Antimicrobial susceptibility, serotype distribution, virulence profile and molecular typing of piliated clinical isolates of pneumococci from east coast, Peninsular Malaysia

Pilus has been recently associated with pneumococcal pathogenesis in humans. The information regarding piliated isolates in Malaysia is scarce, especially in the less developed states on the east coast of Peninsular Malaysia. Therefore, we studied the characteristics of pneumococci, including the piliated isolates, in relation to antimicrobial susceptibility, serotypes, and genotypes at a major tertiary hospital on the east coast of Peninsular Malaysia. A total of 100 clinical isolates collected between September 2017 and December 2019 were subjected to serotyping, antimicrobial susceptibility test, and detection of pneumococcal virulence and pilus genes. Multilocus sequence typing (MLST) and phylogenetic analysis were performed only for piliated strains. The most frequent serotypes were 14 (17%), 6A/B (16%), 23F (12%), 19A (11%), and 19F (11%). The majority of isolates were resistant to erythromycin (42%), tetracycline (37%), and trimethoprim-sulfamethoxazole (24%). Piliated isolates occurred in a proportion of 19%; 47.3% of them were multidrug-resistant (MDR) and a majority had serotype 19F. This study showed ST236 was the most predominant sequence type (ST) among piliated isolates, which was related to PMEN clone Taiwan19F-14 (CC271). In the phylogenetic analysis, the piliated isolates were grouped into three major clades supported with 100% bootstrap values. Most piliated isolates belonged to internationally disseminated clones of S. pneumoniae, but pneumococcal conjugate vaccines (PCVs) have the potential to control them.

Antimicrobial susceptibility profile. Among beta lactam antibiotics, non-meningitis susceptibility breakpoint was used, whereby penicillin resistance was observed in four (4%) isolates and cefotaxime resistance in another isolate; all were from blood. Since data on the disease status of the patients were not accessible, only isolate from cerebrospinal fluid (CSF) was associated with meningitis, while isolates from other isolation sites were assumed as non-meningitis, but none of the isolates in this study were from CSF. For the other antibiotics, the majority of isolates were resistant to erythromycin (n = 42; 42%), tetracycline (n = 37; 37%), and trimethoprim-sulfamethoxazole (n = 24; 24%), with 18% (18/100) of the total collection being MDR. All isolates were susceptible to vancomycin. The distribution of antimicrobial susceptibilities for all isolates in this study is shown in Fig. 1.

Distribution of serotypes.
Seventeen different serotypes with six different serogroups were determined.

Scientific Reports
Phylogenetic analysis. All nucleotide sequences of the seven housekeeping genes for the respective isolates were aligned in the specified order 12 and subjected to phylogenetic analysis among all the 19 piliated isolates (labelled with T followed by numbers) and 11 reference sequences from the MLST database (labelled with REF followed by identity number). The tree was basally rooted to strain numbers T106 and T110, assigned as clade III and branched out into two bigger clades I and II with 100% bootstrap confidence interval values at all branching (Fig. 2). Among the three clades, there was one major group consisting of 12 isolates from this current study (Clade I), while others were small clades comprising two and four isolates (clade II and clade III). The clades also showed clear segregation among the STs, where clade I represented predominantly ST236, followed by ST271, ST2648, ST320, and ST62; clade II for ST700, ST386, ST90, and ST11811; clade III for ST671 for both isolates; and a single ST15604 (novel) in its own lineage. With the exception of the novel ST, all reference sequences of the different STs were clustered according to their respective similar ST of the isolates in this study.

