Abstract

The aim of this retrospective study was to analyze the relationship between occurrence of the serotypes of the Cryptococcus neoformans species complex in tree samples and the climatic conditions registered during samplings in four cities of Colombia, between 1992 and 2004, by means of a logistic regression model and lagged Pearson correlations. During 97 collection dates, 8220 samples from different tree species were taken, of which 2.63% were positive: 56.5% yielded serotype B, 24.7% serotype C and 18.8% serotype A isolates. The prevalence of the serotypes varied among the cities. The results suggest that environmental climatic conditions, mainly humidity, temperature, evaporation and solar radiation, can affect the occurrence of the different serotypes in trees in a differential manner. These different climatic tolerances were reflected in the geographic distribution of the serotypes in Colombia. The climatic conditions for 15 days before the sampling date were correlated with positive or negative isolation of the different serotypes.

Introduction

The Cryptococcus neoformans (C.n.) species complex comprises basidiomycetous yeasts causing life-threatening infection of the lungs and central nervous system in humans and animals (Casadevall & Perfect, 1998). Currently, the strains comprising the C.n. species complex may be considered as belonging to five serotypes: serotype A (C.n. var. grubii), serotypes B and C (C. gattii), serotype D (C.n. var. neoformans) and the hybrid AD (Franzot, 1999; Boekhout, 2001; Kwon-Chung, 2002). Cryptococcus n. var. grubii and C.n. var. neoformans have been isolated worldwide, affecting mainly immunocompromised hosts (Casadevall & Perfect, 1998). Cryptococcus gattii, causing disease mostly in immunocompetent hosts, was originally restricted to tropical and subtropical areas, until a recent outbreak on Vancouver Island, Canada, changed its distribution pattern (Stephen, 2002; Hoang, 2004; Kidd, 2004).

It has been suggested that some abiotic factors, such as pH, humidity, temperature, sunlight and wind, could be important in the distribution and ecology of C.n. and C. gattii in the environment (Ishaq, 1968; Walter & Yee, 1968; Hubalek, 1975; Hubalek & Prikazsky, 1975; Ruiz, 1981; Wang & Casadevall, 1994; Casadevall & Perfect, 1998; Caicedo, 1999; Montenegro & Paula, 2000; Kuroki, 2004; Granados & Castañeda, 2005). Most of these studies have been done for C.n. var. grubii present in avian excreta, where occurrence of this serotype appears to be favored by dry conditions (Ruiz, 1981; Caicedo, 1999; Montenegro & Paula, 2000; Kuroki, 2004; Granados & Castañeda, 2005). In vitro studies have shown differences in thermotolerance between C. n. serotypes A and D (Martinez, 2001). However, studies designed to explore the environmental requirements of the different serotypes of the C.n. species complex in the environment are scarce.

A previous study remarked an apparent seasonal variation in nasal colonization of koalas by C. gattii serotype B in Australia (Krockenberger, 2002). A similar result was found in India, where population density of C.n. var. grubii and C. gattii serotype B declined during the extreme hot summer compared to levels seen in mild spring (Randhawa, 2003). Based on these observations we carried out a study with a constant sampling design that showed an interesting seasonal pattern of occurrence of C. gattii serotype B in trees during a 5-month period in Bogotá city (Granados & Castañeda, 2005). Based on our results, we suggested that rainy months, characterized by high precipitation and humidity, few hours of sunlight, less extreme temperatures and slightly higher values of temperature, favored the occurrence of C. gattii serotype B associated with trees, more than did dry months (Granados & Castañeda, 2005). However, we confirmed in the same study that C.n. var. grubii was more frequently isolated from dry than from fresh pigeon droppings, as reported before (Ruiz, 1981; Caicedo, 1999; Montenegro & Paula, 2000; Kuroki, 2004). These apparently contradictory results led us to hypothesize that different serotypes could have different environmental requirements, which could explain the differences in geographical distribution and ecological niches found between these serotypes.

Our group has succeeded in isolating C.n. var. grubii and C. gattii serotypes B and C strains from different environmental sources in cities with contrasting climates resulting from different altitudes (Duarte, 1994; Callejas, 1998; Escandón, 2005; Granados & Castañeda, 2005; Quintero, 2005). Thus, we took advantage of the climatic diversity and the bimodal thermal regime of Colombia and of the fact that three serotypes can be found in our environment, to analyze the relationship between the frequencies of isolation of C. neoformans and C. gattii serotypes from tree samples and some climatic conditions in four cities of Colombia, between 1992 and 2004 using logistic regression and lagged Pearson correlation models.

