Innovative epitope-based ELISA assay for detecting Group B Streptococcus carriage in pregnant women Authors: Monika Brzychczy-Włoch, Anna Dobrut, Kamil Drożdż, Anna Malska-Woźniak, Wojciech Pabian, Sabina Górska, Ewa Brzozowska, Andrzej Gamian Published: 2026 ABSTRACT AbstractThree epitopes of Streptococcus agalactiae, 41KAFGSPLITN50 from molecular chaperone GroEL (PEP2), 91MIALDGTPNKG101 from enolase (PEP5), and 28LTAAITTVLARRLP41 from EF-Tu (PEP11), were studied as ELISA components for immunodiagnostic detection of infection caused by Streptococcus agalactiae (Group B Streptococcus; GBS) and GBS carriage in pregnant women using IgG, IgM, and IgA antibodies. The current gold standard in GBS detection based on culturing is time-consuming and sometimes yields results that are difficult to interpret. Therefore, the aim of this study was to develop an alternative immunoassay based on strictly defined synthetic epitopes. The study included 143 serum samples divided into a study group (GBS+) and a control group (GBS–). The experiments showed that all peptides (single used and mixed) significantly distinguished between GBS+ and GBS– sera. The best individual candidates were PEP5 and PEP11, as they required lower peptide concentrations and demonstrated higher mean absorbance values measured for the GBS+ samples. In turn, epitopes mixing in all combinations significantly improved immunoreactivity. The results are very promising—and to our best knowledge, pioneering—as there is no similar test on the diagnostic market. We believe that our prototype may constitute a promising alternative to the classic diagnostic methods.Key points• Three synthetic GBS epitopes (from GroEL, enolase, EF-Tu) enable accurate immunodiagnosis of infection.• PEP5 (91MIALDGTPNKG101) and PEP11 (28LTAAITTVLARRLP41) showed highest immunoreactivity with IgG, IgM, and IgA detection.• Combining epitopes significantly improves immunoreactivity and diagnostic potential. Introduction Streptococcus agalactiae (Group B Streptococcus; GBS) is an opportunistic pathogen which asymptomatically colonizes the genitourinary and gastrointestinal tracts in almost one third of healthy women (Rodriguez-Granger et al. 2012). However, for some risk groups, including newborns, pregnant women, the elderly, diabetics, or immunosuppressed patients, GBS infection may lead to several health issues. In newborns, infection may manifest as sepsis, pneumonia, or meningitis depending on whether it is the early onset form of disease (EOD), which appears in the first week of life, or late onset (LOD), which may appear between 1 week and 3 months of life. Most often, it takes the form of meningitis (Rodriguez-Granger et al. 2012). In the 1970s, the mortality rate in newborns diagnosed with GBS was as high as 55% (Dermer et al. 2004). In response to these alarming statistics, the American College of Obstetricians and Gynecologists (ACOG) and the Centers for Disease Control and Prevention (CDC) developed guidelines in the 1990s to minimize this dangerous phenomenon. The procedures included screening recommendations for women between the 35th (35–0/7) and 37th (37–0/7) week of pregnancy by taking swabs from the vaginal introitus and the anal sphincter, followed by microbial cultivation on the appropriate growth medium (Filkins et al. 2021). The introduction of these restricted procedures led to a reduction of morbidity in newborns of over 80%; currently, the incidence of the disease caused by GBS in newborns is 0.23/1000 live births (Verani et al. 2010; Schrag and Verani 2013; Filkins et al. 2021). Although the recommended screening of pregnant women significantly reduced this adverse phenomenon, the gold standard for GBS diagnosis based on culturing on an appropriate medium is not without flaws. One of the limitations is the time-consuming nature of this method. The wait time for reliable results may be up to 7 days, whereas in some cases—for example, preterm birth—time is crucial. Additionally, results obtained by phenotypic methods may be ambiguous in their interpretation, especially when distinguishing S. agalactiae from other beta streptococci (Pyclik et al. 2018; Filkins et al. 2021). Therefore, in the latest guidelines, the CDC recommended MALDI (matrix-assisted laser desorption/ionization) as an alternative for GBS detection. The most notable advances of this mass spectrometry method are shorter wait times for results and the ability to distinguish S. agalactiae from other streptococci, such as Streptococcus halichoeri or Streptococcus pseudoporcinus (Filkins et al. 2021). When the presence of GBS in pregnant women is confirmed, intrapartum antibiotic prophylaxis (IAP) is introduced. Even though IAP significantly reduced the incidence in early onset disease, it did not solve the issue for late onset disease; it also led to an increase in the rate of infections caused by Gram-negative bacteria (Berardi et al. 2010). Another limitation of IAP is the use of antibiotics, which expands the problem of multidrug-resistant bacteria (Filkins et al. 2021). In response to all of these limitations, researchers are carrying out extensive studies to develop innovative diagnostic assays for detection of GBS carriage and/or infection that would provide rapid and unambiguous results. An enzyme-linked immunosorbent assay (ELISA) based on detective synthetic epitopes is considered to be a good prognostic solution (Pyclik et al. 2018). Epitopes (antigenic determinants) are several or dozens of short amino acid fragments of proteins with the ability to directly and specifically bind with compatible sequences of antibodies—paratopes—with the effect of activating immunological reactions and finally leading to the elimination of the pathogens or other external molecules, such as allergens (Greenspan and Di Cera 1999; Trolle et al. 2016; Nowis et al. 2023). This feature has aided scientists in developing new diagnostic methods, improving cancer immunotherapy, and researching innovative subunit vaccines, among other things (Kessler and Melief 2007; Coelho et al. 2015; Mucci et al. 2017; Suprun et al. 2019). Investigations of desired epitopes include both in silico and in vitro methodologies. Potential