Membrane Lipid Composition of the Moderately Thermophilic Ammonia-Oxidizing Archaeon “Candidatus Nitrosotenuis uzonensis” at Different Growth Temperatures

For Thaumarchaeota, the ratio of their glycerol dialkyl glycerol tetraether (GDGT) lipids depends on growth temperature, a premise that forms the basis of the widely applied TEX86 paleotemperature proxy. A thorough understanding of which GDGTs are produced by which Thaumarchaeota and what the effect of temperature is on their GDGT composition is essential for constraining the TEX86 proxy. “Ca. Nitrosotenuis uzonensis” is a moderately thermophilic thaumarchaeote enriched from a thermal spring, setting it apart in its environmental niche from the other marine mesophilic members of its order. Indeed, we found that the GDGT composition of “Ca. Nitrosotenuis uzonensis” cultures was distinct from those of other members of its order and was more similar to those of other thermophilic, terrestrial Thaumarchaeota. This suggests that while phylogeny has a strong influence on GDGT distribution, the environmental niche that a thaumarchaeote inhabits also shapes its GDGT composition.

The reason why the occurrence of cren is limited to the Thaumarchaeota and does not occur in other phyla of Archaea remains unknown. cren is characterized by the presence of an unusual cyclohexane moiety in addition to the presence of four cyclopentane moieties. Molecular modeling has revealed that this cyclohexane moiety disturbs the packing of a GDGT membrane (30), which is an important adjustment to growth temperature. Hence, the acquisition of the trait to produce a cyclohexane moiety in one of the biphytane (BP) chains of the GDGT was interpreted to represent an important step in the evolution of the Thaumarchaeota phylum to conquer the largest biome on Earth, the relatively cold ocean (30). Subsequently, however, cren was also identified in hot springs (51-54) and in thermophilic Thaumarchaeota, i.e., "Ca. figurations are taken into consideration, the BPs detected could lead to 2 possible isomers of GDGT-0, 3 of GDGT-1, 11 of GDGT-2, 9 of GDGT-3, 15 of GDGT-4, and 18 of cren (of which cren= would be one).
Intact polar GDGT distribution and changes with temperature. Five different GDGT polar headgroups were detected in the "Ca. Nitrosotenuis uzonensis" cultures: monohexose (MH), dihexose (DH), two isomers of trihexose (TH1 and TH2), and hexose/phosphohexose (HPH) (see Fig. 2 for structures). Both of the TH isomers underwent the same fragmentation under high-performance liquid chromatographyion trap mass spectrometry (HPLC-ITMS) ( Table 2), which indicates that both sets of isomers have the same distribution of sugars around the GDGT core (as opposed to the constitutional isomers described in reference 57). Therefore, it can be assumed that they are different sugar stereoisomers. The DH-GDGT was assigned a structure with one sugar moiety on each end of the GDGT, based on comparison of the LC-ITMS 2 fragmentation (Table 2) with those reported previously (57). The LC-ITMS 2 fragmentation (Table 2) allowed us to further assign the two TH isomers a structure with two sugar moieties on one side and one on the other.
It should be noted that the intact polar lipids (IPLs) were examined in terms of their MS peak area response, and thus, the relative abundances reported here may not be     abundances of 36% Ϯ 15% and 33% Ϯ 3%, respectively), followed by DH (25% Ϯ 4% and 29% Ϯ 3%), TH1 (22% Ϯ 3% and 23% Ϯ 3%), and TH2 (12% Ϯ 8% and 10% Ϯ 5%), with the least abundant headgroup being MH (4% Ϯ 1% and 5% Ϯ 1%). With the increase in growth temperature from 46°C to 50°C, the headgroup distribution changed noticeably (Table 3); MH increased (from 5% Ϯ 1% to 10% Ϯ 5%), as did DH (from 29% Ϯ 3% to 36% Ϯ 8%). Indeed, DH had become the most abundant GDGT headgroup at 50°C. Considering the variability between replicate cultures, the fractional abundances of TH1 and TH2 did not exhibit an overall change with increasing temperature. The HPH GDGT headgroup decreased with the increase in growth temperature from 46°C to 50°C, from 33% Ϯ 3% to 22% Ϯ 9%. A marked variability in the distribution of core lipids of each intact polar lipid was observed (Table 3 and Fig. 3). The LC-ITMS method utilized did not resolve the isomers Applied and Environmental Microbiology of the GDGTs, nor did it resolve cren from cren=; hence, the IPL-bound GDGT cores are defined as GDGT-0 to GDGT-4 and cren. For example, cren and GDGT-0 were predominantly contained in the HPH IPL, while GDGT-4 and, to a lesser extent, GDGT-3 were predominantly found in the DH, TH1, and TH2 IPLs. Consistent with what was seen for the hydrolysis-derived core lipids, the cultures grown at 46°C and 50°C exhibited similar IPL-bound core lipid distributions, while those grown at 37°C were more distinct (Fig.  3). Overall, in the cultures grown at 46°C and 50°C, cren became more dominant for the MH and HPH IPLs and increased in DH and TH2. For the TH1 IPL, there was no change in the core lipid distribution with increasing temperature.