Materials and methods

Database creation

In this retrospective study we used the information registered by our laboratory about samples taken from trees between 1992 and 2004 in different cities of Colombia, to create as complete as possible a database for the analysis. Data from pigeon dropping samples were not included because most of the records were incomplete. The records used included the date and city of sample collection, the number and species of trees sampled, the number of positive trees, the number and type of samples taken, the number of positive samples, and the serotype yielded on each occasion, if any. Because some of the data were incomplete, the database was reduced to records of only four cities (Bogotá, Medellín, Cali and Cúcuta) that fulfilled the following information: collection date (day, month and year), number of samples taken, number of positive samples and serotype of the isolates. It is important to note that both positive and negative results were included. Only information on samples taken using the conventional technique was included (Staib, 1985).

Region of study

Colombia (4° N, 72° W) is a tropical mountainous country with a great variety of climates, as a result of the split of the Andes Mountain into three mountain chains present along the country (Eslava, 1993). The climate of Colombia is also strongly influenced by the effect of the Alisious winds coming from the northeast and southeast zones (Poveda, 2004). Both wind systems meet in a region denominated the Intertropical Confluence Zone (ITCZ), which moves in a north-south direction over the year and where Colombia is located. As a result, when the central region is under the influence of the ITCZ, its climate is dry and warm, and when the ITCZ moves north, the climate becomes rainy and slightly less warm. This phenomenon explains the bimodal thermal regime of Colombia, consisting of two rainy seasons (April-June and September-November) and two dry seasons (July–August and December–March) (Poveda, 2004).

The four cities studied are between 441 and 1024 km apart (Fig. 1) and are situated at different altitudes (Table 1). These altitudinal differences help explain the climatic differences between the cities. Bogotá is characterized by a temperate climate; Medellín is a city with a mild-warm climate, whereas Cali and Cúcuta are characterized by warm and hot climates, respectively. Some climatic parameters of these cities are shown in Table 1.

1

Location of Colombia in South America and of the four cities under study. Cúcuta 7° 56′ N, 72° 31′ W; Medellín 6° 13′ N, 75° 35′ W; Bogotá 4°43′N, 74°09′W; Cali 3° 24′ N, 76° 32′ W.

1

Location of Colombia in South America and of the four cities under study. Cúcuta 7° 56′ N, 72° 31′ W; Medellín 6° 13′ N, 75° 35′ W; Bogotá 4°43′N, 74°09′W; Cali 3° 24′ N, 76° 32′ W.

1

Historic values of some climatic parameters, altitude and climate of the Colombian cities

Climatic parameter City 
Bogotá Medellín Cali Cúcuta 
Altitude (m above sea level) 2,630 1,490 970 250 
Mean temperature (°C) 13.4 21.8 23.7 27.0 
Maximum temperature (°C) 19.2 27.6 31.9 31.8 
Minimum temperature (°C) 7.4 16.8 16.5 22.5 
Total precipitation (mm) 793.7 1,672.3 908.5 863.3 
Relative humidity (%) 78 68.2 72.8 70.8 
Total evaporation (mm) 88.9 124.0 142.7 184.0 
Total solar radiation (hours) 135.4 157.7 161.8 184.4 
Climate Temperate Mild-warm Warm Hot 
Climatic parameter City 
Bogotá Medellín Cali Cúcuta 
Altitude (m above sea level) 2,630 1,490 970 250 
Mean temperature (°C) 13.4 21.8 23.7 27.0 
Maximum temperature (°C) 19.2 27.6 31.9 31.8 
Minimum temperature (°C) 7.4 16.8 16.5 22.5 
Total precipitation (mm) 793.7 1,672.3 908.5 863.3 
Relative humidity (%) 78 68.2 72.8 70.8 
Total evaporation (mm) 88.9 124.0 142.7 184.0 
Total solar radiation (hours) 135.4 157.7 161.8 184.4 
Climate Temperate Mild-warm Warm Hot 

Climatic data

Climatic data, provided by the National Hydrology, Meteorology and Environmental Studies Institute of Colombia (IDEAM), were integrated into the analysis. Monthly and daily climatic data of these four cities during the study period were evaluated for the following variables: total precipitation (mm), mean relative humidity (%), mean, maximum and minimum temperatures (°C), total evaporation (mm) and total solar radiation (hours).

Statistical analysis

Both samples and climatic data were integrated in a complete database, in which the climatic conditions on collection date and for the 30 days before sampling date were recorded, as well as the proportion of positive samples obtained (positive/total) and the serotype found on each occasion. By integrating all this information, we aimed to explore whether there was a correlation between the climatic conditions on a specific date and serotypes found, if any. Also, we intended to determine whether the climatic conditions on the days prior to the sampling date were determinant factors for finding a certain serotype in trees on a specific day.

Firstly, a logit transformation of the response variable (proportion of positive samples on each sampling date) was performed, to explore small differences in the data and thus to render the analysis more sensitive. Subsequently, a lagged Pearson correlation between response and climatic variables was performed, for each sampling date and for a lag of 1–30 days before the sampling date. Finally, a logistic regression model was applied, with the aim of identifying those climatic variables that were important in the probability of occurrence of each serotype. All statistical analyses were carried out using SAS version 8.02 (SAS Institute Inc., Cary, NC).