epitopes of a particular protein are predicted using bioinformatic tools based on molecular properties, such as hydrophobicity, location of the analyzed sequence within the protein, or the physicochemical properties of the individual amino acids (Emini et al. 1985; Kolaskar and Tongaonkar 1990; Larsen et al. 2006; Artificial Intelligence Research Laboratory 2025). Next, in silico results are verified by appropriate laboratory techniques, including derivation of synthetic peptides with a known amino acid sequence. They are studied in the presence of antibody samples with immunoassays such as ELISA or western blotting (Pyclik et al. 2018; Dobrut et al. 2018). Due to the lack of serological method in the diagnosis of GBS carriage in pregnant women based on an antibody–epitope reaction, the aim of the study was to develop an enzyme-linked immunosorbent assay (ELISA) based on three synthetic epitopes (in a single used and mixed together), derived from three immunoreactive proteins (molecular chaperone GroEL, enolase and EF-Tu), for the detection of GBS carriage and/or diagnosis of the infection caused by these bacteria in pregnant women. Materials and methods Patients The study was approved by Jagiellonian University Bioethical Committee Decision No. 1072.6120.51.2017. Written informed consent was obtained from the participants. The inclusion criteria were pregnant women in the third trimester between 18 and 40 years of age and a written statement of consent to participate in the study. The exclusion criteria were pregnant women below 18 or over 40 years of age (the mean age was 32 years), patients with immunodeficiency or autoimmune diseases, pregnant women with a high-risk pregnancy or perinatal complications, preterm delivery, ongoing antibiotic therapy or antibiotic therapy completed up to 14 days before the examination, ongoing generalized infection or genital infection, and a lack of written consent to participate in the study. The patients were qualified for the study by an obstetrician in accordance with the inclusion/exclusion protocol. The patients underwent microbiological diagnostics for GBS, which included examination for the presence of S. agalactiae on swabs taken from the lower vagina (vaginal introitus) and the rectum (anal sphincter) between the 35th and 37th week of gestation and cultured and assessed in accordance with the current recommendations of the Polish Gynecological Society and CDC protocols (Schrag et al. 2002; Kotarski et al. 2008). Study material and study groups The study material—venous blood serum—was obtained from 100 pregnant women and divided into (1) a study group (GBS+; N = 25) of women with confirmed GBS vaginal/rectal colonization and (2) a control group (GBS–; N = 75) of uncolonized women. Then, based on the information collected in the research questionnaire regarding various risk factors within the groups, four categories of patients were distinguished: patients without underlying diseases (I), patients with thyroid disease (II), patients with gestational diabetes (III), and those with other underlying diseases (IV). In the study group, 17 patients declared no underlying diseases, 4 had been diagnosed with thyroid disease, 1 had been diagnosed with gestational diabetes, and 3 had other underlying diseases. In the control group, 39 patients declared no underlying diseases, 16 had been diagnosed with thyroid disease, 12 had been diagnosed with gestational diabetes, and 8 declared other underlying diseases. Additionally, the study included 43 serum samples collected as part of our previous project (Brzychczy-Wloch et al. 2013; Pyclik et al. 2018; Dobrut et al. 2018) from GBS+ pregnant women, from patients in the postpartum period colonized with GBS, from generalized infections of GBS etiology, and as cross-reactivity control sera from patients with sepsis caused by other agents, such as Streptococcus pyogenes, Staphylococcus hominis, Klebsiella pneumoniae, and Enterococcus faecalis. The collected serum samples were aliquoted and stored at −80 °C until used. Epitope selection and synthesis The detection and characterization of the epitopes, which coated microplate wells and served as the basis for assay, were described in detail in our previous papers (Pyclik et al. 2018; Dobrut et al. 2018). In brief, potential epitopes predicted in bioinformatic analyses were synthesized on polyethylene pins by PEPSCAN—the epitope mapping technique, which is based on a solid-phase synthesis of peptides on polyethylene pins (Geysen et al. 1987; Pyclik et al. 2018; Dobrut et al. 2018) and their immunoreactivity and specificity to the sera obtained from GBS+ patients and GBS– patients, including samples originating from individuals with diagnosed infections caused by other pathogens (cross-reactivity control), were studied with ELISA. From almost 400 amino acid sequences representative of four S. agalactiae proteins—enolase, elongation factor Tu (EF-Tu), inosine 5′-monophosphate dehydrogenase (IMPDH), and molecular chaperone GroEL—three of them were studied as components of the ELISA assay: 41KAFGSPLITN50 from molecular chaperone GroEL (PEP2), 91MIALDGTPNKG101 from enolase (PEP5), and 28LTAAITTVLARRLP41 from EF-Tu (PEP11) (Fig. 1).Fig. 1Example scheme of bioinformatic epitope prediction using appropriate databases (Emini et al. 1985; Kolaskar and Tongaonkar 1990; Larsen et al. 2006; Artificial Intelligence Research Laboratory 2025). A Location of the potential epitope within the 3D model of protein 1HA3 for Thermus thermophilus HB8. B Example consensus epitope prediction results for a given amino acid sequence; C – 3D model of the epitope selected based on bioinformatic analysis Example scheme of bioinformatic epitope prediction using appropriate databases (Emini et al. 1985; Kolaskar and Tongaonkar 1990; Larsen et al. 2006; Artificial Intelligence Research Laboratory 2025). A Location of the potential epitope within the 3D model of protein 1HA3 for Thermus thermophilus HB8. B Example consensus epitope prediction results for a given amino acid sequence; C – 3D model of the epitope selected based on bioinformatic analysis Enzyme-linked immunosorbent assay (ELISA) The ELISA assay for detection