Using the bca Gene Coupled with a Tetracycline and Macrolide Susceptibility Profile to Identify the Highly Virulent ST283 Streptococcus agalactiae Strains in Thailand

Abstract

Invasive infection by Streptococcus agalactiae (GBS) is a significant cause of death in newborn babies. In Thailand, data on strain distribution in GBS, specific virulence genes, and susceptibility patterns are limited. Therefore, our study aimed to establish the sequence type (ST) distribution and to use a specific virulence gene in combination with a susceptibility profile for strain identification. Non-duplicate 277 isolates of GBS were tested for ST, virulence genes, and antimicrobial susceptibility profiles. Twenty-five STs were detected. The ST283 (29.24%) and ST1 (27.07%) were the most common STs. The absence of the bca gene was an excellent marker to rule out ST283. All isolates were susceptible to nearly all tested antibiotics; however, only ST283 revealed 100% susceptibility to tetracycline, while ST1 and other non-ST283 showed 21.33 and 4.96%, respectively. Therefore, combining the alpha-C protein (bca) positive and tetracycline susceptible revealed 100% sensitivity for ST283. However, to identify the ST283, this combination revealed 78.9% specificity, which increased to 80.2% when erythromycin or azithromycin-susceptible was added. The bca positive combined with tetracycline and erythromycin susceptibility results were a simple tool for predicting ST283. The bca negative profile with tetracycline and macrolides resistance was commonly non-ST283. The information gained by this tool would benefit patient management.

1. Introduction

Streptococcus agalactiae (group B Streptococci, GBS) is a Gram-positive coccus in chains that is medically relevant worldwide. Since the 1960s, S. agalactiae has emerged as a leading cause of neonatal abortion or preterm labor [1]. The colonization rate of GBS in the genital tract of pregnant women can be up to 30% [2]. Currently, the American College of Obstetricians and Gynecologists (ACOG) recommends a vaginal-rectal screening for S. agalactiae for all pregnant women with 35–37 weeks gestation [3]. In the elderly, various underlying diseases, such as diabetes, heart disease, cancer, obesity, and immunocompromised individuals, have been shown to increase the risk of S. agalactiae infection [4]. Initially, serotyping based on the structure of capsular polysaccharides was used to identify GBS strains. However, only ten serotypes (Ia, Ib, II, to IX) were classified by this method, and a large proportion of the organism was classified as untypable strains [5,6]. Therefore, other techniques, such as multilocus sequence typing (MLST), which has more discriminative power, are nowadays widely used. Currently, over 2290 STs of GBS have been reported [7]. Although the distribution of GBS STs scatters with different proportions among regions, some specific STs are globally common. At least five STs of S. agalactiae, including ST1, ST12, ST17, ST19, and ST23, are worldwide detected worldwide [7,8,9,10,11]. In addition, certain STs are found in particular regions. For instance, ST283 mainly spreads among Asian countries, such as Hong Kong, Singapore, Lao PDR, Vietnam, and Thailand, ranging from 11 to 76%, and it has yet to be reported from other continents [12]. Interestingly, ST283 has also been reported as a high-risk infection in freshwater fish farms [13].

The severity of GBS infection mainly relies on the ability to colonize and cross tissue walls within the host setting, evading the host defense mechanisms and expressing virulence factors that cause injury to the host [14]. Several virulence factors play roles in the infection. The cylE gene and CAMP factor (cfb) are the most common factors involving pore-forming causing host-cell lysis as a part of GBS pathogenesis [15,16]. Alpha C protein encoded by the bca gene is one of the surface proteins in the Alpha family of proteins [17,18,19]. The alpha C protein has been demonstrated to have the potential to generate systemic infection by action as an adhesin, resulting in an invasive infection [20,21]. The surface-exposed lipoprotein Lmb (encoded by lmb) is involved in colonizing and invading damaged epithelium and also links to adhesion and metal transport in gram-positive bacteria [22,23,24]. The bibA is a novel immunogenic bacterial adhesin that involves anti-phagocytic activity and bacterial colonization. Most strains of GBS express bibA on their surfaces [25,26]. The invasion-associated gene (iagA) encoded a glycosyltransferase involved in producing a cell membrane glycolipid anchor for lipoteichoic acid within the cell wall of GBS [27]. Doran et al. used a meningitis mouse model to determine how this mutation affects GBS pathogenicity in vivo. When the mice were exposed to the iagA mutant strain, their bloodstream survival was comparable to that of the wild-type. However, lower fatality rates (20% vs. 90%) and histological analyses show that pathogenicity and the capacity to cause meningitis were dramatically reduced [28]. However, many studies have substantially controverted these virulence factors and their roles in the pathogenicity of the GBS.

