Antibiotic Resistance and Molecular Epidemiological Characteristics of Streptococcus agalactiae Isolated from Pregnant Women in Guangzhou, South China

Streptococcus agalactiae colonization in pregnant women can cause postpartum intrauterine infections and life-threatening neonatal infections. To formulate strategies for the prevention and treatment of S. agalactiae infections, we performed a comprehensive analysis of antibiotic resistance and a molecular-based epidemiological investigation of S. agalactiae in this study. Seventy-two S. agalactiae strains, collected from pregnant women, were subjected to antibiotic susceptibility tests; then, the screened erythromycin and clindamycin nonsusceptible isolates were used for macrolides and clindamycin resistance genes detection, respectively. Detection of resistance genes, serotyping, and determination of virulence genes were performed by polymerase chain reaction. The clonal relationships among the colonized strains were evaluated by multilocus sequence typing. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) mass peak analysis was performed to discriminate the specific sequence types (STs). In our study, 69.4% and 47.2% of the strains were nonsusceptible to erythromycin and clindamycin, respectively; the multidrug resistance rate was 66.7%. All erythromycin nonsusceptible strains harbored resistance genes, whereas only 52.9% of the clindamycin nonsusceptible strains possessed the linB gene. Erythromycin resistance was mainly mediated by the ermB or mefA/E genes. Four serotypes were identified, and the most common serotype was serotype III (52.8%), followed by Ib (22.2%), Ia (18.0%), and II (4.2%). All the strains were divided into 18 STs that were assigned to nine clonal complexes. Most of the major STs were distributed into specific serotypes, including ST19/serotype III, ST17/serotype III, ST485/serotype Ia, ST862/serotype III, and ST651/serotype III. Analysis of virulence genes yielded seven clusters, of which bca-cfb-scpB-lmb (61.6%) was the predominant virulence gene cluster. Among all ST strains distributed in this region, only the ST17 strains had a mass peak at 7620 Da. The outcomes of this study are beneficial for the epidemiological comparison of colonized S. agalactiae in different regions and may be helpful for developing the strategies for the prevention of S. agalactiae infection in Guangzhou. Furthermore, our results show that MALDI-TOF MS can be used for the rapid identification of the ST17 strains.


