Comparison of microbial community profiles from hydro-processed renewable diesel and F-76 diesel

ABSTRACT

To better understand the microbial communities that drive microbiologically influenced corrosion (MIC), we characterized the bacterial communities present on various surfaces in tanks containing either petroleum-based Navy F76 diesel (F76) or hydro-processed renewable diesel (HRD) collected at the U.S. Naval Research Laboratory, Stennis Space Center, MS, USA.

ANNOUNCEMENT

Fuels in storage tanks are susceptible to microbial intrusion from environmental fungi and bacteria that can result in fuel contamination and microbiologically influenced corrosion (MIC) of the tank material (1). Understanding how to mitigate MIC and which taxa affect MIC requires understanding which taxa are present. Additionally, the shift toward the use of renewable diesel derived from organic material rather than from petroleum sources may increase potential microbial challenges (2). To begin to understand these differences, we investigated the bacterial component of these communities in two fuel tanks: one containing petroleum-based Navy F76 diesel (F76) and one containing hydro-processed renewable diesel (HRD).

We sampled two fuel tanks at the U.S. Naval Research Laboratory, Stennis Space Center, MS, USA. The samples collected in these tanks are described in Table 1. We extracted DNA from samples with the ZymoBIOMICS DNA Microprep Kit (Zymo; D4301) following the standard protocol. For floating mat and excess material samples, we also extracted DNA using the DNeasy PowerBiofilm Kit (Qiagen; 24000-50). We used a one-step amplicon library preparation with V4–V5 primers (515F-Y [5′-GTGYCAGCMGCCGCGGTAA]; 926R [5′-CCGYCAATTYMTTTRAGTTT]) (3) and Thermo Phusion High-Fidelity PCR Mastermix. Some samples were re-run on two other Illumina lanes (“_2” in Table 1) and were prepared using Thermo Phusion Plus PCR Mastermix and the same primers. We cleaned PCRs using an AMPure XP bead cleanup before pooling. We sequenced the pooled library on an Illumina MiSeq at Maryland Genomics to generate 2 × 300 bp (600 cycles) paired-end reads.

TABLE 1.

Description of samples taken from two fuel tanks with different types of fuel including hydro-processed renewable diesel (HRD) and petroleum-based diesel (F76)