of IgG, IgM, and IgA was carried out according to the modified Ivanov’s procedure (Ivanov et al. 1992). Optimization of the procedure included modifying epitopes by conjugation [bovine serum albumin (BSA, Biolim S. A., Gdynia, Poland), keyhole limpet hemocyanin (KLH, Biolim S. A., Gdynia, Poland), a complex of streptavidin-biotinylated peptides (Biolim S. A., Gdynia, Poland), determining the epitopes’ concentration, comparing 96-well plates to improve the peptides’ attachment, and examining the influence that various sera concentration and blocking buffers have on absorbance, temperature, and the time needed for particular steps of the procedure. Finally, 96-well MediSorp plates were coated with 200 µl/well of streptavidin (Merck, Darmstadt, Germany) dissolved in 0.1 M carbonate buffer (Sigma Aldrich, Saint Louis, Missouri, USA, Chempur, Piekary Śląskie, Poland) at a concentration of 2 µg/ml and incubated at 4 °C for 3 days, then at 37 °C for 2 h with shaking. Afterward, unattached protein was removed by triple washing with phosphate buffered saline containing 0.05% Tween-20 (PBST, Thermo Fisher, Waltham, Massachusetts, USA). After any remaining PBST buffer was thoroughly removed from each well, the biotinylated peptides were dissolved in buffer I (0.1 M carbonate buffer; 4% NaCl; 0.1% Tween-20; pH 8.5) at a concentration of 10 µg/ml (100 ml/well) for PEP5 and PEP11 or of 40 µg/ml (100 ml/well) for PEP2 and applied to the titration plates’ wells. Next, the plates were incubated at 4 °C for 3 days and then at 37 °C for 2 h with shaking. Unattached peptides were removed by triple washing with PBST and the plates coated with epitopes conjugated with streptavidin–biotin complex were incubated with 300 µl/well of Protein Free Blocking Buffer (Thermo Fisher, Waltham, Massachusetts, USA) for 1 h at 24 °C. The blocking buffer was removed, and the plates were incubated with 100 µl/well of the sample sera (1:1000 for IgG, IgA and 1:500 for IgM) dissolved in buffer II (0.1 M phosphate buffer; 4% NaCl; 0.2% Tween 20; pH 7.1) for 1 h at 37 °C. Next, the plates were washed three times with PBST and incubated with goat anti-human IgG secondary antibodies conjugated with horseradish peroxidase (BIO-RAD, Hercules, California, USA) for IgG class antibody detection, goat anti-human IgM secondary antibodies conjugated with horseradish peroxidase (Invitrogen, Waltham, Massachusetts, USA) for IgM class antibody detection, and goat anti-human IgA secondary antibodies conjugated with horseradish peroxidase (Invitrogen, Waltham, Massachusetts, USA) for IgA class antibody detection (1:10,000 in buffer II) for 1 h at 37 °C. Afterward, the plates were washed five times with PBST and the wells were dried thoroughly and then incubated with 3,3′,5,5′-tetramethylbenzidine (TMB, Becton Dickinson, Franklin Lakes, New Jersey, USA) for 30 min. To stop the reaction, 50 µl of Stop Solution (Sigma Aldrich, Saint Louis, Missouri, USA) was added to each well. The results of the experiment were obtained by measuring the absorbance value at a wavelength of 450 nm and a reference wavelength of 630 nm using a Tecan Infinity 200Pro apparatus with Magellan software (Pro version 7.3) (Tecan Group Ltd., Männedorf, Swiss). The immunoreactivity of the epitopes (PEP2, PEP5, and PEP11) to IgG, IgM, and IgA antibodies was studied individually. Moreover, the epitopes were mixed in the following proportions: PEP2 + PEP5 = 4:1, PEP2 + PEP11 = 4:1; PEP5 + PEP11 = 1:1; and PEP2 + PEP5 + PEP11 = 4:1:1. All experiments were replicated at least three times. Statistical analysis Statistical analysis was carried out by using Statistica software version 13.3 (TIBCO Software Inc., Palo Alto, CA, USA). The basic descriptive statistics to determine statistical significance in absorbance value among the analyzed groups, the Mann–Whitney U test, Kruskal–Wallis test, and Dunn post hoc test were performed. In order to assess the quality of the classifier and to determine the optimal cut-off point, receiver operating characteristic (ROC) curve analysis assuming equal costs of misclassification was performed, and the optimal cut-off values were determined using the Youden index. The coefficient of variation (CV) was used to check the reproducibility of the assay. Repeatability was considered very good when CV ≤ 10%, good when 10% < CV < 25%, medium when 25% < CV ≤ 35%, and insufficient when CV > 35% (Alajbeg et al. 2017). The effect of mixing the peptides on the absorbance value was checked by two-way analysis of variance (ANOVA) in 2 × 7 between-group designs. p-values of <0.05 were considered significant. Results The determination of protein carrier for ELISA The attempts at determining a stable protein conjugate that strongly attaches epitopes to the bottom of the well and does not demonstrate any reactivity to the studied serum indicated that streptavidin, in opposition to BSA (bovine serum albumin) and KLH (keyhole limpet hemocyanin), was the most reliable conjugate. All serum samples later used in the final ELISA analyses, including GBS-positive, GBS-negative, and cross-reactivity control sera, were included at this optimization stage. Streptavidin demonstrated weak immunoreactivity to the serum samples, both single and those complexed with biotin, and it strongly bound peptides to the surface of the wells. In contrast to BSA (Fig. 2), which showed high immunoreactivity to the sera and KLH, streptavidin barely reacted with the serum and thus did not demonstrate cross-reactivity, though it weakly bound peptides to the plate surfaces, so using it as a conjugate could lead to unreliable results.Fig. 2Comparison of three protein carriers (BSA, KLH, and streptavidin) used for epitope immobilization, expressed as mean absorbance values (± SD) measured in ELISA. Microplate wells were coated with protein carriers at a concentration of 1 µg/mL. Absorbance reflects serum immunoreactivity toward epitope–carrier conjugates. BSA—bovine serum albumin; KLH—keyhole limpet hemocyanin; Strept—streptavidin (component of the streptavidin–biotin complex) Comparison of three protein carriers (BSA, KLH, and streptavidin) used for epitope immobilization, expressed as mean absorbance values (± SD) measured