Cofactor F₄₂₀ tail length distribution in different environmental samples

Abstract

Cofactor F₄₂₀ is an electron carrier playing a crucial role in a variety of microorganisms during redox reactions of the primary and secondary metabolism due to its low redox potential and thus arouses increasing interest. In this study, cofactor F₄₂₀ glutamyl tail length spectra in various habitats like manure, compost, soil, and digester sludge samples and their respective microbial communities were investigated using high performance liquid chromatography and an amplicon sequencing approach A previous in-silico study was used to identify F₄₂₀ producing microorganisms. The highest concentration of cofactor F₄₂₀ could be achieved in the horse manure, digester sludge, and mixed manure samples, which was approximately 100-fold higher than in all the other samples. The high content of the cofactor in the samples with high O₂ availability pointed to the important role of the cofactor not only in redox reactions of anaerobic but also for aerobic microorganisms and indicated its ubiquitous character. The most abundant derivate was F₄₂₀-3 comprising the largest part of the cofactor derivates in seven out of ten samples. The high abundance of F₄₂₀-3 in samples with distinct properties (e.g. O₂ and H₂O availability) showed its important role in redox reactions of the primary and secondary metabolism among prokaryotes.

Graphical abstract

Highlights

• •Analysis of cofactor F₄₂₀ in natural habitats.
• •Annotation of glutamyl tail length spectra in different environments.
• •Analysis of microbial community in natural habitats.
• •Identification of F₄₂₀ producing microorganisms.

1. Introduction

Flavin-dependent enzymes mediate a wide range of redox reactions in the metabolism across all domains of life. Some organisms also synthesize the 5-deazaflavin structure F_420, in which the N-5 atom is substituted through a C atom in the isoalloxazine ring system [1]. The cofactor F₄₂₀ is structurally very similar to flavins but has distinctly different physiochemical properties. It is an obligate two-electron hydride carrier and has a very low standard redox potential and a blue-green intrinsic fluorescence with an absorption maximum at 420 nm [1]. Its properties are very similar to that of nicotinamides and are thus sometimes described as nicotinamides in flavin clothing [2]. With −340 mV, the cofactor has one of the lowest redox potentials of all known redox factors [1]. The redox potential can be reduced even further under certain physiological conditions: In CO₂-reducing methanogens, for example, the redox potential can decrease to – 380 mV when these organisms exhibit a 1:10 ratio of oxidized to reduced cofactor F₄₂₀ [3].

Apart from its important role in redox-reactions of the primary metabolism, biochemical studies have revealed that the cofactor F₄₂₀ can also be part of biosynthesis of various classes of secondary metabolites [4,5]. F₄₂₀H₂ dependent enzymes are involved in the reduction of tetracycline precursors [4,6], are further involved in the production of thiopeptides, lanthipeptides, 4-alkyl-L-proline derivatives, alkaloid guanipiperazine, alchivemycin A, and sulfonamide metabolites [4,5]. Besides, F₄₂₀H₂ depend enzymes are putatively involved in the production of 3-amino-5-hydroxybenzoic acid (AHBA)-containing metabolites, kasugamycin, coronafacoyls and phenazines [4]. The potential of F₄₂₀H₂ dependent enzymes is already used for some medical and industrial biotechnological applications, e. g. the biodegradation of environmental contaminants such as picrate [7]) or nitroaromatic explosives [8]. The antimicrobial prodrugs delamanid and pretomanid used to medicate tuberculosis patients is activated by an F₄₂₀H₂ dependent reductase of Mycobacterium tuberculosis directly within a patient's body [9,10]. However, F₄₂₀ dependent enzymes have a considerable potential for bio catalysis, but according to Shah et al. F₄₂₀ dependent enzymes are under explored and need to be better characterized to use their potential as a source for biocatalysts [11]. For a long time, the cofactor F₄₂₀ could only be chemically identified in the phyla Euryarchaeota and Actinobacteria [12]. In 2017, Ney et al. could show that the five enzymes known for the biosynthesis of the cofactor are incorporated in the genome of at least 653 bacterial and 173 archaeal species and demonstrated that the cofactor is widely synthesized in soil ecosystems and is, thus far, more important in aerobic bacterial metabolism than previously thought [12].

The first discovery of the cofactor occurred in methanogenic Euryarcheota in the year 1972 [13] and after the structure and biosynthesis pathway was established it was assumed that it is universal in all F₄₂₀ producing organisms. Novel investigations of F₄₂₀ biosynthesis proclaim that its biosynthesis pathway differs in various organisms and is not universal [14]. The variations derive from the various substrate utilized to link Fo with its poly-glutamate tail to form the cofactor F₄₂₀. According to the current state of knowledge three different F₄₂₀ biosynthesis pathway are known which could be found in Actinobacteria, Euryarcheota, and Proteobacteria [14]. All three pathways have in common that the Fo headgroup is formed from 5-Amino-6-ribitylamino-2,4(1H,3H)-pyrimidinedione and L-tyrosine. In methanogens this reaction is mediated from the enzymes CofG and CofH, while in mycobacteria the enzyme FbiC is responsible for this reaction [14].

The number of glutamates in the glutamyl chain differs among numerous organisms that express this cofactor. In mycobacteria, the cofactors mostly contain five to seven glutamate residues [15]. Cofactors with two or three glutamate residues seem to prevail in methanogens without cytochromes, whereas four to five glutamate residues are more common in methanogens with cytochromes [16]. Peck [17] and Wunderer et al. [18] also showed that the proportion of various F₄₂₀ cofactor analogs did not remain stable in some methanogenic Archaea during batch cultivation and that even the prevailing carbon source can have an effect on the proportion of cofactor analogs. In Wunderer et al. [18], the proportions of various glutamate tail lengths in Methanosarcina thermophila, a cytochrome-containing methanogen, changed with increasing incubation time, whereas tail length-composition remained the same in Methanoculleus thermophilus, a methanogen without cytochromes. The length of the poly-glutamate tail modulates the binding affinity of F₄₂₀ cofactors with oxidoreductases. Long-chain (five to eight glutamates) F₄₂₀ cofactors bind with a six-to ten-fold higher affinity to these enzymes than short-chain (two glutamates) ones [19]. Additionally, the side chain significantly affects the kinetics of the enzyme, whereby long-chain F₄₂₀ increases the substrate affinity (lower K_m) and reduces the turnover rate (lower k_cat) of these enzymes [19].

