Multilocus Sequence Typing Reveals both Shared and Unique Genotypes of Cryptococcus neoformans in Jiangxi Province, China

Cryptococcosis is a globally distributed infectious fungal disease. However, much remains unknown about its molecular epidemiology in many parts of the world. In this study, we analyzed 86 clinical Cryptococcus neoformans isolates from 14 regions in Jiangxi Province in south central China. Each isolate was from a different patient and 35 of the 86 (40.7%) patients were infected with HIV. All strains belonged to serotype A and mating type α (MATα). Genotyping based on DNA sequences at seven nuclear loci revealed eight sequence types (STs) among the 86 isolates, including two novel STs that have not been reported from other parts of the world. ST5 was the dominant genotype and our comparative analyses showed that these genotypes in Jiangxi likely originated by dispersal from other regions within and outside of China and/or mutations from another genotype within Jiangxi. Though none of the isolates was resistant to the five tested antifungal drugs (flucytosine, amphotericin B, fluconazole, itraconazole, and voriconazole), obvious differences in their minimum inhibitory concentrations were observed, even among isolates of the same ST. Our results suggest that continuous monitoring should be conducted to understand the changing dynamics of C. neoformans in this and other regions.

(MALDI-TOF MS) 7,8 . The isolates were then analyzed for their mating types, multilocus genotypes derived based on sequencing at seven nuclear loci, and susceptibilities to five antifungal drugs.
The taxonomy and systematics of the human pathogenic Cryptococcus have undergone multiple revisions and a broad consensus is still not available. In this study, we follow the commonly used approach that sub-divide the human pathogenic Cryptococcus into two species complexes, the C. neoformans species complex (CNSC) and the C. gattii species complex (CGSC). CNSC includes three serotypes (serotypes A, D, and AD) and four molecular types (VNI, VNB/VNII, VNIII and VNIV). This species complex is mainly responsible for cryptococcal infections in AIDS patients. CGSC has two serotypes (serotypes B and C), four molecular types (VGI, VGII, VGIII and VGIV), and is geographically relatively limited and more commonly diagnosed in immuno-competent than immuno-compromised individuals [9][10][11][12] . Several molecular techniques have been employed for identifying the molecular types and/or genotypes for strains in these two species complexes, including polymerase chain reaction (PCR) fingerprinting, pulsed-field gel electrophoresis (PFGE), amplified fragment length polymorphism (AFLP), multilocus microsatellite typing (MLMT), repetitive-sequence-based PCR, restriction fragment length polymorphism of the URA5 gene (URA5-RFLP), and multilocus sequence typing (MLST) 13 . Among these methods, MLST has become the preferred method by the International Society for Human and Animal Mycology (ISHAM) 11 . The objectives of this study are to analyze the genotypes and antifungal drug susceptibilities of isolates causing cryptococcosis in Jiangxi Province and to compare their genotypes with those from other regions.

Materials and Methods
Cryptococcal Isolates. In this study, 86 isolates of C. neoformans were obtained from patients hospitalized in Jiangxi hospitals from January 2016 to November 2017. These patients came from all major regions of the Province, spanning ~600 km from the south (Ganzhou) to north (Jiujiang) and ~500 km from east (Leping) to west (Pingxiang). The detailed information about each of the samples is presented in Table 1. Request for the clinical isolates and patient information followed institutional guidelines of Nanchang University. The isolates were stored in skim milk at −80 °C until use and were maintained on SDA (Sabouraud Dextrose Agar, 1% yeast extract, 2% peptone, 2% glucose, 1.8% agar) medium at 25 °C during this study for genotyping and MIC testing. DNA Extraction. Genomic DNA was extracted from each isolate following the protocol described by Alessandro et al. 14 , with slight modifications. Briefly, cells were incubated on SDA agar containing 0.5 M NaCl at 30 °C overnight. Protoplasts were generated by incubating cells in 2 ml of urea buffer (8 M urea, 0.5 M NaCl, 20 mM Tris, 20 mM EDTA, 2% SDS (Sigma, USA), pH 8.0) for 3-4 h at room temperature. The protoplasts were collected by centrifugation and vortexed in 400 μl lysis buffer (1% w/v SDS in TE, pH 7.5). After vortexing, 400 μl of phenol-chloroform (1:1, pH 8.0) and 400 μl of 2-μm acid-washed glass beads were added and further vortexed. The mixes were centrifuged and the extracted DNA was washed with 100% ethanol, re-suspended in 100 μl TE, and then stored at −20 °C.

Identification of Species, Lineages, and Mating Types.
