*Article* **Extracellular Enzymatic Activities of Oceanic Pelagic Fungal Strains and the Influence of Temperature**

**Katherine Salazar Alekseyeva 1, \* , Gerhard J. Herndl 1,2 and Federico Baltar 1, \***


**Abstract:** Although terrestrial and aquatic fungi are well-known decomposers of organic matter, the role of marine fungi remains largely unknown. Recent studies based on omics suggest that marine fungi potentially play a major role in elemental cycles. However, there is very limited information on the diversity of extracellular enzymatic activities performed by pelagic fungi in the ocean and how these might be affected by community composition and/or critical environmental parameters such as temperature. In order to obtain information on the potential metabolic activity of marine fungi, extracellular enzymatic activities (EEA) were investigated. Five marine fungal species belonging to the most abundant pelagic phyla (Ascomycota and Basidiomycota) were grown at 5 ◦C and 20 ◦C, and fluorogenic enzymatic assays were performed using six substrate analogues for the hydrolysis of carbohydrates (β-glucosidase, β-xylosidase, and *N*-acetyl-β-D-glucosaminidase), amino acids (leucine aminopeptidase), and of organic phosphorus (alkaline phosphatase) and sulfur compounds (sulfatase). Remarkably, all fungal strains were capable of hydrolyzing all the offered substrates. However, the hydrolysis rate (Vmax) and half-saturation constant (Km) varied among the fungal strains depending on the enzyme type. Temperature had a strong impact on the EEAs, resulting in Q<sup>10</sup> values of up to 6.1 and was species and substrate dependent. The observed impact of temperature on fungal EEA suggests that warming of the global ocean might alter the contribution of pelagic fungi in marine biogeochemical cycles.

**Keywords:** marine fungi; total extracellular enzymatic activity; kinetics; maximum velocity; halfsaturation constant

## **1. Introduction**

Fungi are eukaryotic and osmoheterotrophic organisms depending on organic matter to grow and obtain energy [1]. Osmotrophy involves the secretion of different enzymes to break down complex biological polymers into smaller monomers that can then be taken up through the cell wall [2]. Due to this conversion of organic matter, osmoheterotrophs such as fungi should be major players in the recycling of organic matter. In the marine environment, most of the research on extracellular enzymatic activity (EEA) has been focused on prokaryotes [3,4]. Only a few studies have reported on fungal EEA related to phytoplankton blooms [5] or the degradation of plant-derived matter [6]. As a result, the ecological role of fungi in the biogeochemistry of the oceans, which represents the largest habitat in the Earth's biosphere, remains poorly known [7]. This contrasts with recent evidence of fungal biomass dominating the bathypelagic marine snow [8]. Additionally, recent studies based on omics suggest that pelagic fungi harbor genes indicative of an active role in marine biogeochemical cycles [9]. It has also been shown that a large variety of carbohydrate active enzymes (CAZymes) are expressed in marine pelagic fungi [10]. Still, there is very limited information on the diversity of EEA in pelagic fungi in the ocean, and

**Citation:** Salazar Alekseyeva, K.; Herndl, G.J.; Baltar, F. Extracellular Enzymatic Activities of Oceanic Pelagic Fungal Strains and the Influence of Temperature. *J. Fungi* **2022**, *8*, 571. https://doi.org/ 10.3390/jof8060571

Academic Editor: Wei Li

Received: 27 April 2022 Accepted: 23 May 2022 Published: 26 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

how it might be affected by community composition and/or environmental parameters such as temperature.

Approximately 70% of the Earth's biosphere is composed of persistently cold environments, from the deep sea to polar regions [11,12]. Depending on the optimal growth temperature, organisms living there can be psychrophilic or psychrotrophic [13]. These organisms need to be well adapted to low temperatures, low nutrient availability, and light seasonality [14,15]. Moreover, as low temperatures influence the biochemical reaction rates, organisms must be prepared to overcome those challenges [11].

Here, we investigated the kinetic parameters (Vmax and Km) of the EEA of five oceanic fungal isolates using six different fluorogenic substrate analogues. β-glucosidase and β-xylosidase were used as a proxy for the degradation of plant-derived matter; *N*-acetyl-β-D-glucosaminidase were used for the utilization of animal and fungal chitinous compounds; alkaline phosphatase and leucine aminopeptidase were used for the cleavage of phosphate moieties and peptides, respectively; and sulfatase was used for the degradation of sulfate esters in macromolecules. We used five fungal strains isolated from the oceanic water column, with two strains belonging to the phylum Ascomycota and three to Basidiomycota. We selected these strains because Ascomycota and Basidiomycota are the most abundant pelagic fungal phyla [9,16,17]. Furthermore, the influence of temperature on fungal EEAs was determined.

#### **2. Materials and Methods**

#### *2.1. Culture of Fungi Species*

The fungal species *Blastobotrys parvus* (HA 1620), *Metschnikowia australis* (HA 635), *Rhodotorula sphaerocarpa* (HB 738), and *Sakaguchia dacryoidea* (HB 877) were obtained from the Austrian Center of Biological Resources (ACBR). All these species were isolated from Antarctic Ocean waters at temperatures ranging from -1.24 ◦C to 5.60 ◦C [18–21]. *M. australis* and *R. sphaerocarpa* were isolated close to the South Shetland Islands and Marguerite Bay, respectively. The fungus *Rhodotorula mucilaginosa* was isolated from the Atlantic Ocean at 21.03 ◦C during the Poseidon cruise on board of *RV* Sarmiento de Gamboa in March 2019. The maximum temperature growth reported is 25 ◦C for *B. parvus* [18] and *M. australis* [19], and 30 ◦C for *R. mucilaginosa* [22], *R. sphaerocarpa* [21], and *S. dacryoidea* [20]. In order to have fresh cultures, the pure isolates were cultured on yeast malt extract agar [23,24] for one week. Afterwards, an initial amount of each fungus was diluted in artificial seawater (30 g/L sea salts S9883 Sigma-Aldrich, Vienna, Austria) to obtain an OD<sup>660</sup> ≈ 1 [25]. The optical density (OD) was measured with a UV-1800 Shimadzu spectrophotometer. Then, 10 mL of this fungal culture was inoculated into an autoclaved growth medium containing 2 g/L of glucose, malt extract, peptone, and yeast extract; 35 g/L of artificial sea salts (S9883 Sigma-Aldrich); and 0.50 g/L of chloramphenicol. Afterwards, 150 mL of this medium containing fungi was filled in Schott bottles and to compare the effect of temperature, all strains were grown in triplicate at 5 ◦C and 20 ◦C on a rotary shaker (Jeio Tech ISS-7100 Incubated Shaker, Daejeon, South Chungcheong, Republic of Korea). The culture growth was tracked daily by OD. Once the exponential phase was reached, bottles with similar OD values were chosen in triplicate for further analysis (EEA and biomass).

#### *2.2. Determining Extracellular Enzymatic Activity and Fungal Biomass*

Fluorogenic substrate analogues such as 4-methylumbelliferyl β-D-glucopyranoside (M3633 Sigma-Aldrich), 4-methylumbelliferyl β-D-xylopyranoside (M7008 Sigma-Aldrich), and 4-methylumbelliferyl *N*-acetyl-β-D-glucosaminide (M2133 Sigma-Aldrich) were used to estimate the potential activity of the enzymes β-glucosidase (BGL), β-xylosidase (BXY), and *N*-acetyl-β-D-glucosaminidase (NAG), respectively (Table 1). These enzymes can hydrolyze cellulose [26,27], chitin, and xylan [28,29], respectively, thus mediating carbohydrate degradation by marine fungi. The hydrolysis of 4-methylumbelliferyl phosphate (M8883 Sigma-Aldrich) and N-succinyl-Ala-Ala-Pro-Phe-7-amido-4-methylcoumarin (L2145 Sigma-Aldrich) was used to estimate the potential enzymatic activity of alkaline

phosphatase (ALP) and leucine aminopeptidase (LAP), respectively. ALP is indicative of the capability of microbes to acquire inorganic phosphorus from organic molecules, and LAP is involved in the hydrolysis of proteins and peptides [30]. Finally, 4-methylumbelliferyl sulfate potassium salt (M7133 Sigma-Aldrich) was used to determine the activity of sulfatase (SUL) degrading sulfate esters in macromolecules. The hydrolysis of these fluorogenic substrates analogues were standardized to the corresponding fluorophores. The fluorophores methylcoumaryl amide (MCA) (A9891 Sigma-Aldrich) and methylumbelliferone (MUF) (M1381 Sigma-Aldrich) were dissolved in 2-methoxyethanol to obtain a final concentration of 100, 50, 10, and 1 µM, and 2000, 1000, 100, and 50 µM.

**Table 1.** Targeted enzymes with an analogue fluorogenic substrate and their respective standards, methylumbelliferyl (MUF) and methylcoumaryl (MCA) amide.


Sterile microplates of 96 wells with an F bottom and low protein binding (XT64.1, Carl Roth, Karlsruhe, Baden-Wurtemberg, Germany) were used. The standards were distributed to each biological triplicate to establish a standard calibration curve, and each biological triplicate without any addition was used as a blank to determine the background fluorescence of the medium. Serial dilutions of the fluorogenic substrate were established resulting in 12 final concentrations ranging from 100 to 0.05 µM. The fluorescence was measured with a Tecan Infinite 200 PRO at an excitation wavelength of 365 nm and an emission wavelength of 445 nm. An initial measurement was performed, and then, every hour, a measurement was made over a total period of 3 h. Between measurements, the microplates were incubated in the dark at their respective temperature.