Besides the knowledge of the glutamyl tail length spectrum of certain species, orders, or even classes little is known about the glutamyl tail length spectra of complex microbial communities in various habitats. Despite its promising future applicability in industrial contexts, it is important to take one step back and evaluate the in-situ expression of glutamyl tail length spectra in complex, microbial habitats from a basic-scientific point of view. To our best knowledge, this is the first study dealing with glutamyl tail length distribution in natural communities. Consequently, the aim of the present study was i) to perform an inventory of F₄₂₀ glutamyl tail length in various habitats, ii) to investigate the glutamyl tail length spectra with regard to an overall oxygen content estimate in these habitats, iii) to link physico-chemical parameters with F₄₂₀ glutamyl tail length and iv) to compare these with microbiome data to obtain a first insight into F₄₂₀ production by various microorganisms in various habitats.

2. Material and methods

2.1. Sampling and experimental design

Various environmental samples, including soil, sludge (anaerobic digestion plant), compost (composting plant), and manure were taken and analyzed in triplicates each in and around Innsbruck (Tyrol, Austria, Fig. 1, Table 1) to represent habitats with various oxygen content. The temperature was measured on-site before samples were taken. For the soil samples, a 20 × 20 cm square was cut out and the top layer (horizon A) was removed before sampling. Samples were immediately brought to the laboratory and stored at 4 °C. Soil samples were sieved to 4 mm. Due to the low water content manure samples had to be suspended in water (water: sample ratio 1: 3.5), mixed with an overhead shaker for 30 min, and vented every 10 min. Physico-chemical parameters (pH, dry weight, C:N ratio) were assessed and microbial community (16S rRNA amplicon sequencing) as well as F₄₂₀ tail length analyses were conducted.

Fig. 1.

Map of the Inn Valley near Innsbruck locating sampling spots (scale 1:200 000, created with QGIS 3.34.7). 1 = Mühlau (mixed forest), 2 = Lake Lans (coniferous forest), 3 = Innsbruck (meadow), 4 = Innsbruck (arable soil), 5 = Innsbruck (anaerobic digester), 6 and 7 = Innsbruck (compost), 8 = Lake Lans (swamp), 9 = Innsbruck (mixed manure), and 10 = Polling (horse manure).

Table 1.

List of the various sample types with coordinates of the sample location.

Location	Sample	Longitude	Latitude	Sea level [m]
Mühlau, mixed forest (1)	soil	11°24'9.6"E	47°16'56.2"N	657
Lake Lans, coniferous forest (2)	soil	11°25'5.5"E	47°14'31.6"N	885
Innsbruck, meadow (3)	soil	11°20'20.9"E	47°16'3.6"N	613
Innsbruck, arable soil (4)	soil	11°20'21.2"E	47°16'2.3"N	613
Innsbruck, anaerobic digester (5)	sludge	11°26'47.6"E	47°15'50.7"N	653
Innsbruck, compost (6)	10-day compost	11°27'13.8"E	47°15'58.5"N	582
Innsbruck, compost (7)	6-month compost	11°27'14.1"E	47°15'58.0"N	582
Lake Lans, swamp (8)	soil	11°25'9.1"E	47°14'33.5"N	861
Innsbruck, mixed manure (9)	manure	11°20'14.5"E	47°15'56.7"N	579
Polling, horse manure (10)	manure	11°9'28.7"E	47°16'48.2N	613

2.2. Physico-chemical analyses

For the pH measurement, the various samples were mixed with 0.01 M CaCl₂ (ratio 1: 2.5) and incubated for 3 h. Afterward, the pH was measured using an electrode (Metrohm, Swiss) according to previous protocols [20]. 125 mg heat dried (105 °C, overnight) sample material was used for the evaluation of the C: N ratio [20]. The analysis was done in triplicates on a CN Analyzer CN828-MC (LECO, USA) according to the manufacturer's protocols. Physio-chemical parameters are given in Table 2.

Table 2.

Physico-chemical properties of the various sample locations.

Sample location	Temperature [°C]	pH	Ratio [C:N]	Dry weight [g/g]
Arable Soil	20.80 ± 0.29	7.06 ± 0.11	45.70 ± 8.26	0.87 ± 0.00
Digester sludge	38.07 ± 0.13	7.40 ± 0.01	5.44 ± 0.04	0.04 ± 0.00
Swamp soil	16.27 ± 0.19	6.26 ± 0.17	14.32 ± 0.63	0.40 ± 0.03
10-days compost	53.47 ± 0.27	7.97 ± 0.02	17.84 ± 1.67	0.60 ± 0.01
6-months compost	30.07 ± 0.61	8.04 ± 0.06	14.75 ± 1.22	0.47 ± 0.02
Mixed manure	50.13 ± 0.82	8.22 ± 0.03	14.16 ± 0.35	0.27 ± 0.01
Mixed forest soil	16.43 ± 0.82	6.50 ± 0.08	16.81 ± 1.92	0.92 ± 0.02
Coniferous forest soil	14.77 ± 0.17	3.56 ± 0.19	22.71 ± 1.39	0.76 ± 0.04
Horse manure	30.43 ± 0.53	7.05 ± 0.01	12.31 ± 0.13	0.34 ± 0.03
Meadow soil	19.63 ± 0.21	7.15 ± 0.04	14.32 ± 0.27	0.75 ± 0.01

2.3. Cell disruption, solid-phase extraction, and F₄₂₀ tail length variants analyses via high performance liquid chromatography (HPLC)