In general, although the routine culture process of species identification would provide sufficient information to the physicians to treat individual patients, adding bacterial strain, e.g., ST-283, which is strongly related to invasive infection, would be a good warning sign to help the physician increase awareness and management to avoid any complications from such strains. Nevertheless, current strain detection or identification techniques such as whole genome sequencing require high-technology equipment as well as experienced staff to perform and interpret the results and are time-consuming. The process generally takes days before the result is reported, and most of the clinical microbiology laboratories in the hospitals do not have this facility installed. Therefore, this study aimed to determine the distribution of STs among GBS isolates in Thailand and find alternative tools for strain identification, particularly ST283.

2. Materials and Methods

2.1. Bacterial Isolates and Identification

A total of 277 non-duplicate GBS isolates obtained from clinical specimens, which were sent for culture at the Microbiology Laboratory, Department of Pathology, Faculty of Medicine Ramathibodi Hospital, Thailand, between 2013 and 2022 were used in this study. Overall, 143 isolates (blood = 96, cerebrospinal fluid (CSF) = 8, synovial = 23, body fluid = 4, bone = 6, eye = 4, placenta = 2) were classified as an invasive group and 134 isolates (vaginal = 28, wound = 51, urine = 47, sputum = 8) as a non-invasive group. Any multi-microbial isolates from the same sample were excluded from this study, except the isolates from sputum, which were selected from the cases that predominantly grew with S. agalactiae without any other potential pathogens of the respiratory tract. Before use, isolates were cultured on sheep blood agar at 5% CO₂ at 35 °C for 16–18 h. The isolates were identified by MALDI-TOF MS utilizing the on-target extraction procedure. Briefly, the colony was directly spotted as a thin film on a MALDI steel. The material was overlayed with 1.0 µL of 70% formic acid and dried at room temperature. The spot was then overlaid with 1 µL of a saturated α-cyano-4-hydroxycinnamic acid (HCCA) matrix solution in 50% acetonitrile and 2.5% trifluoroacetic acid (Sigma-Aldrich, Darmstadt, Germany) and air-dried. The target plate was analyzed using MALDI Biotyper Sirius (Bruker Daltonics, Bremen, Germany) with FlexControl 3.0 software. External calibration was performed with a Bruker bacterial test standard (BTS) in mass length between 2000 and 20,000 Da. An identification score of ≥2.0 was reliably identified with at least 2 in 2 consistency for the species level.

2.2. Susceptibility Testing

The antimicrobial susceptibility testing, including penicillin (P), ampicillin (AMP), daptomycin (DAP), linezolid (LZD), vancomycin (VA), chloramphenicol (C), levofloxacin (LEV), erythromycin (E), and tetracycline (TE), was tested by AST panel THAP2F (Sensititre™; Thermo Fisher Scientific, Cleveland, OH, USA) and the protocol was followed according to the manufacturer. The result was interpreted based on recommended breakpoints by the Clinical and Laboratory Standards Institute guidelines M100. Azithromycin (AZ) was also tested using the standard disk diffusion method.

2.3. Genomic DNA Extraction

This DNA was extracted by suspending 3–5 colonies of picked organisms into 200 µL of lysis solution containing 1% Triton X-100, 10 mM Tris-HCl pH 8.0, and 1 mM Disodium ethylene diamine tetra acetate (EDTA). The suspensions were mixed by vortexing for 30 s and heated at 100 °C for 10 min. After that, the suspensions were centrifuged at 15,000× g for 10 min. Finally, the supernatants were taken for PCR testing.

