek1p in virulence using a unique localized murine mastitis model. This localized mastitis model has several advantages over the traditional systemic candidiasis models, particularly with regard to animal welfare, an increasingly important issue when using animal models of infection. The advantages of this model over previous models of infection are that: (1) the route of infection is relevant to spontaneous fungal infections; (2) the dose can be kept low; (3) the mouse remains healthy and immunocompetent; and, most importantly, (4) the infection remains localized throughout the duration of the experiment. In addition, the model will be particularly useful for in vivo virulence studies as well as for examining the efficacy of promising chemotherapeutic substances.
This work was supported by a grant from the Swedish Medical Research Council (K97-99DU-12321-01A). We acknowledge support of the Medical Research Council of Canada for a postdoctoral fellowship (C.C.).