Results

Prevalence of the serotypes of the C.n. species complex

The database created included the results of 97 different samplings carried out between 1992 and 2004, in four climatologically different cities of Colombia (Table 2). During this period, a total of 8220 samples were taken from 5985 trees belonging to at least 35 different species, including native and introduced species (Table 3). Most of the samples were taken from eucalyptus trees (Eucalyptus spp.), followed by bragance trees (Licania tomentosa) and tropical almond trees (Terminalia cattapa). Eucalyptus spp. was the most frequently sampled species in Bogotá, Medellín and Cali, whereas in Cúcuta the majority of the samples were taken from L. tomentosa, followed by T. cattapa and Eucalyptus spp. A complete list of the tree species analyzed in each city is shown in Table 3. The samples were mostly taken from bark, soil and detritus of the trees, although some flower and fruit samples were also taken. Cryptococcus was isolated from 216 (2.6%) samples, even though the frequency of positive samples ranged between 0.33% and 4.18% depending on the city (Table 2).

2

Number of samplings carried out between 1992 and 2004 and prevalence of the Cryptococcus neoformans species complex in the four Colombian cities

City Samplings Samples 
Total Positive 
Bogotá 44 2778 116 4.18 
Cúcuta 45 2772 83 2.99 
Medellín 2067 15 0.73 
Cali 603 0.33 
Total 97 8220 216 2.63 
City Samplings Samples 
Total Positive 
Bogotá 44 2778 116 4.18 
Cúcuta 45 2772 83 2.99 
Medellín 2067 15 0.73 
Cali 603 0.33 
Total 97 8220 216 2.63 
3

Tree species and number of trees sampled. Numbers are expressed in absolute and relative values (%) for each city. Only % values greater than 1% are shown

Tree species n (%) 
Bogotá Cúcuta Medellín Cali Total 
Eucalyptus spp. 1398 (92) 388 (21.5) 1488 (72.0) 511 (84.9) 3785 (63.2) 
Licania tomentosa 519 (28.7) 253(12.2) 772 (12.9) 
Terminalia catappa 429 (23.7) 70 (3.4) 11 (1.8) 510 (8.5) 
Unidentified 90 (5.0) 101 (4.9) 80 (13.3) 271 (4.5) 
Ficus soatensis 55 (3.6) 90 (5.0) 145 (2.4) 
Pinus radiata 10 95 (4.6) 106 (1.8) 
Prosopis juliflora 81 (4.5) 81 (1.4) 
Cassia grandis 72 (4.0) 72 (1.2) 
Meliccoca bijuga 44 (2.4) 44 
Psidium guajava 25 (1.4) 25 
Magnifera indica 21 (1.2) 21 
Acacia decurrens 15 (1.0) 15 
Samanea saman 15 15 
Laphoensis sp. 15 15 
Phitecellobium dulce 13 13 
Guazuma ulmifolia 11 11 
Ficus tequendamae 10 10 
Ceiba pentandra 10 10 
Laurus sp. 10 10 
Cedrela montana 
Cupressus lusitanica 
Coussapoa sp. 
Croton funckianus 
Croton bogotanus 
Ruscus aculeatus 
Tibochina lepidota 
Codiaeum variegatum 
Swinglia gllutinosa 
Citrus sp. 
Hibiscus rosa 
Rubus sp. 
Passiflora foetida 
Gliricida sepium 
Sequoia sempervirens 
Carica papaya 
Annona muricata 
Total 1508 1808 2067 602 5985 
Tree species n (%) 
Bogotá Cúcuta Medellín Cali Total 
Eucalyptus spp. 1398 (92) 388 (21.5) 1488 (72.0) 511 (84.9) 3785 (63.2) 
Licania tomentosa 519 (28.7) 253(12.2) 772 (12.9) 
Terminalia catappa 429 (23.7) 70 (3.4) 11 (1.8) 510 (8.5) 
Unidentified 90 (5.0) 101 (4.9) 80 (13.3) 271 (4.5) 
Ficus soatensis 55 (3.6) 90 (5.0) 145 (2.4) 
Pinus radiata 10 95 (4.6) 106 (1.8) 
Prosopis juliflora 81 (4.5) 81 (1.4) 
Cassia grandis 72 (4.0) 72 (1.2) 
Meliccoca bijuga 44 (2.4) 44 
Psidium guajava 25 (1.4) 25 
Magnifera indica 21 (1.2) 21 
Acacia decurrens 15 (1.0) 15 
Samanea saman 15 15 
Laphoensis sp. 15 15 
Phitecellobium dulce 13 13 
Guazuma ulmifolia 11 11 
Ficus tequendamae 10 10 
Ceiba pentandra 10 10 
Laurus sp. 10 10 
Cedrela montana 
Cupressus lusitanica 
Coussapoa sp. 
Croton funckianus 
Croton bogotanus 
Ruscus aculeatus 
Tibochina lepidota 
Codiaeum variegatum 
Swinglia gllutinosa 
Citrus sp. 
Hibiscus rosa 
Rubus sp. 
Passiflora foetida 
Gliricida sepium 
Sequoia sempervirens 
Carica papaya 
Annona muricata 
Total 1508 1808 2067 602 5985 