Discussion
Molecular studies involving pneumococcal isolates from the east coast states of Peninsular Malaysia are scarce. The current study is the first report to address piliated pneumococci in this eastern region. Based on the 2018 Malaysia economic report 13 , Terengganu has a lower gross domestic product (GDP) than some states on the west coast, but it is famous as a local and international tourist destination owing to its beautiful beach. This could affect the dynamics of pneumococcal dissemination, which is why the genetic background of pneumococcal isolates from this area is of interest. In accordance with a previous study 14 , our finding also showed sputum (40%) and blood (38%) as the most common sources with certain serotype distribution pattern. Meanwhile, a slight male predominance was noted in the study population. Nonetheless, statistical analysis showed that the isolation sites were not significantly correlated with gender, age, vaccine serotype, multidrug-resistance, and pilus genes. Subsequently, the analysis addressed the molecular aspects, particularly the piliated elements. Serotype 14 was the most predominant serotype among pneumococcal isolates in our collection. A previous study showed serotype 19F as the common serotype among isolates from different sites of isolation from various Malaysian hospitals 15 . On the other hand, the frequency of occurrence of serotype 19F was lower in this current study; it is not known whether this is the serotype pattern in the east or potentially a natural serotype shift despite the lack of pneumococcal immunization. The PCV has only been made available privately and is probably more broadly available and affordable in some developed states in the west. However, since 2020, it has been part of the national immunization program. Meanwhile, the current study showed that serotype 15B/C predominated among non-vaccine serotypes. A previous study reported that serotype 15B/C emerged as the predominant non-vaccine serotype and was widely disseminated 16 . Additionally, serotype predominance in relation to age group was observed in a previous study that revealed serotype 14 as the most common serotype in children less than 48 months of age in Russia 17 . This was in agreement with our finding that serotype 14 and 6A/B was predominant among ≤ 5 year-old group. Serotype 14 was also found to be the most prevalent serotype among invasive paediatric isolates in a recent study in Malaysia 7 .
The incidence of antibiotic resistant pneumococci has been increasing in Asia 18,19 . However, a lower resistance frequency was observed for penicillin (4%) in this current study, consistent with a previous report from Malaysia that also demonstrated a low prevalence of penicillin resistance at only 5.6% 20 . On the other hand, a higher prevalent frequency (26.8%) of penicillin resistance was reported in Korea 21 . Meanwhile, among the 100 isolates in this study, a high incidence of resistance was observed against erythromycin (42%), tetracycline (37%), and trimethoprim-sulfamethoxazole (24%). A similar trend was documented by a recent study in Malaysia, with resistance to both erythromycin and tetracycline being 42.9% 7 . Co-resistance against these antibiotics was the common MDR pattern among the isolates, which was also found in other Asian countries 22,23 . Nevertheless, the isolates remained largely susceptible to ceftriaxone, cefotaxime, and vancomycin, which were consistently reported in a previous study in Malaysia as well 7 .
The pneumococcal regiments of virulence genes tested in this study are commonly present in pneumococci worldwide, as also observed in this study with the exception of pili 24 . Pilus is a sex organ, but it has been suggested that pili additionally play a significant role in pathogenesis by promoting adhesion and colonization of host tissue, as well as cellular invasion to provide an extra advantage to the isolates 9 . In previous studies, PI-1 was found to occur in 16.6% -35% 25,26 and PI-2 in approximately 9.5-20% of pneumococcal isolates 27,28 . The prevalence of piliated isolates in this study was within the range, with only 19% of isolates possessing at least one of the pilus genes. In similar fashion, a previous local Malaysian study revealed 20.6% of isolates amplified for PI-1 and 14.0% for PI-2 29 . Interestingly, despite their lower occurrence, 6 out of 19 of the piliated isolates in this current study were collected from blood (31.6%). Likewise, a Canadian study reported that blood was the common sources for piliated isolates 30 . In addition, about 47.3% of piliated isolates in our study were also MDR, being resistant to erythromycin, tetracycline, and trimethoprim-sulfamethoxazole. Earlier evidence showed that MDR was strongly associated with piliated isolates 31 .
While serotype 14 was the most frequent in our study, serotype 19F was predominant, being associated with the presence of PI-1 alone and both PI-1 + PI-2. Similarly, a study in China found that serotype 19F was the most frequently associated with PI-1 32 . Overall, the majority of piliated isolates in our study belonged to PCV13 exhibiting serotypes 19F, 19A, 6A/B, and 14. This is in agreement with previous reports regarding a correlation between piliated pneumococci isolates and vaccine-serotypes, commonly 19F, 19A, 14, and 6A/6B, which are included in PCV13 25,33 . On the other hand, Regev-Yochay and colleagues reported the emergence of PI-1 piliated isolates of the non-vaccine type after the implementation of PCV7 in the USA 34 . This served to highlight the implications of PCV for the advent of a new clonal expansion, which could be possibly facilitated by the pili properties as well, especially in establishing successful colonization.
Although we characterized only 19 piliated pneumococcal isolates for MLST, we revealed 11 different STs, including one newly assigned ST in the present study. From the MLST perspective, PI-1 and PI-2 are frequently found in pneumococcal lineages that are related to international PMEN clones. ST236 was the most prevalent among piliated isolates in our collection and belonged to CC271. ST236 is also the dominant lineage in Southeast Asia and many parts of the world, but there are limited studies discussing the association of ST236 with the appearance of pilus genes. In our study, the majority of PI-1 alone and both PI-1 + PI-2 isolates associated with ST236 were related to PMEN clone Taiwan 19F -14 (CC271), at the same time being vaccine type (19F and 19A) with MDR phenotype. This is in concordance with a previous study in Japan that reported the incidence of ST236 associated with PI-1 isolates and the widely distributed multidrug-resistant serotype 19F clone Taiwan 19F -14 35 . Meanwhile, MLST analysis in our study showed that ST271 and ST320 of CC271, harbouring PI-1 alone and both PI-1 + PI-2 genes, were also associated with MDR clone Taiwan 19F -14. CC271 has been associated with the multidrug-resistant pneumococcal clone (Taiwan 19F -14), which has been described worldwide as carrying a piliated element 32 .

