Pulmonary Aspergillosis in Humboldt Penguins—Susceptibility Patterns and Molecular Epidemiology of Clinical and Environmental Aspergillus fumigatus Isolates from a Belgian Zoo, 2017–2022

Aspergillus fumigatus is the main causative agent of avian aspergillosis and results in significant health problems in birds, especially those living in captivity. The fungal contamination by A. fumigatus in the environment of Humboldt penguins (Spheniscus humboldti), located in a Belgian zoo, was assessed through the analysis of air, water, sand and nest samples during four non-consecutive days in 2021–2022. From these samples, potential azole-resistant A. fumigatus (ARAF) isolates were detected using a selective culture medium. A total of 28 veterinary isolates obtained after necropsy of Humboldt penguins and other avian species from the zoo were also included. All veterinary and suspected ARAF isolates from the environment were characterized for their azole-resistance profile by broth microdilution. Isolates displaying phenotypic resistance against at least one medical azole were systematically screened for mutations in the cyp51A gene. A total of 14 (13.6%) ARAF isolates were identified from the environment (n = 8) and from Humboldt penguins (n = 6). The TR34/L98H mutation was observed in all resistant environmental strains, and in two resistant veterinary strains. To the best of our knowledge, this is the first description of this mutation in A. fumigatus isolates from Humboldt penguins. During the period 2017–2022, pulmonary aspergillosis was confirmed in 51 necropsied penguins, which reflects a death rate due to aspergillosis of 68.0%, mostly affecting adults. Microsatellite polymorphism analysis revealed a high level of diversity among environmental and veterinary A. fumigatus isolates. However, a cluster was observed between one veterinary isolate and six environmental strains, all resistant to medical azoles. In conclusion, the environment of the Humboldt penguins is a potential contamination source of ARAF, making their management even more complex.


Introduction
The saprophytic fungus Aspergillus fumigatus is responsible for opportunistic infections affecting birds and mammals, including humans [1]. It can affect a wide variety of species such as domestic, free-ranging or captive wild animals [2]. Aspergillosis is the most common fungal infectious disease affecting penguins in zoos, with up to 99% of cases attributed to Aspergillus section Fumigati. It represents a major limiting factor for the

Environmental Aspergillus fumigatus Sampling inside the Penguin Enclosure
Cultures obtained on a malt extract + chloramphenicol (MC) medium from environmental air sampling at four predefined locations( Figure 1) revealed a mean A. fumigatus burden in spring and summer of 41.5 and 56.5 CFU/1000 L, respectively, and 15.75 and 22.75 CFU/1000 L in autumn and winter, respectively. These differences were, however, not significant and the overall A. fumigatus contamination (log 10 CFU/1000 L) statistically similar between the different study time points for all four study locations (p-value = 0.27) ( Figure 2). All A. fumigatus colonies isolated from a malt extract + chloramphenicol + 4 mg/L tebuconazole (MC + T) medium (n = 68) were stored and further analyzed in this study. All isolates were confirmed as A. fumigatus using MALDI-TOF MS with a score ≥ 2.0. The samples from sand, water and nest swabs were negative for A. fumigatus.
Antibiotics 2023, 12, x FOR PEER REVIEW 3 of 16 assess the impact of environmental A. fumigatus contamination on the clinical incidence of avian aspergillosis in Humboldt penguins in a Belgian zoo. The susceptibility patterns towards medical azoles and the mutations in the cyp51A gene were studied, alongside the genotyping of the strains, in order to estimate the epidemiology of the infections.

Environmental Aspergillus fumigatus Sampling inside the Penguin Enclosure
Cultures obtained on a malt extract + chloramphenicol (MC) medium from environmental air sampling at four predefined locations( Figure 1) revealed a mean A. fumigatus burden in spring and summer of 41.5 and 56.5 CFU/1000 L, respectively, and 15.75 and 22.75 CFU/1000 L in autumn and winter, respectively. These differences were, however, not significant and the overall A. fumigatus contamination (log10 CFU/1000 L) statistically similar between the different study time points for all four study locations (p-value = 0.27) ( Figure 2). All A. fumigatus colonies isolated from a malt extract + chloramphenicol + 4 mg/L tebuconazole (MC + T) medium (n = 68) were stored and further analyzed in this study. All isolates were confirmed as A. fumigatus using MALDI-TOF MS with a score ≥ 2.0. The samples from sand, water and nest swabs were negative for A. fumigatus.   Four sample types were taken: red cross = air sample; blue triangle = water; green diamond = sand; yellow circle = nest swab. assess the impact of environmental A. fumigatus contamination on the clinical incidence of avian aspergillosis in Humboldt penguins in a Belgian zoo. The susceptibility patterns towards medical azoles and the mutations in the cyp51A gene were studied, alongside the genotyping of the strains, in order to estimate the epidemiology of the infections.

