ITS, SOD, DHFR, DHPS, CYB and mt LSU rRNA genes of P. jirovecii in this study were amplified, cloned, sequenced and classified respectively. Meanwhile, we analyzed the combinations of the multilocus genotype and assessed the association between the genotypes and clinical conditions of patients. The correlations of the clinical characteristics such as the severity of the diseases and geographical origins with the multilocus gene sequences of the P. jirovecii were reported in the previous studies [2, 10, 12, 13]. However, it is difficult to accurately determining their association mentioned above based on the small sample sizes in this study.
The ITS region of P. jirovecii is the most polymorphic loci and widely used for genotyping . In this study, we assessed and analyzed the P. jirovecii ITS genotype of 3 clinical samples from Shanxi province according to the new, simplified genotype nomenclature system . We found ITS 4, ITS10 and ITS 59 genotypes, which were also identified in our previous study . Furthermore, ITS 59, the GenBank accession number is MK300661, was the second occurrence of ITS genotypes in China. ITS 2, ITS10 and ITS 16 were also found in this study. There were two BALF specimens were detected with more than one ITS genotypes, while only single ITS genotype was found in the remaining one patient. These findings showed that there were mixed infections with two different P. jirovecii strains in the 2 patients mentioned above, similar to the previous reports of the ITS genotyping of P. jirovecii in China [14, 15]. This study suggested that a PJP individual might be co-infected with more than one genotype of P. jirovecii. Furthermore, the genotypes of P. jirovecii ITS region plays important roles in understanding its epidemiology, transmission and pathogenesis [2, 15–17]. So far, the epidemiological studies on P. jirovecii ITS genotypes in Europe, Africa, America and other Asian countries showed that there have been at least 62 ITS genotypes reported in GenBank . Genotype 1 was the most common ITS genotypes based on worldwide studies [18–23]. The most prevalent genotype of ITS in Japan was Genotype 10 , which was observed in our study. In previous studies in China, Genotype 22 was most common in Beijing, Tianjin and Liaoning [2, 15], whereas, Genotype 1 in Guangzhou . However, it is difficult to accurately assess and statistically analyze the correlation between geographic diversity and P. jirovecii ITS genotypes based on the small sample size in this study. Notwithstanding the foregoing, our study identified and analyzed the ITS genotype of P. jirovecii isolates from Shanxi province in PR China for the first time. Meanwhile, the results in this study provided the partial epidemiological data of P. jirovecii ITS genotypes in China and further demonstrated the new ITS genotype nomenclature system previously reported is a simpler and useful tool to analyze and interpret.
SOD mutations of P. jirovecii were commonly observed at two different positions of nucleotide bases (i.e., 110 and 215) and genotypes SOD1 and SOD2 were detected most frequently in previously studies [12, 25–28]. Genotypes SOD3 and SOD4 were the second most common . In the present study, both SOD1 and SOD2 were observed, but other SOD genotypes were not. Mixed genotypes containing SOD1 and SOD2 were found in the isolate from the HIV-negative patient. Previous studies showed that the clinical outcomes of the patients infected with P. jirovecii SOD1 genotype were poor, indicating that the virulence of P. jirovecii might be closely correlated with genotype SOD1 [10, 12, 29, 30] .
The combination of trimethoprim and sulfamethoxazole (TMP-SMX), the first-line drug for treatment against PJP, is well tolerated and significantly effective . DHFR and DHPS are respectively targets of TMP and SMX. With the widespread use of TMP-SMX, the development of drug resistance by P. jirovecii has been emerged as a concerning problem. Several previous studies have demonstrated that treatment failures of PJP are closely correlated with the high mutation rates of DHPS and DHFR gene and the frequencies of mutations in DHPS gene are greatly higher than that in DHFR, indicating that the DHPS mutations are caused by the pressure of drug selection but not random occurrences [31, 32]. The DHPS wild-type (WT) sequence was more frequent displaying in P. jirovecii DHPS genotype, while WT and DHFR 312 synonymous mutations (nucleotides at position 312C) were common in DHFR genotype [7, 33, 34]. The synonymous mutations in DHFR genotypes are probably based on polymorphisms rather than the pressure for drug selection [1, 34]. Furthermore, the rates of DHPS and DHFR mutations were still very low [12, 35–37]. In this study, DHPS WT and DHFR 312C were observed at the DHPS and DHFR locus and no other mutations were detected, consistent with the prior studies reported in PR China [12, 34, 37]. This low prevalence of DHPS and DHFR mutations was not only demonstrated in China but also in other developing countries such as Brazil, South Africa and Thailand [38–40]. The reason for no mutations in DHPS and DHFR genes in this study were was probably due to the PJP subjects included without TMP-SMZ prophylaxis.