Among the positive samples, 119 (55.1%) yielded serotype B, 60 (28.2%) serotype C and 36 (16.7%) serotype A isolates (Table 4). We found differences in the prevalence of each serotype when comparing the four cities, with serotype B being predominant in Bogotá and Cali, serotype A in Medellín, and serotype C in Cúcuta. All C. gattii serotype C isolates were obtained from T. cattapa trees, excepting one strain recovered from an Eucalyptus tree in Bogotá.

4

Serotype distribution in positive samples from trees in the cities. Numbers are expressed in absolute and relative values (%) for each city

City Serotype A Serotype B Serotype C Total 
Bogotá 4 (3.4) 111 (95.7) 1 (0.9) 116 
Cúcuta 24 (29.0) 0 (0) 59 (71.0) 83 
Medellín 8 (53.3) 6 (40.0) 1 (6.7) 15 
Cali 0 (0) 2 (100) 0 (0) 
Total 36 (16.7) 119 (55.1) 61 (28.2) 216 
City Serotype A Serotype B Serotype C Total 
Bogotá 4 (3.4) 111 (95.7) 1 (0.9) 116 
Cúcuta 24 (29.0) 0 (0) 59 (71.0) 83 
Medellín 8 (53.3) 6 (40.0) 1 (6.7) 15 
Cali 0 (0) 2 (100) 0 (0) 
Total 36 (16.7) 119 (55.1) 61 (28.2) 216 

The Cryptococcus n. species complex and climatic conditions

By integrating sampling and climatic data, we wanted to explore if there was a correlation between the climatic conditions on a specific date and the serotypes that were found. The initial analysis suggested that daily values of climatic variables were more informative than monthly values, as the former characterize more rigorously the climatic conditions for a specific date. Moreover, a monthly value represents a mean or cumulative value (in the case of rainfall) of daily values, so it was not informative in cases where the sample was collected in the first week of the month, for example. We therefore used only daily climatic values.

The lagged Pearson correlation was performed between the response variable (proportion of positive samples of each serotype) and all climatic variables (temperature, precipitation, relative humidity, etc), for each sampling date and for a lag of 1 to 30 days before the sampling date. This analysis was performed by integrating the data from all cities. The results obtained for most climatic variables showed that a time span as long as 15 days preceding the sampling date could have an influence on the probability of isolating C.n. and C. gattii from trees (data not shown). Thus, we decided to carry out the lagged Pearson correlation for each climatic variable, using the climatic data for a period of 15 days before sampling.

Based on our data, some climatic variables appear to be related to the probability of isolating one specific serotype from the environment. Relative humidity and mean, maximum and minimum temperatures were either directly or inversely related to the presence of all serotypes. Figure 2 shows the correlation values between the isolation of one specific serotype from trees and relative humidity values for the sampling date (day 0) and the 15 previous days. Relative humidity was inversely correlated with the occurrence of environmental isolates of serotypes A and C on the sampling day, as well as on the 15-day period before sampling (Fig. 2a and c), so low values of relative humidity favored the occurrence of serotypes A and C in trees. It is important to note that serotype C seems to be more sensible to relative humidity conditions than serotype A, as correlation values were more constant and slightly more negative for serotype C than for serotype A. By contrast, correlation values between relative humidity and the presence of C. gattii serotype B were positive, implying that serotype B required relatively high values of humidity for survival (Fig. 2b).

2

Lagged Pearson correlation between relative humidity and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) Serotype A; (b) serotype B; (c) serotype C.

2

Lagged Pearson correlation between relative humidity and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) Serotype A; (b) serotype B; (c) serotype C.

Temperature appeared also to be a determining factor for survival of the C.n. species complex in trees. The occurrence of C.n. var. grubii was favored by high values of mean, maximum and minimum temperatures (Fig. 3a), although our results suggest that its occurrence depended slightly more on high mean temperature than on high minimum and maximum temperatures. The occurrence of C. gattii serotype C was correlated with higher values of minimum and maximum temperatures, but not with mean temperature, suggesting that mostly extreme temperatures were restricting its distribution (Fig. 3c). By contrast, the occurrence of C. gattii serotype B was high when mean, maximum and minimum temperatures values were low (Fig. 3b). Besides that, maximum and minimum temperatures appeared to have a stronger effect on the occurrence of this serotype than did mean temperature.