For fungal biomass determination, combusted (450 ◦C; for 6 h) Whatman GF/F filters (WHA1825047 Sigma-Aldrich, 47 mm filter diameter) were individually wrapped in aluminum foil and weighed. Then, 40 mL of each fungal culture triplicate was gently filtered onto a combusted and weighed filter, and was dried at 80 ◦C for 3 days. Thereafter, the sample was weighed again to determine the fungal biomass as dry weight.

#### *2.3. Determination of the Kinetic Parameters of the Extracellular Enzymatic Activity (EEA)*

The increase in fluorescence over time (3 h) was transformed into hydrolysis rate [µmol L−<sup>1</sup> h −1 ] using the equation obtained from the standard calibration lines of MCA and MUF. The resulting hydrolysis rates were fitted directly with the Michaelis–Menten equation using nonlinear least-squares regression analysis with R software [31]. The enzymatic kinetic parameters maximum velocity (Vmax) and half-saturation constant (Km) were calculated. To obtain the biomass-specific activity, the Vmax was normalized to the fungal biomass obtained from the dry weight. Additionally, the Q<sup>10</sup> value was calculated to identify the dependence of enzyme activity on temperature [32].

#### *2.4. Statistical Analyses*

For the kinetic parameters Vmax and Km, a one-way analysis of variance (ANOVA) was performed to determine differences among species and temperature. Tukey's Honestly Significant Difference (Tukey's HSD) was used for a multiple and simultaneous comparison between species and to identify significance at the species level. A Student's T-test was used to test the normal distribution of the data. A principal component analysis (PCA) was performed to analyze the fungal enzymatic activity by species and substrate. For this

purpose, the prcomp command of the R software was used. To maximize the sum of the variance of the squared loadings, Varimax rotation was performed.

#### **3. Results**

Remarkably, all the fungal strains were hydrolyzing all the fluorogenic substrates offered (Figures 1 and 2). The kinetic parameters Vmax and K<sup>m</sup> of the EEA varied, however, among the different fungal strains and substrates (Figures 3 and 4). Additionally, the response of EEA to temperature was species and substrate dependent (Figures 5 and 6).

**Figure 1.** Vmax in µmol/g biomass/h obtained from the total enzymatic activity for the substrates representing carbohydrates such as (**A**) β-glucosidase (BGL), (**B**) β-xylosidase (BXY), and (**C**) *N*-acetylβ-D-glucosaminidase (NAG) of five marine fungal isolates *B. parvus, M. australis*, *R. mucilaginosa*, *R. sphaerocarpa*, and *S. dacryoidea* at 5 ◦C and 20 ◦C. According to Tukey's HSD, bars denoted by a different letter (a, b, and c) are significantly different (*p* < 0.05), whereas bars denoted by a common letter (ab) are not significantly different.

**Figure 2.** Vmax in µmol/g biomass/h obtained from the total enzymatic activity for the substrates representing proteins, phosphorus, and sulfur, such as (**A**) leucine aminopeptidase (LAP), (**B**) alkaline phosphatase (ALP), and (**C**) sulfatase (SUL) of the five marine fungal isolates *B. parvus*, *M. australis*, *R. mucilaginosa*, *R. sphaerocarpa*, and *S. dacryoidea* at 5 ◦C and 20 ◦C. According to Tukey's HSD, bars denoted by a different letter (a, b, c, and d) are significantly different (*p* < 0.05), whereas bars denoted by a common letter (ab and bc) are not significantly different.

**Figure 3.** K<sup>m</sup> in µM obtained from the total enzymatic activity for the substrates representing carbohydrates, such as (**A**) β-glucosidase (BGL), (**B**) β-xylosidase (BXY), and (**C**) *N*-acetyl-β-Dglucosaminidase (NAG) of the five marine fungal isolates (*B. parvus, M. australis*, *R. mucilaginosa*, *R. sphaerocarpa*, and *S. dacryoidea*). Measurements were performed at 5 ◦C and 20 ◦C in the exponential phase. According to Tukey's HSD, bars denoted by a different letter (a, b, and c) are significantly different (*p* < 0.05), whereas bars denoted by a common letter (ab) are not significantly different.

**Figure 4.** Km in µM obtained from the total enzymatic activity for the substrates representing proteins, phosphorus, and sulfur, such as (**A**) leucine aminopeptidase (LAP), (**B**) alkaline phosphatase (ALP), and (**C**) sulfatase (SUL) of the five marine fungal isolates (*B. parvus*, *M. australis*, *R. mucilaginosa*, *R. sphaerocarpa*, and *S. dacryoidea*). Measurements were performed at 5 ◦C and 20 ◦C in the exponential phase. According to Tukey's HSD, bars denoted by a different letter (a, b, and c) are significantly different (*p* < 0.05), whereas bars denoted by a common letter (ab and bc) are not significantly different.

β β β **Figure 5.** Q<sup>10</sup> of the normalized total enzymatic activity (Vmax) with the biomass (dry weight) for the substrates β-glucosidase (BGL), β-xylosidase (BXY), *N*-acetyl-β-D-glucosaminidase (NAG), leucine aminopeptidase (LAP), alkaline phosphatase (ALP), and sulfatase (SUL) and the species *B. parvus*, *M. australis*, *R. mucilaginosa*, *R. sphaerocarpa*, and *S. dacryoidea*. β β β

β β β **Figure 6.** PCA plot from the normalization of the total enzymatic activity (V ) with the β β β **Figure 6.** PCA plot from the normalization of the total enzymatic activity (Vmax) with the biomass (dry weight) at 5 ◦C and 20 ◦C for all the species. This was compared for each substrate, which corresponds to alkaline phosphatase (ALP), β-glucosidase (BGL), β-xylosidase (BXY), leucine aminopeptidase (LAP), *N*-acetyl-β-D-glucosaminidase (NAG), and sulfatase (SUL).

#### β *3.1. Carbohydrate Active Enzymes*

#### β 3.1.1. β-Glucosidase (BGL)

μ μ μ μ *S. dacryoidea* exhibited a significantly higher Vmax for BGL than the other fungal strains (*t*-test; *p* < 0.001) at 5 ◦C (6.4 ± 3.2 µmol/g biomass/h) and 20 ◦C (7.4 ± 3.0 µmol/g biomass/h) (Figure 1A). The other fungal strains exhibited generally low BGL activity,

particularly at 5 ◦C. At 20 ◦C, the BGL activity was slightly higher than at 5 ◦C, except for *R. mucilaginosa,* which maintained a low BGL activity at both temperatures. Consequently, Vmax was a species- and temperature-specific value, with Q<sup>10</sup> values varying between 1.4 and 2.4, except for *R. mucilaginosa* (Figure 5).

The K<sup>m</sup> was significantly higher in *S. dacryoidea* than in the other fungal strains (*t*-test; *p* < 0.001) and ranged between 139.6 ± 19.1 µM and 78.6 ± 11.9 µM (Figure 3A). Moreover, the K<sup>m</sup> of *S. dacryoidea* was significantly higher at 20 ◦C than at 5 ◦C (*t*-test; *p* < 0.001). Nonetheless, the other two species of the phylum Basidiomycota, *R. sphaerocarpa* and *R. mucilaginosa*, showed low K<sup>m</sup> values, especially the latter. Finally, for the Ascomycota species, the K<sup>m</sup> was highest at 5 ◦C (*t*-test; *p* = 0.01) with 12.9 ± 3.4 µM for *B. parvus* and 38.7 ± 14.6 for *M. australis*.

#### 3.1.2. β-Xylosidase (BXY)

All fungal strains tested were capable to cleave xylose, however, at low rates (Figure 1B). *B. parvus* exhibited a significantly higher Vmax for BXY than the other strains (*t*-test; *p* < 0.001). In *B. parvus*, the Vmax was 0.8 ± 0.3 µmol/g biomass/h at 5 ◦C and 1.0 ± 0.5 µmol/g biomass/h at 20 ◦C. In the other strains, the hydrolysis rates were higher at 20 ◦C than at 5 ◦C, amounting to 0.5 and 0.1 µmol/g biomass/h, respectively. When all the fungal species were compared, the temperature had a greater effect in *M. australis* and *S. dacryoidea* with Q<sup>10</sup> values of 2.1 and 1.8, respectively, than in the other fungal strains (Figure 5). For *B. parvus* and both species of the genus *Rhodotorula*, Q<sup>10</sup> values close to 1 suggested that Vmax was independent of the temperature.

The K<sup>m</sup> varied significantly between both temperatures (*t*-test; *p* = 0.90) (Figure 3B). At 5 ◦C, *S. dacryoidea* showed a high K<sup>m</sup> (49.7 ± 27.6 µM) followed by *B. parvus* with a K<sup>m</sup> value of 37.2 ± 8.5 µM. In contrast, at 20 ◦C, the K<sup>m</sup> decreased in both *B. parvus* and *S. dacryoidea* but increased in *M. australis*. The other two species, *R. sphaerocarpa* and *R. mucilaginosa*, showed low K<sup>m</sup> values corresponding to their low Vmax.