Environmental Aspergillus fumigatus Sampling inside the Penguin Enclosure
Cultures obtained on a malt extract + chloramphenicol (MC) medium from environmental air sampling at four predefined locations( Figure 1) revealed a mean A. fumigatus burden in spring and summer of 41.5 and 56.5 CFU/1000 L, respectively, and 15.75 and 22.75 CFU/1000 L in autumn and winter, respectively. These differences were, however, not significant and the overall A. fumigatus contamination (log10 CFU/1000 L) statistically similar between the different study time points for all four study locations (p-value = 0.27) ( Figure 2). All A. fumigatus colonies isolated from a malt extract + chloramphenicol + 4 mg/L tebuconazole (MC + T) medium (n = 68) were stored and further analyzed in this study. All isolates were confirmed as A. fumigatus using MALDI-TOF MS with a score ≥ 2.0. The samples from sand, water and nest swabs were negative for A. fumigatus.

Broth Microdilution Antifungal Susceptibility Testing and cyp51A Sequencing
All clinical (n = 35) and environmental (n = 68) A. fumigatus isolates were subjected to broth microdilution antifungal susceptibility testing. A total of six clinical (17.14%) and eight environmental (11.76%) A. fumigatus isolates displayed resistance against at least one medical azole (Table 3). Table 3. Antifungal susceptibility testing results of the isolates displaying antifungal resistance against at least one medical azole, and associated cyp51A mutation. All A. fumigatus isolates showing resistance against at least one medical azole were further characterized by cyp51A sequencing. Two clinical isolates showed the presence of the TR34/L98H mutation (Table 3) and one displayed several nucleotide mutations, resulting in three amino acid substitutions F46Y, M172V, E427K. Three other clinical isolates did not show any mutations known to cause resistance in the cyp51A gene. All environmental isolates showed the presence of the TR34/L98H mutation and one (21-0503) showed an additional G54R nucleotide mutation.

Microsatellite Genotyping of the Aspergillus fumigatus Isolates
Genotyping was performed on a selection of 45 A. fumigatus isolates ( Figure 3, Table A1). The discriminatory power of the combined markers reached 98.8. A total of 21 veterinary isolates were included-18 from Humboldt penguins and 3 from other avian species (Psarocolius decumanus, Spheniscus demersus and Rollulus rouloul) to investigate if transmission between bird species was possible. No identical genotype was shared between the veterinary isolates. Twenty-four environmental isolates, representing both susceptible and resistant strains, were also analyzed and resulted in 20 different genotypes. The same genotype was found in four environmental strains (21-0515, 21-0516, 21-0517 and 21-0518), and were closely related to two other genotypes: a first one shared by two environmental strains (21-0503 and 21-0510), and a second one corresponding to an isolate (21-0524) originating from a Humboldt penguin ( Figure 3, Table A1). Genotypes within this cluster differed by only one marker (STRAf2A). All seven isolates in this cluster were resistant against at least one medical azole ( Figure 3). The six environmental strains originated from the study time point in autumn (October 2021). The Humboldt penguin (21-0524) died during the same period, in August 2021 (Table 2). and were closely related to two other genotypes: a first one shared by two environmental strains (21-0503 and 21-0510), and a second one corresponding to an isolate (21-0524) originating from a Humboldt penguin ( Figure 3, Table A1). Genotypes within this cluster differed by only one marker (STRAf2A). All seven isolates in this cluster were resistant against at least one medical azole (Figure 3). The six environmental strains originated from the study time point in autumn (October 2021). The Humboldt penguin (21-0524) died during the same period, in August 2021 (Table 2).

