A pangenome approach-based loop-mediated isothermal amplification assay for the specific and early detection of Bordetella pertussis

Abstract

Despite widespread vaccination, Bordetella pertussis continues to cause pertussis infections worldwide, leaving infants at the highest risk of severe illness and death, while people around them are likely the main sources of infection and rapidly spread the disease. Rapid and less complex molecular testing for the specific and timely diagnosis of pertussis remains a challenge that could help to prevent the disease from worsening and prevent its transmission. We aimed to develop and validate a colorimetric loop-mediated isothermal amplification (LAMP) assay using a new target uvrD_2 informed by the pangenome for the specific and early detection of B. pertussis. Compared to that of multitarget quantitative polymerase chain reaction (multitarget qPCR) using a large clinical DNA specimen (n = 600), the diagnostic sensitivity and specificity of the uvrD_2 LAMP assay were 100.0% and 98.6%, respectively, with a 99.7% degree of agreement between the two assays. The novel colorimetric uvrD_2 LAMP assay is highly sensitive and specific for detecting B. pertussis DNA in nasopharyngeal swabs and showed similar diagnostic accuracy to complex and high-cost multitarget qPCR, but it is faster, simpler, and inexpensive, which makes it very helpful for the reliable and timely diagnosis of pertussis in primary health care and resource-limited settings.

Figure 1

Screening of new pangenome-based diagnostic targets for specific detection of B. pertussis. (a) Heatmap and hierarchical clustering analysis of candidate diagnostic genes informed by the classical Bordetella pangenome (n = 977). (b) Functional annotation of candidate diagnostic genes and their coverage (presence or absence) in the Bordetella bronchiseptica and Bordetella pertussis genomes. DCG, diagnostic candidate gene; 1, presence; 0, absence. Figure 1a was created using pheatmap v1.0.12 and dendextend v1.16.0 packages for R v4.0.5. Figure 1b was generated using ggplot2 v3.3.6 package for R v4.0.5.

Figure 2

Effect of temperature and uvrD_2, MaoC, and 10,409 hyp primer sets on colorimetric LAMP assay performance. (a) Box plot depicting the amplification time (time to positivity) at the optimized temperature by the LAMP primer set. Boxes indicate the first to third quartile (interquartile range, IQR) with median (central line), and whiskers encompass 95% of the data. *P ≤ 0.001 shown at the top of the boxes. Significance was analyzed by one-way analysis of variance (ANOVA) of the mean amplification time from independent experiments performed in triplicate. (b) Colorimetric LAMP assays using uvrD_2, MaoC, and 10,409 hyp primer sets, in triplicate, at the optimized temperature for the detection of Bordetella pertussis DNA. (c) Evaluation of the uvrD_2 primer set at different temperatures. Error bars represent the standard deviations of the mean amplification time from independent experiments performed in triplicate. **P < 0.001. Significance was analyzed by one-way analysis of variance (ANOVA). (d) Turbidity of the NTC LAMP reaction analyzed in (c). The dotted line shows the threshold turbidity at 650 nm. A LAMP-positive reaction was considered when the turbidity increased above 0.05 within 60 min and the color changed from brown to green. LAMP assays were performed with B. pertussis DNA (positive control) obtained from B. pertussis strain Tohama 1 (ATCC BAA-589) at 10⁶ copies of genome equivalent per microliter (gEq × μl⁻¹). NTC negative nontemplate control.

Figure 3

Evaluation of the uvrD_2-based LAMP assay using preheated and nonpreheated template DNA from a clinical sample panel with a bacterial load gradient. Preheated DNA consisted of denaturing template DNA at 96 °C for 5 min prior to loading of the sample for the LAMP reaction. Reactions were considered positive for LAMP products if they had both a color change from brown to green and a ladder-like banding pattern on agarose gel after electrophoresis. S1–S4, clinical samples positive for B. pertussis DNA with different cycle threshold (CT) values obtained by multitarget qPCR; NS, clinical samples negative for B. pertussis DNA; bp, base pair; µ_Tp, mean time to positivity (Tp) in minutes (min) in the LAMP reaction (mean ± SD). Original gels are presented in online Supplementary Fig. S2.

Figure 4

Analytical sensitivity of the colorimetric uvrD_2 LAMP assay using nasopharyngeal swab specimens spiked with Bordetella pertussis DNA. The limit of detection (LoD) was assessed using B. pertussis DNA serially diluted tenfold from 10⁵ to 1 copy of genome equivalent per microliter (gEq × μl-1) using nasopharyngeal swabs that were negative for B. pertussis (Bp), B. parapertussis, and B. holmesii DNA. All log-dilutions were analyzed in parallel by multitarget qPCR based on (a) IS481 and (b) ptxS1 and (c) colorimetric LAMP assays. The LoD of the colorimetric uvrD_2 LAMP assay was 10 copies gEq × μl-1, which was defined as the lowest concentration of B. pertussis DNA at which ≥ 95% of replicates (n = 10) showed positive results. LAMP reactions were considered positive for B. pertussis DNA if they had both a color change from brown to green and a ladder-like banding pattern on agarose gel after electrophoresis. NDP, DNA pool obtained from nasopharyngeal swabs that were negative for B. pertussis, B. parapertussis, and B. holmesii DNA; NTC negative nontemplate control, BPC B. pertussis positive control, bp base pair.

Figure 5

Analytical specificity of the colorimetric LAMP assay using (a) bacterial and (b) viral pathogens associated with respiratory infections. Bpp Bordetella parapertussis, Bh Bordetella holmesii, Bb Bordetella bronchiseptica, Nm Neisseria meningitides, Spn Streptococcus pneumoniae, Hinf Haemophilus influenza, Kpn Klebsiella pneumoniae, Cdif Corynebacterium diphtheria, Sa Staphylococcus aureus, Hpi Haemophilus parainfluenzae, Pae Pseudomonas aeruginosa; Ec Escherichia coli, Flu influenza virus, Vict Victoria, Yam Yamagata, SARS-CoV-2 severe acute respiratory syndrome coronavirus 2, hMPV human metapneumovirus, RSV respiratory syncytial virus, RhV rhinovirus, NTC negative nontemplate control, BPC B. pertussis positive control.

Figure 6

Detection based on colorimetric LAMP of Bordetella pertussis DNA compared to multitarget qPCR using clinical samples. Scatter plots show a comparison between the testing results of LAMP and multitarget qPCR assays for DNA samples obtained from nasopharyngeal swabs (n = 600). The turbidity of the LAMP reaction in all samples was measured at OD 650 nm. The dotted black lines indicate the threshold value of turbidity (0.05) at which the LAMP reaction was considered positive. LAMP assays were based on target uvrD_2, while the qPCR targets for the multitarget assays included (a) IS481 and (b) ptxS1.

Figure 7

Colorimetric LAMP-positive samples over the time from symptom onset and antibiotic treatment from individuals with pertussis. (a) Violin plots show the time to positivity of the LAMP reaction over the days from symptom onset to sampling, (b) individuals with antibiotic treatment and (c) azithromycin doses administered to individuals with pertussis. (d) The dot plot shows the distribution of CT values for ptxS1 with qPCR according to the administered azithromycin dose. NA not available.