#### 3.1.3. *N*-acetyl-β-D-glucosaminidase (NAG)

*B. parvus* exhibited similar NAG hydrolysis rates at 5 ◦C and 20 ◦C (*t*-test; *p* = 0.39) (Figure 1C). At 5 ◦C, the Vmax was 0.8 ± 0.4 µmol/g biomass/h and at 20 ◦C 0.7 ± 0.2 µmol/g biomass/h. Although the other strains exhibited low NAG enzymatic activity, it increased from 5 ◦C to 20 ◦C, particularly in *M. australis* and *S. dacryoidea*. In *M. australis* the Vmax increased from 0.1 ± 0.08 µmol/g biomass/h to 0.5 ± 0.2 µmol/g biomass/h and in *S. dacryoidea* from 0.04 ± 0.02 µmol/g biomass/h to 0.1 ± 0.06 µmol/g biomass/h. The Q<sup>10</sup> values for NAG in *M. australis* and *S. dacryoidea* were 2.1 and 1.1, respectively (Figure 5).

The high NAG enzymatic activity detected in *B. parvus* coincided with a high K<sup>m</sup> (Figure 3C). At 5 ◦C, the K<sup>m</sup> was 2.9 ± 1.8 µM, while at 20 ◦C it was 12.1 ± 5.6 µM. At 20 ◦C, the K<sup>m</sup> of *M. australis* increased to 8.7 ± 4.6 µM. Thus, the K<sup>m</sup> values of these two Ascomycota species increased with temperature, but they remained low in the three Basidiomycota species.

## *3.2. Extracellular Enzymes Targeting Proteins, Phosphorus, and Sulfur Compounds* 3.2.1. Leucine Aminopeptidase (LAP)

Generally, the LAP activity was higher at 20 ◦C than at 5 ◦C in all the fungal strains (*t*-test; *p* < 0.001), except in *M. australis* (*t*-test; *p* = 0.46) (Figure 2A). The LAP activity in *R. sphaerocarpa* at 20 ◦C was 4.3 ± 1.5 µmol/g biomass/h; however, it was only 1.1 ± 0.2 µmol/g biomass/h (*t*-test; *p* < 0.001) at 5 ◦C resulting in a Q<sup>10</sup> of 6.1 (Figure 5). All the other fungal strains had Q<sup>10</sup> values ranging from 1.9 to 1.1. Thus, the LAP activity was the only extracellular enzymatic activity tested that showed a temperature dependency in all the fungal strains (Figure 5).

*B. parvus* and *S. dacryoidea* exhibited significantly higher K<sup>m</sup> values than the other fungal strains (*t*-test; *p* < 0.001) (Figure 4A). The K<sup>m</sup> values varied between 195.4 µM and 41.6 µM in *B. parvus* and between 122.3 µM and 42.2 µM in *S. dacryoidea*. While in *B. parvus,* the K<sup>m</sup> decreased with increasing temperature, and in *S. dacryoidea,* the K<sup>m</sup> value incremented with increasing temperature.

#### 3.2.2. Alkaline Phosphatase (ALP)

At 5 ◦C, *M. australis* exhibited a significantly higher Vmax (8.8 ± 3.7 µmol/g biomass/h) than the other fungal isolates (*t*-test; *p* = 0.03) (Figure 2B). At 20 ◦C, however, the Vmax was about 10-fold lower than at 5 ◦C (0.6 ± 0.2 µmol/g biomass/h). In contrast, in all the other fungal strains, Vmax slightly increased with temperature (Figure 2B). The Q<sup>10</sup> values ranging from 2.6 to 1.0 indicated a moderate temperature dependency of ALP in all the fungal strains examined, except in *M. australis* (Figure 5).

The K<sup>m</sup> was significantly higher in the Ascomycota than in the Basidiomycota strains (*t*-test; *p* = 0.01) (Figure 4B). *M. australis* exhibited a higher K<sup>m</sup> at 5 ◦C (101.5 ± 43.5 µM) than at 20 ◦C, whereas the K<sup>m</sup> in *B. parvus* was higher at 20 ◦C (85.7 ± 10.5 µM) than at 5 ◦C. The other fungal species belonging to the Basidiomycota phylum exhibited K<sup>m</sup> values ranging from 55.1 µM to 4.1 µM (Figure 4B).

#### 3.2.3. Sulfatase (SUL)

Sulfatase (SUL) activity was generally low compared to the other extracellular enzymatic activities (Figure 2C). Higher Vmax values were determined for *B. parvus* and *S. dacryoidea* at 5 ◦C (1.1 and 0.9 µmol/g biomass/h, respectively) (*t*-test; *p* = 0.01) than at 20 ◦C (0.4 and 0.3 µmol/g biomass/h, respectively). In contrast, for *M. australis*, *R. mucilaginosa*, and *R. sphaerocarpa*, the SUL activity increased with temperature. The Q<sup>10</sup> value in *M. australis* was 1.9 while all the other fungal strains were <1.0 (Figure 5). With the exception of *R. sphaerocarpa,* all the fungal strains exhibited higher K<sup>m</sup> values at 5 ◦C than at 20 ◦C (*t*-test; *p* ≤ 0.002) (Figure 4C).

#### *3.3. Relation between Enzyme Kinetic Parameters, Enzyme Types, and Phylogeny*

PCA analysis allowed for the comparison of the five studied marine fungal species and the different extracellular enzymatic activities determined. The explained variance of the dataset was 63.9% (Figure 6). The minor angle between *R. mucilaginosa* and *R. sphaerocarpa* belonging to the same genera (*Rhodotorula*) indicated very similar activity levels and extracellular enzyme characteristics. In contrast, the other fungal strains substantially differed in their extracellular enzymatic activity and enzyme characteristics, independent of their taxonomic affiliation.

#### **4. Discussion**

All the studied marine pelagic fungi species, *B. parvus*, *M. australis*, *R. mucilaginosa*, *R. sphaerocarpa*, and *S. dacryoidea*, produced extracellular enzymes to degrade substrate analogues of carbohydrates such as cellulose, chitin, and xylan. Additionally, they all produced enzymes to cleave off amino acids, phosphate, and sulfate esters from organic compounds.

#### *4.1. Extracellular Enzymatic Activities of Pelagic Fungal Isolates*

#### 4.1.1. Influence of Taxonomy/Diversity on the Different EEAs

Since each organism has specific enzymatic capabilities and substrate preferences [33], extracellular enzymatic activities (EEA) can be used as functional traits to investigate functional diversity [34]. Hence, the different EEAs detected in the studied marine fungi can be used to infer their influence on marine ecological processes. Interestingly, we found that the two species belonging to the genus *Rhodotorula* (*R. mucilaginosa* and *R. sphaerocarpa*) exhibited similar kinetic parameters (Vmax and Km) for the majority of extracellular enzymes (Figures 1–4 and 6). This might indicate some degree of trait conservation among organisms on the genus level, although more species of this genus need to be investigated before a firm conclusion can be drawn. Differences were observed, however, in the EEAs of all the other fungal strains.

Polysaccharides, as the most abundant organic compound class, but also the most complex one, require a wide range of enzymes to degrade them [35,36]. We measured three EEAs responsible for the cleavage of cellulose, xylan, and chitin (Figure 1). In terrestrial environments, the plant cell wall is composed mainly of cellulose and protected by lignin [37]. In marine ecosystems, cellulose is present in the algae cell wall and covered by distinct polymers [38]. Nonetheless, as marine cellulose is more accessible, but less frequent than other substrates, only specific microorganisms are capable of degrading cellulose [39]. Some studies have reported cellulose degradation by marine fungi like *Arthrinium saccharicola* [40] and *Lulworthia floridana* [41]. Other studies have also described cellulose hydrolysis by wider distributed fungal species, for instance, *Aspergillus niger* [42] and *Trichoderma virens* [43]. Vaz, et al. [44] showed that 76% of the studied marine fungi exhibit cellulolytic activity. In this case, even though all the marine fungal species used were able to cleave cellulose, *S. dacryoidea* dominated this EEA (Figure 1A). Hudson [45] stated that each fungi species have different capacities to decompose cellulose due to diverse enzymatic machinery. *S. dacryoidea* exhibited a high K<sup>m</sup> indicating a low affinity to the substrate [31]. This also suggests that, even though the overall enzymatic activity is high, the substrate dissociates easily from the enzyme [46].

Xylan is a polysaccharide formed by residual monosaccharides called xylose [29]. Similar to cellulose, in terrestrial environments, xylan can be found in plants [47], whereas in the ocean, xylan can be present in algae [48,49]. Fungi can degrade xylose via the oxidoreductase pathway [50] as it is a primary carbon source [28]. Even though the general EEA was low (Figure 1B), we can deduce that the studied marine fungi are capable of releasing enzymes related to the hydrolysis of xylose. Raghukumar, et al. [51] identified low xylanase activity rates of fungal coastal strains. Duarte, et al. [52] also reported xylanase activity of Antarctic fungal strains but highlighted a higher activity of Basidiomycota over Ascomycota strains. In this study, we could not identify a clear difference between these two phyla. Nonetheless, the low K<sup>m</sup> suggests a high affinity of the enzyme to the substrate at low concentrations (Figure 3B). Thus, it seems likely that pelagic marine fungi might use xylose as a carbon source even when present at low concentrations.