Atovaquone, a ubiquinone analog that inhibits electron transport at the cytochrome bc1 complex, targets the mitochondrial CYB gene. Atovaquone is effective for PJP prophylaxis in at-risk groups. Previous studies supported the point that CYB mutations in Q0 region (i.e., coenzyme Q binding site) are significantly associated with atovaquone [41–43]. In this present work, CYB mutations were not only including genotype CYB 1 and genotype CYB 2, which are the common genotypes documented in previous studies [7, 12], but also containing genotype CYB 7 and genotype CYB 8. However, the patients collected in our study had not received atovaquone for treatment or prophylaxis against PJP, implying a negative pressure for drug selection. Therefore, CYB mutations had no correlation with atovaquone in this study, but P. jirovecii mutations were detected, consistent with previous studies . Additionally, genotype CYB 1 and genotype CYB 8 were observed together in the patient without HIV, suggesting the presence of mixed infection.
We found that the Genotype 3 of mt LSU-rRNA gene was observed in the PJP patients with HIV in this study. A previous study in China suggested that the most common mt LSU-rRNA genotype was Genotype 3 (15/30) in AIDS-PJP patients , which were consistent with our study. Medrano F.J. reported that Genotype 3 was most commonly detected in AIDS-PJP patients, while Genotype 2 in the groups without HIV in Spain . In previous studies, Genotype 2 was most common in patients with chronic respiratory diseases and autoimmune diseases, similar to this, Genotype 2 was also observed in the patients with ILD in our study. This patient was co-infected with genotypes 1 and 2. These studies indicated that the virulence of Genotype 3 was stronger than other genotypes of mt LSU-rRNA gene and would be more likely to invade and damage the lung tissues among immunocompromised patients. Further studies are required to confirm the above findings. Additionally, unlike the studies mentioned above, Genotype 2 was the highest frequent genotype in patients with HIV, but Genotype 3 was the most common genotype in HIV-negative patients with malignant diseases in India . The predominant genotype of mt LSU-rRNA gene was Genotype 1 in the patients from Portugal, Spain and some other regions [22, 45]. Genotype 3 was the most common genotype in patients with HIV in eastern China  and in this study. These previous epidemiological surveys suggested that the distribution of P. jirovecii mt LSU-rRNA genotypes might be associated with geographical origins.
Moreover, coinfections with two genotypes of P. jirovecii in the same patient without HIV were detected for different genetic loci including ITS, SOD, CYB and mt LSU rRNA in this study, while coinfections in HIV positive patients were uncommon. These mixed-genotypes were confirmed by sequencing of plasmid clones and the suspected mutations were detected in at least 2 clones for each patient. The pulmonary concurrent infection with other fungi and bacteria was detected in the HIV-negative patient included in this work who was co-infected with multiple strains of P. jirovecii. Combined with some serum biochemical parameters especially 1, 3-β-D-Glucan (BDG) and lactate dehydrogenase (LDH) and clinical as well as laboratory characteristics and of the patients, it seems that the HIV-negative patient was more serious, suggesting that coinfections with multiple strains of P. jirovecii and other pulmonary pathogens were correlated with the severities of the diseases. Severe hypoxemia was presented in this non-HIV patient with PJP. According to the data of clinical studies, the clinical symptoms the PJP patients without HIV are commonly atypical, rapid and much more serious than the AIDS-PJP patients, with the poorer clinical outcomes even death [46–49]. However, it is unknown the varieties of clinical characteristics whether caused by the mixed genotypes based on the small sample sizes. Furthermore, despite the relatively high prevalence of coinfection with multiple P. jirovecii strains in humans were reported in vast majority of the strain variation studies [1, 2, 22, 24, 40, 45, 50], it is also unacknowledged that the occurrence mechanism, potential clinical outcomes and influences on P. jirovecii virulence and transmission. Further studies are needed to elucidate these unknown fields on coinfection.
We assessed the combinations of P. jirovecii genotype at ITS, mt LSU-rRNA, DHFR, DHPS, CYB, and SOD loci. This is the first time to analyze the multilocus genotypes of P. jirovecii isolated from Shanxi. The combination of SOD 1, DHFR 312C, DHPS WT, mt3 genotypes separately occurred in the two HIV-positive patients. A recent similar study indicated combination of SOD1, DHFR312T, DHPS WT, mt3, CYB1 genotypes is associated with the poorest prognosis . Except the DHFR 312 synonymous mutations and CYB1, the combination of SOD 1, DHFR 312C or DHFR 312T, DHPS WT, mt3 are possible to be far more common among HIV-positive populations. Polymorphic combinations of mt85A, SOD110C/215T, CYB C and mt85T, SOD110C/215T, CYB C in P. jirovecii were the most common genetic types previously reported, obviously correlated with more severe PJP cases and with high rates of mortality . The combination type of mt85A, SOD110C/215T, CYB C was observed in the HIV-negative-PJP case with a more serious clinical consequence, whereas, it was also detected in one AIDS-PJP patient in our study. Therefore, the larger-scale and more intensive studies on the genetic polymorphisms of P. jirovecii were needed to reveal its associations with clinical characteristics and outcomes.