3

Lagged Pearson correlation between mean, maximum and minimum temperatures and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) serotype A; (b) serotype B; (c) serotype C.

3

Lagged Pearson correlation between mean, maximum and minimum temperatures and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) serotype A; (b) serotype B; (c) serotype C.

Precipitation was strongly correlated with the occurrence of serotype A on sampling day and one day before, but not so in the preceding days (Fig. 4). No significant correlation was found between precipitation and occurrence of serotypes B and C (data not shown).

4

Lagged Pearson correlation between precipitation and occurrence of Cryptococcus neoformans var. grubii on the sampling day (0) and the 15 days preceding sampling.

4

Lagged Pearson correlation between precipitation and occurrence of Cryptococcus neoformans var. grubii on the sampling day (0) and the 15 days preceding sampling.

A significant correlation was found between evaporation values and occurrence of the three serotypes (Fig. 5). The occurrence of serotypes A and C was related with high evaporation values (Fig. 5a and c), whereas serotype B was more prevalent when evaporation was low (Fig. 5b).

5

Lagged Pearson correlation between evaporation and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) serotype A; (b) serotype B; (c) serotype C.

5

Lagged Pearson correlation between evaporation and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) serotype A; (b) serotype B; (c) serotype C.

A similar general pattern was observed when analyzing solar radiation, as occurrence of serotypes A and C was correlated with high values of solar radiation for almost all days (Fig. 6a and c), whereas prevalence of serotype B was correlated with low values of solar radiation (Fig. 6b).

6

Lagged Pearson correlation between solar radiation and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) serotype A; (b) serotype B; (c) serotype C.

6

Lagged Pearson correlation between solar radiation and the occurrence of Cryptococcus neoformans species complex on the sampling day (0) and the 15 days preceding sampling. (a) serotype A; (b) serotype B; (c) serotype C.

Discussion

This retrospective study compiled the data about environmental strains of C.n. var. grubii and C. gattii serotypes B and C isolated from tree samples in four cities in Colombia during a 12-year period. As far as we know, this is the first survey aiming to compile such environmental information for a country. The large number of records, involving 97 samplings of 5985 trees belonging to at least 35 different species and 8220 samples collected during 12 years, allows us to establish some trends for the prevalence of the different serotypes associated with trees in Colombia.

The overall prevalence of C.n. species complex in trees in Colombia was 2.6%, which is low compared to the prevalence reported in other regions (Randhawa, 2003; Randhawa, 2005; Soares, 2005). It is worth noting that our survey involved only environmental samples taken from trees and that the overall prevalence of C.n. species complex in the environment could be higher if other niches such as pigeon droppings were also taken into account. It remains to be determined if this low prevalence of C.n. species complex in trees could have an influence on causing clinical cases of cryptococcosis in Colombia.

Bogotá and Cúcuta showed a higher prevalence of the C.n. species complex in trees than Medellín and Cali (4.18% and 2.99% vs. 0.73% and 0.33%, respectively). Moreover, this was not an effect of differences in the tree species studied, as Eucalyptus spp. were the most sampled species in all cities except for Cúcuta. These results are in agreement with the clinical prevalence of this pathogen reported by Lizarazo (2005): the prevalances per 106 clinical cases per year were 3.6 in the Bogotá region, 3.8 in Santanderes (Cúcuta and surrounding regions), 3.0 in Antioquia, Caldas, Quindío and Risaralda region (including Medellín) and 2.7 in the Pacific region (including Cali). Thus, clinical incidence in the different regions could partially be explained in terms of environmental prevalence, although the pigeon dropping niche was not explored in the current survey.

Based on our previous study (Granados & Castañeda, 2005), we hypothesize that different serotypes of the C.n. species complex could have different climatic requirements, which could explain the differences in geographical distribution and ecological niches found between serotypes. The present survey confirms our hypothesis, as it is clear that the environmental requirements of C. gattii serotype B are completely different from those of C. gattii serotype C and C.n. var. grubii. The occurrence of serotypes A and C in trees was favored when relative humidity was low and temperature, evaporation and solar radiation were high. In contrast, the occurrence of C. gattii serotype B was correlated with high relative humidity and low temperatures, evaporation and solar radiation. Even though almost all climatic variables were correlated with the occurrence of the serotypes, precipitation was not clearly correlated with the presence of serotypes B and C, as was expected based on previous observations and on the fact that relative humidity was evidently correlated. The reasons for this discrepancy are not known.