Introduction
Streptococcus agalactiae is a microorganism that commonly colonizes in the gastrointestinal and genitourinary tracts of healthy women [1,2]. S. agalactiae colonization in pregnant women is the main reason for intrauterine infection or transmission during parturition [3], which can cause serious neonatal infectious diseases, such as pneumonia, sepsis, and meningitis [4,5]. Many countries have already issued guidelines for the prevention of maternal and neonatal S. agalactiae infections [6][7][8]. ese guidelines generally recommend that pregnant women with S. agalactiae colonization should receive intrapartum antibiotic prophylaxis (IAP) to reduce the incidence of these infections [9,10]. Currently, there are no official guidelines for S. agalactiae screening and prevention in China [5,11]. However, in recent years, S. agalactiae has become an important pathogen in Chinese obstetrics and neonatology [12]. In mainland China, the S. agalactiae colonization rate of pregnant women ranged from 3.7 to 14.5%, and the incidence of invasive neonatal S. agalactiae diseases was 0.55-1.79 per 1000 live births, with a case fatality risk ranging from 6.5 to 7.1% [13]. e first-line antibiotic for IAP is penicillin [9,10]; however, since some people are allergic to it, macrolides (e.g., erythromycin) and lincosamides (e.g., clindamycin) are used as the alternative antibiotics [14]. Unfortunately, the resistance of S. agalactiae to macrolides and lincosamides is increasing worldwide [4,9,11,15,16]. Previous studies have shown that the ermB, ermTR, and mefA/E genes are involved in the resistance to macrolides [17], and the linB gene has been linked to the resistance to lincosamides and moderate sensitivity to macrolides [18].
Capsular polysaccharide (CPS) is an important virulence factor of S. agalactiae, resulting in invasive infections [16,19]. According to the different antigenicity of the CPS [20,21] or by the multiplex polymerase chain reaction (PCR) assay [9,22], S. agalactiae can be divided into 10 serotypes (Ia, Ib, and II to IX). Five of them (Ia, Ib, II, III, and V) are the predominant types of S. agalactiae that cause invasive infections [16,23]. e understanding of the serotype distribution is pivotal in developing CPS-based polyvalent vaccines [16,21]. In recent years, there have been a few reports on the serotype distribution of colonized S. agalactiae in China [11,16,24,25]. Although there are some regional differences in the serotype distribution of S. agalactiae in these studies, similar common serotypes (III, Ia, V, and Ib) were observed. Several studies have shown that the virulence of S. agalactiae depends on not only the CPS, but also the surface-exposed bacterial proteins [26][27][28][29]. Several molecular typing methods have been applied to epidemiological studies, such as pulsed-field gel electrophoresis (PFGE), random amplification of polymorphic DNA, and multilocus sequence typing (MLST), of which MLST is a robust and convenient method that results in exchangeable data in different labs [30]. e combination of MLST and serotyping is a valuable method for S. agalactiae typing. For instance, the ST17 strain, belonging to serotype III, was found to be "hypervirulent" S. agalactiae, apparently associated with neonatal invasive diseases [31,32]. Recently, whole-genome sequencing (WGS) has been used in genomic research of S. agalactiae [9,33], such as accurate serotyping [34], in silico virulence investigation and antimicrobial susceptibility testing [33,35], and speciation and evolutionary analysis [36]. However, at present, due to the high cost of WGS, it is not suitable for large sample analysis [34].
In recent years, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), described as a "revolution in clinical microbiology" technology [32], has been widely used for bacterial identification based on protein fingerprints [37,38], due to its accuracy, speed, and high throughput. is technology can also find characteristic proteomic biomarkers in some phylogenetic lineages of species and subspecies [38,39]. For example, it was reported that "hypervirulent" S. agalactiae ST17 and emerging ST-1 clones could be rapid detected by MALDI-TOF MS [39]. e purpose of this study was to investigate the antibiotic resistance and resistance genes, serotype distribution, virulence, and genotyping of S. agalactiae, isolated from pregnant women in Guangzhou and establish rapid screening methods for some STs, which may be beneficial for the prevention and control of S. agalactiae infection.

Strains Collection and
Identification. 72 nonrepeating S. agalactiae strains were collected from pregnant women (35-37 weeks of pregnancy) in Guangdong Provincial Maternity and Children's Hospital, which is one of the major women's medical centers in Guangzhou, between January 2017 and December 2017. 1013 pregnant women were screened for colonized S. agalactiae during that time, and the carriage rate was 7.1% (72/1013). e vaginal-rectal swabs collected from pregnant women were inoculated in Todd-Hewitt broth (bioMérieux, France), and the broth was incubated at 37°C in a 5% CO 2 environment for 24 h. Afterwards, the samples were subcultured on Columbia agar, containing 5% sheep blood (bioMérieux, France), and the agar was incubated in the same conditions for 18-24 h. e suspected isolates were initially identified according to the following points: β-hemolysis (or nonhemolysis), colony morphology, and CAMP test [6]; the presumed S. agalactiae isolates were then confirmed by MALDI-TOF MS (VITEK ® MS, bioMérieux, France).

DNA Extraction.
e pure S. agalactiae cultures were harvested on Columbia agar, containing 5% sheep blood. e genomic DNA was extracted from each isolate using the MiniBEST Bacteria Genomic DNA Extraction Kit (TaKaRa, China) according to the manufacturer's instructions.

Serotyping.
Serotyping by the multiplex PCR assay, as previously reported by Imperi et al. [22], was carried out for each isolate of S. agalactiae in this study. ose strains, which could not be classified into any serotype, were designated as nontypeable (NT).