Discussion
Aspergillosis in captive birds plays an important role which can pose problems in their management. All avian aspergillosis cases in our study arose by A. fumigatus, confirming its role as the main causative fungal agent of avian aspergillosis in penguins, as previously described [12,40]. The naturally elevated body temperature of the birds, ranging from 39 to 41 • C, combined with the thermotolerance of A. fumigatus, favor its incidence. The second most common agent of bronchopulmonary aspergillosis in birds is A. flavus, accounting for about 5% of avian aspergillosis cases [11], but this species was not detected in this study.
In the present study, the mortality due to aspergillosis reached 86.7% in 2017 and 76.5% in 2018. In 2018, we see a clear rise in aspergillosis incidence compared to the other years). This might be explained by the extreme weather conditions in 2018, especially due to the high temperatures in summer: 65% of the deaths in 2018 occurred during the summer months. Additionally, the population size at that time was large and a breeding program was operative in 2018. The effect of overcrowding was demonstrated in a study performed over 6 years in Magellanic penguins (Spheniscus magellanicus), where 65% of aspergillosis cases took place during the year with the highest population density [5]. Measures were taken in the Belgian zoo in 2019 to lower the mortality of aspergillosis: ventilation holes were installed in the nests, monitoring of the water temperature below 20 • C to mimic their natural habitat, shade was provided in summer, breeding program was interrupted and direct contacts between visitors and penguins were discontinued. The breeding program started again in 2020 with 4 breeding couples in 2020 and 8 in 2021. The mortality in penguins increased again when breeding restarted. Over the period 2017-2021, it appeared that the aspergillosis incidence was 3-fold higher in animals rearing young (n = 22, 12 female and 10 male) than those that did not (n = 7, 5 female and 2 male). The highest mortality was observed in adults, which is in contrast with previous findings, where most of the cases of aspergillosis occur in juveniles [5,13,41]. Avian aspergillosis in adult penguins is mostly due to the chronical form and is linked to immune suppression [6]. Noteworthy, juveniles younger than 2 months (n = 20) were not tested in this study for the presence of aspergillosis.
Environmental sampling in the penguin housing did not show significant differences between the four study time points. However, mean fungal loads in the environment were slightly more elevated in spring and summer. This is in line with the findings of Cateau et al. reporting higher fungal burden in September compared to April and December [13]. Similarly, a study performed on African Penguins in the Maryland Zoo in Baltimore, described higher environmental fungal load during the warmer period from the end of spring to the beginning of autumn [42].
Artificial nests made up of plastic in the synthetic rock formations did not reveal the presence of A. fumigatus in this study. Here, only two nests were consistently sampled on 3 out of the 4 sampling days, which might explain the absence of A. fumigatus. However, unpublished data from the zoo obtained on a larger sampling campaign of all the nests (n = 25), confirmed that in the period of 2017-2018, A. fumigatus colonies were found in all nests, with higher average counts from May to September, as compared to October to March. This is in agreement to the study of Cateau et al. which found a high fungal load in the nests [13].
Azole resistance in A. fumigatus represents an emerging problem in human and veterinary medicine and was detected in 14 isolates, both from the environment and from penguins. Most of the mutations conferring azole resistance in A. fumigatus are found in the cyp51A gene encoding for lanosterol-C14-α-demethylase, the target protein of azole drugs [27]. Cyp51A mutations can be tandem repeats (TR) in the promotor region of the gene, single-nucleotide polymorphisms (SNPs), or both [27]. Many articles describe the TR34/L98H mutation conferring resistance to azole drugs, which has been linked to the intense use of agricultural azoles for crop protection [31]. The zoo is located near a larger city, however, it is also surrounded by many agriculturally cultivated plots. In this study, the TR34/L98H mutation was found in both clinical and environmental samples. This mutation usually leads to pan-azole resistance phenotype in human clinical samples, which can be seen in several samples in this study (Table 3). In contrast, the most common mutations in cyp51A leading to azole-resistance that develop during antifungal treatment, occur in amino acid sites G54, G138, M220, and G448 [43]. Prophylaxis with azoles in distressed penguins, as well as the treatment of aspergillosis in captive penguins, are very common [25]. Several penguins included in our study received treatment for extended periods due to an increased chance of disease development.
