Optimizing bacterial DNA extraction in urine

Abstract

Urine is an acceptable, non-invasive sample for investigating the human urogenital microbiota and for the diagnosis of sexually transmitted infections. However, low quantities of bacterial DNA and PCR inhibitors in urine may prevent efficient PCR amplification for molecular detection of bacteria. Furthermore, cold temperatures used to preserve DNA and bacteria in urine can promote precipitation of crystals that interfere with DNA extraction. Saline, Dulbecco's Phosphate Buffered Saline, or Tris-EDTA buffer were added to urine from adult men to determine if crystal precipitation could be reversed without heating samples beyond ambient temperature. Total bacterial DNA concentrations and PCR inhibition were measured using quantitative PCR assays to compare DNA yields with and without buffer addition. Dissolution of crystals with Tris-EDTA prior to urine centrifugation was most effective in increasing bacterial DNA recovery and reducing PCR inhibition. DNA recovery using Tris-EDTA was further tested by spiking urine with DNA from bacterial isolates and median concentrations of Lactobacillus jensenii and Escherichia coli 16S rRNA gene copies were found to be higher in urine processed with Tris-EDTA. Maximizing bacterial DNA yield from urine may facilitate more accurate assessment of bacterial populations and increase detection of specific bacteria in the genital tract.

Fig 1. Images and microscopy of urine crystal precipitation and dissolution.

(a) Pellet from refrigerated urine sample with crystal precipitation. After centrifugation, pellet is pink in appearance. (b) Refrigerated urine sample with crystal precipitation before and (c) three minutes after Tris-EDTA addition. (d) Microscopic view (20X) of crystals from urine pellet without and (e) with Tris-EDTA addition.

Fig 2. Bacterial DNA concentrations in urine samples before and after treatment with saline, DPBS, or Tris-EDTA.

Data shown as box and whisker plots with whiskers representing maximum and minimum values (16S rRNA gene copies/mL of urine). Lines within each box represent median values. Mean bacterial concentrations in post-treatment samples treated with Tris-EDTA were found to be significantly higher.

Fig 3. Bacterial DNA recovery from spiked urine with and without Tris-EDTA addition.

Data shown as box and whisker plots with whiskers representing maximum and minimum values (16S copies/mL urine). Lines within each box represent median values. Dotted lines represent 100% yield based on spiked positive controls. Mean Escherichia coli and Lactobacillus jensenii DNA recovery in samples treated with Tris-EDTA were found to be higher than in samples not treated with Tris-EDTA.

Fig 4. Distribution of bacterial DNA extracted from urine samples treated with 10% v/v Tris-EDTA.

Data shown for 955 urine specimens including 143 specimens that were from NGU+ men and 812 specimens that were from NGU- men. An additional 43 samples had bacterial concentrations below the lower limit of detection (16 = NGU+ men and 27 = NGU- men). Bacterial concentrations were significantly lower in men with NGU.