in ELISA. Microplate wells were coated with protein carriers at a concentration of 1 µg/mL. Absorbance reflects serum immunoreactivity toward epitope–carrier conjugates. BSA—bovine serum albumin; KLH—keyhole limpet hemocyanin; Strept—streptavidin (component of the streptavidin–biotin complex) Immunoreactivity in ELISA of PEP2–41KAFGSPLITN50 from molecular chaperone GroEL In order to obtain the optimal concentration of PEP2 for the detection of IgG antibodies, 10 µg/ml, 20 µg/ml, and 40 µg/ml were tested in ELISA in the presence of GBS+ serum samples, which originated from patients with ongoing infection caused by GBS, GBS carrier women during childbirth, and pregnant GBS carriers and GBS– obtained from non-carrier pregnant women and from infections caused by S. hominis (cross-reactivity control). For the 10 µg/ml concentration of PEP2, the mean absorbance measured for the GBS+ sera was significantly higher than for the GBS– samples. In the first step, immunoreactivity to all GBS+ sera (GBS infection and GBS carriage in pregnancy and childbirth) and GBS– sera was compared; the differences were significant (GBS– A450 = 0.087 ± 0.025; GBS + A450 = 0.161 ± 0.047; p < 0.001) (Fig. 3A). In the next step, all groups were individually analyzed. The sensitivity determined by ROC curve of the assay reached 76.5%, while the specificity was 100.0% (Fig. 3B). For the PEP2 concentration of 20 µg/ml, differences in immunoreactivity were also significant (GBS– A450 = 0.090 ± 0.025; GBS + A450 = 0.147 ± 0.048; p < 0.001) (Fig. 3A). Based on the analysis of the ROC curve, the sensitivity of the test was 76.5%, while the specificity was 85.7% (Fig. 3B). Due to the significant differences between GBS– and GBS + absorbance at concentrations of 10 µg/ml and 20 µg/ml, the concentration of 40 µg/ml was included to improve the assay’s sensitivity and specificity. This led to a doubling of the mean absorbance value measured for GBS + sera (A450 = 0.268 ± 0.095) with no increase for the GBS– group (A450 = 0.092 ± 0.017) (p < 0.001) (Fig. 3A), as well as an improvement in sensitivity and specificity: 94.1% and 100%, respectively (Fig. 3B). Because 40 µg/ml was the most sufficient concentration, the investigation of PEP2 was limited to the detection of IgG antibodies due to the high costs of the experiments and the limited amount of synthesized peptide.Fig. 3Mean absorbance values measured with standard deviation (whiskers) for PEP2 at concentrations of 10 µg/ml, 20 µg/ml, and 40 µg/ml in the presence of IgG antibodies (A) and receiver operating characteristic curves for PEP2 at concentrations of 10 µg/ml, 20 µg/ml, and 40 µg/ml for IgG antibodies (B). Legend: GBS−—sera obtained from healthy women (without GBS carriage and/or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or infection Mean absorbance values measured with standard deviation (whiskers) for PEP2 at concentrations of 10 µg/ml, 20 µg/ml, and 40 µg/ml in the presence of IgG antibodies (A) and receiver operating characteristic curves for PEP2 at concentrations of 10 µg/ml, 20 µg/ml, and 40 µg/ml for IgG antibodies (B). Legend: GBS−—sera obtained from healthy women (without GBS carriage and/or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or infection Immunoreactivity in ELISA of PEP5–91MIALDGTPNKG101 from enolase In order to study the usability of PEP5 (10 µg/ml) as a biomarker in ELISA for the detection of IgG antibodies, the following serum samples (GBS+) were used in the investigation: sera from patients with ongoing infection caused by GBS, samples from GBS carriers during childbirth, pregnant GBS carriers, and samples from patients with sepsis of S. hominis and S. pyogenes etiology (cross-reactivity control). The GBS– control samples included sera obtained from pregnant non-carrier women. It was found that the assay based on PEP5 significantly distinguished (p < 0.001) GBS + (mean A450 = 0.371 ± 0.146) and GBS– sera (mean A450 = 0.117 ± 0.073) (Fig. 4A). For IgG antibodies, the PEP11-based assay showed significantly higher immunoreactivity in GBS+ sera compared with GBS– controls (p < 0.001). Mean absorbance values were A450 = 0.441 ± 0.163 for GBS+ samples and A450 = 0.179 ± 0.061 for GBS– sera (Fig. 4B). The Kruskal–Wallis test for differences in absorbance among the groups (infection, carriers during childbirth, pregnant carriers, and cross-reactivity group) showed statistically significant results (p < 0.001); Dunn’s post hoc test indicated that the assay can significantly distinguish GBS carriage (p < 0.001) and GBS infection (p < 0.001) from sera obtained from healthy individuals. It is worth highlighting that no cross-reactivity was observed. In turn, the analysis of the ROC curve for PEP5 showed that at the cut-off point, which was A450 = 0.184, the sensitivity of the test was 100.0%, while the specificity was 84.3% (Fig. 5A, Table 1). For PEP11, ROC curve analysis indicated a cut-off value of A450 = 0.301, yielding a sensitivity of 92.3% and a specificity of 97.2% (Fig. 5B, Table 1).Fig. 4Mean absorbance values measured with standard deviation (whiskers) for PEP5 and PEP11. Subfigures present IgG responses to PEP5 (A) and PEP11 (B), IgM responses to PEP5 (C) and PEP11 (D), and IgA responses to PEP5 (E) and PEP11 (F). Legend: GBS− —sera obtained from healthy women (without GBS carriage or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or infection; red line—cut-off pointFig. 