Our results suggest that serotypes A and C are more thermotolerant and hygrophobic than serotype B. These trends are supported by a large amount of data over a long period of time, involving both dry and wet seasons. Most of these observations are in agreement with our previous study (Granados & Castañeda, 2005). On that occasion, we suggested that rainy months, characterized by high precipitation and humidity, few hours of sunlight, less extreme temperatures and slightly higher temperatures, favored the occurrence of C. gattii serotype B in trees in the temperate climate of Bogotá, more than dry months. On the same survey, we confirmed that C. n. var. grubii was more frequently isolated from dry than from fresh pigeon droppings, as reported by other authors (Ruiz, 1981; Caicedo, 1999; Montenegro & Paula, 2000; Kuroki, 2004; Granados & Castañeda, 2005). The present results could help explain these apparently contradictory observations, as they could be due to different climatic preferences of these two species.

Cryptococcus gattii was thought to be restricted to tropical and subtropical regions (Sorrell, 2001), until the recent outbreak of cryptococcosis due to C. gattii serotype B on the southern part of Vancouver Island, British Columbia, Canada, changed its geographical distribution (Stephen, 2002; Hoang, 2004; Kidd, 2004). It was suggested that the large-scale colonization reported in this temperate climate zone could be reflecting a striking change in the distribution of this serotype, due to a global warming and to the Vancouver Island microclimate (Kidd, 2004). Furthermore, this serotype was recently detected for the first time in another temperate zone, Buenos Aires (Argentina), during mild spring (Davel, 2003). Also, a recent study carried out in Colombia showed that C. gattii serotype B was present in different altitudinal zones and thus in different climates, including temperate and cold climates (Quintero, 2005).

In the current study, serotype B showed an intermediate prevalence in Medellín, a city characterized by a mild warm climate; but it was the most prevalent serotype in a temperate city, Bogotá, and a warm city, Cali. Nevertheless, the prevalence found in Cali could be biased, as only four samplings were undertaken in this city. Our results could help explain the occurrence of serotype B at temperate latitudes such as Canada and Argentina and in temperate climates of tropical regions typical of the Andes Mountains of Colombia.

One could argue that the high prevalence of serotype B in Bogotá and Cali could not be reflecting a climatic preference, but a strong association of this serotype with Eucalyptus trees, as this species was the most sampled one in both cities. However, Eucalyptus sp. was also the most studied species in Medellín, but in this case serotype A (and not B) was the most prevalent serotype. Moreover, the fact that no C. gattii serotype B isolates were found in Cúcuta, despite the fact that samples were collected from 388 Eucalyptus trees, allows us to support our hypothesis, without discarding any effect of biased sampling on the isolates recovered.

Despite the genetical proximity of serotypes B and C (Boekhout, 2001; Latouche, 2003), these serotypes showed drastic differences regarding climatic preferences. Cryptococcus gattii serotype C had the highest correlation values for all variables tested, suggesting that serotype C is more demanding regarding climatic conditions: a high minimum and maximum temperature, evaporation and solar radiation, and low relative humidity. In fact, the environmental samples that were positive for this serotype were completely dry and sandy. Furthermore, this serotype was mostly isolated from Cúcuta, where it is associated with almond trees. As shown on Table 1, this city has the hottest climate of the cities studied. This serotype has also been found in the environment in Neiva and Barranquilla, other Colombian cities with similar climates, as well as in Puerto Rico (E. Quintero, personal communication). Hence, higher thermal and desiccation tolerance of C. gattii serotype C could confer a survival advantage in extremely hot climates and could help explain the worldwide scarcity of environmental isolates of this serotype.

As all but one of the C. gattii serotype C isolates were obtained from almond trees, one could argue that this serotype was mostly isolated from Cúcuta because of its close relation with host almond trees and not because of a climate preference. This could be true to some extent as this species, originally native to India but naturalized in tropical and subtropical regions, is well adapted to harsh conditions characterized by intense sunlight, salty air and drought prone soils. Nonetheless, we further analyzed all samplings of almond trees in Cúcuta in terms of frequency of isolation and found that C. gattii serotype C was not isolated from these trees in all samplings. Interestingly, most negative almond trees samples were taken on dates when the minimum and maximum temperatures were below the average historical value (data not shown) and thus when climate was less hot, supporting the proposed relationship between C. gattii serotype C occurrence and hot climate. Thus, both almond trees and C. gattii serotype C appear to share the same climatic requirements and this could explain to some extent the preference of this serotype for almond trees. Studies aimed at obtaining serotype C isolates from tree species other than almond trees whose geographic distribution matches the climatic conditions mentioned above would be of great interest to resolve this issue.

The climatic pattern of serotype A was similar to that of serotype C, although it apparently tolerates lower temperatures and more humidity than serotype C. Its occurrence was correlated with low humidity and high solar radiation and evaporation. Similarly, previous studies have reported that dry pigeon droppings are more suitable for colonization by C.n. var. grubii than fresh droppings (Ruiz, 1981; Caicedo, 1999; Montenegro & Paula, 2000; Kuroki, 2004; Granados & Castañeda, 2005). Although in this study serotype A occurred in hot, warm and temperate climates, it was most prevalent in Medellín, a city with mild warm climate. The climate of Medellín could reflect the optimal climatic conditions for survival of this serotype.