MLST.
Seven housekeeping genes (adhP, pheS, atr, glnA, sdhA, glcK, and tkt) were amplified to implement MLST, as described previously [26]. e obtained products were purified and sequenced using sanger sequencing by the Sangon Biotech (Shanghai) Co., Ltd. e allele numbers and STs were obtained using online data (http://pubmlst.org/ sagalactiae/). e phylogenetic tree based on the concatenated sequences of the seven loci was constructed by the MEGA6 software using the UPGMA method. e relationships among different STs were estimated using eBURST version 3.1 (http://eburst.mlst.net/v3/). Note that only those shared six identical alleles were defined as a clonal complex (CC) [40,41]. Additionally, to understand the location of the STs for this study, the eBURST diagram of the S. agalactiae population was constructed using all STs found around the world, which are reported in the international database (http://pubmlst.org/sagalactiae/) (see Table S1 in the Supplementary Materials) as of August 2018. . e D-zone test was carried out for the strains nonsusceptible to erythromycin but susceptible to clindamycin. Streptococcus pneumoniae ATCC 49619 was used as a quality control strain to ensure the credibility of the results. Multidrug resistance of S. agalactiae was defined as acquired nonsusceptibility to at least three classes of antibiotics as previously described [42,43].

Determination of Macrolides and Lincosamides Resistance Genes.
e macrolides resistance genes ermB, ermTR, and mefA/E were tested by PCR in all erythromycin nonsusceptible isolates, while the lincosamides resistance gene linB was detected in each clindamycin-resistant strain. e primers that were used and the conditions in which the reactions were performed were as described previously [17,44,45].

Detection of Virulence Genes.
Five major virulence genes that encode the surface-proteins, including toxins CAMP factor (cfb), α-(bca) and β-subunits (bac) of C protein, C5a peptidase (scpB), and laminin-binding protein (lmb), were analyzed using PCR. e specific experimental procedures used in the present study have been published earlier [29].

Characteristic Mass Peak
Analysis. MALDI-TOF MS characteristic mass peak analysis was performed as briefly described below. First, single colonies from the overnightcultured isolates were picked and coated duplicate on a target plate (VITEK ® MSDS, bioMérieux, France). α-Cyano-4-hydroxycinnamic acid (CHCA; bioMérieux, France) (1 μL) was added to the tested samples, and the target plate was left to dry. e RUO Mode (VITEK ® MS RUO, Shimadzu, France) was then calibrated with Escherichia coli ATCC 8739 strain, and the mass spectrometry data of each strain were collected in a positive linear mode within a mass range from 3,000 Da to 20,000 Da [46]. e data were imported to SARAMIS premium software (bioMérieux, France), and the comparison of the mass peaks was carried out by Lanchpad software (Shimadzu Biotech, USA).

Statistical and Data Analyses.
Fisher's exact test was used to evaluate the differences in antibiotic sensitivity and distribution of resistance genes among different serotypes. e relationships between CCs, serotypes, and related virulence gene profiles were analyzed by a correspondence analysis. All statistical analyses were performed with SPSS software version 22.0. A P-value <0.05 was considered statistically significant.

Ethical Statement.
is study protocol was in accordance with the Helsinki Declaration of 1964. Data collected in this study did not include the information about patients, which was exempted from the formal medical ethical approval by the ethics committee of the hospital where the research was carried out (ZM2016-280).
Using the eBURST tool, a population snapshot was constructed to show the clusters of all known STs (number of STs � 1311; see Table S1 in the Supplementary Materials) in the entire S. agalactiae MLST database, and the relationship between the STs classified in this study and worldwide prevalent STs (see Figure 2) was elucidated. By eBURST analysis, 18 STs were assigned to 9 CCs (CC8, CC17, CC19, CC23, CC24, CC61, CC103, CC328, and CC485), with CC19 being the most prevalent (45.8%). e eBURST analysis displayed that CC17, CC61, CC103, CC328, and CC485 were close to CC19. Interestingly, CC19 and CC8 were close to CC1, while CC24 was close to CC23.