5Receiver operating characteristic (ROC) curves determined for PEP5 and PEP11. Subfigures show IgG responses to PEP5 (A) and PEP11 (B), IgM responses to PEP5 (C) and PEP11 (D), and IgA responses to PEP5 (E) and PEP11 (F)Table 1ELISA parameters determined for PEP5 (91MIALDGTPNKG101 from enolase) and PEP11 (28LTAAITTVLARRLP41 from EF-Tu)Study peptidesPEP 5PEP 11Test parametersIgGIgMIgAIgGIgMIgACut-off0.1840.4010.3490.3010.8700.507AUC0.97(0.94–1.00)0.96(0.93–1.00)0.94(0.9–0.98)0.98(0.96–1.00)0.98(0.94–1.00)1(0.99–1.00)TP332329242526FP11517201FN030210TN596853697170Sensitivity (95% CI)100%(89.4%–100.0%)88.5%(69.9%–97.6%)100%(88.7%–100.0%)92.3%(74.9%–99.1%)96.2%(80.4%–99.9%)100.0%(86.77%–100.0%)Specificity (95% CI)84.3%(73.6%–91.9%)93.2%(84.7%–97.7%)75.7%(64.0%–85.2%)97.2%(90.2%–99.7%)100.0%(94.9%–100.0%)98.6%(92.4%–100.0%)PPV(95% CI)0.75(0.60–0.87)0.82(0.63–0.94)0.63(0.48–0.77)0.92(0.75–0.99)1.00(0.86–1.00)0.96(0.81–1.00)NPV(95% CI)1.00(0.94–1.00)0.96(0.88–0.99)1.00(0.93–1.00)0.97(0.90–1.00)0.99(0.93–1.00)1.00(0.95–1.00)PLR(95% CI)6.36(3.7–10.95)12.92(5.48–30.44)4.12(2.72–6.23)32.77(8.32–129.05)-71.0(10.14–497.10)NLR(95% CI)-0.12(0.04–0.36)-0.08(0.02–0.3)0.04(0.01–0.26)-ACC(95% CI)89.3%(81.69%–94.55%)91.9%(84.7%–96.5%)82.8%(73.9%–89.7%)95.9%(89.8%–98.9%)98.9%(94.4%–100.0%)98.9%(94.4%–100%)Youden index0.840.820.760.890.960.99Legend: cut-off—cut-off point determined by Youden Index on the basis of ROC curve, AUC—area under the curve, TP—true positive, FP—false positive, FN—false negative, TN—true negative, PPV—positive predictive value, NPV—negative predictive value, PLR—likelihood ratio positive, NLR—likelihood ratio negative, ACC—accuracy, 95% CI—95% confidence interval Mean absorbance values measured with standard deviation (whiskers) for PEP5 and PEP11. Subfigures present IgG responses to PEP5 (A) and PEP11 (B), IgM responses to PEP5 (C) and PEP11 (D), and IgA responses to PEP5 (E) and PEP11 (F). Legend: GBS− —sera obtained from healthy women (without GBS carriage or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or infection; red line—cut-off point Receiver operating characteristic (ROC) curves determined for PEP5 and PEP11. Subfigures show IgG responses to PEP5 (A) and PEP11 (B), IgM responses to PEP5 (C) and PEP11 (D), and IgA responses to PEP5 (E) and PEP11 (F) Table 1ELISA parameters determined for PEP5 (91MIALDGTPNKG101 from enolase) and PEP11 (28LTAAITTVLARRLP41 from EF-Tu)Study peptidesPEP 5PEP 11Test parametersIgGIgMIgAIgGIgMIgACut-off0.1840.4010.3490.3010.8700.507AUC0.97(0.94–1.00)0.96(0.93–1.00)0.94(0.9–0.98)0.98(0.96–1.00)0.98(0.94–1.00)1(0.99–1.00)TP332329242526FP11517201FN030210TN596853697170Sensitivity (95% CI)100%(89.4%–100.0%)88.5%(69.9%–97.6%)100%(88.7%–100.0%)92.3%(74.9%–99.1%)96.2%(80.4%–99.9%)100.0%(86.77%–100.0%)Specificity (95% CI)84.3%(73.6%–91.9%)93.2%(84.7%–97.7%)75.7%(64.0%–85.2%)97.2%(90.2%–99.7%)100.0%(94.9%–100.0%)98.6%(92.4%–100.0%)PPV(95% CI)0.75(0.60–0.87)0.82(0.63–0.94)0.63(0.48–0.77)0.92(0.75–0.99)1.00(0.86–1.00)0.96(0.81–1.00)NPV(95% CI)1.00(0.94–1.00)0.96(0.88–0.99)1.00(0.93–1.00)0.97(0.90–1.00)0.99(0.93–1.00)1.00(0.95–1.00)PLR(95% CI)6.36(3.7–10.95)12.92(5.48–30.44)4.12(2.72–6.23)32.77(8.32–129.05)-71.0(10.14–497.10)NLR(95% CI)-0.12(0.04–0.36)-0.08(0.02–0.3)0.04(0.01–0.26)-ACC(95% CI)89.3%(81.69%–94.55%)91.9%(84.7%–96.5%)82.8%(73.9%–89.7%)95.9%(89.8%–98.9%)98.9%(94.4%–100.0%)98.9%(94.4%–100%)Youden index0.840.820.760.890.960.99Legend: cut-off—cut-off point determined by Youden Index on the basis of ROC curve, AUC—area under the curve, TP—true positive, FP—false positive, FN—false negative, TN—true negative, PPV—positive predictive value, NPV—negative predictive value, PLR—likelihood ratio positive, NLR—likelihood ratio negative, ACC—accuracy, 95% CI—95% confidence interval ELISA parameters determined for PEP5 (91MIALDGTPNKG101 from enolase) and PEP11 (28LTAAITTVLARRLP41 from EF-Tu) 0.97 (0.94–1.00) 0.96 (0.93–1.00) 0.94 (0.9–0.98) 0.98 (0.96–1.00) 0.98 (0.94–1.00) 1 (0.99–1.00) 100% (89.4%–100.0%) 88.5% (69.9%–97.6%) 100% (88.7%–100.0%) 92.3% (74.9%–99.1%) 96.2% (80.4%–99.9%) 100.0% (86.77%–100.0%) 84.3% (73.6%–91.9%) 93.2% (84.7%–97.7%) 75.7% (64.0%–85.2%) 97.2% (90.2%–99.7%) 100.0% (94.9%–100.0%) 98.6% (92.4%–100.0%) PPV (95% CI) 0.75 (0.60–0.87) 0.82 (0.63–0.94) 0.63 (0.48–0.77) 0.92 (0.75–0.99) 1.00 (0.86–1.00) 0.96 (0.81–1.00) NPV (95% CI) 1.00 (0.94–1.00) 0.96 (0.88–0.99) 1.00 (0.93–1.00) 0.97 (0.90–1.00) 0.99 (0.93–1.00) 1.00 (0.95–1.00) PLR (95% CI) 6.36 (3.7–10.95) 12.92 (5.48–30.44) 4.12 (2.72–6.23) 32.77 (8.32–129.05) 71.0 (10.14–497.10) NLR (95% CI) 0.12 (0.04–0.36) 0.08 (0.02–0.3) 0.04 (0.01–0.26) ACC (95% CI) 89.3% (81.69%–94.55%) 91.9% (84.7%–96.5%) 82.8% (73.9%–89.7%) 95.9% (89.8%–98.9%) 98.9% (94.4%–100.0%) 98.9% (94.4%–100%) Legend: cut-off—cut-off point determined by Youden Index on the basis of ROC curve, AUC—area under the curve, TP—true positive, FP—false positive, FN—false negative, TN—true negative, PPV—positive predictive value, NPV—negative predictive value, PLR—likelihood ratio positive, NLR—likelihood ratio negative, ACC—accuracy, 95% CI—95% confidence interval In the experiments carried out on IgM class antibodies, we observed higher mean immunoreactivity in general, but the differences in mean absorbance value between GBS + (mean A450 = 0.756 ± 0.360) and GBS– (mean A450 = 0.227 ± 0.116) were significant (p < 0.001) (Fig. 4C). For IgM antibodies, a statistically significant difference in immunoreactivity was also observed between the analyzed groups (p < 0.001). GBS + sera demonstrated higher mean absorbance values (A450 = 1.478 ± 0.478) compared with GBS– samples (A450 = 0.507 ± 0.150) (Fig. 4D). The Kruskal–Wallis test showed significant differences among the groups (p < 0.001), whereas Dunn’s post hoc test revealed differences in mean absorbance values between both GBS carriage (p < 0.001) and GBS infection (p = 0.048) in comparison with healthy individuals (GBS–), as well as between GBS carriage (p = 0.002) and GBS infection (p = 0.015) and serum samples obtained from patients with ongoing sepsis of S. hominis and S. pyogenes etiology. The ROC curve, used to determine the cut-off point for PEP5, which was A450 = 0.401, indicated that the sensitivity of the test was 88.5% and the specificity was 93.2% (Fig. 5C, Table 1). For PEP11, ROC curve analysis indicated a cut-off value of A450 = 0.870, yielding a sensitivity of 96.2% and a specificity of 100.0% (Fig. 5D, Table 1). The usability of PEP5 in the detection of IgA antibodies was confirmed because the differences in mean absorbance values between the GBS + (mean A450 = 0.534 ± 0.177) and GBS– (mean A450 = 0.307 ± 0.074) groups were significant (p < 0.001) (Fig. 4E). For IgA antibodies, the PEP11-based ELISA significantly differentiated GBS + and GBS– sera (p < 0.001). The mean absorbance value for GBS + samples was A450 = 0.931 ± 0.291, whereas for GBS– sera it was A450 = 0.317 ± 0.078 (Fig. 4F). Significant differences in mean absorbance tested by the