For cofactor F₄₂₀ analyses, cells were disrupted through heat treatment (autoclaving) and the cofactor purified by solid-phase extraction: 5 g of the various sample materials were filled up to 10 mL with distilled water and 10 mL of cell disruption buffer was added (200 mM potassium dihydrogen phosphate, 50 mM ethylenediaminetetraacetic acid, and 1% (v/v) polysorbate 80, pH 7) and subjected to a temperature-pressure treatment applying 121 °C and 1.2 bar pressure for 30 min. Subsequently, the cell lysate was centrifuged for 5 min at 10 000×g and 8 mL of the supernatant was used for solid-phase extraction according to Markt et al. [21] The various tail length variants of the cofactor F₄₂₀ were separated via reversed-phase ion-pair high performance liquid chromatography. The separation occurred on a prominence HPLC system (Shimadzu, Japan) equipped with a Gemini-NX C18 5μ 110 A 150 × 3 mm column (Phenomenex, Germany), an UV/VIS (@210 nm) and a fluorescence detector (absorption 420 nm, emission 475 nm). The column was heated at 40 °C and a flow rate of 1.0 mL/min was applied. The mobile phase was a gradient of solvent A (aqueous solution of 10 mM tetrabutylammonium hydroxide, 20 mM di-ammonium hydrogen phosphate, pH 7 adjusted with 85% phosphoric acid) and solvent B (100% acetonitrile) and varied as described in [18].

2.4. Amplicon sequencing

DNA was extracted in triplicates for each sampling location with the NucleoSpin® Soil Kit (MCHEREY&NAGEL, Germany) according to manufactures recommendations using 250–1000 mg fresh sample. Buffer SL1 was used as lysis buffer and elution occurred in 50 μL. The DNA was quantified fluorometrically using Quant-iT™ PicoGreen™ dsDNA Assay Kit (Invitrogen, USA) and an Anthos-Zenyth Multimode Detector [22]. Obtained DNA concentrations can be found in Table S2.

The small subunit (SSU) rRNA gene primers 515f and 806r [23] were used to target the V4 region, according to the Earth Microbiome Project [24]. The NGS library was prepared in-house according to well established protocols [25] with following modifications: The PCR mix contained 5 μL Q5® reaction buffer 5x (NEB, Germany), 0.25 μL Q5® Hot Start High-Fidelity DNA Polymerase (NEB, Germany), 10 mM dNTPs, 500 nM forward and reverse primer, 20% enhancer (NEB, Germany), 5 ng template and PCR-grade water to fill up to 25 μL for each reaction. Thermocycling conditions: 25 cycles of denaturation for 10 s, annealing at 57 °C (second PCR step: 56 °C) for 20 s, and elongation for 20 s. PCR products were purified using the Monarch PCR & DNA Cleanup Kit (NEB, Germany). The DNA was checked for contaminants via the NanoDrop 2000c™ system and the DNA quantity was measured via QuantiFluor® dsDNA Dye (Promega, USA). A final ready-to-load sample pool of 13 ng/μL, with a 260/280 absorbance ratio of 1.86 [25,26], was sent to Microsynth AG (Switzerland) for sequencing on a MiSeq™ System (Illumina®, USA) according to the company's protocols.

The raw reads were processed with mothur version 1.45.2 [27]. A contig file was created with the paired-end reads resulting in 6 396 503 sequences across all 32 samples (three parallels of the ten sample locations and two MOCK samples). Two different defined MOCK communities were used for the NGS validation. The ZymoBIOMICS™ Microbial Community standard (containing eight bacterial and two yeast microorganisms, further referred as Mock1) and the archaeon Methanosarcina thermophila DSM 1825 (DSMZ, German Collection of Microorganisms and Cell Cultures, further referred to as Mock2). Unique sequences were aligned with the SILVA V138.1 database [28]. Chimeric amplicons were removed with the VSEARCH algorithm and the sequences were classified with the k-nearest neighbor algorithm. The minimal sampling depth was 42 778 sequences per sample. All microorganisms of the MOCK community could be found after the amplicon sequencing.

2.5. Statistical and graphical analyses

All spreadsheets and the transformation (log 10) of the F₄₂₀ data were done with Excel® (Microsoft®). The map of the sample locations was created with QGIS 3.34.7 Prizren (QGIS Development Team, Swiss). Graphs and analyses of variance (ANOVA) for F₄₂₀ data were created with Statistica™ 13 (TIBCO® Software Inc., USA), whereas RStudio 2023.09.1 + 494 with R version R-4.4.1 (Posit PBC, USA) was used for further analyses and graphical illustrations: The heatmap, showing the most abundant F₄₂₀ producing taxa (≥50reads) over all samples, was done with the packages readxl [29], pheatmap [4,30] and extrafont [31]. Hierarchical clustering was used for defining clusters. Stacked bar plots of metagenomic data were done with the packages as described earlier as well as the packages microbiome [32], randomcoloR [33] and dplyr [34], tidyverse [35], and ggeasy [36]. Nonmetric multidimensional scaling was done with the packages vegan version 2.6–6.1 [37], readxl and extrafont. F₄₂₀ tail length data were Hellinger transformed prior ordination using Bray distance. Oxygen availability was dummy coded (1: low, 2: medium, 3: high) and environmental data were overlaid as vectors using the vegan command envfit. Stress value was 0.8 and the non-metric fit showed an R² of 0.99.

$$F_{420}averagelength=\frac{{\sum }_{k=F_{420}-2}^{F_{420}-10}{F}_{420}k\ast k}{\sum _{k=F_{420}-2}^{F_{420}-10}{F}_{420}k}$$ (1)

2.6. Identification of potentially F₄₂₀ producing microorganisms

The results of the in silico analysis part from Ney et al. [12] were used to identify microorganisms potentially able to produce cofactor F₄₂₀. Ney et al. provided in their study a list of the 653 bacterial and 173 archaeal species (whose genomes were available on NCBI in 2017) which incorporated all five known enzymes for F₄₂₀ biosynthesis in their genome. This list was compared with the results of the amplicon sequencing of the presented study to identify potential F₄₂₀ producing microorganisms at genus level and to compare them with the glutamyl tail length spectra of the respective sample. The genome of many potentially F₄₂₀ producing microorganisms is still unknown and could therefore not be included in this study. For this reason, the identified F₄₂₀ producing genera likely represent an underestimation of the actual F₄₂₀ producing microbes.