2.4. Specific Virulence Gene Detection

Five virulent genes (lmb, bca, bibA, iagA, and cylE) were detected by PCR. The reactions were performed with the SensoQuest lab-cycler instrument (SensoQuest GmbH, Berlin, Germany). Each amplification consisted of 2× Prime Taq Premix (Biotechrabbit GmbH, Berlin, Germany), 10 µM of specific forward and reverse primers of each tested gene (BioDesign, Bangkok, Thailand), 19 µL of distilled water, and 100–250 ng of template DNA in a final volume of 50 µL. After an initial denaturation at 95 °C for 10 min, the 35 cycles of amplification were heated at 94 °C for 1 min. Then, they were annealed at 51 °C for 45 s, extension at 72 °C for 1 min, and final extension at 72 °C for 7 min, respectively. The PCR products were analyzed using 1% agarose gel. Approximately 5% of the products were sent for sequencing (Celemics, Inc., Geumcheon-gu, Seoul, Republic of Korea). Confirmation of each sequence was performed using the Basic Local Alignment Search Tool (BLAST) [National Center for Biotechnology Information (NCBI), Bethesda, MD, USA].

2.5. Multilocus Sequence Typing (MLST)

This MLST panel of S. agalactiae includes seven housekeeping genes as the alcohol dehydrogenase (adhP), phenylalanyl tRNA synthetase (PheS), glutamine transporter protein (atr), glutamine synthetase (glnA), serine dehydratase (sdhA), glucose kinase (glcK), and transketolase (tkt). Each amplification mixture consisted of 2× Prime Taq Premix (biotechrabbit GmbH, Berlin, Germany), 10 µM of each forward and reverse primers (BioDesign, Bangkok, Thailand), 19 µL of distilled water, and 100–250 ng of template DNA in a final volume at 50 µL. Following an initial denaturation at 95 °C for 10 min, the 35 cycles amplification was heated at 94 °C for 1 min. Then, annealing was conducted at 50 °C for 45 s, extension at 72 °C for 1 min, and final extension at 72 °C for 7 min. The PCR products were analyzed using 1% agarose gel and sequenced using a NGS-based approach. The allelic data were identified using the database from the website https://pubmlst.org/organisms/streptococcus-agalactiae (accessed on 23 August 2023).

2.6. Statistical Analysis

This project used Fisher’s exact test to calculate the association between (i) sequence type and type of infection, (ii) between ST283 and ST1 of tetracycline susceptible, and (iii) between ST283 and a non-ST1 non-ST283 group of tetracycline susceptible at a significant level p-value < 0.05. A statistical software program for biomedical research, MedCalc (version 23.0.5), is available online at https://www.medcalc.org/. The Ramathibodi Institutional Review Board granted ethics approval (COA number: MURA2022/245 and MURA2023/278).

3. Results

3.1. Distribution of GBS Isolates and Sequence Types

Of the 277 S. agalactiae isolates, 64% were from female patients. The age of patients ranged from 2 days to 102 years. However, the majority of the samples were from adults and elderly cases, which covered over 80%. Among these, a total of 25 STs from 8 Clonal Complex (CC) were detected in this study. Nearly 60% of the isolates were CC1 and CC283 (

Table 1. Epidemiology data of S agalactiae by MLST.

Clonal Complex (CC)	Sequence Type (ST)	The Total Number of Each ST (%)
CC1	1	75 (27.1%)
	14	1 (0.4%)
	1626	2 (0.7%)
CC12	12	10 (3.6%)
	41	3 (1.1%)
	361	1 (0.4%)
	509	1 (0.4%)
	652	3 (1.1%)
CC17	17	18 (6.5%)
CC19	19	22 (7.9%)
	28	3 (1.1%)
	335	2 (0.7%)
	861	4 (1.4%)
	1167	1 (0.4%)
CC23	23	16 (5.8%)
	249	2 (0.7%)
CC283	283	81 (29.2%)
	739	1 (0.4%)
	751	1 (0.4%)
CC327	485	8 (2.9%)
CC459	196	2 (0.7%)
Non-CC	103	6 (2.2%)
	314	6 (2.2%)
	651	3 (1.1%)
	889	5 (1.8%)
Total		277 (100%)