Kruskal–Wallis test were also found among all groups (p = 0.001) and Dunn’s post hoc test indicated usability in the detection of GBS infection (p < 0.001) and GBS carriage in pregnancy (p < 0.001) and childbirth (p = 0.026). Samples from the cross-reactivity control group (patients with ongoing sepsis of S. hominis and S. pyogenes etiology) demonstrated absorbance below the detection threshold (cut-off). The ROC curve, used to determine the cut-off point for PEP5, which was A450 = 0.349, indicated that the sensitivity of the test was 100% and the specificity was 75.7% (Fig. 5E, Table 1). For PEP11, ROC curve analysis indicated a cut-off value of A450 = 0.507, yielding a sensitivity of 100.0% and a specificity of 98.6% (Fig. 5F, Table 1). Additionally, for PEP5, the CV was used to check the repeatability of the assay for three classes of antibodies. IgG and IgA repeatability was found to be very good and IgM repeatability was good. Immunoreactivity in ELISA of PEP11–28LTAAITTVLARRLP41 from EF-Tu The immunoreactivity and selectivity of PEP11 at a concentration of 10 µg/ml was investigated as a biomarker in ELISA for the detection of GBS carriage and/or GBS infection in the presence of serum samples obtained from patients with ongoing infection caused by GBS and samples taken from GBS carriers. In turn, the negative control included serum samples isolated from non-carrier pregnant women. The detection of IgG antibodies in the sera samples showed significant (p < 0.001) differences in immunoreactivity between GBS+ (mean A450 = 0.441 ± 0.163) and GBS– (mean A450 = 0.179 ± 0.061) (Fig. 4). This may indicate that PEP11 can be considered a good biomarker in the assay. Moreover, the Kruskal–Wallis test revealed significant differences among the study groups (p < 0.001), whereas Dunn’s post-hoc test demonstrated significant differences for GBS infection (p = 0.044) and GBS carriage (p < 0.001) in comparison with the samples from non-carriers. It was also found that ELISA based on PEP11 demonstrated high sensitivity (92.3%) and specificity (97.2%) (cut-off point 0.301) (Fig. 5, Table 1). Interesting results were obtained for the assay for detecting IgM class antibodies: very high mean absorbance values for both GBS+ (mean A450 = 1.478 ± 0.478) and GBS– (mean A450 = 0.507 ± 0.150) (Fig. 4). However, the differences were significant (p < 0.001). Further analyses, including the Kruskal–Wallis test followed by Dunn’s post hoc test, demonstrated significant differences in absorbance among GBS infections (p = 0.031) as well as GBS carriage (p < 0.001) in comparison with healthy individuals. As the cut-off point was high (A450 = 0.870), the specificity and sensitivity were very high (almost 100%) (Fig. 5, Table 1). Very promising and repeatable results were obtained for IgA class antibodies. The experiments showed high mean absorbance for GBS+ samples (mean A450 = 0.931 ± 0.291) with low absorbance for the GBS– sera (mean A450 = 0.317 ± 0.078) (Fig. 4). ANOVA test followed by Dunn’s post hoc test showed significant differences in immunoreactivity between GBS carriers and non-carriers (p < 0.001). Moreover, the sensitivity and specificity were almost 100% with a cut-off point of 0.507 (Fig. 5, Table 1). In turn, CV analysis showed very good reproducibility for all classes of antibodies. Therefore, in summary, the best markers in the ELISA assay for detecting GBS carriage and/or infection caused by GBS are PEP5 and PEP11. The use of peptides in combination For this paper, we also investigated the impact of mixing the detective epitopes on immunoreactivity and selectivity. For this purpose, PEP2 (40 µg/ml), PEP5 (10 µg/ml), and PEP11 (10 µg/ml) were mixed in the following proportions: PEP2 + PEP5 = 4:1, PEP2 + PEP11 = 4:1; PEP5 + PEP11 = 1:1; and PEP2 + PEP5 + PEP11 = 4:1:1. They were studied in the presence of GBS+ sera and GBS– sera. The results are very promising, as they not only indicate statistically significant differences between the tested peptides and their combinations in the tested groups, but they also showed that regardless of the peptide mixture used, the average absorbance value of the GBS+ sera significantly increased the mean absorbance value. It is worth highlighting that in the GBS– group, regardless of the peptide or combination of peptides, the absorbance was at a similar level (Fig. 6). Therefore, the use of peptides in combination significantly increases the power of discriminating GBS+ from GBS– sera, which can be considered an improving factor in the development of the assay.Fig. 6Mean absorbance values measured with standard deviation (whiskers) for single epitopes (PEP2, PEP5, and PEP11) and combinations (PEP2 + PEP5, PEP2 + PEP11, PEP5 + PEP11, and PEP2 + PEP5 + PEP11) in the presence of GBS− and GBS+ serum samples for IgG (A), IgM (B), and IgA (C) antibody classes. Legend: GBS− —sera obtained from healthy women (without GBS carriage or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or GBS infection; strept—streptavidin (control) Mean absorbance values measured with standard deviation (whiskers) for single epitopes (PEP2, PEP5, and PEP11) and combinations (PEP2 + PEP5, PEP2 + PEP11, PEP5 + PEP11, and PEP2 + PEP5 + PEP11) in the presence of GBS− and GBS+ serum samples for IgG (A), IgM (B), and IgA (C) antibody classes. Legend: GBS− —sera obtained from healthy women (without GBS carriage or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or GBS infection; strept—streptavidin (control) Differences in interpretation Among both GBS+ and GBS– samples, the pool of patient samples included ones from patients with underlying diseases (thyroid disease, gestational diabetes, and others). It was found that the presence of one of them had a significant impact raising the absorbance value for PEP5 in GBS– IgG antibody class (p < 0.001) and GBS– IgA antibody class (p = 0.011). Surprisingly, for PEP11, underlying diseases did not lead to higher immunoreactivity in the GBS– group. Moreover, the differences were statistically significant (p = 0.021 for IgG, p = 0.004 for IgM, and p < 0.001 for IgA). Even though no significant impact of underlying diseases on mean absorbance in the GBS + group was observed, the