3. Results and discussion

3.1. Distribution of F_420-n in different habitats

The total concentration of the F₄₂₀ cofactor derivates of the various sample locations were logarithmized (log 10) and are shown in Fig. 2. Horse manure samples showed the highest amount of F₄₂₀ cofactor derivates, followed by digester sludge samples and mixed manure samples. In all other samples, cofactor concentrations were rather similar (Fig. 2) and approximately two orders of magnitude lower.

Fig. 2.

Total concentration of F₄₂₀ derivates (F₄₂₀-2 - F₄₂₀-10; F₄₂₀-2:10) for various sample locations, shown as log 10 peak area count/g dry weight [pac/g dw].

The high concentration of F₄₂₀ in samples with limited or very low O₂ availability (digester sludge, mixed manure, and horse manure) indicates the important role of the redox cofactor F₄₂₀ during the absence of O₂ as a terminal electron acceptor. However, the swamp soil and the two compost types, also characterized by limited O₂ availability at least in specific niches, showed significantly lower concentrations of cofactor F₄₂₀. However, considerable amounts of the cofactor could also be isolated from samples with a better O₂ availability (arable soil, mixed forest soil, coniferous forest soil, and meadow soil). This confirms the findings from Ney et al. [12] that the cofactor F₄₂₀ is far more important in aerobic habitats than previously thought.

The proportions of the various cofactor F₄₂₀ derivates (F₄₂₀-2: F₄₂₀-10) of the total F₄₂₀ derivates in the respective sample location are shown in Fig. 3 and the standard deviation of each cofactor derivate in Table S1. In the arable soil samples, all F₄₂₀ cofactor derivates could be proven except the cofactor with two glutamate residues (F₄₂₀-3 to F₄₂₀-10). The cofactor with six glutamate residues accounted for the largest share reflecting 28.9% of total F₄₂₀, followed by F₄₂₀-5 at 25.5%, F₄₂₀-7 at 16.9%, F₄₂₀-8 at 10.2%, and F₄₂₀-4 at 9.6% (Fig. 3). The cofactors with three, nine, and ten glutamate residues contributed less than 5% of total F₄₂₀ derivates. The glutamyl tail length spectrum in the meadow soil samples was very similar to the glutamyl tail length spectrum of the arable soil. The most abundant cofactor was F₄₂₀-5 at 26.7%, followed by F₄₂₀-6 at 25.3%, F₄₂₀-7 at 15.6%, F₄₂₀-4 at 11.3%, F₄₂₀-8 at 9.2%, and F₄₂₀-3 at 8.1%; F₄₂₀-2 was also not detected (Fig. 3). The cofactors with nine and ten glutamate residues again were least abundant and contributed less than 5%.

Fig. 3.

Proportions [%] of the various F₄₂₀ derivates of total F₄₂₀ in the respective sample.

In digester sludge, all of the F₄₂₀ cofactors (F₄₂₀-2: F₄₂₀-10) were found, whereby cofactors with two and three glutamate residues constituted the majority: Cofactor F₄₂₀-2 contributed 50.1% and F₄₂₀-3 44.9% to total F₄₂₀ derivate pool. The cofactor with four glutamate residues reached 2.6% and the cofactors with five to ten glutamate residues stayed under 1% of total F₄₂₀ (Fig. 3). In the swamp soil, all F₄₂₀ cofactors could be evidenced except for F₄₂₀-7. The most abundant cofactor was F₄₂₀-3 at 32%, followed by F₄₂₀-5 at 24%, F₄₂₀-9 at 18.9%, F₄₂₀-4 at 13.3%, and F₄₂₀-6 at 5.5%. The cofactors with two, eight, and ten glutamate residues were the least abundant with less than 5%.

In the 10-days compost, all cofactors were found except the cofactor with two glutamate residues; predominant was cofactor F₄₂₀-3 reflecting more than half of total F₄₂₀ (53.3%). The second most abundant cofactor was F₄₂₀-8 at 11.1%, followed by F₄₂₀-5 at 8.3%, F₄₂₀-7 at 8.2%, F₄₂₀-4 at 8%, and F₄₂₀-6 at 6.3% (Fig. 3). The least abundant cofactors were those with nine and ten glutamate residues (less than 5%). The glutamyl tail length spectrum in the 6-months compost samples was very similar to the glutamyl tail length spectrum in the 10-days compost samples. In the 6-months compost, cofactors with three and four glutamate residues became more abundant, while at the same time, a decrease of cofactors with six to ten glutamate residues was observed. It could be shown that during the maturation of the compost (10-days to 6-months), the glutamyl tail length spectrum is slightly shifting towards F₄₂₀ cofactors with shorter glutamyl tails. However, in the 6-months compost samples, the cofactors with two and nine glutamate residues could not be found, whereas cofactor with three glutamate residues was the most abundant at 66.3%, followed by F₄₂₀-4 at 12.1%, and F₄₂₀-5 at 8.3% (Fig. 3). Cofactors with six, seven, eight, and ten glutamate residues were reflecting less than 5% of total F₄₂₀ derivates.