). The most common ST of S. agalactiae in Thailand was ST283, which occupied approximately 30%, followed by ST1 (27.1%). The ST283 was strongly associated with the invasive infection (p < 0.0001), in which all isolates were detected from invasive specimen types. The remaining 23 STs ranged from 0.4–7.9%, and more than half were ST12, ST17, ST19, and ST23. Most of these isolates were associated with non-invasive type, particularly ST19, in which 95.5% were non-invasive (p < 0.0001). In contrast, ST1 was significantly related to invasive and non-invasive infections, p < 0.05 (

Table 2. The proportions of the invasive and non-invasive types.

Sequence Type (ST)	Number of Isolates	Invasive Infection (%)	Non-Invasive Infection (%)	p Value
283	81	81 (100%)	0 (0%)	p < 0.0001 *
1	75	31 (41.33%)	44 (58.67%)	p < 0.05 */**
19	22	1 (4.55%)	21 (95.45%)	p < 0.0001 **

p-value < 0.05 indicates statistical significance (Fisher exact test) invasive *, non-invasive **.

). The number of isolates from other STs was less than 20; therefore, the statistics were not calculated. The distribution of each ST is shown in Table 1.

3.2. Association of bca Gene and the Type of Infections

The association between the presence of virulence genes and the type of S. agalactiae infections (invasive and non-invasive groups) was determined by performing gene-specific PCR. The data of tested virulence genes revealed that only the bca gene was associated with invasiveness (

Table 3. The proportions of the bca detected of each ST.

Sequence Type	Total Isolates	Invasive		Non-Invasive
		bca + (%)	bca − (%)	bca + (%)	bca − (%)
ST283	81	81 (100)	-	-	-
ST1	75	24 (32)	7 (9.3)	30 (40)	14 (18.6)
ST19	22	-	1 (4.5)	-	21 (95.5)
ST17	18	-	9 (50)	-	9 (50)
ST23	16	-	4 (25)	-	12 (75)
ST12	10	2 (20)	-	8 (80)	-
ST485	8	-	2 (25)	-	6 (75)
ST103	6	-	1 (16.7)	-	5 (83.3)
ST314	6	3 (50)	1 (16.7)	-	2 (33.3)
ST889	5	-	2 (40)	-	3 (60)
ST861	4	-	2 (50)	-	2 (50)
ST28	3	-	1 (33.3)	-	2 (66.7)
ST41	3	-	1 (33.3)	-	2 (66.7)
ST651	3	-	-	-	3 (100)
ST652	3	-	-	3 (100)	-
ST196	2	-	-	-	2 (100)
ST249	2	-	-	1 (50)	1 (50)
ST335	2	-	-	-	2 (100)
ST1626	2	-	-	1 (50)	1 (50)
ST14	1	-	-	-	1 (100)
ST361	1	1 (100)	-	-	-
ST509	1	1 (100)	-	-	-
ST739	1	-	-	1 (100)	-
ST751	1	-	-	1 (100)	-
ST1167	1	-	-	-	1 (100)
Total	277	112 (40.4)	31 (11.2)	45 (16.2)	89 (32.1)

). Therefore, the bca was selected as the candidate gene for further association study. Of 277 isolates, 157 (56.68%) carried the bca gene; among these, 51.59% were ST283. In this study, 11 STs contained the bca gene, and only 6 STs, including ST283, ST12, ST652, ST509, ST739, and ST751, identified the gene in all isolates. However, the number of isolates of these STs, except ST283, was low, and most were associated with non-invasive types. Therefore, detection of the bca gene alone was insufficient to classify the sequence type or type of infection (Table 3). Still, it was shown to be an excellent rule-out marker for ST283 detection.