It is also noteworthy that climatic conditions of a period as long as 15 days before sampling were correlated with the probability of finding isolates of the C.n. species complex. This information should be taken into account for future sampling designs.

In summary, the current results suggest that C.n. var. grubii, C. gattii serotype B and C. gattii serotype C have different climatic tolerances, which could help explain geographic differences in the prevalence of serotypes. Nevertheless, we do not discard the possibility that other abiotic factors such as wind, pH and chemical composition of the substrate, as well as biotic factors, also influence the environmental distribution of C.n. species complex. Finally, we plan the inclusion of other environmental sources such as pigeon droppings in further analysis and the corroboration of our hypothesis in vitro, by simulating the climatic requirements of each serotype.

Acknowledgements

We thank Eduardo Granados for his valuable and excellent statistical assistance, IDEAM (National Hydrology, Meteorology and Environmental Studies Institute of Colombia) as source of the climatic information and Colciencias for financial support (Young Researchers Program Code 90-2003).

References

Boekhout
T
Theelen
B
Diaz
M
Fell
JW
Hop
WCJ
Abeln
ECA
Dromer
F
Meyer
W
(
2001
)
Hybrid genotypes in the pathogenic yeast Cryptococcus neoformans
.
Microbiology
 
147
:
891
907
.
Caicedo
LD
Alvarez
MI
Delgado
M
Cardenas
A
(
1999
)
Cryptococcus neoformans in bird excreta in the city zoo of Cali, Colombia
.
Mycopathologia
 
147
:
121
124
.
Callejas
A
Ordóñez
N
Rodríguez
MC
Castañeda
E
(
1998
)
First isolation of Cryptococcus neoformans var. gattii, serotype C, from the environment in Colombia
.
Med Mycol
 
36
:
341
344
.
Casadevall
A
Perfect
JR
(
1998
)
Cryptococcus neoformans
 .
American Society for Microbiology
,
Washington, D.C.
Davel
G
Abrantes
R
Brudny
M
Cordoba
S
Rodero
L
Canteros
CE
Perrotta
D
(
2003
)
1st Environmental isolation of Cryptococcus neoformans var. gattii in Argentina
.
Rev Argent Microbiol
 
35
:
110
112
.
Duarte
A
Ordóñez
N
Castañeda
E
(
1994
)
Association of yeasts of the Cryptococcus genus with Eucalyptus species in Santafe de Bogotá
.
Rev Inst Med Trop Sao Paulo
 
36
:
125
130
.
Escandón
P
Quintero
E
Granados
DP
Huérfano
S
Ruiz
A
Castañeda
E
(
2005
)
Isolation of Cryptococcus gattii serotype B from detritus of Eucalyptus trees in Colombia
.
Biomedica
 
25
:
390
397
.
Eslava
JA
(
1993
)
Climatología y diversidad climática de Colombia
.
Rev Acad Colomb Cienc
 
18
:
507
538
.
Franzot
SP
Salkin
IF
Casadevall
A
(
1999
)
Cryptococcus neoformans var. grubii: separate varietal status for Cryptococcus neoformans serotype A isolates
.
J Clin Microbiol
 
37
:
838
840
.
Granados
DP
Castañeda
E
(
2005
)
Isolation and characterization of Cryptococcus neoformans varieties recovered from natural sources in Bogotá, Colombia, and study of ecological conditions in the area
.
Microb Ecol
 
49
:
282
290
.
Hoang
LM
Maguire
JA
Doyle
P
Fyfe
M
Roscoe
DL
(
2004
)
Cryptococcus neoformans infections at Vancouver Hospital and Health Sciences Centre (1997–2002): epidemiology, microbiology and histopathology
.
J Med Microbiol
 
53
:
935
940
.
Hubalek
Z
(
1975
)
Distribution of Cryptococcus neoformans in a pigeon habitat
.
Folia Parasitol (Praha)
 
22
:
73
79
.
Hubalek
Z
Prikazsky
Z
(
1975
)
Growth of Cryptococcus neoformans in UV-irradiated excreta of pigeons
.
Folia Microbiol (Praha)
 
20
:
231
235
.
Ishaq
CM
Bulmer
GS
Felton
FG
(
1968
)
An evaluation of various environmental factors affecting the propagation of Cryptococcus neoformans
.
Mycopathol Mycol Appl
 
35
:
81
90
.
Kidd
SE
Hagen
F
Tscharke
RL
Huynh
M
Bartlett
KH
Fyfe
M
MacDougall
L
Boekhout
T
Kwon-Chung
KJ
Meyer
W
(
2004
)
A rare genotype of Cryptococcus gattii caused the cryptococcosis outbreak on Vancouver Island (British Columbia, Canada)
.
Proc Natl Acad Sci USA
 