Antibiotic Susceptibility and Resistance
Genes. All isolates were sensitive to penicillin, ampicillin, vancomycin, and linezolid. As shown in Table 1    susceptible to clindamycin, and eight of them were positive for the D-zone test. e macrolides resistance genes were determined for 50 isolates of erythromycin nonsusceptible S. agalactiae, and at least one macrolides resistance gene was detected in each strain.

Characteristic Mass Peak Analysis. MALDI-TOF MS
analysis revealed the differences in the mass peaks among the different STs. Relative to the other STs, all ST17 strains (n � 8) were identified at the mass peak of 7620 Da (see Figure S1 in the Supplementary Materials). In other words, the mass peak of 7620 Da was the characteristic mass peak of the ST17 strains.
3.6. Relationships among the Various Molecular Characteristics of S. agalactiae Strains. According to the UPGMA dendrogram, the S. agalactiae strains isolated from pregnant women in Guangzhou were divided into two main branches (see Figure 2), namely, CC1 and CC23 (see Figure 1), which were further divided into different subbranches, representing the different STs (see Figure 2). Interestingly, some STs were closely related to specific serotypes (see Figure 2). For instance, the ST17, ST651, and ST862 strains were attributed to serotype III, while the ST485 and ST885 were strains classified as serotypes Ia and II, respectively. e correspondence analysis indicated that there was a significant correspondence between the CCs and virulence gene profiles (χ 2 � 106.52, P-value <0.001) and the serotypes and virulence gene profiles (χ 2 � 51.61, P-value � 0.001) (see Figure 3). For example, the bca-cfb-scpB-lmb profile was distributed in CC17 and CC23, the bca-cfa-bac profile was distributed in CC103, the bca-cfb profile was distributed in serotype Ia, and the bca-scpB-lmb profile was distributed in serotype II.
A systematic review and meta-analysis revealed that the carriage rate of S. agalactiae in pregnant women was 10% (95% CI 8,12), with 7% in Asia and 19% in non-Asian countries [24]. In this study, the carriage rate of S. agalactiae in pregnant women in Guangzhou was found to be 7.1%. Although the carriage rate was not obviously different from the average rate in Asia, it is necessary to investigate the prevalence of S agalactiae in different regions of China, considering the huge population of China and limited research data of colonized S. agalactiae. Previous reports have shown differences in the phenotypic and genotypic characteristics of colonized S. agalactiae from different geographical regions [16,28,47]. Accordingly, determining the population structure of colonizing isolates (e.g., genetic diversity, virulence factors, and antibiotic resistance) is the key to understand S. agalactiae disease in a region [9]. e present study is the first comprehensive investigation of the molecular epidemiological characteristics of colonized isolates in pregnant women in Guangzhou, South China.
Serotyping of S. agalactiae is crucial in determining the pathogenicity of the isolates [48]. In this study, four serotypes (Ia, Ib, II, and III) were identified by molecular    (2), CC103 (1), and CC485 (6) bca-cfb-bac 7 (9.7%) III (7) CC19 (1) serotyping. Overall, the three primary serotypes among the colonized isolates in Guangzhou were III, Ib, and Ia (93.0%), which were also reported as predominant serotypes in other Chinese cities, such as Dongguan (85.7%) [11] and Beijing (66.1%) [16]. However, the predominant serotypes in many other areas in other reports, including serotypes Ia, II, and Ib in Brazil (65.4%) [49] and serotypes III, V, and Ia in Toronto (67.0%) [9] and Shanghai (79.6%) [21], are different from those in our study. It has been reported that the colonization of serotype III isolates in pregnant women is a high risk factor for neonatal infections [12,21]. Notably, the proportion of the most common serotype (serotype III) in our region (52.8%) was similar to that reported in adjacent Dongguan (54.9%) [11], but apparently higher than that reported in Shanghai (East China, accounting for 35.9%) [21] and Beijing (North China, accounting for 32.1%) [16]. ese data suggest that the prevalence of serotype III isolates in South China may be higher than that in other parts of China.
