Optimization of DNA extraction from human urinary samples for mycobiome community profiling

Abstract

Introduction: Recent data suggest the urinary tract hosts a microbial community of varying composition, even in the absence of infection. Culture-independent methodologies, such as next-generation sequencing of conserved ribosomal DNA sequences, provide an expansive look at these communities, identifying both common commensals and fastidious organisms. A fundamental challenge has been the isolation of DNA representative of the entire resident microbial community, including fungi.

Materials and methods: We evaluated multiple modifications of commonly-used DNA extraction procedures using standardized male and female urine samples, comparing resulting overall, fungal and bacterial DNA yields by quantitative PCR. After identifying protocol modifications that increased DNA yields (lyticase/lysozyme digestion, bead beating, boil/freeze cycles, proteinase K treatment, and carrier DNA use), all modifications were combined for systematic confirmation of optimal protocol conditions. This optimized protocol was tested against commercially available methodologies to compare overall and microbial DNA yields, community representation and diversity by next-generation sequencing (NGS).

Results: Overall and fungal-specific DNA yields from standardized urine samples demonstrated that microbial abundances differed significantly among the eight methods used. Methodologies that included multiple disruption steps, including enzymatic, mechanical, and thermal disruption and proteinase digestion, particularly in combination with small volume processing and pooling steps, provided more comprehensive representation of the range of bacterial and fungal species. Concentration of larger volume urine specimens at low speed centrifugation proved highly effective, increasing resulting DNA levels and providing greater microbial representation and diversity.

Conclusions: Alterations in the methodology of urine storage, preparation, and DNA processing improve microbial community profiling using culture-independent sequencing methods. Our optimized protocol for DNA extraction from urine samples provided improved fungal community representation. Use of this technique resulted in equivalent representation of the bacterial populations as well, making this a useful technique for the concurrent evaluation of bacterial and fungal populations by NGS.

Fig 1. Optimization of microbial dna extraction requires multiple disruption steps.

(A) Eight variations in the protocol (at left) were noted to increase yields as determined by quantitative PCR. Relative fungal DNA yields were calculated from quantitative PCR using the Fungiquant pan-fungal PCR primer pair and normalized to a mixed fungal DNA standard. The negative control samples were processed in parallel, but did not have any input cellular material. Multiple protocol variations, such as enzymatic pre-digestion (B) or carrier DNA use during DNA column binding (C), were tested individually in triplicate for multiple subjects (minimum n = 4), both male and female, before incorporating.

Fig 2. Large-scale confirmation of optimization of microbial DNA purification.

The individual conditions noted to increase yields were tested in aggregate in a larger-scale optimization panel. DNA was concurrently isolated from 60 identical 1 ml urine samples from each of 4 subjects with variations in cell wall disruption methods (as indicated at the top), enzymatic pre-treatment methods (bottom), and centrifugation speeds (rows indicated at right adjacent to heat map). Fungal DNA yields from these 60 variations in isolation methodology were calculated from Fungiquant qPCR as described in Fig 1, then scaled across all samples. Values are expressed as a heat map, with bright red signifying the highest yields and black the lowest yields across all samples.

Fig 3. Fungal community representation is influenced by specimen volume.

(A) DNA was isolated from a range of urinary volumes in male and female subjects (n = 3 each) and assessed by qPCR for fungal DNA. Calculated fungal DNA concentrations were calculated by normalization to a fungal standard. The optimal concentrations were achieved using 25 ml urine specimens. *: P<0.05 in comparison to 1 mL yields. (B) Following fungal DNA amplification by qPCR using broad-spectrum fungal primers, products from 25, 5 and 1 ml samples were assessed by 2% agarose gel electrophoresis. Standards indicating the PCR product size are shown on the left. Each band represents unique taxa within the urinary fungal population. NPC: no primer control.

Fig 4. Urine storage and centrifugation conditions impact DNA extraction efficiency.

In a subset of urine samples, both refrigeration and high-speed centrifugation were associated with precipitation of varying crystals that interfered with DNA purification. (A) A single urine specimen before and after refrigeration and centrifugation at 5000 rcf. In the post-centrifugation specimen, a red-orange, sandy pellet was observed after centrifugation consistent with the “brick-layer’s dust” characteristic of amorphous urates. (B) The pellet seen in A was examined by light microscopy (x400 magnification), revealing disorganized amorphous urate crystals. (C) Amorphous phosphates from alkaline urine. (D) The “envelope”-type crystals characteristic of calcium oxalate could also be identified in urine (magnified in the inset picture), but did not constitute the majority of the crystalline material.

Fig 5. Our optimized DNA extraction method outperforms commercial methods.

We compared fungal (left) and bacterial (right) extraction and characterization after our optimized protocol in comparison to three commercial DNA preparation kits using standardized urine and vaginal swab samples. (A) Individual urine specimens were divided into equal aliquots of 30 ml each. DNA was isolated from each aliquot using the specified methods; this process was repeated in quadruplicate. Samples were assessed by qPCR for fungal (left) and bacterial (right) DNA. Calculated DNA concentrations were determined by normalization to a mixed fungal and bacterial standard with a known DNA concentration. *: P<0.001, **: P<0.005. (B, C) Samples were sequenced in quadruplicate by next generation sequencing for the ITS1 (left) and 16S (right) primers. (B) The stacked bar plots represent the mean relative abundances for the fungal (left) and bacterial (right) populations in individual sequencing runs. (C) Shannon diversity indices were calculated from the microbial populations resulting from NGS for each purification method.