However, only the TR34/L98H mutation was detected, in two strains isolated from Humboldt penguins, suggesting that the resistance of these strains was acquired from the environment. The isolate 21-680 harbored several other amino acid substitutions in the cyp51A gene: F46Y, M172V, E427K. The combination of these amino acid substitutions was reported in approximatively 10% of all A. fumigatus isolates tested worldwide [44], including in patients receiving azole treatment [45]. Generally, they display elevated MIC values for the medical azoles compared to the wild-type (WT) cyp51A [45]. However, their susceptibility profiles are inconsistent and were described as both azole-susceptible or resistant by different authors [31,[44][45][46][47][48][49][50][51]. No known resistance conferring mutations in the cyp51A gene were found in the remaining veterinary isolates. Other mechanisms, such as mutations in other genes [32,36,40] or efflux pumps [34] could be involved in the decreased susceptibility of these isolates to POSA. All eight environmental azole-resistant strains harbored the TR34/L98H mutation. Three of them displayed the typical pan-azole phenotype, whereas the remainder were resistant against ISA and POSA, but had MIC values in the area of technical uncertainty (ATU) for VOR and ITC.
Genotyping of A. fumigatus isolates in this study revealed a broad diversity in both environmental and veterinary strains, suggesting independent events of contamination.
Additionally, the azole-resistant strains were not all closely related to each other, indicating that the resistance was acquired multiple times and has different origins. No relation was observed between the veterinary isolates included in this study. In contrast, Cateau et al. found identical genotypes among veterinary isolates, but their sampling was performed within a short timespan [13]. A veterinary strain from a Humboldt penguin (which died in August 2021), however, clustered with six environmental isolates from the study time point in autumn. Moreover, all seven isolates of this cluster were resistant against at least one medical azole and harbored the TR34/L98H mutation. This suggests that the Humboldt penguin acquired the strain from the environment. Interestingly, isolates 21-0503 and 21-0510 had identical genotypes, but differed in MIC values and cyp51A sequencing. Isolate 21-0503 indeed had an additional G54R amino acid mutation, alongside the TR34/L98H mutation. The observed MIC values were higher for 21-0510 than for 21-0503 for ISA and POSA. Additionally, two strains (21-0428 and 21-0488) originating from the same animal had different unrelated genotypes, indicating that Humboldt penguins can be infected by multiple A. fumigatus strains. Altogether, within this study we were able to capture a small proportion of the large diversity present in the environment and veterinary A. fumigatus strains present in the Belgian zoo.
This research has several limitations. The first is the timeframe of environmental sampling which was performed on 4 non-consecutive days. This gives a limited view on the seasonality since single time point measurements depend on many different aspects such as temperature; wind or humidity. Future studies should therefore consider continuous long-term sampling. Secondly, environmental sampling was only performed in 2021 and 2022; while the majority of the veterinary strains were isolated during the previous years.
There is limited evidence in humans that the infection can be spread from patient to patient [52]. There is, however, no evidence reported yet of such events in animals. However, considering the anatomy of the respiratory system of birds, and the high fungal burden in their air sacks/lungs, it could also be true for birds. This would need more research with multiple time point sampling of penguins living in the same habitat. However, this might be a challenge considering the invasive nature of sampling living penguins and the poor reliability of diagnostic tools ante mortem.
The One Health concept envisages a tripartite health system based on the environment, animals and humans and their mutual interactions. This concept directs us towards a more holistic approach in the surveillance of infectious diseases on a global scale. The interactions between the environment, the ubiquitous mold A. fumigatus, and the birds, illustrate such complex ecosystem. The latter is changing in an accelerated manner due to global warming, which should be considered when addressing research on infectious diseases. In this paper, we were able to show a probable interaction between the environment and animals, but transmission between animals was not evidenced. However, we observed the same resistance profile and gene mutations in the environment and in the animals, which are also observed in human isolates [31]. The effect of extreme weather conditions was demonstrated. We could predict these phenomena to occur more often with the effects of global warming.
In conclusion, this report contributes to a better understanding of the molecular epidemiology of avian aspergillosis in penguins, dominated by A. fumigatus. It is also the first report of a TR34/L98H mutation in A. fumigatus isolates obtained from penguins, showing the relevance of monitoring azole resistance of A. fumigatus in veterinary science. Genotyping revealed infection by multiple A. fumigatus strains in the same penguin individual, as well as a clustering between environmental and veterinary isolates. More frequent sampling could provide more insight on the diversity and possible transmission of A. fumigatus between Humboldt penguins and their environment. The high mortality rates due to aspergillosis observed in this study also question the best combination of practices in captive penguins management.