To confirm that the isolates all belonged to the human pathogenic Cryptococcus species complexes, we used the MALDI-TOF MS (BioMerieux, Marcy L'Etoile, France), following the protocols described in Mctaggart et al. using the on-plate protein extraction method 8 . Briefly, isolates were first cultured on SDA and incubated at 30 °C for 24 h. One single colony of each isolate was smeared onto each MALDI-TOF MS analysis plate, and the proteome of each isolate was extracted via 0.5 μl formic acid and 1.0 μl matrix liquid. The protein profile was automatically generated for those proteins with molecular weights ranging from 2 to 20 kDa. Escherichia coli ATCC8739 was used as a quality control.
To identify whether the isolates belonged to either the C. neoformans species complex or the C. gattii species complex, we plated all isolates on L-canavanine-glycine-bromothymol blue agar 15 , followed by sequencing of all isolates at the SOD1 gene 16 . The sequences were then compared with those of five strains representing the known molecular types of C. neoformans: WM148 (serotype A, VNI), WM 626 (serotype A, VNII), Bt63 (serotype A, Botswana), WM 628 (serotype D, VNIII), and WM629 (serotype AD, VNIV) as well as those in the GenBank and the ISHAM MLST database. Furthermore, the standard strains H99 (serotype A, MATα), JEC21 (serotype D, MATα), and JEC20 (serotype D, MATa) were used as references to determine the serotype and mating type for each of the isolates, following the methods described by Yan et al. using serotype and mating type-specific primers at the STE20 gene for PCR 17 . These primers target the STE20Aa, STE20Aα, STE20Da and STE20Dα alleles. After amplification, all the PCR products were electrophoresed on 1% agarose gels in 0.5xTBE buffer at 100 V for 60 min and then visualized under UV light by comparison with their reference strains.
MLST Analysis. Aside from obtaining the SOD1 gene sequence for each of the isolates for species identification, we also obtained the sequences at six other genes as suggested by the ISHAM consensus MLST scheme for C. neoformans and C. gattii 11 . Briefly, these six DNA fragments were located in the following genes CAP59, GPD1, IGS1, LAC1, PLB1, and URA5. Primers and PCR conditions followed that described in Hiremath et al. 18 . All sequences were submitted to the National Center for Biotechnology Information (NCBI) database to acquire GenBank accession numbers and the C. neoformans/C. gattii species complex database (http://mlst.mycologylab. com) to obtain sequence type (ST) numbers.
Phylogenetic Analysis. Phylogenetic analysis of the concatenated sequences of seven MLST loci was performed using MEGA version 7.0 software 19 . A phylogenetic tree was produced by the Neighbor-Joining algorithm using alignments of the concatenated sequences at the seven gene loci from our isolates and two reference strains H99 and WM148. we extracted all the published genotype information for all the Chinese isolates at the seven sequenced loci from the Cryptococcus MLST database (http://mlst.mycologylab.com). These Chinese populations were then analyzed using the GenAlEx software (version 6.5) 20 . Two analyses were performed. In the first, the overall genetic variation was partitioned into within and between geographic populations through AMOVA. In the second, the genetic differences between all pairwise geographic populations were analyzed. All regional populations with a sample size of greater than five isolates were included in the above analyses. Statistical significance of the observed genetic differences was determined by 1000 permutations using the GenAlEx software 20 .
Antifungal Susceptibility Testing. The in vitro antifungal susceptibility testing of all 86 isolates of amphotericin B (AMB), flucytosine (5FC), fluconazole (FLU), voriconazole (VOR), itraconazole (ITR) was performed using the ATB TM FUNGUS-3 kit (BioMerieux, Marcy L'Etoile, France). The minimal inhibitory concentrations (MIC) were determined following instructions provided by the User's Manual. Candida krusei ATCC6258 and Candida parapsilosis ATCC22019 were used as reference quality controls. The obtained MIC values were compared to those recommended breakpoints to determine whether the strains were susceptible or resistant to specific antifungal drugs. The MIC breakpoints for fluconazole and flucytosine were ≥16 μg/ml and ≥32 μg/ml respectively as suggested based on the User's Manual of ATB TM FUNGUS-3. For amphotericin B, we followed the resistance breakpoint of ≥2 μg/ml as suggested by CLSI document M27-A3 21 and Nguyen et al. 22 . At present, there are no consensus interpretive breakpoints of ITR and VOR based on the ATB TM FUNGUS-3 system for C. neoformans. Here we follow previous studies and used a MIC ≥ 1 μg/ml as the resistance breakpoint for both ITR and VOR 23,24 . Data availability statement. All the data described in this manuscript are presented in the paper (for genotype information and MIC values of all 86 isolates) as well as deposited in the publicly accessible database (http://mlst.mycologylab.com) for all nucleotide sequences.