3.3. Specific Susceptibility Pattern of ST283 Versus Other STs

According to the recommendation in the CLSI M39 guideline on generating accumulative susceptibility patterns, which required at least 30 isolates of each organism group for statistical analysis, the antimicrobial susceptibility data of this study were divided into three major groups: ST1, ST283, and other STs. All 277 isolates were 100% susceptible to penicillin, ampicillin, daptomycin, linezolid, and vancomycin. Other antibiotics, including chloramphenicol, erythromycin, azithromycin, levofloxacin, and tetracycline, gave various susceptible percentages (

Table 4. Susceptibility pattern between ST1, ST283, and other ST groups.

Antibiotic	Susceptible Percentage
	ST1 (N = 75)	ST283 (N = 81)	Other ST (N = 121)
Penicillin	100	100	100
Ampicillin	100	100	100
Erythromycin	85	100	69
Azithromycin	85	100	69
Tetracycline	21 *	100 */**	5 **
Daptomycin	100	100	100
Chloramphenicol	100	96	87
Levofloxacin	100	100	89
Linezolid	100	100	100
Vancomycin	100	100	100

*/** p-value < 0.0001 indicates statistical significance (Fisher exact test).

). Interestingly, all ST283 isolates were susceptible to tetracycline and macrolide drugs (E and AZ), while most other STs were resistant to tetracycline. Therefore, the result of tetracycline resistance was implied to be an excellent rule-out marker for ST283.

3.4. Using the bca Gene, Tetracycline and Macrolide Susceptibility for ST283 and Non-ST283 Detection

Based on the available baseline information, we would like to propose a simple solution tool for ST283 detection using the bca gene, tetracycline-susceptible, and macrolide-susceptible versions. As the previous result showed that all ST283 isolates carried the bca gene, the bca negative became an excellent marker to rule out ST283 with 100% sensitivity and specificity. In addition, for those bca positive isolates, adding tetracycline and macrolide susceptibility testing results was required for the rule-out criteria. Any not susceptible (intermediate or resistant) result of either tetracycline or macrolide showed 100% sensitivity and specificity for identification as non-ST283 strains (

Table 5. Proposing a recommendation for ST283 detection.

	Testing Results			Sensitivity/Specificity
	bca	Tetracycline	Macrolide	Sensitivity	Specificity
ST283	+	S	NT	100%	78.9%
	+	S	S	100%	80.2%
Non-ST283	−	V	V	100%	100%
	+	S	R	100%	100%
	+	I	S	100%	100%
	+	R	V	100%	100%

Abbreviations: S = Susceptible, I = Intermediate, R = Resistant, NT = Not tested, V = either S or I or R.

). However, to identify an isolate as ST283, if the test results were revealed as bca positive and susceptible to tetracycline alone, the specification was 78.9%. Moreover, if the isolate was also susceptible to macrolides, the specificity slightly increased to 80.2%.

4. Discussion

Three significant findings were demonstrated in this study. Firstly, ST283 and ST1 were Thailand’s most common sequence types, occupying 56% of the isolates, of which both STs shared nearly equal proportions, approximately 28%. The ST283 is strongly associated with invasive infection since it was found in only invasive cases. It is worth mentioning here that in this study, there were two cases in which the organism was isolated from two sites in the same patients. The samples from both cases were synovial fluid and blood, and all isolates were ST283. However, to avoid confusion in selecting the isolates, these two cases were not included in this study. The ST1, on the other hand, was isolated from both groups of samples. A previous study in Thailand from 2012 to 2016 revealed a high proportion of ST283 at 62% (34/55), followed by ST1, ST19, and other STs at 15, 11, and 12%, respectively [29]. In contrast, another study in 2017–2018 on 31 GBS isolates from blood culture showed the opposite proportion, in which ST1 versus ST283 was 65% and 16%, respectively [30]. The inconsistent proportion of ST1 and ST283 between these two studies and our study may be influenced by the different number of isolates and protocols of each study or the period of strain collection. Nevertheless, these studies support our finding of the high prevalence of ST1 and ST283 among GBS in Thailand. In 2021, the Food and Agriculture Organization of the United Nations (FAO) Bangkok launched the GBS ST283 risk profile in freshwater fish farms as a higher virulence ST than other sequence types. Although this biological process by which eating food contaminated with GBS can cause sepsis is unknown, as is the route of transmission, any potential fecal–vaginal transmission of ST283 requires close surveillance [13]. Apart from Thailand, the ST283 was also reported in other Asian countries (Lao PDR, Vietnam, Cambodia, Hong Kong, and Singapore [12,31]. The ST1, on the other hand, was widespread worldwide. Many countries from various regions, including Australia, Canada, France, Ireland, Israel, Japan, Kenya, Malaysia, New Zealand, Portugal, Serbia, Sweden, Taiwan, and the United Kingdom, have reported ST1 as a dominant strain causing invasive and non-invasive infections [7,8,9,10,11].