Environmental Aspergillus fumigatus Sampling within the Penguin Enclosure
Environmental sampling was conducted between April 2021 and January 2022 at a Belgian zoo. The maximal population size of the group was 52 in 2021 and 28 in 2022. Since the start of the Humboldt penguin program in 2013, the zoo welcomed a total of 214 animals. The penguin habitat, which is exclusively outdoors, includes a temperature-monitored swimming pool (<20 • C) and artificial rock formations with 25 built-in nests (Figure 4). The entire habitat is roofed with a steel wired net to prevent the entry of other birds.

Isolation of Aspergillus fumigatus from Environmental Samples
All samples were plated on two media: malt extract agar + chloramphenicol (MC) and MC supplemented with 4 mg/L of tebuconazole (MC + T). The plates were incubated at 48 °C ± 1 °C for 48 h ± 2 h to prevent the growth of most environmental fungi. For the air samples, a total of 1000 L of air was impacted on each medium using the MAS-100 NT™ impactor (Merck ® , Darmstadt, Germany) with a rate of 100 L/min. Quantitative results were expressed as log10 CFU/1000 L air. From each sand sample, 1 g was dissolved in 9 mL of 0.85% NaCl + 0.01% Tween 20 solution. After thorough vortexing, 100 microliters were plated onto the two media. A total of 100 mL of water from the swimming pool was filtered using a nitrocellulose membrane filter (0.45 µm, Sartorius, Göttingen, Duitsland) in two times, after which each filter was placed on a different medium. For the surface sampling of the nests, swabs were taken on the inside of the nests, (floor and walls) and inserted in 1 mL of Amies liquid (eSwab, Copan, Menen, Belgium). Following a 1 min vortexing step, 100µL of each suspension was seeded onto the two media.
The MC medium was used to determine the total number of A. fumigatus isolates. Fungal colonies on MC + T were isolated and identified for further analysis to detect potential azole-resistant A. fumigatus (ARAF) isolates. The microscopic and macroscopic features of each colony were used to identify every A. fumigatus isolate. Matrix-assisted laser desorption ionization-time of flight mass spectrophotometry (MALDI-TOF MS) was used to confirm the identity of the suspected ARAF isolates that grew on MC + T [53].

Clinical Incidence of Avian aspergillosis in Humboldt Penguins
Aspergillosis incidence in the Humboldt penguins in the zoo was assessed for the period 2017-2022. Necropsy was performed on all dead animals older than 2 months with suspected aspergillosis (Figure 5). Confirmed cases of pulmonary aspergillosis were verified by light microscopy using a lactophenol cotton blue stain and culture of A. fumigatus of the lungs and/or the surface of the air sacs. Four study time points of environmental sampling were performed on April 1st 2021, 29 June 2021, 13 October 2021 and 12 January 2022, covering spring, summer, autumn and winter, respectively. At each sampling day, four types of samples were taken: air, water, sand and nest surface (except for nest swabs on April 1st since breeding couples were present and could not be disturbed). Sand and air were taken at four predefined locations ( Figure 1). Two nest samples and one water sample of the pool were taken (Figure 1).

Isolation of Aspergillus fumigatus from Environmental Samples
All samples were plated on two media: malt extract agar + chloramphenicol (MC) and MC supplemented with 4 mg/L of tebuconazole (MC + T). The plates were incubated at 48 • C ± 1 • C for 48 h ± 2 h to prevent the growth of most environmental fungi. For the air samples, a total of 1000 L of air was impacted on each medium using the MAS-100 NT™ impactor (Merck ® , Darmstadt, Germany) with a rate of 100 L/min. Quantitative results were expressed as log10 CFU/1000 L air. From each sand sample, 1 g was dissolved in 9 mL of 0.85% NaCl + 0.01% Tween 20 solution. After thorough vortexing, 100 microliters were plated onto the two media. A total of 100 mL of water from the swimming pool was filtered using a nitrocellulose membrane filter (0.45 µm, Sartorius, Göttingen, Duitsland) in two times, after which each filter was placed on a different medium. For the surface sampling of the nests, swabs were taken on the inside of the nests, (floor and walls) and inserted in 1 mL of Amies liquid (eSwab, Copan, Menen, Belgium). Following a 1 min vortexing step, 100µL of each suspension was seeded onto the two media.
The MC medium was used to determine the total number of A. fumigatus isolates. Fungal colonies on MC + T were isolated and identified for further analysis to detect potential azole-resistant A. fumigatus (ARAF) isolates. The microscopic and macroscopic features of each colony were used to identify every A. fumigatus isolate. Matrix-assisted laser desorption ionization-time of flight mass spectrophotometry (MALDI-TOF MS) was used to confirm the identity of the suspected ARAF isolates that grew on MC + T [53].