101
:
17258
17263
.
Krockenberger
MB
Canfield
PJ
Malik
R
(
2002
)
Cryptococcus neoformans in the koala (Phascolarctos cinereus): colonization by Cryptococcus neoformans var. gattii and investigation of environmental sources
.
Med Mycol
 
40
:
263
272
.
Kuroki
M
Phichaichumpon
C
Yasuoka
A
et al
. (
2004
)
Environmental isolation of Cryptococcus neoformans from an endemic region of HIV-associated cryptococcosis in Thailand
.
Yeast
 
21
:
809
812
.
Kwon-Chung
KJ
Boekhout
T
Fell
JW
Diaz
M
(
2002
)
Proposal to conserve the name Cryptococcus gattii against C. hondurianus and C. bacillisporus (Basidiomycota, Hymenomycetes, Tremellomycetidae)
.
Taxon
 
51
:
804
806
.
Latouche
GN
Huynh
M
Sorrell
TC
Meyer
W
(
2003
)
PCR-restriction fragment length polymorphism analysis of the phospholipase B (PLB1) gene for subtyping of Cryptococcus neoformans isolates
.
Appl Environ Microbiol
 
69
:
2080
2086
.
Lizarazo
J
Linares
M
Restrepo
A
Castañeda
E
Colombian Cryptococcosis Study Group
(
2005
)
Cryptococcosis in Colombia: Surveillance results 1997–2004
.
Sixth International Conference on Cryptococcus and Cryptococcosis
,
Boston, Massachusetts
,
June 24–28
, E-5:115.
Martinez
LR
Garcia-Rivera
J
Casadevall
A
(
2001
)
Cryptococcus neoformans var. neoformans (serotype D) strains are more susceptible to heat than C. neoformans var. grubii (serotype A) strains
.
J Clin Microbiol
 
39
:
3365
3367
.
Montenegro
H
Paula
CR
(
2000
)
Environmental isolation of Cryptococcus neoformans var. gattii and C. neoformans var. neoformans in the city of Sâo Paulo, Brazil
.
Med Mycol
 
38
:
385
390
.
Poveda
G
(
2004
)
La hidroclimatología de Colombia: una síntesis desde una escala inter-decadal hasta la escala diurna
.
Rev Acad Colomb Cienc
28
:
201
222
.
Quintero
E
Castañeda
E
Ruiz
A
(
2005
)
Environmental distribution of Cryptococcus neoformans in the department of Cundinamarca-Colombia
.
Rev Iberoam Micol
 
22
:
93
98
.
Randhawa
HS
Kowshik
T
Khan
ZU
(
2003
)
Decayed wood of Syzygium cumini and Ficus religiosa living trees in Delhi/New Delhi metropolitan area as natural habitat of Cryptococcus neoformans
.
Med Mycol
 
41
:
199
209
.
Randhawa
HS
Kowshik
T
Khan
ZU
(
2005
)
Efficacy of swabbing versus a conventional technique for isolation of Cryptococcus neoformans from decayed wood in tree trunk hollows
.
Med Mycol
 
43
:
67
71
.
Ruiz
A
Fromtling
RA
Bulmer
GS
(
1981
)
Distribution of Cryptococcus neoformans in a natural site
.
Infect Immun
 
31
:
560
563
.
Soares
MC
Paula
CR
Dias
AL
Caseiro
MM
Costa
SO
(
2005
)
Environmental strains of Cryptococcus neoformans variety grubii in the city of Santos, SP, Brazil
.
Rev Inst Med Trop Sao Paulo
 
47
:
31
36
.
Sorrell
TC
(
2001
)
Cryptococcus neoformans variety gattii
.
Med Mycol
 
39
:
155
168
.
Staib
F
(
1985
)
Sampling and isolation of Cryptococcus neoformans from indoor air with the aid of the Reuter Centrifugal Sampler (RCS) and Guizotia abyssinica creatinine agar. A contribution to the mycological-epidemiological control of Cr. neoformans in the fecal matter of caged birds
.
Zentralbl Bakteriol Mikrobiol Hyg [B]
 
180
:
567
575
.
Stephen
C
Lester
S
Black
W
Fyfe
M
Raverty
S
(
2002
)
Multispecies outbreak of cryptococcosis on southern Vancouver Island, British Columbia
.
Can Vet J
 
43
:
792
794
.
Walter
JE
Yee
RB
(
1968
)
Factors that determine the growth of Cryptococcus neoformans in avian excreta
.
Am J Epidemiol
 
88
:
445
450
.
Wang
Y
Casadevall
A
(
1994
)
Decreased susceptibility of melanized Cryptococcus neoformans to UV light
.
Appl Environ Microbiol
 
60
:
3864
3866
.

Author notes

Editor: Stuart Levitz