Besides digester sludge, mixed manure was the only sample where all cofactors (F₄₂₀-2: F₄₂₀-10) could be verified (Fig. 3). F₄₂₀-3 was predominant at 71%, followed by F₄₂₀-2 at 9.2%, and F₄₂₀-4 at 5.3%. The cofactors with five to ten glutamate residues were below 5% of total F₄₂₀. Horse manure had a similar glutamyl tail length spectrum when compared to mixed manure regarding short chained F₄₂₀ cofactors. The most abundant cofactor was F₄₂₀-3 at 89.5%, F₄₂₀-2 s at 9.7% and F₄₂₀-4 still accounted for 0.8%. Mid- and long-chained F₄₂₀ cofactors (five to ten glutamate residues) were completely missing in horse manure samples.

In the mixed forest soil, the cofactor with three glutamate residues was the most abundant at 31.3% of total F₄₂₀, followed by F₄₂₀-6 at 23.5%, and F₄₂₀-5 at 16.6%. F₄₂₀-4 and F₄₂₀-7 accounted for 11.1% and F₄₂₀-8 for 6.3% (Fig. 3). The cofactors with two, nine, and ten glutamate residues could not be detected in the mixed forest soil. The glutamyl tail length spectrum of cofactor F₄₂₀ in the coniferous forest soil was very similar to that in the mixed forest soil, except for some differences regarding cofactors F₄₂₀-3 and F₄₂₀-8. In the coniferous forest soil, the predominant cofactor was F₄₂₀-3 at 36.8%, followed by F₄₂₀-5 at 17.8%, F₄₂₀-6 at 17.2%, F₄₂₀-4 at 15.9%, and F₄₂₀-2 at 6.5%. The cofactors with seven, eight, and nine glutamate residues contributed with less than 5% to the total F₄₂₀ derivate pool, and the cofactor with ten glutamate residues could not be proven in the coniferous forest soil.

The average glutamyl tail length of the cofactor F₄₂₀ was calculated for each sample and is shown in Fig. 4. The shortest average glutamyl tail length could be proven in the digester sludge samples with 2.62 ± 0.005 glutamate residues while the second shortest tail length could be verified in horse manure with 2.91 ± 0.008, followed by mixed manure with 3.49 ± 0.009, 6-months compost with 3.85 ± 0.471, coniferous forest soil with 4.22 ± 0.701, 10-days compost with 4.62 ± 0.391, mixed forest with 4.91 ± 0.293, swamp soil with 5.12 ± 0.134, meadow soil with 5.73 ± 0.104, and arable soil with 6.02 ± 0.056. However, according to the KRUSKAL-WALLIS ANOVA, the average F₄₂₀ glutamyl tail length in the digester sludge samples was significantly shorter than in the arable soil and meadow soil samples. The average glutamyl tail length of the cofactor in the horse manure was also significantly shorter than in the arable soil, whereas no significant differences could be found between all other sample locations. Additionally, the average F₄₂₀ glutamyl tail length for samples with various oxygen availability (low, medium, and high) was calculated (Fig. S1). In samples with low oxygen availability the shortest average glutamyl tail length could be proven with 3.01 ± 0.362 glutamate residues, which was significantly shorter than in the samples with medium and high oxygen availability. Samples with medium oxygen availability exhibited a mean glutamyl tail length of 4.53 ± 0.634 whereas samples with higher oxygen availability showed a mean glutamyl tail length of 5.22 ± 0.806.

Fig. 4.

Mean F₄₂₀ glutamyl tail length of samples from various locations. Significant differences (p < 0.05) are indicated by various characters.

The cofactor F₄₂₀ has one of the lowest redox potentials of all currently known redox factors [1] and is therefore perfectly suited to take up electrons which accrue during the oxidation of various compounds (also various environmental pollutants). Ney et al. [19] could demonstrate that the length of the glutamyl tail is decisive for the binding affinity and turnover rate of the cofactor with oxidoreductases. Long-chain (five to eight glutamates) F₄₂₀ binds with a six-to ten-fold higher affinity to oxidoreductases than short-chain F₄₂₀ (two glutamates) [19]. Conversely, this suggests that a shorter glutamyl chain length of the cofactor F₄₂₀ leads to a lower binding affinity and higher turnover rates and indicates the high potential for bioremediation of environments with shorter F₄₂₀ glutamyl tail length as observed in the present study for e. g. digester sludge.

NMDS analysis (Fig. 5) demonstrates F₄₂₀ tail length in the relationship to environmental data. While habitats with low oxygen concentrations (digester sludge, mixed manure, horse manure) tended to result in shorter tail length spectra, longer ones were found with higher oxygen availability, thus, confirming results from analysing mean F₄₂₀ tail length (Fig. 4).

Fig. 5.

Nonmetric multidimensional scaling (NMDS) ordination of F₄₂₀ tail length data of all investigated habitats. Blue vectors show environmental data and F₄₂₀ cofactor derivates are plotted in red. Each point represents a sample and dashed circles enclose all points of the respective habitat. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.2. Microbial community composition

The relative abundance of the 50 most abundant genera across all samples is depicted in Fig. 6 to microbiologically characterize the samples briefly. Additionally, a relative abundance plot of 15 most abundant genera was made and is shown in Fig. S2. The abundance of microorganisms in the various soils was very similar except for the coniferous forest soil. In the arable soil, Nitrososphaeraceae unspecified genus was the most abundant taxa, followed by Vicinamibacterales genus as well as another unspecified genus of Vicinamibacteraceae (Fig. 6). In the mixed forest soil, Vicinamibacterales genus was the most abundant microorganism, followed by not specified genera of Rokubacteriales and Vicinamibacteraceae (Fig. 6), whereas meadow soil contained the same genera as the arable soil, with Vicinamibacterales genus being the most abundant one. Also, in the swamp soil the Vicinamibacterales genus was the most abundant one, followed by the genera Rokubacteriales and Subgroup_17 which belongs to the class Vicinamibacteria (Fig. 6). In the coniferous forest soil, which showed the greatest differences in microbial abundance among the studied soils, the genus Acidothermus, which belongs to the class Actinobacteria, was most abundant. The second and third most abundant genera were the Isosphaeraceae and the Gemmataceae with no cultivated representatives (Fig. 6).