Chitin is one of the most abundant naturally occurring polysaccharides, and it is an essential component of the cell wall of fungi, the exoskeleton of arthropods, the radula of mollusks, and the beak of cephalopods [53,54]. Although the overall chitinase activity was low, we can infer that the studied marine fungi are capable of degrading chitin (Figure 1C). The chitin degradation by marine fungi has been reported in species such as *Lecanicillium muscarium* [55] and *Verticillium lecanii* [56]. Interestingly, the number of fungal enzymes involved in the degradation of chitin is related to the fungal chitin content, which varies strongly between fungal species and is dependent on the growth mode [57]. For instance, the hyphae-like fungal cell wall consists of 10 to 20% of chitin [54], whereas yeast-like fungi have a rather low chitin content of 0.5 to 5% [58]. The filamentous fungus *B. parvus* exhibited high chitinase activity (Figure 1C). Moreover, the low K<sup>m</sup> obtained for this species suggests that only a low substrate concentration is needed to saturate its chitinase. Thus, pelagic marine fungi likely use chitin as a carbon source even when present at only low concentrations.

Leucine aminopeptidase is a critical biological enzyme due to its key role in the degradation of proteins [59]. In the present study, this enzymatic activity was different for each fungal species. Despite the majority of microbial LAP being intracellular, extracellular enzymes have been reported in filamentous fungi [60]. All the species we tested showed LAP activity with *R. sphaerocarpa* exhibiting the highest Vmax. The substrate concentration needed to achieve half Vmax varied among species (Figure 4A). Although *B. parvus* had one of the lowest Vmax, its K<sup>m</sup> was the highest (Figures 2A and 4A). In contrast, *R. sphaerocarpa* exhibited high substrate affinity and a low K<sup>m</sup> (Figure 4A) Thus, as the kinetics for LAP varied among the fungal species, the protein hydrolysis in the ocean might be species dependent.

Inorganic phosphate (Pi) is the preferred phosphorus source for microbial uptake. In surface waters, however, Pi frequently limits phytoplankton productivity [61]. To overcome this P-limitation, microorganisms use dissolved organic phosphorus (DOP) [62]. For prokaryotic microorganisms, Baltar et al. [63] reported that irrespective of the phosphate bioavailability, the activation of alkaline phosphatase was related to sporadic pulses of organic matter. Thus, a high K<sup>m</sup> might be beneficial to allow for high cleavage rates when organic substrate availability is high. The species *R. sphaerocarpa*, *R. mucilaginosa*, and *S. dacryoidea*, belonging to the phylum Basidiomycota, exhibited a low enzymatic activity but a high K<sup>m</sup> (Figures 2B and 4B). In contrast, the species of the phylum Ascomycota (*B. parvus* and *M. australis*) seem to be more suitable to overcome this P-limitation, but a higher amount of substrate, and hence of organic matter, might be needed for this purpose.

In living organisms, sulfur is the sixth most abundant element as it can be found in amino acids, such as cysteine and methionine, but also in polysaccharides and proteoglycans [35]. In contrast to terrestrial polysaccharides, several marine polysaccharides, especially in the cell wall of macroalgae, are highly sulfated [35]. The terrestrial fungus *Fusarium proliferatum* was reported to produce sulfatase for the assimilation of sulfated fucoidans of brown algae [64]. In the present study, the species of the genus *Rhodotorula* maintained a low sulfatase activity, whereas it varied among the other species. The Vmax and K<sup>m</sup> were high in *S. dacryoidea,* indicating fast hydrolysis and low substrate affinity. As all the fungal species showed different sulfatase activity, we can deduce that these marine fungi might be capable to use sulfated amino acids as well as carbohydrates.

In this study, even though we analyzed just a few hydrolysis possibilities, the functional diversity seems to be broad in marine fungi. As suggested by Berlemont [65] and Baltar et al. [10], marine fungi are potentially involved in carbohydrates' degradation. Based on our results, we can infer that marine pelagic fungi are actively utilizing carbohydrates such as cellulose, chitin, and xylose, as well as some carbohydrates with sulfur content, potentially of algal origin [47].

The ocean is a complex and diverse ecosystem composed of microorganisms such as archaea, bacteria, fungi, protists, and viruses. As osmoheterotrophic organisms, fungi might provide intermediate decomposition products needed for other microorganisms. These also might lead to the proliferation or inhibition of other microbes as well as enzymatic activities. As marine fungi can utilize a wide range of organic substrates [66], nutrient availability can impact the magnitude and distribution of extracellular enzymatic activities [33].

#### 4.1.2. Temperature Influence

Temperature is considered one of the most important abiotic factors because it influences essentially all biochemical reactions [67]. Enzymes are sensitive to temperature [46] influencing the kinetics along with the substrate binding property and stability [33,68]. According to the Van't Hoff rule, a temperature increase of 10 ◦C can double a reaction rate. Q<sup>10</sup> values lower than 1.0 would indicate a reaction rate completely independent of temperature, whereas values above 1 indicate thermodependency [69].

The results obtained in this study indicate that the majority of enzymatic activities were lower at 5 ◦C than at 20 ◦C (Figures 1 and 2). In general terms, at 5 ◦C the species *R. mucilaginosa* and *R. sphaerocarpa* maintained Vmax values as low as 0.1 µmol/g biomass/h, whereas the other species exhibited only half of the activity at 5 ◦C, particularly for BGL and ALP (Figures 1 and 2). This reduced enzymatic synthesis at a low temperature might be due to limited transcriptional and translational activity, limited protein folding, and DNA and RNA secondary structures' stabilization [11]. *B. parvus*, for all the substrates except ALP, had a higher K<sup>m</sup> at 5 ◦C, which suggests that the affinity of this species for some substrates increases with increasing temperature. Taken together, the effect of temperature on the characteristics of extracellular enzymes depends on the fungal species and the type of enzyme.

Psychrophiles microorganisms have evolved a complex range of adaptation strategies, such as production of antifreeze proteins [70] and exopolysaccharides (EPS) [71], high

levels of unsaturated fatty acids to maintain the membrane fluidity [72], and certain enzymes adapted to those temperatures [11]. Gerday et al. [73] described a peculiar type of extracellular enzymes known as "cold-adapted enzymes" produced by microorganisms living at low temperatures. For these enzymes, the reaction rate is dependent on the encounter rate of the enzyme and substrate, so it is controlled mainly by diffusion, and it is temperature independent [14,74].

Aghajari et al. [75] suggested that the main structural feature of these "cold-adapted enzymes" is flexibility or plasticity. The structures involved in the catalytic cycle are more flexible, whereas other structures that do not participate in the catalytic cycle might be more rigid [67,76]. For instance, the chitinase of *Glaciozyma antarctica* presented fewer salt bridges and hydrogen bonds, which increased its flexibility [12,77]. Another key structural feature of these enzymes is stability [73,74], with for example, amino acids modifications in key regions of the protein [77–80]. Nonetheless, there is not a single strategy, as each cold-adapted enzyme can perform different ways to enhance its activity at low temperatures [12,74].

Cold-adapted enzymes have been reported from a wide variety of marine fungi [12,44,52,55,56,77,81,82]. *M. australis* and *R. sphaerocarpa* were one of the few species that showed a noticeable enzymatic activity at 5 ◦C for ALP and SUL, respectively (Figure 2B,C). *M. australis* is an endemic species of Antarctic waters [19,83], whereas *R. sphaerocarpa* was originally isolated close to Marguerite Bay on the west side of the Antarctic Peninsula but has a wider distribution including the Caribbean Sea [84] and the Andaman Sea [85], among others. Low temperatures exert high selective pressure on endemic organisms [86], such as alkalinity phosphatase in *M. australis*. For this species, the substrate-binding affinity was lower at 5 ◦C, whereas for *R. sphaerocarpa*, K<sup>m</sup> was lower at this temperature (Figure 4). This suggested that the enzyme–substrate complex of *M. australis* ALP is more effective at higher substrate concentration typical for Antarctic waters known as a major high-nutrient low-chlorophyll region in the global ocean [87].

Fungal cold-adapted chitinases have been previously reported [77,88,89]. Ramli et al. [77] found that the chitinase sequence of *G. antarctica* had a low sequence identity with other chitinases. Moreover, they found that the enzyme flexibility was due to certain amino acids substitutions in the surface and loop regions. In this study, at 5 ◦C, we could only identify a higher chitinase activity for *B. parvus*. For the rest of the species, there was a positive enzymatic activity, but higher at 20 ◦C.

Microorganisms isolated from cold environments can also display kinetic parameters similar to those of their mesophilic counterparts [69,90]. Ito et al. [91] deduced that a high Q<sup>10</sup> value (>2) is due to a conformational change in proteins and indicates the need for high activation energy. We found that only for LAP, all the examined fungal species expressed Q<sup>10</sup> values higher than 1. Generally, the temperature where an enzyme can achieve its highest activity does not match the optimal growth temperature of the microorganism that is producing it [78]. Apparently, cold-adapted species, such as *B. parvus*, *M. australis*, *R. sphaerocarpa*, and *S. dacryoidea,* can respond to a temperature rise by increasing enzymatic activity. According to the Arrhenius equation, temperature can influence the activation energy needed to initiate a chemical reaction, and hence, its rate. At a higher temperature, the molecules gain energy to move faster, which also increases the collisions between enzymes and substrates. As a result, elevated extracellular enzymatic activity at increasing temperatures in the surface waters might lead to changes in cleavage and uptake rate of organic matter in oceanic fungi.

#### **5. Conclusions**

In the present study, we have shown that different marine fungal strains exhibit varying extracellular enzyme characteristics with Vmax and K<sup>m</sup> values varying over a range of one order of magnitude. Although the fungal species were isolated from coastal Antarctic waters (*B. parvus*, *M. australis*, *R. sphaerocarpa*, and *S. dacryoidea*), and hence are potentially adapted to low temperatures, they exhibited higher extracellular enzymatic activity at 20 ◦C than at 5 ◦C, with some exceptions. While in some fungal strains the K<sup>m</sup> values for specific extracellular enzymes were higher at low temperatures, for other enzymes they were lower. Additionally, there was considerable species-specific variability in the extracellular enzymatic activity. Taken together, our study indicates that temperature might be one of most important physical factors controlling marine fungal extracellular enzymatic activity. Thus, species composition and temperature determine the role of marine fungi in organic matter cleavage in the global ocean.