The remaining uncommon 23STs were isolated at a much lower frequency, and the majority were associated with non-invasive infections. Interestingly, ST12, ST17, ST19, and ST23, which were occasionally detected in this study, were reported as the top five STs in Serbia, Sweden, the United Kingdom, the United States, Israel, New Zealand, Ireland, and Japan [8,10,32,33].

Secondly, of the selected virulence genes, lmb, bibA, iagA, and cylE were detected in nearly all isolates; therefore, these genes were unreliable for classifying infection types. However, the prevalence of these genes varied among regions [5,25,26,30]. The bca gene was found in 56.7% of the isolates. Although the bca gene was revealed in 11 of 25 STs, these STs were not specific to the type of infections. Therefore, the existence of the bca gene could not be used to discriminate against STs. However, it is worth noting that all ST283 isolates carried the bca gene, which, on the other hand, demonstrated that the absence of the bca gene showed to be an excellent marker to rule out ST283. This finding was concordant with other ST283 isolates documented in the PubMLST database [5].

The antimicrobial susceptibility testing results provided this study’s third finding. Generally, all isolates were susceptible to beta-lactam antibiotics, daptomycin, linezolid, and vancomycin and variable outcomes occurred in macrolides and tetracycline. Interestingly, all ST283 isolates were susceptible to tetracycline and macrolides, while approximately 90% of non-ST283 were resistant to tetracycline. Tulyaprawat et al. reported a similar susceptibility result from five isolates of ST283 that were 100% susceptible to tetracycline and erythromycin, while ST1 and other STs were resistant [30]. Another study by Neemuchwala et al. found that ST1 from blood culture was mainly resistant to tetracycline and erythromycin [9]. Similar to the bca gene, tetracycline or macrolides that are not susceptible (intermediate or resistant) are shown to be excellent markers for ruling out ST283. A combination of tetracycline susceptible and bca gene positive profiles revealed 100% sensitivity and 78.9% specificity. Adding macrolide susceptibility slightly increased the specificity to 80.2% for ST283. Since all ST283 isolates were found from invasive cases, using this new approach for rapid screening of ST283 colonization in the genital tract of pregnant women would lead to an appropriate prophylaxis regimen and reduce the incidence of vertical transmission in newborns. On the other hand, improving specificity may require additional tests or alternative approaches, such as detecting a specific peak of the protein profile by MALDI-TOF, which had been proposed in other organisms [34,35].

Apart from the above major findings, some interesting points are worth noting. A very rare ST361 was isolated from the blood sample. The information on this ST was only found in the MLST database; to our knowledge, it has not been reported elsewhere. Other strains which have yet to be reported in Thailand, such as ST41, ST485, ST861, and ST889, were reported in Ireland [33], while ST103, ST314, ST485, and ST509 were found in United States [36]. Our finding demonstrated the worldwide spreading of GBS strains.

5. Conclusions

Using simple markers such as the bca gene in combination with the susceptibility profile of tetracycline and macrolides can be an excellent marker for rapid rule-out markers of an invasive ST283 strain in Thailand. This would lead to more appropriate treatment in either infected cases or the GBS carriers, particularly pregnant women. However, since the specificity for identifying ST283 using these markers was approximately 80%, this gap may require additional tests to improve the outcome.