Scientific Reports
| (2021) 11:8220 | https://doi.org/10.1038/s41598-021-87428-z www.nature.com/scientificreports/ The molecular analysis of S. pneumoniae in this study revealed several genetic lineages, namely, ST90, ST386, and ST671, which were associated with PI-1 alone and the majority of them from the invasive site carrying the MDR phenotype; those STs were also related to the widely disseminating PMEN clones, including Spain 6B -2 (CC90), Poland 6B -20 (CC386), and Spain 9V -3 (CC156), respectively. A previous study reported that PI-1 was primarily associated with PMEN clones, such as Spain 6B -2 (CC156) and Taiwan 19F -14 (CC320), with all of them being vaccine-serotype covered by PCV13 31 . In our collection, the only PI-2 isolate was found in ST62 and associated with Netherland 8 -33 (CC53) clone. Zahner and colleagues also reported that PI-2 isolates from the invasive site in their US study were linked to Netherland 8 -33 clone (ST62) but had serotype 11A, which is a non-vaccine type 28 . This indicated that the presence of pili was attributed to homogenous clonal expansion and the majority of the MDR clones were occupied by those piliated strains.
The phylogenetic tree indicated a close genetic lineage among several piliated isolates, especially those carrying both pilus genes. The most dominant clade I consisted of isolates that exhibited mainly ST236, ST271, ST320, and ST2648 of CC271 carrying PI-1 alone and both PI-1 + PI-2. Most isolates from this clade showed a similar MDR pattern and vaccine-serotype as well. There were a few strains in the dominant clade I manifesting different serotypes, which could potentially be due to capsular switches of serotype 36 . A notable observation was two piliated isolates related to ST236 expressing serotypes 19A and 6A/B. The only PI-2 isolate in clade I had serotype 20, which was a non-vaccine type. Clade II consisting of PI-I alone isolates was the second major clade that represented the diversity of STs, including ST700, ST386, ST90, and ST11811. Most of the piliated strains from this clade were of serotype 6A/B, and only one isolate had serotype 19A. ST386 and ST90 isolates had the MDR phenotype. The emergence and expansion of the MDR serotype 6C-CC386 lineage in Brazil was reported after universal use of PCV 37 , while our study revealed serotype 6A/B for ST386, suggesting a potential capsular switch event. Such an observation suggested that pili might help this lineage to emerge after the selective pressure of the vaccination.
Meanwhile, clade III consisted of two ST671 isolates that had PI-I alone. They had similar serotype 14 and non-MDR and they were both from the invasive site (blood), which reflected their close association. The new ST15604 belonged to its own lineage in the phylogenetic tree, having PI-1 alone and serotype 8, which was a nonvaccine type. ST15604 could possibly emerge differently from others as it was located more basally and distinct in the tree structure. All in all, the analysis showed that the majority of the piliated pneumococcal isolates in this study displayed close genetic correlation due to being clustered together, with some exhibiting similar serotypes and STs. It can thus be deduced that they could have initially derived from a common origin and subsequently disseminated within the population.
Nevertheless, the phylogenetic analysis relies on mutation as per the MLST principle. Gene ddl that is used in the S. pneumoniae MLST scheme had been previously linked to sequence replacement due to the 'hitch hiking' effect. The reason for this is that it is located near the penicillin-binding protein 2b gene (pbp2b) that is wellknown with its mosaic structure due to sequence recombinational replacement leading to penicillin resistance 38 . Such a mutational event may also affect the nearby ddl gene to end up with similar sequence diversity. Our study did not take into consideration such potential sequence replacement in ddl gene that could affect the phylogenetic output due to supporting data limitations. In addition, because of research scarcity, the frequency of such ddl gene-diversity-related event is unknown. Further analysis utilizing whole genome sequencing is highly warranted to elucidate this matter and to provide a better understanding of the genetic organization and evolution of the piliated isolates.