Fig. 6.

Relative abundance plot of the 50 most anundant genera across all samples.

In digester sludge samples, a genus from the Anaerolineaceae was the most abundant one and accounted for almost 50% of the sequences, followed by Methanosaeta spp. and Bacteriodetes_vadinHA17 (Fig. 6). In the 10-days compost samples, the genus SBR 1031, which belongs to the class Anaerolineae, was the predominant genus, followed by Anaerolineaceae genus which was the most abundant genus in the digester sludge samples. The third most abundant genus in the 10-days compost was OM190 which belongs to the phylum Planctomycetota (Fig. 6). The 6-months compost samples were very similar to the 10-days compost samples. The predominant genus was again SBR1031, followed by Anaerolineaceae genus and Vicinamibacterales genus.

The most abundant genus in horse manure samples was Lysinibacillus belonging to the order Bacillales and second most abundant was the genus WCHB1-41 which belongs to the class Kiritimatiellae, followed by Akkermansia spp. belonging to the order Verrucomicrobiales (Fig. 6). By contrast, in the mixed manure the most abundant genus was the Hydrogenispora which accounted for approximately 50% of the sequences. The second most abundant genus was Truepera, followed from the genus Lysinibacillus (Fig. 6).

This short description of the dominant microbial genera in the various samples serves as an overview and general characterization of the samples but cannot inform on microbial F₄₂₀ producing microorganisms. In a next step, microbiome data were used to compare microorganisms found in this study with known F₄₂₀ producers that were shown to be capable of synthesizing F₄₂₀-n genetically [12].

3.3. F₄₂₀ producing community

The results of the in silico analysis of Ney et al. [12] were used to identify microorganisms capable of producing cofactor F₄₂₀ in various environmental samples. Fig. 7 is linking the abundance of F₄₂₀ producing microorganisms (showing an abundance of more than 50) with the origin of sample. Regarding the F₄₂₀ producing microbial community soil samples clearly separated from those of manure, compost and sludge habitats (Fig. 7). While in soil abundant bacterial genera were mainly responsible for the composition of the F₄₂₀-n pool, in manure and sludge, but also in compost, methanogens accounted for F₄₂₀ production. In the following, the most important producers of F₄₂₀ in the various habitats are pinpointed.

Fig. 7.

Heatmap (log 10) of F₄₂₀ producing genera according to in-silico analysis of Ney et al. [12]. Genera with an abundance of n ≥ 50 across all samples were included.

The bacterial genus Acidothermus was the most abundant F₄₂₀ producer in the coniferous forest soil, followed by the genera Bradyrhizobium, Conexibacter, and Mycobacterium (Fig. 7). Nothing is known about the glutamyl tail length spectra of the genus Acidothermus, but it belongs to the class of Actinobacteria and according to Daniels et al. [38] they should produce only small amount of the cofactor F₄₂₀ compared to methanogenic Archaea. Bradyrhizobium belongs to the class of Alphaproteobacteria and the genus Conexibacter to Thermoleophilia; however, details on their glutamyl tail length spectra are missing so far. Members of the genus Mycobacterium should tend to attach mainly five to seven glutamate residues to the cofactor F₄₂₀ [15]. Nevertheless, the cofactor F₄₂₀-3 was the most abundant cofactor derivate in the coniferous forest soil, followed from the cofactors with five and six glutamate residues (Table 3). In the arable soil, Nocardioides was most abundant genus, followed by the bacterial genera Solirubrobacter, Pseudonocardia, Blastococcus, Sphingomonas, and Rubrobacter (Fig. 7). The genera Nocardioides, Pseudonocardia, Blastococcus are aerobic members of the class Actinobacteria and should only produce small amounts of the cofactor F₄₂₀ [38]. Moreover, the genus Solirubrobacter belongs to the class Thermoleophilia, Sphingomonas to Alphaproteobacteria, and Rubrobacter to Rubrobacteria. So far, their glutamyl tail length spectra have not been investigated. In the arable soil, F₄₂₀ with six and five glutamate residues were predominant in the present study (Table 3). In the meadow soil, the genera Mycobacterium, Solirubrobacter, Pseudonocardia, Rubrobacter, and Nocardioides were predominant (Fig. 7). The genus Mycobacterium, a member of the class Actinobacteria, is a known F₄₂₀ producer and was described to attach mainly five to seven glutamates to the cofactor according to Bair et al. [15]. Solirubrobacter is a genus of the class Thermoleophilia, Pseudonocardia and Nocardioides belong to Actinobacteria and the genus Rubrobacter to the class Rubrobacteria. Their F₄₂₀ glutamyl tail length spectra have not been investigated so far. In meadow soil samples, F₄₂₀ with five and six glutamate residues were dominating which is consistent with the high abundance of the genus Mycobacterium in these samples (Table 3). In the mixed forest, the genera Mycobacterium, Solirubrobacter, and Bradyrhizobium were predominant (Fig. 7). The genus Mycobacterium belongs to the class Actinobacteria and was described to mainly attach five to seven glutamate residues to the cofactor F₄₂₀ [15]. The genus Solirubrobacter belongs to the class Thermoleophilia and Bradyrhizobium is a member of the order Rhizobiales which belongs to the class Alphaproteobacteria. So far, nothing is known about the glutamyl tail length of these two genera. However, in the present study the cofactors with three, six, and five glutamate residues were the most abundant in the mixed forest soil which is consistent with the high abundance of members of the genus Mycobacterium (Table 3). In the swamp soil no clear dominant genus could be found (Fig. 7). The most abundant genera in these samples were Mycobacterium, Conexibacter, Solirubrobacter, and Rubrobacter (Fig. 7). As already mentioned in the genus Mycobacterium the cofactors with five to seven glutamate residues should be dominant [15]. Whereas, the genera Conexibacter and Solirubrobacter are members of the class Thermoleophilia and the genus Rubrobacter belongs to the Rubrobacteria. These three genera have in common that the glutamyl tail spectra of their cofactor F₄₂₀ derivates have not been investigated so far. The dominant cofactor derivates in the swamp soil were the F₄₂₀-3, F₄₂₀-5, and F₄₂₀-9. The high abundance of the F₄₂₀-5 in these samples concurs with the high abundance of the genus Mycobacterium known for predominant F₄₂₀-5 to F₄₂₀-7 production (Table 3).