**Author Contributions:** Conceptualization, K.S.A. and F.B.; data curation, K.S.A.; formal analysis, K.S.A.; funding acquisition, G.J.H. and F.B.; investigation, K.S.A.; methodology, K.S.A. and F.B.; project administration, F.B.; software, K.S.A.; supervision, F.B.; visualization, K.S.A.; writing—original draft, K.S.A.; writing—review and editing, K.S.A., G.J.H., and F.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** F.B. was supported by the Austrian Science Fund (FWF) projects OCEANIDES (P34304-B), ENIGMA (TAI534), and EXEBIO (P35248). G.J.H. was supported by the Austrian Science Fund (FWF) project ARTEMIS (P28781-B21), the projects I486-B09 and P23234-B11 by the European Research Council (ERC) under the European Community's Seventh Framework Programme (FP7/2007–2013)/ERC grant agreement 268595 (MEDEA project). Open Access Funding by the Austrian Science Fund (FWF).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation to any qualified researcher.

**Acknowledgments:** We would like to thank Charlotte Doebke, Marilena Heitger, Marina Montserrat Díez, and Katarína Tamášová for their valuable help in the laboratory work.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Xuefeng Peng 1,2, \* and David L. Valentine 1,3**


**\*** Correspondence: xpeng@seoe.sc.edu

**Abstract:** Fungi in terrestrial environments are known to play a key role in carbon and nitrogen biogeochemistry and exhibit high diversity. In contrast, the diversity and function of fungi in the ocean has remained underexplored and largely neglected. In the eastern tropical North Pacific oxygen minimum zone, we examined the fungal diversity by sequencing the internal transcribed spacer region 2 (ITS2) and mining a metagenome dataset collected from the same region. Additionally, we coupled <sup>15</sup>N-tracer experiments with a selective inhibition method to determine the potential contribution of marine fungi to nitrous oxide (N2O) production. Fungal communities evaluated by ITS2 sequencing were dominated by the phyla *Basidiomycota* and *Ascomycota* at most depths. However, the metagenome dataset showed that about one third of the fungal community belong to early-diverging phyla. Fungal N2O production rates peaked at the oxic–anoxic interface of the water column, and when integrated from the oxycline to the top of the anoxic depths, fungi accounted for 18–22% of total N2O production. Our findings highlight the limitation of ITS-based methods typically used to investigate terrestrial fungal diversity and indicate that fungi may play an active role in marine nitrogen cycling.

**Keywords:** marine fungi; oxygen minimum zone; nitrous oxide; diversity; <sup>15</sup>N tracer; size-fractioned; eastern tropical North Pacific; metagenome

#### **1. Introduction**

Oceanic oxygen minimum zones (OMZs) are characterized by a sharp oxycline and redox gradient in the water column [1]. As a result, OMZs support diverse microbial communities that directly impact the global biogeochemical cycling of nitrogen, carbon, sulfur, and trace metals [2–6]. As in many other types of marine environments, bacteria and archaea have been the focus of microbial ecology research in OMZs, whereas microbial eukaryotes, in particular fungi, have received much less attention [7,8].

An early cultivation-based survey of marine fungi in the Indian Ocean that included the Arabian Sea OMZ found *Rhodotorula rubra* and *Candida atmosphaerica* to be cosmopolitan, and the yeast population densities ranged from 0–513 cells per liter of seawater [9]. In the eastern tropical South Pacific (ETSP) OMZ off the coast of Chile, high summertime fungal biomass in the water column have been reported [10], including diatom parasites from the phylum Chytridiomycota [11]. In the third and the largest open ocean OMZ, the eastern tropical North Pacific (ETNP), protists diversity has been investigated by sequencing the V4 region of 18S small subunit rRNA genes [7], but little is known about the diversity and function of fungi in this environment.

Fungi in the water column are generally thought to contribute to organic matter recycling, particularly in particle-associated environments [12–14]. High hydrolytic activity on proteinaceous substrates in large size fractions (>25 µm and >90 µm) have been reported in the water column of the ETSP and attributed to fungi given low bacterial biomass in those size fractions [15]. However, it remains unclear if fungi in the water column of OMZs play

**Citation:** Peng, X.; Valentine, D.L. Diversity and N2O Production Potential of Fungi in an Oceanic Oxygen Minimum Zone. *J. Fungi* **2021**, *7*, 218. https://doi.org/ 10.3390/jof7030218

Academic Editor: Federico Baltar

Received: 2 February 2021 Accepted: 15 March 2021 Published: 17 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

a role in nitrogen cycling. Discovered in the early 1990s, fungal denitrification is known as a process that reduces nitrate or nitrite with nitrous oxide (N2O) as the end-product [16,17]. This adds to the multiple other pathways and processes (ammonia oxidation, bacterial denitrification, and chemodenitrification) that can produce N2O [18], a potent greenhouse gas and ozone-depleting agent [19]. Many fungal strains have been found to have the ability to produce N2O [20], including an *Aspergillus terreus* strain isolated from the Arabian Sea OMZ [21]. In marine environments, fungal denitrification with N2O as the end-product has been reported from coastal marine sediment in India and Germany [22,23], but its potential contribution in the water column remains unclear.

We investigated the fungal community composition in the eastern tropical North Pacific oxygen minimum zone by sequencing the internal transcribed spacer region 2 (ITS2) and classifying shotgun metagenome reads. To estimate the fungal contribution to N2O production in the water column, we used a selective inhibition method combined with <sup>15</sup>N-labeled tracer incubation experiments. Fungal communities evaluated by ITS2 sequencing were dominated by the phyla *Basidiomycota* and *Ascomycota* at most depths. The metagenome dataset showed that early-diverging fungi accounted for about one third of the fungal community, and the subsurface peaks of fungal abundance coincided with both cyanobacterial abundance and eukaryotic algal abundance. Incubation experiments suggest a possible role of fungi in N2O production in the water column of the ETNP OMZ.

#### **2. Materials and Methods**

#### *2.1. Site Description and Seawater Filtration*

In March 2018, aboard the R/V Sally Ride in the eastern tropical North Pacific oxygen minimum zone, two stations were visited to study fungal diversity and the potential fungal contribution to N2O production (Figure 1a). Dissolved oxygen concentration was determined using the SBE 43 dissolved oxygen sensor attached to the conductivity, temperature, and depth (CTD) rosette. Seawater was collected at multiple depths spanning from the oxycline to the anoxic depths (Figure 1b) using 30 L Niskin bottles.

**Figure 1.** (**a**) Sampling locations in the eastern tropical North Pacific (ETNP) oxygen minimum zone. Color contour shows oxygen concentrations at 100 m depth from World Ocean Atlas 2013 (March average from 1955–2012) [24]. (**b**) Depths sampled for nitrous oxide (N2O) production experiments are marked by filled symbols. Color contour shows oxygen concentrations measured during this cruise using a Seabird SBE 43 dissolved oxygen sensor.

μ μ To collect particulate material at different size fractions, seawater was sequentially filtered through a 47 mm Whatman Grade 541 acid-hardened cellulose filter paper (22 µm nominal particle retention rating, GE Healthcare 1541–047, Marlborough, MA, USA), a 47 mm polycarbonate filter (2.0 µm nominal pore size, Millipore Isopore TTTP-04700, Burlington, MA, USA), and a Sterivex filter (0.22 µm nominal pore size, Millipore SVGP01050, Burlington, MA, USA), using a peristaltic pump filtration at a flow rate < 50 mL/min. For

each sample set, 23 to 55 L of seawater was filtered (Table S1). Each 47 mm filter was stored in a 47 mm petri dish and flash-frozen in liquid nitrogen before storage at −80 ◦C.

#### *2.2. DNA Extraction*

In the laboratory, DNA was extracted using the DNeasy Plant Mini Kit (QIAGEN Cat No. 69104, Germantown, MD, USA) following the DNeasy Plant Handbook [25], except that the cell disruption step was customized for our samples. Filter paper from the Sterivex filters were extracted from the plastic case using a snap-blade knife. All filters were first cut into 2 by 2 mm pieces using sterilized scissors and transferred into 2 mL screw cap tubes containing 1 mL of 0.5 mm zirconia/silica beads (Biospec products #11079105z, Bartlesville, OK, USA), 600 mL of buffer AP1, and 6 µL of RNase A. Bead beating of the samples was performed for 90 s using a Biospec Mini-BeadBeater-16 and was followed by incubation at 65 ◦C for 10 min. After centrifugation at 20,000× *g* for 5 min, the supernatant in each sample tube was transferred to a fresh 2 mL microcentrifuge tube and neutralized with 195 µL of Buffer P3. The remaining DNA extraction steps followed the DNeasy Plant Handbook without modifications. An extraction blank was included for each batch of extraction procedure. DNA yield was quantified using a Qubit fluorometer (ThermoFisher Scientific, Waltham, MA, USA) following the manufacturer's instructions. All extracted DNA samples were stored at −80 ◦C until amplicon library construction.

