Environmental Factors Associated With Soil Prevalence of the Melioidosis Pathogen Burkholderia pseudomallei: A Longitudinal Seasonal Study From South West India

Abstract

Melioidosis is a seasonal infectious disease in tropical and subtropical areas caused by the soil bacterium Burkholderia pseudomallei. In many parts of the world, including South West India, most cases of human infections are reported during times of heavy rainfall, but the underlying causes of this phenomenon are not fully understood. India is among the countries with the highest predicted melioidosis burden globally, but there is very little information on the environmental distribution of B. pseudomallei and its determining factors. The present study aimed (i) to investigate the prevalence of B. pseudomallei in soil in South West India, (ii) determine geochemical factors associated with B. pseudomallei presence and (iii) look for potential seasonal patterns of B. pseudomallei soil abundance. Environmental samplings were performed in two regions during the monsoon and post-monsoon season and summer from July 2016 to November 2018. We applied direct quantitative real time PCR (qPCR) together with culture protocols to overcome the insufficient sensitivity of solely culture-based B. pseudomallei detection from soil. A total of 1,704 soil samples from 20 different agricultural sites were screened for the presence of B. pseudomallei. Direct qPCR detected B. pseudomallei in all 20 sites and in 30.2% (517/1,704) of all soil samples, whereas only two samples from two sites were culture-positive. B. pseudomallei DNA-positive samples were negatively associated with the concentration of iron, manganese and nitrogen in a binomial logistic regression model. The highest number of B. pseudomallei-positive samples (42.6%, p < 0.0001) and the highest B. pseudomallei loads in positive samples [median 4.45 × 10³ genome equivalents (GE)/g, p < 0.0001] were observed during the monsoon season and eventually declined to 18.9% and a median of 1.47 × 10³ GE/g in summer. In conclusion, our study from South West India shows a wide environmental distribution of B. pseudomallei, but also considerable differences in the abundance between sites and within single sites. Our results support the hypothesis that nutrient-depleted habitats promote the presence of B. pseudomallei. Most importantly, the highest B. pseudomallei abundance in soil is seen during the rainy season, when melioidosis cases occur.

Keywords: Burkholderia pseudomallei; South West India; environmental surveillance; melioidosis; physicochemical factors; seasonal variation; soil.

FIGURE 1

Number and field quality of soil samples collected for the detection of Burkholderia pseudomallei. Each sector of the circle stands for a respective type of field. Sample counts are given as bold numbers; the number of sites is written in brackets.

FIGURE 2

Map of soil sampling locations in the main soil surveillance study from July 2016 to January 2019. Black dots mark the sampling sites located in Udupi (14 fields) and triangles represent sites in Shimoga (six fields).

FIGURE 3

Analyses procedure of soil samples collected close to diagnosed melioidosis cases from July 2016 to January 2019. A total of 1,704 samples were processed and analyzed by direct TTSS1-qPCR (1) and cultivated using two different culture approaches: the consensus method (3) and a drainage filter enrichment (2). The number of B. pseudomallei positive (Bps+) and negative (Bps–) samples detected are given as bold numbers below the method. From July 2017 onward, 968 samples that were negative by direct qPCR and in the culture methods (2) and (3) were selected for an additional molecular analysis of the filter enrichment broth of method (2). A total of 60 TTSS1-qPCR positive and 60 negative soil samples were subjected to a physicochemical analysis.

FIGURE 4

Seasonal prevalence of B. pseudomallei qPCR-positive soil samples from Udupi and Shimoga districts. (A) B. pseudomallei positive detected samples collected in Udupi in different seasons. (B) Percentage of positive soil samples detected in Udupi separated by sites. (C) B. pseudomallei positive samples collected in Shimoga in different seasons. (D) Percentage of positive detected soil samples in Shimoga separated by sites. The two starlets above the bars of site S3 and S14 mark culture positive sampling sites (Fisher’s exact test: *p < 0.05, **p < 0.01, ****p < 0.0001).

FIGURE 5

Seasonal differences of B. pseudomallei burden in soil samples. (A,B) The B. pseudomallei burden of qPCR positive samples throughout the seasons of Udupi and its single sites (Kruskal–Wallis, *p < 0.05, ***p < 0.001, ****p < 0.0001). (C,D) Burden of B. pseudomallei in positive detected soil samples collected at Shimoga and its sites. Samples that became positive in subsequent enrichment culture are marked with an asterisk in the data set. The measured molecular load of the culture positive soil samples was 3.76 × 10⁶ genome equivalents in the sample of site S3 and 1.95 × 10⁵ genome equivalent for site S14. Boxes line values from the 25th to 75th percentiles of a data set; the vertical line in a box plot marks the median value. Whiskers from the lower and upper quartile represent 1.5 times the interquartile range.

FIGURE 6

Positivity rate of B. pseudomallei soil samples of paddy and other locations. (A) Positivity of soil samples based on direct qPCR in paddy and other fields depicted for all 1,704 samples and separated for Udupi and Shimoga. (B) B. pseudomallei positivity rate of soil samples from paddy fields (left) and other locations of Udupi (right). Dots in the bars represent the percentage of positive samples for single sites (Fisher’s exact test: *p < 0.05, **p < 0.01).

FIGURE 7

Seasonal difference in the burden of B. pseudomallei positive soil samples in paddy and other field types. Dot plots show B. pseudomallei genome equivalents (GE) per gram of soil detected in paddy and other fields sampled (Kruskal–Wallis; ns, not significant; *p < 0.05, ***p < 0.001, ****p < 0.0001). Samples that became positive in subsequent enrichment culture are marked with an asterisk in the data set. The measured molecular load of the culture positive soil samples was 1.95 × 10⁵ genome equivalent for the paddy site S14 and 3.76 × 10⁶ genome equivalents in the uncultivated site S3. Boxes line values from the 25th to 75th percentiles of a data set; the vertical line in a box plot marks the median value. The whiskers from the lower and upper quartile represent 1.5 times the interquartile range.

FIGURE 8

Difference in the iron and manganese concentration between B. pseudomallei positive (Bps+) and negative (Bps–) soil samples. Concentration of (A) iron in the monsoon (left side) and dry season (right site) and of (B) manganese. The term ‘dry season’ refers to the post-monsoon and summer season (Kruskal–Wallis, Dunn’s test; ns, not significant; *p < 0.05, **p < 0.01). Boxes line values from the 25th to 75th percentiles of a data set; the vertical line in a box plot marks the median value. The whiskers from the lower and upper quartile represent 1.5 times the interquartile range.

FIGURE 9

Core genome MLST (cgMLST)-based genomic comparison of the isolated Indian B. pseudomallei strains with a diverse, global collection of B. pseudomallei strains. (A) The UPGMA tree shows clustering of soil isolates from this study with other B. pseudomallei isolates from different parts of the world. The line in the lower left corner illustrates 0.1% cGLMST column difference (% of core genome target differences). The geographic origin of strains is given as country code prior to the strain name (THA Thailand, BRA Brazil, ECU Ecuador, BFA Burkina Faso, MDG Madagascar, IND India, LKA Sri Lanka, AUS Australia). (B) A cgMLST-based minimum spanning tree illustrates the genetic diversity of Indian isolates. Isolate names are given as short forms: IND-60 refers to ERR298760, IND-59 to ERR298759 and IND-37 to RR3110337. The numbers on the connecting lines refer to the number of allele differences. Strains included with accession numbers of respective databases are listed in Supplementary Table 1.