finding for GBS– should be considered when interpreting the results of the ELISA test (Fig. 7).Fig. 7Comparison of the influence of underlying diseases on the immunoreactivity of PEP5 and PEP11 expressed as mean absorbance values. Subfigures show IgG responses to PEP5 (A) and PEP11 (B), IgM responses to PEP5 (C) and PEP11 (D), and IgA responses to PEP5 (E) and PEP11 (F). Legend: none—patients with no underlying diseases; with—patients with underlying diseases. GBS−—sera obtained from healthy women (without GBS carriage or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or infection. The whiskers on the bars represent the standard deviation Comparison of the influence of underlying diseases on the immunoreactivity of PEP5 and PEP11 expressed as mean absorbance values. Subfigures show IgG responses to PEP5 (A) and PEP11 (B), IgM responses to PEP5 (C) and PEP11 (D), and IgA responses to PEP5 (E) and PEP11 (F). Legend: none—patients with no underlying diseases; with—patients with underlying diseases. GBS−—sera obtained from healthy women (without GBS carriage or infection); GBS+—sera obtained from patients diagnosed with GBS carriage and/or infection. The whiskers on the bars represent the standard deviation Comparison of PEP5 and PEP11 immunoreactivity The immunoreactivity of PEP5 and PEP11 was directly compared in GBS− serum samples. Statistically significant differences were detected for IgG and IgM antibody classes (both p < 0.001). In the IgG class, PEP11 exhibited higher mean absorbance values (A450 = 0.179 ± 0.061) than PEP5 (A450 = 0.117 ± 0.073). A similar pattern was observed for IgM antibodies, with substantially higher immunoreactivity for PEP11 (A450 = 0.507 ± 0.150) compared with PEP5 (A450 = 0.226 ± 0.117). In contrast, no statistically significant differences were detected between the peptides for IgA antibodies. The mean absorbance values were comparable for both peptides, with A450 = 0.308 ± 0.074 for PEP5 A450 = 0.317 ± 0.078 for PEP11 (Fig. 8A).Fig. 8Mean absorbance values measured with standard deviation (whiskers) for PEP5 and PEP11 in GBS− (A) and GBS+ (B) serum samples in the presence of IgG, IgM, and IgA antibodies. Legend: GBS− —sera obtained from healthy women (without GBS carriage or infection); GBS+ —sera obtained from patients diagnosed with GBS carriage and/or GBS infection Mean absorbance values measured with standard deviation (whiskers) for PEP5 and PEP11 in GBS− (A) and GBS+ (B) serum samples in the presence of IgG, IgM, and IgA antibodies. Legend: GBS− —sera obtained from healthy women (without GBS carriage or infection); GBS+ —sera obtained from patients diagnosed with GBS carriage and/or GBS infection In GBS + serum samples, a different pattern of peptide-specific immunoreactivity was observed. Statistically significant differences between PEP5 and PEP11 were detected for IgM and IgA antibody classes (both p < 0.001). In the IgM class, PEP11 exhibited substantially higher mean absorbance values (A450 = 1.478 ± 0.478) compared with PEP5 (A450 = 0.756 ± 0.360). Similarly, IgA immunoreactivity was significantly higher for PEP11 (A450 = 0.931 ± 0.291) than for PEP5 (A450 = 0.555 ± 0.184). In contrast, no statistically significant differences were observed between the peptides for IgG antibodies, although mean absorbance values were slightly higher for PEP11 (A450 = 0.441 ± 0.163) than for PEP5 (A450 = 0.371 ± 0.146) (Fig. 8B). Discussion The introduction of GBS screening in pregnant women followed by intrapartum antibiotic prophylaxis (IAP) significantly reduced morbidity and mortality in newborns, especially in early onset disease. However, the problem still exists in the case of late onset disease and in adult patients, particularly the elderly and diabetics. Additionally, the overuse of antibiotics leads to a higher incidence of multidrug-resistant bacteria (Libster et al. 2012). Therefore, rapid diagnosis of GBS carriage and/or infection could shorten the wait time for results, which in turn would accelerate the initiation of targeted antibiotic therapy or would help avoid unnecessary administration of antibiotics following a negative result. As there is no such quick, precise, affordable assay on the diagnostic market, the prototype ELISA based on the detective epitopes of three immunoreactive proteins could improve the GBS diagnostic market, at least as a supplementary option. In recent years, epitopes have been broadly studied—particularly as a vaccine component, but also as a component of immunoassays and drug targets (Cheong et al. 2016; Oscherwitz 2016; Hajissa et al. 2019; Suprun et al. 2019). High chemical stability, knowledge of amino acid sequences, ease of synthesis, high immunogenicity induction, and low cross-reactivity are cited as the greatest advantages of synthetic epitopes (Fritze and Anderson 2000; Hajissa et al. 2019). Moreover, due to their minimal immunogenic region, it is possible to precisely predict the directed immune responses (Oli et al. 2020; Soleymani et al. 2022). Because the last feature is mostly related to vaccinology, the remaining ones can be useful in immunoassay development. Immunoreactive proteins and their epitopes are widely studied in microbiology, including for S. agalactiae (Cheong et al. 2016; Oscherwitz 2016; Hajissa et al. 2019; Suprun et al. 2019; Dobrut and Brzychczy-Włoch 2021). In our study, we demonstrated that certain sequences—41KAFGSPLITN50 from molecular chaperone GroEL (PEP2), 91MIALDGTPNKG101 from enolase (PEP5), and 28LTAAITTVLARRLP41 from EF-Tu (PEP11)—can be considered biomarkers in the ELISA assay for diagnosis of GBS infection and/or carriage because they significantly distinguished GBS+ from GBS– samples and demonstrated high specificity and selectivity, both in single used and in combination. To our best knowledge, the epitopes described in this paper have not been studied in this context, underscoring the pioneering nature of this research. Nevertheless, immunoreactive proteins and their epitopes, as well as immunodiagnostic assays, are of interest to other researchers, especially in the context of an anti-GBS vaccine. It should be emphasized that the present study was designed as a proof-of-concept