Clinical Incidence of Avian aspergillosis in Humboldt Penguins
Aspergillosis incidence in the Humboldt penguins in the zoo was assessed for the period 2017-2022. Necropsy was performed on all dead animals older than 2 months with suspected aspergillosis ( Figure 5). Confirmed cases of pulmonary aspergillosis were verified by light microscopy using a lactophenol cotton blue stain and culture of A. fumigatus of the lungs and/or the surface of the air sacs.

Broth Microdilution Antifungal Susceptibility Testing and cyp51A Sequencing
All strains able to grow on MC + T were tested by broth microdilution method according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines [54]. Briefly, a cell suspension of 1-5 × 10 6 CFUs (colony-forming unit) per ml was prepared in 10 mL of saline water (8.5 g/L NaCl) from a 5 day old subculture on Sabouraud chloramphenicol agar tube. Subsequently, 1 mL of the cell suspension was added and mixed with 10 mL of RPMI-1640 medium (Sigma-Aldrich, Saint-Louis, MO, USA). A total of 100 µL of this cell suspension was added to each well of a 96-well plate containing 100 µL of serial dilutions of the antifungals and a control. The plates were in-

Broth Microdilution Antifungal Susceptibility Testing and cyp51A Sequencing
All strains able to grow on MC + T were tested by broth microdilution method according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines [54]. Briefly, a cell suspension of 1-5 × 10 6 CFUs (colony-forming unit) per ml was prepared in 10 mL of saline water (8.5 g/L NaCl) from a 5 day old subculture on Sabouraud chloramphenicol agar tube. Subsequently, 1 mL of the cell suspension was added and mixed with 10 mL of RPMI-1640 medium (Sigma-Aldrich, Saint-Louis, MO, USA). A total of 100 µL of this cell suspension was added to each well of a 96-well plate containing 100 µL of serial dilutions of the antifungals and a control. The plates were incubated at 35 • C ± 1 • C for 48 h. The minimal inhibitory concentration (MIC) of four medical azoles (itraconazole (ITC), voriconazole (VRC), posaconazole (POSA) and isavuconazole (ISA)) was determined on suspected ARAF strains. The MIC was determined visually as the lowest concentration of antifungal drugs causing complete inhibition of fungal growth. Candida krusei (IHEM 9560 = ATCC 6258), Candida parapsilosis (IHEM 3270 = ATCC 22019) and A. fumigatus (IHEM 28944 = ATCC 204305) were used as quality control strains. Azole resistance was defined according to the EUCAST clinical breakpoints (v10.0) [55], for ITC and VOR with MIC > 1 mg/L, POSA MIC > 0.25 mg/L and ISA MIC > 2 mg/L).

Genotyping
A selection of environmental and clinical strains were analyzed by microsatellite polymorphism genotyping using three multiplex PCRs. A total of nine microsatellite markers consisting of di-, tri-, or tetranucleotide short tandem repeats (STR) were used [56]. Fungal DNA extraction was performed by freeze-drying the cultures and mechanically breaking the cells by bead-beating. DNA was then extracted using the ZR Fungal/Bacterial DNA MiniPrep Kit (Zymo Research) following the manufacturer's instructions. Genotyping was performed by Genoscreen (Lille, France) using the PCR conditions described by De Valk et al. [56] with the fluorophores 6FAM/HEX/NED. The size of the amplicons was determined with a ABI 3730XL genetic analyzer using the GeneScan 500 ROX size standard (ABI) and the GeneMapper v5.0 software. The size of each microsatellite fragment was measured to determine the number of repetitions for each marker according to de Valk et al. [56]. All results are reported as repeat numbers. The relatedness of the strains was estimated by a minimum spanning tree analysis in Bionumerics 8.0 (Applied Maths, St-Martens-Latem, Belgium). The discriminatory power of the microsatellite markers was calculated using the Simpson index of diversity (Hunter 1990).

Statistics
Statistical analysis was performed to achieve global comparison between each study time point, using the non-parametric Friedman test. The statistical and graphics software R (version 4.2.0) was used. The significance level was set at p-value < 0.05.  Data Availability Statement: All available data are displayed in this paper.

Acknowledgments:
The authors thank the laboratory experts for their help. They are also grateful for the valuable advice from Agustin Resendiz-Sharpe, Dirk Stubbe and the statistical analysis by Inge Van Damme.