Table 3.

Summary of the dominant F₄₂₀ producing genera and F₄₂₀ cofactors of the respective sampling area.

Sample location	Dominant F420 producing genera	Dominant F420-n [%]
Coniferous forest soil	Acidothermus, Bradyrhizobium, Conexibacter, Mycobacterium	F420-3 [36.8]
		F420-5 [17.8]
		F420-6 [17.2]
Arable soil	Nocardioides, Solirubrobacter, Pseudonocardia, Blastococcus, Sphingomonas, Rubrobacter	F420-6 [28.9]
		F420-5 [25.5]
		F420-7 [16.9]
Meadow soil	Mycobacterium, Solirubrobacter, Pseudonocardia	F420-5 [26.7]
		F420-6 [25.3]
		F420-7 [15.6]
Swamp soil	Mycobacterium, Conexibacter, Solirubrobacter, Rubrobacter	F420-3 [32.0]
		F420-5 [24.0]
		F420-9 [18.9]
Mixed forest soil	Mycobacterium, Solirubrobacter, Bradyrhizobium	F420-3 [31.3]
		F420-6 [23.5]
		F420-5 [16.6]
Mixed manure	Pseudomonas, Methanosarcina, Halomonas, Methanoculleus	F420-3 [71.0]
		F420-2 [9.2]
		F420-4 [5.3]
10 days-compost	Methanosarcina, Longispora, Thermobispora	F420-3 [53.3]
		F420-8 [11.1]
		F420-5 [8.3]
6-months compost	Methanosarcina, Longispora, Micromonospora	F420-3 [66.3]
		F420-4 [12.1]
		F420-5 [8.3]
Digester sludge	Methanosaeta, Methanobacterium, Novosphingobium	F420-2 [50.1]
		F420-3 [44.9]
		F420-4 [2.6]
Horse manure	Methanocorpusculum, Methanobrevibacter, Arthrobacter	F420-3 [89.5]
		F420-2 [9.7]
		F420-4 [0.8]

In the various soil samples, similar F₄₂₀ genera popped up e.g. like Mycobacterium, Solirubrobacter, Pseudonocardia, or Rubrobacter. The appearance of mainly aerobic mircoorganisms in coniferous and mixed coniferous forest-, arable-, meadow-, and swamp soil samples being responsible for F₄₂₀ production proofs the in-silico analysis of Ney et al. [12] stating that the ability of F₄₂₀ production is more wide-spread in bacterial genomes also in aerobic habitats than previously assumed for a long time. Regarding F₄₂₀ production anaerobic microorganisms colonizing anaerobic soil niche in the commonly as aerobic considered habitat did not seem to contribute significantly to the F₄₂₀ pool when looking at 16S rRNA gene-based abundance data.

Another picture was obtained for habitats that are considered as (mainly) anaerobic. In the mixed manure samples the genera Pseudomonas, Methanosarcina, Halomonas, and Methanoculleus were the predominant F₄₂₀ producing microorganisms (Fig. 7). Pseudomonas ssp. and Halomonas ssp. belong to the class of Gammaproteobacteria and to our best knowledge nothing is known about the glutamyl tail length of the cofactor F₄₂₀ in members of these genera. In contrast, according to Gorris and van der Drift [16] and Wunderer et al. [18] members of the archaeal genus Methanosarcina should tend to attach mainly four to five glutamate residues to the cofactor, whereas in Methanoculleus ssp. cofactors with three and four glutamate residues should be predominant [18]. The cofactor with three glutamate residues was clearly the most abundant cofactor in these samples, reflecting 71% of all F₄₂₀ cofactor derivates (Table 3). This could indicate a major influence of hydrogenotrophic methanogens (Methanoculleus sp.) towards acetoclastic methanogens (Mehtanosarcina sp.) on the glutamyl tail length spectra in these samples. Wunderer et al. [18] could already show that the cofactor F₄₂₀ content in Methanoculleus thermophilus was 100-fold higher than in Methanocarcina thermophila. Regarding F₄₂₀ producing microorganisms 10-days and 6-months compost samples were very similar. The archaeal genus Methanosarcina was the predominant genus in both composts, followed by the bacterial genus Longispora. The third most abundant F₄₂₀ producer in the 10-days and 6-months compost was Thermobispora spp. and Micromonospora spp., respectively (Fig. 7). The genus Methanosarcina can use a broader spectrum of carbon sources and can perform acetoclastic, methylotrophic, or even hydrogenotrophic methanogenesis [39]. However, members of the genus Methanosarcina are specialized in methanogenesis at high acetate concentrations [40], and are therefore indicators for high acetate concentrations. According to Gorris and van der Drift [16] and Wunderer et al. [18], F₄₂₀ with four and five glutamate residues are predominant in Methanosarcina spp. Regarding the genera Longispora, Thermibospora, and Micromonospora, to our best knowledge, nothing is known about their glutamyl tail length spectra, however, as all three belong to the class Actinobacteria these organisms should only have contributed small amounts of the cofactor F₄₂₀ to the overall F₄₂₀ pool [38]. Nevertheless, in both compost types F₄₂₀-3 was the dominant F₄₂₀ cofactor, reflecting more than 50% of total F₄₂₀ (Table 3). In digester sludge, the archaeal genus Methanosaeta, a strictly acetoclastic methanogenic group [41], was clearly the most abundant F₄₂₀ producing microorganism based on 16S rRNA gene analysis, followed by archaeal Methanobacterium and bacterial Novosphingobium (Fig. 7). According to Gorris and van der Drift [16] and Wunderer et al. [18] methanogens with cytochromes mainly tend to attach four to five glutamates to the cofactor F₄₂₀. Wunderer et al. [18] could also show that in Methanosarcina thermophila, a facultative acetoclastic methanogen, the cofactors with four and five glutamate residues dominated regardless of the used substrate. The genus Methanobacterium is described as hydrogenotrophic and can use CO₂/H₂ or formate as substrate [42], and according Wunderer et al. [18], hydrogenotrophic methanogens mainly produce F₄₂₀ with two to four glutamate residues. The cofactor with two and three glutamate residues constituted the largest fraction of the F₄₂₀ cofactor spectrum (95%) in digester sludge samples (Table 3), indicating that hydrogenotrophic methanogens like Methanobacterium sp. and Methanobrevibacter spp. (Fig. 7) had the greatest impact on the glutamyl tail length spectrum in the digester sludge even though the genus Methanosaeta, an obligate acetoclastic methanogen turned out to be more abundant in these samples. However, Wunderer et al. [18] showed that the cofactor F₄₂₀ concentration during acetoclastic methanogenesis was 100-fold lower than during hydrogenotrophic methanogenesis possibly explain the high abundance of Methanosaeta spp. and concurrent high concentrations of F₄₂₀-2:3. During hydrogenotrophic methanogenesis, cofactor F₄₂₀ is directly involved as a redox cofactor for the methylene-H₄MPT-dehydrogenase and reductase reaction in the methanogenesis process, while it is not directly involved in acetoclastic methanogenesis and therein only required for other metabolic purposes, like the reduction of NADP or the detoxification of O₂ [1]. These preliminary results show that the DNA-based, relative abundance of microbial taxa do not necessarily reflect physiological properties and activities [43]. This seems especially true when comparing DNA approaches with F₄₂₀ glutamyl tail length profiles which have not been considered so far in natural habitats.