## *2.3. Sequencing and Analysis of the ITS2 Region*

The second region of the internal transcribed spacer (ITS2) with flanking regions in the 5.8S and 28S ribosomal RNA was targeted for amplicon library preparation following the Illumina 16S Metagenomic Sequencing Library Preparation [26] with the following modifications. The amplicon PCR was performed using primers ITS3tagmix [27] and ITS4tag001 [28] (Table S2) with Phusion® high-fidelity DNA polymerase (New England BioLabs, M0530, Ipswich, MA, USA). The thermal cycle started with 30 s at 98 ◦C, followed by 30 cycles of 15 s at 98 ◦C, 30 s at 55 ◦C, and 45 s at 72 ◦C. The final elongation at 72 ◦C was 10 min long. The quantity and quality of final PCR products were determined using a Qubit (ThermoFisher Scientific, Waltham, MA, USA) and TapeStation 2200 (Agilent, Santa Clara, CA, USA), respectively. Identical quantities of each sample were pooled, and the products were sequenced on an Illumina MiSeq 2 × 250 PE platform in the Biological Nanostructures Lab at UCSB.

Raw sequence reads were first trimmed using ITSxpress to contain only the ITS2 region [29]. Trimmed reads were merged, quality filtered, dereplicated, and denoised following the USEARCH pipeline [30] to generate amplicon sequence variants (ASVs). Taxonomic assignment was performed using a combination of Naïve Bayes classifier implemented in QIIME2 [31] and BLASTn [32] against full UNITE + INSD dataset v02.02.2019 [33] and curated manually. To determine putative fungal denitrifiers in our samples, we performed a closed reference OTU picking [34] of our ASVs and ITS2 sequences from fungi tested for N2O production [20] using UCLUST [35] implemented in QIIME v1.9.1 [36] against the UNITE database [33]. Raw reads generated in this study are available at the National Center for Biotechnology Information (NCBI) under BioProject PRJNA623945.

#### *2.4. Analysis of Fungal Diversity and Function from Metagenomes*

As an independent approach to evaluate the fungal diversity in the eastern tropical North Pacific oxygen minimum zone, we investigated a metagenome dataset sampled at a nearby station in March 2012 when the hydrographic conditions were highly similar (Figure S1) [37]. Raw reads were first filtered using the tool BBDuk Version 38.73 [38] with the options "ktrim=r ordered minlen=51 minlenfraction=0.33 mink=11 tbo tpe rcomp=f k=23 ftm=5". Adapters were trimmed from the BBDuk-filtered reads using the tool Trimmomatic Version 0.39 [39] with the options "ILLUMINACLIP:\$adapters:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:100". Reads that passed both quality filtering and adapter trimming were queried against the NCBI nr database using DIA-

MOND [40] with an e-value threshold of 1 × 10−<sup>5</sup> and the options "–sensitive –min-orf 20". The resultant NCBI taxonomy ID was used to assign taxonomy to each read. In search for the presence of the gene diagnostic for fungal N2O production, the cytochrome P450 nitric oxide reductase (*P450nor*) [41], we searched the metagenome assemblies under the same BioProject (PRJNA350692) for putative *P450nor* using HMMER v3.2.1 [42] against a *P450nor* profile, which includes both eukaryotic and prokaryotic cytochrome P450 genes [43]. All putative *P450nor* hits were checked against the NCBI nr database using blastp [32] to determine taxonomy.

## *2.5. Measurements of Potential N2O Production Rates*

To measure potential rates of N2O production, parallel incubation experiments were performed by adding 0.1 mL of 5 mM 99% pure <sup>15</sup>N-labeled potassium nitrate or 0.1 mL of 0.8 mM 99% pure <sup>15</sup>N-labeled ammonium chloride (Cambridge Isotopes, Cambridge, MA, USA) to 120 mL glass serum bottles containing freshly collected seawater. To minimize the introduction of atmospheric oxygen, each bottle was overflown three times its volume with water directly from Niskin bottles before filling and crimp-sealing it with grey butyl rubber stopper and aluminum caps. Headspace was created in each bottle by replacing 2 mL of seawater with ultra-high pure helium (Airgas HE UHP300, Radnor, PA, USA). End points were taken by adding 1 mL of 50% (*w*/*v*) zinc chloride to separate parallel incubations approximately 0 and 24 h after incubations began at Station 1 and 0, 12, and 24 h after incubations began at Station 2. The concentration of ammonium in seawater was measured onboard according to the fluorometric method of Holmes et al. [44], with a detection limit of 15 nmol L−<sup>1</sup> . Nitrate concentrations were assayed using a Lachat Flow Injection Analyzer at the Analytical lab at the Marine Science Institute, University of California, Santa Barbara following standard analytical methods [45]. The detection limit of nitrate was 0.2 µmol L−<sup>1</sup> .

The potential contribution of fungal N2O production to total N2O production was determined by incubations with <sup>15</sup>N-labeled nitrate (15NO<sup>3</sup> <sup>−</sup>) and chloramphenicol (87.7 mg L−<sup>1</sup> final concentration), which was applied to each incubation one hour before the addition of <sup>15</sup>N tracers to inhibit all prokaryotic activities. All solutions added to the incubation bottles were purged by ultra-high pure helium for one hour at a flow rate of 40 mL min−<sup>1</sup> . The difference between fungal N2O production and total N2O production from incubations with <sup>15</sup>NO<sup>3</sup> <sup>−</sup> is attributed to bacterial denitrification (Figure S2). N2O production rates measured in incubations with <sup>15</sup>N-labeled ammonium (15NH<sup>4</sup> + ) are attributed to archaea and/or bacterial nitrification.

The quantity and isotopic composition of dissolved N2O was determined using a Delta XP isotope ratio mass spectrometer coupled to a purge-and-trap front end. The detection limit was 1.0 nmol N, and the precision for δ <sup>15</sup>N was 2.0‰ (*n* ≥ 3). The rate of N2O production (RN2O) was calculated from the equation [46]:

$$\mathrm{R\_{N\_2O}} = \frac{\mathrm{d^{15}N\_2O/dt}}{\mathrm{f^{15}} \times \mathrm{V}}$$

where d15N2O/dt is the rate of <sup>15</sup>N2O production determined from linear regression of the amount of <sup>15</sup>N2O against time, f<sup>15</sup> is the fraction of <sup>15</sup>N labeled substrate, and V is the volume of the incubation. The amount of <sup>15</sup>N2O at each time point is calculated from the equation:

$$\mathrm{N\_2O} = \mathrm{N\_2O} \times \frac{\left(\frac{\mathrm{\delta^{15}N\_{\mathrm{-}N\_2O}}{1000} + 1\right) \times \mathrm{R\_{ref}}}{1 + \left(\frac{\mathrm{\delta^{15}N\_{\mathrm{-}N\_2O}}{1000} + 1\right) \times \mathrm{R\_{ref}}}$$

where N2O is the amount of nitrous oxide determined from in-house N2O concentration standards (Figure S3), δ <sup>15</sup>N\_N2O is the bulk isotopic composition of sample N2O, and Rref is isotopic composition of reference gas. The linearity effect for the range of N2O measured was negligible compared to the enriched δ <sup>15</sup>N\_N2O measured from our incubation samples (Figure S4).

#### **3. Results**

#### *3.1. Fungal Diversity Assessed by Sequencing the ITS2 Region*

Assessment of the fungal community in the eastern tropical North Pacific using the internal transcribed spacer region 2 (ITS2) revealed that taxa from the phyla *Basidiomycota* and *Ascomycota* dominated at most depths at both stations (Figure 2). The relative abundance of *Basidiomycota* was higher than *Ascomycota* at nearly all depths (Table S3). The most prevalent and abundant taxon is the basidiomycetous yeast family Sporidiobolaceae, primarily consisting of the genera *Rhodotorula*, *Rhodosporidiobolus*, and *Sporobolomyces* (Table S4). Sporidiobolaceae tend to have a higher relative abundance in the 0.2–2 µm size fraction. On the other hand, *Aureobasidium* (Ascomycota) and Exobasidiomycetes (Basidiomycota, primarily *Meira*) were enriched in the larger size fractions (2–22 and >22 µm). In contrast, the basidiomycetous yeast *Malassezia*, when detected, were enriched only in the 2–22 µm size fraction. At Station 2, most of the fungal community cannot be classified even at the phylum level based on ITS2 sequences, indicating the presence of novel fungal lineages in the oxycline of ETNP oxygen minimum zone.

#### *3.2. Fungal Diversity Assessed from Metagenomes*

Fungal community composition assessed by metagenomic reads also showed the dominance of Dikarya fungi, but in contrast to the results from ITS2 sequencing, the relative abundance of Ascomycota was consistently higher than that of Basidiomycota (Figure 3). Surprisingly, over one third of the fungal community belong to early-diverging phyla including Mucoromycota, Zoopagomycota, Chytridiomycota, Blastocladiomycota, Cryptomycota, and Microsporidia. The fungal community composition from 60 to 300 m was uniform, while there was a trend of increasing relative abundance for Ascomycota with depth.

**Figure 3.** The relative abundance of each fungal phylum identified in the metagenomes [37] collected from a location near Station 2 in this study.