Tank	Fuel type	Sample date	Sample type	Filter type	Replicate	Storage temp (°C)	Sample ID	Extraction kit	Illumina lane	Library name	Reads (unfiltered)	Reads post QC	Reads post dada2	Reads post chimera	ASVs	Relative abundance of α-proteobacteria	SRA accession
1	HRD	7/13/23	Interface	Unknown volume (leaking manifold), 0.22 µm FGLP (hydrophobic)	A half filter	−80	F1	Zymo	2	F1_2	90	83	55	55	0	0	SRR31601067
						−80			3	F1	1,066	459	275	275	0	0	SRR31601066
					B half filter	−80	F2	Zymo	1	F2_2	71	63	58	58	0	0	SRR31601054
						−80			3	F2	26,588	710	455	455	0	0	SRR31601044
				208 mL through 0.22 µm GS filter (hydrophilic)	A quarter filter	−80	F3	Zymo	1	F3_2	5,459	5,425	5,348	5,348	99	65.44	SRR31601052
						−80			3	F3	8,743	2,146	1,905	1,899	73	63.82	SRR31601053
					B quarter filter	−80	F4	Zymo	2	F4_2	4,815	4,785	4,685	4,685	102	61.05	SRR31601036
						−80			3	F4	59,999	35,260	34,711	34,372	186	64.70	SRR31601035
					C quarter filter	−80	F5	Zymo	2	F5_2	1,131	1,126	1,075	1,075	0	0	SRR31601056
						−80			3	F5	92,206	66,282	65,749	65,632	214	59.66	SRR31601055
					D quarter filter	−80	F6	Zymo	2	F6_2	5,850	5,831	5,740	5,740	112	55.07	SRR31601065
						−80			3	F6	105,315	74,785	73,918	73,503	233	51.33	SRR31601064
			Coupon	One swab	A	−80	F16	Zymo	3	F16	25,963	19,739	19,166	19,166	186	61.16	SRR31601063
				One swab	B	−80	F17	Zymo	3	F17	126,095	95,625	94,092	89,688	219	56.66	SRR31601062
				One swab	C	−80	F18	Zymo	3	F18	146,549	112,122	110,412	104,755	231	53.48	SRR31601061
			Swabs from coupon tower	One swab	Single sample; No replicate	4.0 until 10/4/2023, then −80	F46	Zymo	3	F46	18,326	10,553	10,244	10,149	130	71.65	SRR31601047
			Excess material	~4 mL	Single sample; No replicate	4.0 until 10/4/2023, then −80	F48	Qiagen	3	F48_Q	170,279	128,701	126,703	112,145	228	63.72	SRR31601048
								Zymo	3	F48	99,292	75,655	74,648	71,879	268	59.57	SRR31601049
2	F76	7/14/23	Interface	Unknown volume through 0.22 µm GS (hydrophilic)	A quarter filter	−80	F23	Zymo	1	F23_2	56,917	56,728	56,586	55,954	121	81.98	SRR31601060
									3	F23	4,658	2,038	1,623	1,604	31	87.84	SRR31601031
					B quarter filter	−80	F24	Zymo	1	F24_2	431	429	391	391	0	0	SRR31601032
									3	F24	2,177	1,716	1,237	1,237	35	82.12	SRR31601033
			Floating mat	Collected by sterile serological pipette into separatory funnel where water was drained off the bottom. Mat fraction was collected into a 2 mL tube, centrifuged at 14k for 1 min, fuel layer was pipetted off the top volume 0.5–1.5 mL	B	−80	F28	Qiagen	3	F28_Q	141,833	108,700	107,755	96,774	124	82.41	SRR31601034
								Zymo	1	F28_2	53,281	53,075	52,917	52,384	90	90.79	SRR31601057
									3	F28	173,824	132,224	130,818	110,948	134	77.39	SRR31601058
					C	−80	F29	Qiagen	3	F29_Q	139,370	106,530	105,374	91,587	125	73.24	SRR31601059
								Zymo	1	F29_2	29,962	29,852	29,736	29,523	71	93.60	SRR31601037
									3	F29	160,793	119,537	118,413	100,925	114	91.42	SRR31601038
					D	−80	F30	Zymo	3	F30	67,907	50,709	50,164	47,327	100	93.33	SRR31601039
								Qiagen	3	F30_Q	148,248	113,169	112,125	98,234	108	84.70	SRR31601040
					E	−80	F31	Qiagen	3	F31_Q	64,669	48,627	48,062	44,134	95	87.19	SRR31601041
								Zymo	3	F31	107,587	81,149	80,232	73,356	130	92.41	SRR31601042
			Coupon	One swab	A	−80	F40	Zymo	3	F40	32,596	23,362	23,055	21,359	65	91.37	SRR31601043
				One swab	B	−80	F41	Zymo	3	F41	52,251	38,531	38,128	35,475	75	82.53	SRR31601045
				One swab	C	−80	F42	Zymo	3	F42	80,789	62,137	61,678	59,744	81	81.19	SRR31601046
			Excess material	2 mL	Single sample; No replicate	−80	F52	Qiagen	3	F52_Q	87,822	67,009	66,154	60,203	136	91.77	SRR31601050
								Zymo	3	F52	215,143	157,972	156,849	130,552	140	93.18	SRR31601051

We analyzed samples with the R statistical Software v4.4.1. Due to poor reverse read quality, we only analyzed forward reads here. We ran the dada2 pipeline in R trimming at 240 bp and increased the maxEE to 4 with dada2 v1.32.0 (4). As we had run these samples across several lanes, we merged lanes after identifying ASVs and then removed chimeric reads. We assigned taxonomy using Silva v138.1 (5). Further pre-processing of ASV tables included removing mitochondrial and chloroplast sequences and eliminating samples with fewer than 1,000 reads after processing in R using phyloseq v1.48.0 (6). Analyses included compositional diversity analyzes calculating the centered log-ratio, performing a PCA, and identifying differences among groups with adonis2 using the R-packages vegan v2.6.8 (7) and compositions v2.0.8 (8). We also examined taxa pooled at different taxonomic levels using tip_glom to examine differences at the Class and Family levels (6). All visualizations were performed in R with ggplot v3.5.1 (9).

In these analyses, we highlight differences among bacterial communities in HRD and F76 diesels. We found significant differences between the two fuel types (PERMANOVA R² = 0.3518, P < 0.001, Fig. 1B). Notably, the taxa in HRD were more diverse, while those in F76 were dominated by the class Alphaproteobacteria, family Sphingomonadaceae. We also found differences in other families—the HRD samples had a higher relative abundance of Bacteroidia (including Chitinophagaceae and Flavobacteriaceae), and F76 samples had a higher relative abundance of Obscuribacteraceae.

Fig 1.

(A) Picture of sample collected from F76 tank showing visible black microbial mats in the reddish fuel layer and orange corrosion products at the bottom of the water layer. (B) PCA plot summarizing microbial community differences among samples. (C) Taxa plot, condensed to class indicating relative abundance of all classes within samples. (D) Taxa plot, collapsed to family showing relative abundance of all families present at ≥5% relative abundance in one sample. In (C) and (D), we have highlighted higher taxonomic levels in color and order (e.g., all blues in (C) are Proteobacteria and all blues in (D) are Alphaproteobacteria).

DATA AVAILABILITY

All raw reads were deposited in GenBank SRA under the BioProject PRJNA1193007, and links to individual samples (SRR31601031-SRR31601067) are in Table 1.