e CPS of S. agalactiae is not only a basis for serotyping, but also an important target for vaccine research. In the development of conjugated multivalent vaccines, only five serotypes (Ia, Ib, II, III, and V) were covered [50,51]. Fortunately, all isolates in this study, except two NT strains, were hypothetically covered by a vaccine combination containing polysaccharide conjugates of the five serotypes.
NT strains have also been reported in previous studies [2,21,52] and may be related to mutations in the CPS locus [9,53].
Molecular epidemiological studies using MLST reveal the characteristics of the distribution of colonized S. agalactiae lineages. In this study, 18 distinct STs were identified, and these were then grouped into nine CCs. e colonized isolates from Guangzhou show relatively higher genetic diversity. Nonetheless, the three prevalent CCs, that is, CC19, CC485, and CC17, accounted for 73.6%. CC19 is the most common CC among asymptomatic pregnant women in our region, a finding similar to the global trend [2,14,47]. Interestingly, MLST analysis revealed that some bovine dominant subtypes had begun to spread in humans. ST485 was previously thought to be undetectable in humans, but recently, it has dramatically increased in pregnant women in China (2.5%-14.13%) [16,33,54]. In addition, some studies have shown that ST485 is highly pathogenic and may cause threatened abortion, premature rupture of membranes, and early-onset diseases (EOD) [33,55]. It is worth noting that ST485 has become one of the leading STs in our region (accounting for 8.6%). Although all subtypes of CC61 and CC67 were previously reported as bovine strains [47,53], ST929, which can cause threatened abortion in pregnant women, was first reported as a new subtype of CC67 in 2017  Canadian Journal of Infectious Diseases and Medical Microbiology [33]. Furthermore, 2.8% of the colonized isolates in our study belong to ST929. e eBURST population snapshot (as of August 2018, number of STs � 1311) (see Figure 2) showed different CCs related to ST485 and ST929, which were previously grouped into CC103 and CC67, respectively [33]. is difference can be explained by the high genetic diversity observed and the dynamic nature of S. agalactiae in the MLST database [53].
One of the major objectives of this study was to investigate the gene-based antibiotic resistance of colonized S. agalactiae. In the current study, all strains were sensitive to penicillin, which was the first-line antibiotic for the treatment and prevention of S. agalactiae infection [10,11]. Nevertheless, it is worth noting that our region has a high proportion of MDR strains. Erythromycin and clindamycin have long been considered as effective alternatives in IAP for penicillin-allergic women [9,17]. However, the resistance rates for erythromycin and clindamycin in our region were 66.7% and 47.2%, respectively. Recently, such high resistance rates have also been observed in other Chinese cities, such as Beijing [16], Shanghai [29], and Dongguan [11]. Overall, the resistance rates for erythromycin and clindamycin in China were higher than those in other regions [2,9,18,56]. erefore, if erythromycin or clindamycin is used as an alternative antibiotic, antimicrobial susceptibility test should be performed in advance. Erythromycin resistance is usually associated with three macrolide resistance genes: ermB, ermTR, and mefA/E [17,57]. Although the erythromycin resistance rates of S. agalactiae in different Chinese cities were similar, the resistance mechanisms were not the same. In this study, erythromycin resistance was mediated mainly by the ermB or mef A/E genes, and no isolates were found to carry the ermTR gene alone. e erythromycin-resistant strains carrying resistance genes in our study were different from those in previous reports [57]. e linB gene, encoding a lincosamide nucleotidyltransferase, confers moderate sensitivity to erythromycin and resistance to clindamycin [17,18]. However, the linB gene was detected in two erythromycin-sensitive strains. is may be because the linB gene was not expressed, but further research is needed to confirm this inference.
In the current study, five main virulence genes of S. agalactiae, which encode the surface proteins that involved in adhesion, invasion, or immune evasion [58], were investigated. Similar to previous reports [27,28,49,59], our report showed that the cfb, scpB, and lmb genes were distributed in most strains. However, the prevalence of the bca and bac genes in our study was different from that in previous reports. All strains in our study harbored the bca gene, indicating that the number of bca-positive strains was significantly higher in our region than that reported in other regions [29,49,59]. Our results indicated that the bac gene was identified only in serotype III (Table 2). is result was different from those reported around 2000, which showed that the bac gene was predominantly present in serotypes Ia, Ib, and II [27,59]. ese results suggested that the epidemiological distribution of S. agalactiae virulence genes might vary from period to period and region to region. Dmitriev A et al. used PFGE to show that the bac-positive strains were genetically homogeneous [59]. Instead, we found the that bac gene was distributed in different STs, suggesting that the bac-positive strains were genetically heterogeneous. e result of a previous study that used dotblot hybridization was consistent with our inference [27].
is indicated that when PFGE was used as the major method for determining diversity within a population, the measure of diversity was missed [27]. We found that the bcacfb-scpB-lmb profile was the most common gene cluster (Table 2), distributed in all serotypes and most CCs. Nevertheless, the correspondence analysis showed that the distribution of virulence gene clusters between the CCs and serotypes was different. e ST17 strain, belonging to serotype III, was defined as "hypervirulent" S. agalactiae, associated with early onset meningitis and late-onset diseases (LOD) in neonates [39]. With the implementation of extensive screening and IAP strategies, the incidence of EOD has been greatly reduced [12,39]. However, the morbidity rates of LOD remain unchanged [12,39], partly because the ST17 strains are particularly capable of persisting in the vaginal flora [39]. Detection of "hypervirulent" ST17 strains in vaginal samples or in neonates should permit the identification of a neonate population that presents with high risk for S. agalactiae infection [60], and in such a population, antimicrobial therapy and strict follow-up are absolutely necessary [60]. In this study, ST17 was one of the most prevalent STs, and all ST17 strains belonged to serotype III. It is necessary to establish a rapid screening method for ST17 strains in pregnant women in this region. MALDI-TOF MS can rapidly distinguish among some bacterial subtypes using characteristic mass spectrum peaks [37]. Previous studies using MALDI-TOF MS (Bruker Daltonics) confirmed the characteristic mass spectrum peak of the ST17 strain at 7625 Da [2,39]. However, we found this peak at 7620 Da using VITEK ® MS. We speculated that the characteristic mass peaks of different detection systems are slightly different. ST106 was also found to harbor this peak [2], yet we did not identify this ST in this study. We hypothesized that MALDI-TOF MS is suitable for the rapid screening of the ST17 strains in our region.
With the continuous reduction in the cost of WGS and the rapid development of bioinformatic infrastructures [34], it is expected that WGS will be performed in the future to reveal more subtle characteristics of S. agalactiae. is study was limited in its small sample size, single-center design, and retrospective nature. Our next experiment will include an increased sample size. Overall, this study added new information on the antibiotic resistance and molecular characteristics of colonized S. agalactiae in China.

Conclusion
e data obtained in this study indicated that the main molecular epidemiological characteristics of colonized S. agalactiae in Guangzhou are similar to those in other regions. However, some regional characteristics are also shown, especially relatively high prevalence of highly pathogenic ST485 and ST929. Moreover, the high carrying 8 Canadian Journal of Infectious Diseases and Medical Microbiology rate of the bca gene and low carrying rate of the bac gene are also prominent features. is study facilitated the epidemiological comparison of different regions and the prevention of postpartum intrauterine and neonatal infections. Furthermore, due to the high multidrug resistance rate of the colonized strains in our region, anti-infective treatment based on antibiotic resistance monitoring is necessary. Finally, we hypothesized that MALDI-TOF MS is suitable for the rapid screening of the ST17 strains in Guangzhou.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request. Disclosure e funding bodies played no role in the research design and implementation or dissemination of findings of the study.