#### *3.3. Relative Abundance of Fungi Compared to Other Taxa*

Mining the metagenomes allowed us to estimate the relative abundance of fungi and other taxa as part of the overall microbial community. In each metagenome, 23–57% of the reads had a positive hit against the NCBI nr database with an e-value threshold of 1 × 10−<sup>5</sup> (Figure S5a). Fungi accounted for 0.02–0.22% of all classifiable reads, with a subsurface peak at 70 m (Figure 4a). The relative abundance of fungal reads decreased with depth below 70 m but showed a small increase at 140 m, which was the top of the oxygen deficient layer of the water column. Ciliophora, Oomycetes, and Dinophyceae were similar to fungi in both abundance and distribution. The abundance of reads classified as eukaryotic algae also showed a subsurface peak at 70 m (Figure 4b). The abundance of cyanobacteria decreased with depth overall, but there was an increase at 110 m, coinciding with the deep chlorophyll maximum situated immediately above the anoxic depths.

−

μ μ <sup>−</sup> μ <sup>−</sup> **Figure 4.** (**a**) Dissolved oxygen concentration (µM) and the relative abundance of reads classified as Fungi, Ciliophora, Oomycetes, and Dinophyceae from the metagenomes [37] collected from a location near Station 2 in this study; (**b**) dissolved oxygen concentration (µM), nitrate (NO<sup>3</sup> <sup>−</sup>) concentration (µM), fluorescence (mg m−<sup>3</sup> ) [47], and the relative abundance of reads classified as Cyanobacteria and eukaryotic algae. Eukaryotic algae include Haptophyta, Bacillariophyta, Rhodophyta, Cryptophyta, and Viridiplantae.

#### *3.4. Potential Contribution of Fungal N2O Production*

− − − − − − N2O production rates from incubation experiments with either <sup>15</sup>NO<sup>3</sup> <sup>−</sup> or <sup>15</sup>NH<sup>4</sup> + were detected at multiple oxycline depths of the ETNP oxygen minimum zone (OMZ), and the maximum rate was found at the oxic–anoxic interface (Figure 5). The rates of N2O production from incubations with <sup>15</sup>NO<sup>3</sup> <sup>−</sup> and chloramphenicol were used as an approximation of fungal N2O production, and they ranged from 0% of total N2O production from <sup>15</sup>NO<sup>3</sup> <sup>−</sup> at 275 m at Station 2 to 56% of total N2O production from <sup>15</sup>NO<sup>3</sup> <sup>−</sup> at 90 m at Station 1. N2O production from incubations with <sup>15</sup>NO<sup>3</sup> <sup>−</sup>, both with and without chloramphenicol, was lower at elevated in situ oxygen (O2) concentration. When integrated from the oxycline (60 m at Station 1 and 90 m at Station 2) to the oxic–anoxic interface, fungal N2O production approximated by incubations with <sup>15</sup>NO<sup>3</sup> <sup>−</sup>, and chloramphenicol accounted for 18–22% of total N2O production (Table S5).

− − **Figure 5.** N2O production rates measured from incubation experiments with <sup>15</sup>NO<sup>3</sup> <sup>−</sup> only (orange and black bars combined), with <sup>15</sup>NO<sup>3</sup> <sup>−</sup> and chloramphenicol (orange bars), and with <sup>15</sup>NH<sup>4</sup> + (green bars). Chloramphenicol was intended to inhibit bacterial and archaeal activity.

#### *3.5. Search for Fungal Denitrifiers and Functional Genes*

μ To search for putative fungal denitrifiers in the water column of the eastern tropical North Pacific OMZ, we first analyzed the ITS2 amplicon dataset generated in this study. Of all 237 ASVs, only one (ASV211) was clustered at 97% similarity level with the ITS2 sequences from fungal strains previously shown to produce N2O (*Chaetomium* sp.) [20]; this ASV was present only in the 0.2–2 µm size fraction at 83 m from Station 1 and at a relative abundance of 0.2% (Table S6). However, one ASV (ASV8) was 96.95% similar to the ITS2 region of the N2O-producing *Penicillium melinii* [20], which was isolated from seawater [48]. The relative abundance of ASV8 ranged from 0.1 to 8.5% at hypoxic and

anoxic depths (Table S6), and it belongs to the same family (Aspergillaceae) as the N2Oproducing fungal strain isolated from the Arabian Sea OMZ [21]. While this putative denitrifier could potentially produce N2O in the water column of the OMZ, its potential contribution to the total N2O production is estimated to be less than 0.3% given the low N2O yield by *P. melinii* in the laboratory (Appendix A). Continuing the search for putative fungal denitrifiers, we analyzed a previously published March 2012 metagenome dataset from a nearby station [37] for the occurrence of *P450nor*. All *P450nor* hits identified by a hidden Markov model (HMM) profile search [43] were prokaryotic genes, most of which from the phylum Actinobacteria (Table S7).

#### **4. Discussion**

#### *4.1. Fungal Diversity in Different Size Fractions*

In the eastern Tropical North Pacific (ETNP) oxygen minimum zone, the most prevalent and abundant fungal taxa in our ITS2-based survey was the basidiomycetous yeast from the family Sporidiobolaceae, which are usually characterized by the production of carotenoids that color their cells red, pink, or orange and, hence, have the name "red yeasts". The high relative abundance of red yeasts in the 0.2–2 µm size fraction (Figure 2) is consistent with their typically small cell size [49]. A previous study on yeasts in the Indian Ocean, which includes the Arabian Sea oxygen minimum zone in the northern part, has also found red yeasts to be the predominant taxa [9]. On the other hand, *Aureobasidium* (Ascomycota) and Exobasidiomycetes (Basidiomycota, primarily *Meira*) were enriched in the larger size fractions (2–22 µm and >22 µm), consistent with a larger size [50,51], and indicating a likely association with particles. In contrast, the basidiomycetous yeast *Malassezia*, when detected, were enriched only in the 2–22 µm size fraction, suggesting that they were not associated with particles larger than 22 µm.

There is increasing recognition of marine fungi as a key component of the marine microbiome and biogeochemical cycles [12,13,15,52,53], but our knowledge about their diversity and function is far less compared to other microbial eukaryotes [54]. This is particularly the case in the water column of the open ocean, where metabarcoding surveys of fungal diversity are often unable to classify most of the fungal community [55,56]. In one of the two stations sampled in this study, we also could not resolve the taxonomy for most of the fungal community (Figure 2), highlighting the limitation of an ITS-based approach to study fungal diversity. Because the public databases for ITS sequences are primarily based on studies of terrestrial environments or fungal strains, this suggests that the open ocean water column harbors previously undiscovered lineages of fungi.

#### *4.2. Ecology of Marine Fungi in the Oxygen Minimum Zone*

By classifying metagenome reads, we showed that about a third of the fungal community were from early-diverging lineages [57], including Mucoromycota, Zoopagomycota, Chytridiomycota, Blastocladiomycota, Cryptomycota, and Microsporidia (Figure 3). Such diversity was previously undiscovered due to both the difficulty in cultivation and the consequent lack of representation in ITS databases. It is unclear what ecological roles these early-diverging fungi play, except for Chytridiomycota, which are typically associated with phytoplankton such as diatoms [58]. Nevertheless, the depth profile of the relative abundance of fungi in relation to other microbial taxa provides a hint (Figure 4). The cooccurrence of subsurface peaks (at 70 m) of the relative abundance of fungi and eukaryotic algae suggests that fungi are directly associated with eukaryotic algae, perhaps as parasites. A secondary peak of the relative abundance of fungi (at 140 m) was observed below that of the deep chlorophyll maximum (at 110 m) typically found at the oxic–anoxic interface of oxygen deficient waters [59]. Hence, the fungal communities at the secondary peak (at 140 m) are likely predominantly saprotrophic, feeding on the particles formed at the base of the mixed layer.

Surprisingly, we did not observe a pronounced effect of oxygen concentration on the fungal community composition, evaluated by either ITS2 amplicon sequencing or

metagenomes. In contrast, the protist community in the anoxic depth of the ETNP and eastern tropical South Pacific oxygen minimum zone was enriched in Syndiniales, euglenozoan flagellates, and acantharean radiolarians [7,8]. This may be a result of the versatile respiratory/fermentative metabolisms fungi possess, but it could also be due to the inability of the methods used in this study to detect novel fungal lineages from the oxygen deficient waters. The metagenome-based approach avoids the typical biases associated with amplicon sequencing such as primer bias, but it is limited by the sequences available in the chosen database (NCBI nt in this study).

## *4.3. Fungal N2O Production in the Oxygen Minimum Zone*

In order to distinguish the N2O production by fungi from bacteria and archaea, we combined <sup>15</sup>N tracer incubation experiments with selective inhibition using the antibiotic chloramphenicol (Figure S2). Chloramphenicol is a broad-spectrum antibiotic that has been used to isolate anaerobic gut fungi from the rumen microbiome [60], and its final concentration used in our incubations was scaled by the typical bacterial cell density in the rumen vs. seawater. Nonetheless, we did not make direct measurements of the specificity and effectiveness of chloramphenicol on inhibiting bacterial and archaeal activities in seawater. Therefore, it is possible that the N2O production rates we measured from incubations with <sup>15</sup>NO<sup>3</sup> <sup>−</sup> and chloramphenicol include N2O from partially inhibited bacterial denitrification. Consequently, we conservatively interpret those rates as an upper limit of potential fungal N2O production.