The genus Methanocorpusculum was predominant in horse manure, which otherwise showed the highest similarity with digester sludge samples regarding F₄₂₀ producing microorganisms. The second and third most abundant F₄₂₀ producer based on the findings of Ney et al. [12] were Methanobrevibacter spp. and Arthrobacter spp., respectively (Fig. 7). The members of the genus Methanocorpusculum reduce CO₂ to methane with different electron donators depending on the respective species [44]. Similarly, the genus Methanobrevibacter uses the CO₂ reducing pathway and therefore tends to mainly attach two to four glutamate residuals [1,18] to the cofactor F₄₂₀. Nothing is known about the glutamyl tail length spectrum of the cofactor F₄₂₀ in the genus Arthrobacter. However, this genus belongs to the class Actinobacteria and Daniels et al. [38] showed that several members of this class contained F₄₂₀ although to a less extent (at maximum 1% of the total amount of the cofactor that can be synthetized by methanogens). This indicates that the influence of these organisms on the glutamyl tail length spectrum should be rather low. In the horse manure samples, only F₄₂₀ cofactors with two, three, and four glutamates could be proven (Table 3), suggesting that mainly hydrogenotrophic methanogens were responsible for the determination of the glutamyl chain length spectrum in these samples [16,18]. This is in accordance with the obtained metagenomic data (Fig. 6).

4. Conclusion

The highest concentration of all F₄₂₀ cofactor derivates could be isolated from digester sludge, mixed manure, and horse manure going along with the predominantly anoxic characteristics of these habitats and indicating the important role of the cofactor for redox reactions in the absence of O₂ as a terminal electron acceptor. The cofactor content in all other samples was significantly lower. The swamp soil and the two compost samples had at least niches with limited O₂ availability but showed about two orders of magnitude lower cofactor contents than the strictly anaerobic habitats. A possible reason might be the higher contribution of hydrogenotrophic methanogens in digester sludge and horse manure samples, whereas, with regard to 16S rRNA gene sequencing data, acetoclastic methanogenesis (Methanosarcinales) was more abundant in compost samples. Hydrogenotrophic methanogenic cells need substantially more F₄₂₀ than those of acetoclastic methanogens as it is directly involved in hydrogenotrophic methanogenesis and thus lower F₄₂₀ concentrations were found in those samples. By contrast, the high cofactor content in the various aerobic soils confirms that F₄₂₀ is far more important in aerobic habitats than previously assumed and proved its ubiquitous abundance. The high concentration of the F₄₂₀-3 in samples with distinct properties like fluctuating O₂ and H₂0 availability points out its important role as electron carrier during redox reactions of the primary and secondary metabolism among prokaryotes.

These results give a preliminary overview of F₄₂₀ tail length spectra in various natural habitats with complex microbial interaction and highlight variations between data derived from metagenomic analyses and physiological parameters. However, further investigations are needed to enlighten the role of F₄₂₀ producing Archaea and Bacteria in various, natural habitats.

CRediT authorship contribution statement

Mathias Wunderer: Writing – review & editing, Writing – original draft, Visualization, Formal analysis. Rudolf Markt: Resources, Methodology, Investigation, Data curation, Conceptualization. Eva Maria Prem: Writing – review & editing, Visualization. Nico Peer: Data curation. Andja Mullaymeri: Data curation. Andreas O. Wagner: Writing – review & editing, Validation, Supervision, Resources, Project administration, Funding acquisition.

Data availability statement

The raw reads for the 16S amplicon sequencing are available on NCBI (National Center for Biotechnology Information) under the BioProject ID: PRJNA997793. Non-sequencing data will be made available on request. These data were not deposited into a publicly available repository.

Submission declaration

The present work has not been published previously, is not under consideration for publication elsewhere, and its publication is approved by all authors.

Ethics declarations

This publication does not deal with human embryos and fetuses, vulnerable individuals or groups, patients, invasive techniques, personal data or sensitive personal data or dual use issues. According to the Ethics Self-Assessment of the H2020 Program, there are no ethical concerns connected to this publication.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is/are the supplementary data to this article: