Optimizing DNA Extraction Methods for Nanopore Sequencing of Neisseria gonorrhoeae Directly from Urine Samples

Abstract

Empirical gonorrhea treatment at initial diagnosis reduces onward transmission. However, increasing resistance to multiple antibiotics may necessitate waiting for culture-based diagnostics to select an effective treatment. There is a need for same-day culture-free diagnostics that identify infection and detect antimicrobial resistance. We investigated if Nanopore sequencing can detect sufficient Neisseria gonorrhoeae DNA to reconstruct whole genomes directly from urine samples. We used N. gonorrhoeae-spiked urine samples and samples from gonorrhea infections to determine optimal DNA extraction methods that maximize the amount of N. gonorrhoeae DNA sequenced while minimizing contaminating host DNA. In simulated infections, the Qiagen UCP pathogen mini kit provided the highest ratio of N. gonorrhoeae to human DNA and the most consistent results. Depletion of human DNA with saponin increased N. gonorrhoeae yields in simulated infections but decreased yields in clinical samples. In 10 urine samples from men with symptomatic urethral gonorrhea, ≥92.8% coverage of an N. gonorrhoeae reference genome was achieved in all samples, with ≥93.8% coverage breath at ≥10-fold depth in 7 (70%) samples. In simulated infections, if ≥10⁴ CFU/ml of N. gonorrhoeae was present, sequencing of the large majority of the genome was frequently achieved. N. gonorrhoeae could also be detected from urine in cobas PCR medium tubes and from urethral swabs and in the presence of simulated Chlamydia coinfection. Using Nanopore sequencing of urine samples from men with urethral gonorrhea, sufficient data can be obtained to reconstruct whole genomes in the majority of samples without the need for culture.

Keywords: DNA extraction; Nanopore sequencing; Neisseria gonorrhoeae; whole-genome sequencing.

FIG 1

N. gonorrhoeae-simulated infections; limit of detection using Nanopore sequencing. (A) The proportion of sequenced reads classified as human, bacterial, or viral. (B) The proportion of bacterial bases classified as N. gonorrhoeae, and (C) the proportion of the NCCP11945 N. gonorrhoeae reference genome covered at ≥10-fold depth. For WHO V (orange markers), the actual spike concentration achieved was lower than the targeted concentration (see Fig. S3).

FIG 2

Sequencing speed in 18 N. gonorrhoeae and C. trachomatis NAAT-negative urine samples spiked with N. gonorrhoeae. The actual spiking concentration achieved is rounded to the nearest order of magnitude for the purposes of the legend. Estimated coverage depth is calculated as the number of bases of N. gonorrhoeae DNA sequenced divided by the length of the reference genome.

FIG 3

Performance in clinical samples positive for N. gonorrhoeae; yield of bacterial, human, and viral DNA sequenced. Results from 10 participants are shown, including where available reads obtained from a urethral swab, a urine sample collected into a cobas lysis buffer (processed with and without mechanical lysis with beads), and a urine sample collected in a universal container with a boric acid additive processed with and without treatment with saponin. The results present numbers of human reads and total reads, including data demultiplexed by porechop and guppy; all other analyses are based on guppy data alone (see Materials and Methods for details).

FIG 4

Performance in clinical samples positive for N. gonorrhoeae, including coverage breadth and depth. (A) The mean coverage depth achieved for samples processed by one the five methods tested, i.e., the total number of bases of sequence generated divided by the length of the NCCP11945 N. gonorrhoeae reference genome. (B) The proportion of the NCCP11945 N. gonorrhoeae reference genome covered by at least one read. (C) Data from urine samples processed without saponin treatment.

FIG 5

Performance in clinical samples positive for N. gonorrhoeae; relative proportions of species sequenced per sample. The Z-score, denoted by shade, for each taxon is the number of standard deviations above the mean number of bases per taxon for each sample.