Statements on study approvals. We confirm that all methods used in this study were carried out in accordance with relevant guidelines and regulations. In addition, all experimental protocols were approved by Nanchang University and that informed consent was obtained from all subjects for the Cryptococcus neoformans isolates analyzed in this study. and Ganzhou (ST31) in the south. The allelic assignments of our individual gene sequences in the MLST database for each of the eight multilocus sequence types are presented in Table 2. Table 2 also shows the distributions of the individual alleles in the Jiangxi population among all the known sequence types in the MLST database. In total, 18 alleles were found at the seven loci in Jiangxi. Two loci, SOD1 and URA5 were monomorphic in Jiangxi and their alleles (#1 at both loci) were distributed broadly in many other STs within and outside of China. The remaining five loci were polymorphic, with allele numbers ranging from two (CAP59) to six (GPD1). At each of the remaining five loci, the Jiangxi population shared alleles with a diversity of known STs from other geographic regions (Tables 2 and 3). Among these 18 alleles at the seven loci, none was specific to Jiangxi and all have been found elsewhere (Tables 2 and 3).
Among the eight multilocus STs in our Jiangxi population of C. neoformans, six (ST5, ST31, ST32, ST139, ST186, and ST359) have been reported previously from other geographic areas (Table 3). The remaining two genotypes (ST226 and ST319) have only been found in our study population. The geographic distributions of these eight STs are shown in Table 3. Of the six shared STs between Jiangxi and other regions, three (ST5, ST31, and ST32) have been found in multiple continents/countries. For example, ST5 has been reported from the US, Europe, Brazil, South Africa, and several countries in eastern and southeastern Asia. The high prevalence of ST5 in the Jiangxi population is consistent with what has been reported previously from other parts of China and China's neighbouring countries, such as Korea, Japan, and Vietnam. The remaining three STs (ST139, ST186, and ST359) were geographically unique, had been reported so far only from Africa, Shanghai and Sichuan Province in China, respectively. Phylogenetic Analysis. To further reveal the relationships among the isolates and genotypes, we conducted a Neighbour-joining analysis of the concatenated gene sequences at the seven MLST loci (Fig. 1). Here, only one representative strain of each of the eight sequence types was included in this analysis to allow better visualization. Two reference strains H99 and WM148, both of the VNI molecular type group, were also included (Fig. 1). Our analysis showed that ST5, ST186, and ST359 were genetically very similar, differ from each other by one to a few  nucleotides at one (GPD1) of the seven sequenced loci ( Fig. 1 and Table 2). Similarly, ST31 and ST32 were very close to each other, differed from each other by a few nucleotides at the PLB1 locus. Overall, these five STs formed a tight cluster with each other. In contrast, the other three STs (ST139, ST226, and ST319) were more distantly related to each other and to those five STs described above based on the concatenated gene sequences at these seven loci.  27 provinces/municipalities in China have been deposited in the database, including the 86 isolates from Jiangxi Province in our study. Among the 27 geographic populations, 12 had isolates of less than five each (most of these 12 populations had only 1-2 isolates each!) and these populations were excluded from our population genetic comparisons. The remaining 15 populations included a total of 364 isolates (Table 4). Our analyses revealed that overall, geographic separation contributed significantly to the total observed genetic variations of the Chinse population of C. neoformans. Specifically, AMOVA result showed that about 65% of the observed genetic variation were due to geographic separation while 35% was found within geographic populations (P < 0.001). Among the seven loci, five (GPD1, IGS1, LAC1, SOD1, and URA1) showed significant geographic differentiations while the remaining two (CAP59 and PLB1) showed no significant differentiations (detailed data not shown). Our further analyses identified that the observed genetic differentiations were mostly due to the genetic uniqueness of the population from Beijing (Table 4). Of the remaining 91(14 × 13/2) pairwise comparisons, only the Jiangxi-Sichuan populations showed statistically significant genetic differentiation (Table 4).

Relationships Among Geographic Populations of C. neoformans in China.
Antifungal Susceptibility. The antifungal drug susceptibility results are presented in

Discussions
In this study, we analyzed the genotypes and drug susceptibility profiles of 86 isolates obtained from across Jiangxi Province in China. Our analyses identified eight multilocus sequence types, with five of which represented by only one isolate each. Of the eight STs found in our sample from Jiangxi, six have been reported from other geographic regions while two were novel, identified so far only in Jiangxi Province. The dominant sequence type in Jiangxi, ST5, is a broadly distributed genotype and has been commonly found in other parts of China as well as in the Far East. These eight genotypes show several types of allelic and phylogenetic relationships. Our antifungal drug susceptibility test results showed that none of the 86 strains were resistant to the five tested antifungal drugs. However, some of the strains showed relatively high MIC values. Below we discuss the relevance of our results to earlier studies and the potential implications of these results to the management of cryptococcosis in Jiangxi Province.