When integrated from the oxycline to the oxic–anoxic interface, fungal N2O production accounted for 18–22% of total N2O production (Table S5). While we interpret this as the maximum possible contribution of fungi to N2O production, we suggest that fungal denitrification could be an important pathway for N2O production in oceanic oxygen minimum zones. Denitrification, the sequential reduction in NO<sup>3</sup> <sup>−</sup> to N2, is known to be inhibited by trace amounts of O<sup>2</sup> [61]. In OMZs, it was demonstrated that 297 nM of O<sup>2</sup> repressed 50% of total N2O production at the oxic–anoxic interface [62]. In this study, N2O produced in incubation with <sup>15</sup>NO<sup>3</sup> <sup>−</sup> (primarily via denitrification) was inhibited by increasing levels of in situ O<sup>2</sup> concentration. However, the inhibitory effect of O<sup>2</sup> appeared to be less pronounced on fungal denitrification than on bacterial denitrification at low levels, revealing a potential niche (0.0 < O<sup>2</sup> < 0.93 µM) for fungi capable of denitrification. Since this potential niche is shallower than the oxygen deficient waters, the potential fungal N2O production can have a higher chance of reaching the ocean–atmosphere interface than N2O produced at deeper depths. Alternatively, N2O produced by fungi can be taken up by bacteria with the atypical ("Clade II") nitrous oxide reductase [63] (primarily Flavobacteria and Chloroflexi), which was shown to have a peak in relative abundance at depths immediately above the oxic–anoxic interface [37].

## *4.4. Molecular Evidence for Fungal N2O Production*

The search for putative fungal denitrifiers using ITS2 sequence identity suggests an extremely low abundance of known N2O-producing fungi in the water column of oxygen minimum zones. However, it should be noted that the collection of N2O-producing fungi used to identify these putative denitrifiers consists of soil fungi exclusively [20], so fungal lineages capable of N2O production from the open ocean were likely excluded. Therefore, there may be other N2O-producing fungi from the ETNP OMZ unidentified by this approach.

The absence of fungal *P450nor* in the metagenomic dataset we queried may be attributed to insufficient sequencing depth, given the low percentage (0.02–0.22%) of fungal reads classified by DIAMOND against the NCBI nr database [40] (Figure 4a). Even under the most simplistic assumption that there was only one fungal species present and it possessed *P450nor* in its genome, the sequencing depth in most samples was insufficient to recover just one copy of fungal *P450nor* (Table S8). Additionally, most DNA extraction protocols applied in published metagenome studies (including [37]) are not customized for

disrupting chitinous cell walls. This likely resulted in under-sampling DNA from fungi, of which the biomass is low in the ocean water column (0.01–0.12 µg of carbon L−<sup>1</sup> ) [10], especially compared to bacteria (estimated average of 10.5 µg of carbon L−<sup>1</sup> ) [64]. Finally, it should be noted that the detection of fungal *P450nor* genes in metagenomes does not necessarily imply fungal denitrification, as *P450nor* in certain fungal genomes appear to be involved in secondary metabolisms instead [43].

#### **5. Conclusions**

Our findings highlight the previously unrecognized fungal diversity in the eastern tropical North Pacific oxygen minimum zone, particularly from the early-diverging taxa as revealed by analysis of shotgun metagenomes. The depth distribution pattern of fungi in relation to cyanobacteria and eukaryotic algae suggest direct association in the mixed layer of the water column and indirect feeding below the deep chlorophyll maximum. Given the limitations of the selective inhibition method using chloramphenicol, we estimate that fungi contribute no more than 18–22% to total N2O production from the oxycline to the oxic–anoxic interface. It remains challenging for omics-based approaches to provide molecular evidence for fungal denitrification, partially because current databases and recent studies have primarily focused on terrestrial fungi. Overcoming the shortfalls of existing methods necessitates approaches that can target fungal diversity and function, such as the use of RNA-seq combined with eukaryotic messenger RNA enrichment or the combination of fluorescence-activated cell sorting and (meta)genome sequencing. These new methods can increase the likelihood of capturing genetic evidence for fungal activities and functional diversity.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2309-6 08X/7/3/218/s1, Table S1. Volume of seawater filtered for particulate matter collection. Table S2. Primers designed for PCR. The parts in bold font are standard Illumina adapters. The five ITS3tag primers were combined in equimolar as the forward primer. Table S3. Fungal community composition at the phylum level assessed by the internal transcribed spacer region 2 (ITS2) sequencing. Table S4. (Separate spreadsheet) Taxonomy and read count of each amplicon sequence variant (ASV) in each sample. Table S5. N2O production rates integrated from the oxycline to the oxic–anoxic interface depths and the percentage of potential fungal contribution. Table S6. Relative abundance of putative denitrifiers as part of the total fungal community assessed by the internal transcribed region 2 (ITS2). Non-zero values are highlighted in bold. Table S7. (Separate spreadsheet) Scientific names and kingdoms of the top blastp result against NCBI nr database for genes identified as putative *P450nor* by a hidden Markov model search. All genes were from previously published metagenome assemblies sampled at a station near Station 2 in this study (see Figure S1). Table S8. Comparison of actual number of reads and the estimated minimum number of reads required to recover one copy of *P450nor* in each sample from the Fuchsman et al. (2017) study. The estimated minimum number of reads required were calculated assuming an average fungal genome size of 40,973,539 bp and the length of the *P450nor* gene as 1056 bp. Figure S1: (a) Sampling location from March 2012 ("Fuchsman17") where the metagenome samples investigated in this study were collected, and sampling locations in this study ("Station 1" and "Station 2"). Color contour shows oxygen concentrations at 100 m depth from World Ocean Atlas 2013 (March average from 1955–2012). (b) Comparison of potential density (σθ), oxygen concentration (O<sup>2</sup> ), and fluorescence measured by Seabird profiling sensors from this study (2018) and from the "Fuchsman17" study (2012). Figure S2. Incubation scheme using a selective inhibition method combined with <sup>15</sup>N tracer technique to determine the fungal contribution to N2O production. In each incubation, the final concentration of <sup>15</sup>NO<sup>3</sup> <sup>−</sup> was 3 µM and the final concentration of <sup>15</sup>NH<sup>4</sup> <sup>+</sup> was 0.5 µM. Chloramphenicol was added to a final concentration of 87.7 mg L−<sup>1</sup> . Figure S3. Calibration curve used to calculate the amount of N2O present in each sample based on the total areas under m/z of 44, 45, and 46 ("Area All"). Figure S4. The amount (nmol) and bulk δ <sup>15</sup>N (‰) of N2O measured from incubation samples in which N2O production rates were non-zero (black empty circles) and of N2O concentration standards (blue filled diamonds). Error bars represent standard deviations (*n* > 5). Figure S5. (a) The fraction of metagenome reads with a DIAMOND hit. (b) The percentage of metagenome reads that were classified as archaea, bacteria, viruses, and algae. The samples were collected during the same month

in 2012 at a station close to the station 2 in this study (Figure S1). Read classification was performed using DIAMOND against the NCBI nr database with an e-value threshold of 1 × 10−<sup>5</sup> .

**Author Contributions:** Conceptualization, X.P. and D.L.V.; methodology, X.P.; formal analysis, X.P.; resources, X.P. and D.L.V.; data curation, X.P.; writing—original draft preparation, X.P.; writing review and editing, X.P. and D.L.V.; visualization, X.P.; funding acquisition, X.P. and D.L.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Simons Foundation Postdoctoral Fellowship in Marine Microbial Ecology (No. 547606), NSF Grants OCE-1635562, OCE-1756947, OCE-1657663, and the C-BRIDGES program. Supercomputing resources were provided by the Center for Scientific Computing at UCSB, which is supported by NSF Grant CNS-0960316.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Raw reads generated in this study are available at NCBI under BioProject PRJNA623945.

**Acknowledgments:** We are indebted to members of Bess Ward's and Karen Casciotti's Labs, the crew of R/V Sally Ride, and Frank Kinnaman for general assistance. We thank Alyson Santoro's lab and the Analytical Lab at the Marine Science Institute in UCSB for advice on mass spectrometry method development.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

## *Extrapolation of N2O Production by a Putative Fungal Denitrifier*

ASV8 was 96.95% similar to the ITS2 region of the N2O-producing *Penicillium melinii* [20], which was isolated from seawater [48]. The relative abundance of ASV8 ranges from 0.1 to 8.5% at hypoxic and anoxic depths (Table S4), and it belongs to the same family (Aspergillaceae) as the N2O-producing fungal strain isolated from the Arabian Sea OMZ [21]. We estimated the maximum potential contribution of ASV8 to N2O production by scaling its mass-dependent N2O production rate (assumed to be identical to *Penicillium melinii* [20]) by its estimated mass-based abundance, assuming the maximum observed (ITS-based) fractional abundance of 8.5% and fungal carbon content from a similar location. The following equation was used:

$$\mathbf{R\_{N\_2O}} = \frac{\mathbf{P\_{pm}} \times \mathbf{m\_f} \times \mathbf{f\_{ASV8}}}{\mathbf{T}}$$

where Ppm is the measured N2O production rate for *Penicillium melinii* [20], m<sup>f</sup> is the estimated fungal carbon mass (0.12 µg L−<sup>1</sup> ) as measured for the eastern tropical South Pacific oxygen minimum zone [10], fASV8 is the largest fractional contribution of ASV8 to our data set (8.5%), and RN2<sup>O</sup> is the estimated maximum rate of N2O production by ASV8. To make RN2<sup>O</sup> an upper bound, we used the largest value of fungal carbon content and an N2O production rate by *P. melinii* (2 mg N2O-N g−<sup>1</sup> week−<sup>1</sup> ) reported previously [20]. By this approach, the RN2<sup>O</sup> extrapolated for this putative fungal denitrifier is 2.1 × 10−<sup>4</sup> nmol L−<sup>1</sup> d −1 . This is accounts for at most 0.3% of any fungal N2O production rate measured in this study.

#### **References**