Conclusion
This study carried out a phenotypic and genotypic analysis of piliated pneumococcal isolates at a major tertiary hospital on the east coast of Peninsular Malaysia. This tourist attraction area was chosen as studies on the topic in question were scarce there. We found that the pneumococcal pilus islet was associated with clonal spread involving many serotypes such as 19F, 19A, and 6A/B, which are mostly covered by PCVs. The predominant ST of the piliated isolates was ST236. This was linked to the clone Taiwan 19 -14. ST236, ST271, ST320, ST90, ST386, and ST2648 of piliated isolates, which have also been associated with a high frequency of MDR. Based on these findings, it can be suggested that global pneumococcal lineages have been disseminated in this area and pili could play a role in the spread of antibiotic resistant clones, as supported by the phylogenetic analysis as well. Fortunately, the vaccine-serotypes exhibited by these clones can be controlled by PCVs, but a potential serotype switch may rule out isolates from the vaccine coverage. This warrants continuous monitoring, particularly when the PCV has been fully implemented.

Materials and methods
Bacterial isolates. Clinical S. pneumoniae isolates were collected from the Microbiology Laboratory, Department of Pathology, Hospital Sultanah Nur Zahirah (HSNZ), Kuala Terengganu, Terengganu between September 2017 and December 2019. This is a tertiary hospital with more than 800 beds serving the capital city of the state of Terengganu and nearby areas on the east coast of Peninsular Malaysia. Each clinical isolate was obtained from a different individual from invasive (blood) and non-invasive sites (sputum, eye, pus, bronchial aspirate and swab of a non-sterile area). The ages of the patients ranged from one month to 82 years old. The information on disease and admission status of patients was not accessible and therefore was excluded from analysis. S. pneumoniae was confirmed by bile solubility and susceptibility to ethylhydrocupreine disc (optochin). S. pneumoniae ATCC 49,619 was included as a reference strain in all analyses.
Determination of capsular types. Pneumococcal capsular types were deduced by multiplex PCR using published primers recommended by the Center for Disease Control and Prevention (CDC) 47 . Primers were divided into six multiplex sets named A, B, C, D, E, and F, as previously described 29,48 . The cpsA gene found in all known pneumococcal serotypes was used as the positive control, while the 100 bp plus DNA ladder marker (GeneDirex) was used for molecular weight reference.
The internal fragments of seven housekeeping genes, namely, aroE, gdh, gki, recP, spi, xpt, and ddl, were amplified by PCR, as previously described 12 . Sequences were submitted to the MLST database (http:// spneu moniae. mlst. net) for the assignment of allelic profiles and sequence type (ST). New alleles and ST were submitted to the curator of the MLST website for verification. The PHYLOViZ software was used for assigning the isolates for clonal complexes (CC) defined as cluster sharing at least five out of seven alleles. ST profiles were inferred to Pneumococcal Molecular Epidemiology Network (PMEN) clones in the PMEN database (http:// www. sph. emory. edu/ PMEN/ pmen_ table1. html) to identify the close lineage of circulating clones. MLST phylogenetic analysis based on the seven housekeeping genes 12 was determined by using Molecular Evolutionary Genetics Analysis version 7 (MEGA7). Appropriate reference sequences ID 1592, 345, 6665, 4312, 13506, 40817, 30311, 19496, 38687, 32260, and 1254 for respective STs were retrieved from the MLST database and included in the analysis as control. The phylogenetic tree was constructed by MEGA7 using the maximumlikelihood method based on the Tamura-Nei model, while the reliability of the tree was estimated via bootstrap analysis with 1000 replicates. Statistical analysis. Chi-square was used to compare the demographic characteristics of the patients with phenotypic and genotypic variables of the pneumococcal isolates. Statistical significance was indicated by p < 0.05. Ethical considerations. Ethical approval for this study was granted by the Medical Research and Ethics Committee of the Malaysian Ministry of Health, National Medical Research Register (approval no. NMRR 17-1025-35696). The study used de-identified pneumococcal isolates collected at the microbiology laboratory of the hospital. Since it did not meet the definition of research involving human subjects, informed consent was not required for this study; only data related to isolation site, age, and gender of the patients associated with the isolates were provided and were not traceable to the sampled individuals. Additionally, all methods were carried out in accordance with the relevant guidelines and regulations.