investigation aimed at evaluating the feasibility of using strictly defined synthetic epitopes as components of an ELISA-based immunodiagnostic assay. Although the selected epitopes demonstrated high specificity and strong immunoreactivity, the application of these antigens within an ELISA format represents a distinct methodological step that requires careful optimization and validation. Similar epitope-based ELISA approaches have been reported in the literature for diagnostic or vaccine-related purposes, demonstrating the potential of synthetic peptides as reliable immunoassay components, while also highlighting the necessity of rigorous assay validation (Chalifour et al. 1992; Mucci et al. 2017; Suprun et al. 2019). In this context, the results presented herein should be interpreted as an initial validation performed using well-characterized serum samples under controlled laboratory conditions. Independent inter-laboratory validation using external cohorts would constitute an essential next step toward clinical implementation of the assay. Such validation was beyond the scope of the current study and would require collaboration with additional diagnostic laboratories. Nevertheless, the high sensitivity, specificity, and reproducibility observed in the present work support the robustness of the proposed approach and justify further translational research. The best-known GBS antigen is capsular polysaccharide (CPS), which is a component of anti-GBS vaccines (monovalent and trivalent) being investigated in phase I/II clinical studies (Madhi et al. 2016, 2017; Madrid et al. 2017). Clinical studies on anti-GBS vaccines are challenging to design because they require a large number of participants; therefore, studies on immunoassays may deliver significant information for vaccine development, such as antibody levels against particular antigens (Buffi et al. 2019). As an anti-GBS vaccine component, GroEL protein—whose epitopes were part of our study—was investigated for GBS infection in Nile tilapia by Li et al. (2019). In turn, Pumchan et al. (2020). investigated 11 potential B-cell epitopes from four immunogenic GBS proteins that can be used as components in a chimeric multiepitope vaccine, with promising results. Research using 3D modeling of immunogenic epitopes derived from ScaAB, a surface lipoprotein of S. agalactiae belonging to the LraI family and involved in metal ion transport and contributing to bacterial virulence, has delivered promising results. ScaAB was evaluated as a component of an experimental recombinant vector vaccine (Fedorova et al. 2018). ELISA is a rapid, affordable, and sensitive method that allows quantitative and semi-quantitative determination of antibody concentration in all antibody classes. Therefore, this method can be used in vaccine development, but above all for diagnosing various diseases. In the literature, most papers describing epitopes concern vaccine development (Pumchan et al. 2020; Carboni et al. 2020; Campisi et al. 2021). Studies on ELISA for GBS detection based on epitopes were carried out by Chalifour et al. (1992), who in a two-site variant showed that with the anti-trirhamnoside monoclonal antibody to capture the antigen and the polyclonal antibody as an enzyme-conjugated second antibody reacted with only Group B Streptococcus were able to detect this pathogen in the vaginal microbiota at 3 × 104 CFU/test (colony forming unit) for GBS. Other studies carried out with ELISA on a multi-epitope subunit vaccine against Group B Streptococcus infection, consisting of 15 immunoreactive proteins, have delivered very promising results showing 100% protection from fatal events (Zhang et al. 2022). For optimization of the assay, several conjugates were tested. It was important to increase specific immunoreactivity with GBS+ sera and to assure repeatable results of the tests, which is a common issue in ELISA (Guttormsen et al. 1996, 2002; Bhushan et al. 1998; Kasper et al. 1999). Our studies showed that wells coated with streptavidin that covalently linked peptides through biotin significantly improved absorbance values measured for GBS+ serum samples. The results were repeatable, and no unspecific reactions were observed. BSA was used as both a carrier and a blocker, resulting in high and unspecific reactivity. This can be explained by the fact that BSA is a strong antigen and samples isolated from patients with accompanying diseases resulted in unspecific and high absorbance values compared to the remaining samples (Moura et al. 2014; Damoiseaux et al. 2019). The third carrier under study was KLH, a highly immunogenic T cell-dependent antigen that is increasingly used in research. It is particularly recommended as a carrier for peptides, small proteins, and non-antigenic haptens to induce or enhance an immune response (Greenfield et al. 2021). Although KLH demonstrates a number of advantages, technical issues associated with very weak binding of KLH-conjugated peptides to titration plates nevertheless forced us to reject this protein in further studies. A new diagnostic approach based on defined, highly specific epitopes supported by numerous studies seems to be the future of innovative diagnostics. In our study, we attempted to prove the usability of three epitopes as markers in ELISA for the detection of infections caused by S. agalactiae and/or GBS carriage in pregnant women. The mean absorbance values measured in 143 venous serum samples—both GBS+ and GBS− —indicated that these epitopes significantly distinguished carriers from healthy individuals, though the best individual candidates were PEP5 and PEP11. PEP2 required a high concentration of peptide, and its immunoreactivity was not very high; therefore, it would be better as a component in a mixture of epitopes. The ROC curve showed that the assay based on those epitopes would deliver reliable results, as both its selectivity and specificity reached even 100%. We are aware that a larger sample size would confirm our results and could indicate the impact of accompanying diseases on false positive results. However, the results clearly show a high potential of this enzyme-linked immunosorbent assay based on these three epitopes.