Although a considerable amount of information exists on the epidemiology and molecular typing of C. neoformans strains in China, there is very little data on cryptococcosis from Jiangxi Province. Studies from the Chinese Mainland, Taiwan, and Hong Kong indicated that the prevalence of cryptococcosis in HIV/AIDS patients ranged from 12.9% to 24.7%, which is significantly lower than that of many other regions in the world 25 . While the total number of HIV-positive patients are not known in Jiangxi Province, HIV-positive patients account for over 40% of the sources of our strains in this study. In contrast, the percentage of isolates from individuals without obvious predisposing risk factors was significantly lower (9/86, 10.5%) than those reported before from other parts of China 26 , but more similar to those from regions outside of China. In Jiangxi Province, cryptococcal infection was more commonly found in middle-aged people, the main group with HIV infections in our samples, than in other age groups. Furthermore, unlike previous studies that found no prominent gender bias in the incidence of cryptococcosis in China 26 , our data showed that the male-female gender ratio was 2.3:1. The ratio in Jiangxi is similar to those reported from Brazil and Europe 27,28 , in which the male to female gender ratio was about 2.9:1.
Similar to observations from other parts of China and other Asian countries, such as Korea, Japan and Thailand [19][20][21][22][23][24][25][26][27][28][29][30][31][32] , our data showed that the 86 clinical isolates of C. neoformans from Jiangxi Province had relatively limited amount of genetic variation. All isolates were of the same mating type and the same genotype group VNI. The genotype group VNI is globally the dominant lineage of C. neoformans responsible for cryptococcosis 27,28 . A previous study by Fang et al. 25 . indicated that serotype A, molecular type VNI, and MATα strains of C. neoformans predominate HIV-negative patients in China. Our study suggests this genotype group also predominates the HIV-infected patients in Jiangxi.
To date, seventeen STs of C. neoformans var. grubii have been identified in China. They include ST5, ST31, ST38, ST53, ST57, ST63, ST93, ST186, ST191, ST194, ST195, ST295, ST296, ST359, and ST360 in Mainland China, while ST4 and ST6 are found in Hong Kong 13,[33][34][35] . In this study, eight STs (i.e. ST5, ST31, ST32, ST139, ST186, ST226, ST319, and ST359) were founded in Jiangxi Province and only four of these eight STs overlap with those reported from other parts of China. This result suggests that there is likely abundant unique genetic diversity of C. neoformans in Jiangxi Province. Among the shared STs between Jiangxi and outside of Jiangxi, the majority belonged to ST5, the most common ST in all East Asian countries where epidemiology data are available, including China, Japan, and South Korea 30,36,37 . Interestingly, two other STs, ST31 and ST32, found in Jiangxi were also broadly distributed. According to Khayhan et al. 33 , ST139 has so far been found only in Africa. However, its relative frequency in Africa is not known. The geographic distribution patterns of these six shared STs found in Jiangxi suggest that both long-and short-distance dispersals are common in C. neoformans. Consistent with this hypothesis, aside from the Beijing population, we found limited evidence for genetic differentiation between most pairs of geographic populations of C. neoformans in China. At present, the reason(s) for the genetic distinctiveness of the Beijing population is not known. However, as suggested previously 1,18 , a diversity of factors such as wind, animals such as pigeons, and anthropogenic activities could have contributed to the dispersals of genotypes between Jiangxi Province and other regions both within and outside of China.
In this study, all 86 cryptococcal isolates were susceptible to 5FC, AMB, FCA, ITR, and VRC. Our results suggest that the standard initial therapy for cryptococcosis, AMB combined with 5FC, should still work for patients in Jiangxi Province 38 . However, variations in MICs were observed among the isolates. For each of the drugs, there were isolates showing high MIC values. At present, there was no apparent relationship between MIC to any of the drugs and geographic origins and/or strain genotypes. We would like to note that some of the strains showed high MIC values to multiple drugs. Our results thus call for close monitoring of drug susceptibilities of cryptococcal strains in Jiangxi Province.
In conclusion, our study revealed both shared and divergent genotypes and patterns of cryptococcal epidemiology between Jiangxi Province and other parts of China. Specifically, in both Jiangxi and other parts of China,  ST5 was the predominant sequence type. In addition, both unique STs and evidence for long distance dispersals were found among most surveyed regions in China. However, different from previous studies in China, our results identified that most patients in Jiangxi Province with cryptococcosis had underlying risk factors associated with compromised immunity. At present, the mechanism for the predominance of ST5 in East Asian populations is not known. One possibility is that ST5 is more virulent than other sequence types to East Asians. Another possibility is that ST5 was the founder clone in East Asia that has adapted to the local ecological niches. Additional investigations are needed in order to test these possibilities.  Table 5. Susceptibilities of the 86C. neoformans isolates from Jiangxi Province against five common antifungal drugs. 1 ST: sequence type as determined based on the combined sequences at the seven loci. 2 5FC: 5-Flucytocine.