DISCUSSION
"Ca. Nitrosotenuis uzonensis" is a moderate thermophile enriched from a thermal spring (55), which sets it apart from other cultured members of the Nitrosopumilales (NP) (group 1.1a), generally considered to be a predominantly marine/aquatic, mesophilic order (2). The genus Nitrosotenuis is the only genus within the family Nitrosotenuaceae within the NP, and members of this genus can be found widely distributed in soils, freshwater, hot springs, the subsurface, and activated sludge (58).
It has been reported that cren and cren= exist in a wide range of hot spring environments (51)(52)(53)(54)(59)(60)(61)(62) and in cultures of thermophilic Thaumarchaeota, i.e., "Ca. Nitrososphaera gargensis" (36) and N. yellowstonensis (33), contradicting the previous hypothesis that the production of cren was linked to the radiation of Thaumarchaeota in mesophilic environments (30). The GDGT distribution of the moderate thermophile "Ca. Nitrosotenuis uzonensis" further reinforces the idea that cren and cren= are general biomarkers for Thaumarchaeota rather than representing an adaptation of members of this phylum to mesophilic temperatures.
Specific changes in "Ca. Nitrosotenuis uzonensis" membrane lipid composition as a response to growth temperature. The relative abundance of the core lipids of "Ca. Nitrosotenuis uzonensis" varied with increasing growth temperature, with more cren and cren= and less GDGT-0 to GDGT-4 (including isomers). It is well established that Thaumarchaeota increase their cren and cren= proportions at higher temperatures, which explains the fundamental role of cren= in the sea surface temperature (SST) proxy TEX 86 (tetraether index of tetraethers consisting of 86 carbons) (63), particularly at relatively high temperatures of Ͼ20°C. The TEX 86 SST proxy, and its low-temperature (Ͻ15°C) and high-temperature (Ͼ15°C) versions TEX L 86 and TEX H 86 , respectively, have been applied to temperature reconstructions in a wide range of marine and lacustrine settings (see references 29 and 64 for reviews). We calculated the TEX 86 values for the replicate stationary "Ca. Nitrosotenuis uzonensis" cultures (  16°C Ϯ 3.4°C lower than the actual growth temperatures (Table 4). Previous studies have reported poor correlations between TEX 86 values and temperature, and, hence, inaccurate temperature estimates, in thermophilic Thaumarchaeota cultures and in samples from thermal environments (33,51,53,61). In this context, it is important to note that neither TEX 86 nor TEX H 86 was designed to be used in terrestrial thermal environments such as hot springs. In this context, it is also interesting to keep in mind that TEX 86 was found to correlate with the concentration of bicarbonate, not temperature, in a range of Nevada hot springs (51). Our results now further demonstrate that TEX 86 does not reflect well lipid membrane adaptation for thermophilic Thaumarchaeota. As culturing conditions other than temperature were kept constant in this study, we cannot examine the relationship that variables such as growth phase, bicarbonate concentration, ammonium oxidation rates, and pH would have on the GDGT distribution in the "Ca. Nitrosotenuis uzonensis" cultures.
It is also well established that temperature is a primary factor controlling the number of GDGT cyclopentane moieties with increasing temperature leading to an increasing number of cyclopentane moieties (63,(66)(67)(68)(69). However, this effect was only minor for "Ca. Nitrosotenuis uzonensis," as the average number of cyclopentane moieties increased from 3.1 Ϯ 0.3 at 37°C to 3.7 Ϯ 0.3 at 46°C and 50°C (Table 4).
As the temperature increased, specific GDGTs changed in their abundance relative to the abundance of their isomers (Table 5). With the increase in temperature from 37°C to 50°C, GDGT-2 and GDGT-2= went from a distribution of 60:40 to 52:48, GDGT-3 and GDGT-3= went from 71:29 to 54:46, and GDGT-4, GDGT-4=, and GDGT-4Љ went from 83:16:1 to 66:30:5 (Table 5). Interestingly, while the overall percentage of cren= increased with temperature, it remained constant relative to cren: the ratios of cren to cren= were 84:16 at 37°C and 85:15 at 50°C (Table 5). It has recently been revealed that cren= has a stereochemically different cyclopentane ring than that of cren, a difference in stereochemistry that has been postulated to have an effect on membrane fluidity, therefore playing a role in maintaining membrane homeostasis (31). In the temperature range examined for "Ca. Nitrosotenuis uzonensis," cren= was not upregulated relative to cren as a membrane adaption to increasing temperature. However, GDGT-2= was upregulated relative to GDGT-2, GDGT-3= was upregulated relative to GDGT-3, and GDGT-4= and GDGT-4Љ were upregulated relative to GDGT-4. As explained in Results, we are not able to determine the stereochemistry of the different GDGT isomers detected in this study; however, the biphytanes (BPs) released from the GDGTs of "Ca. Nitrosotenuis uzonensis" grown at 46°C (31) included three different BPs with two cyclopentane rings and two BPs with two cyclopentane rings and one cyclohexane ring. Combinations of these BPs can give rise to a wide range of isomers. Here we hypothesize that the change in the composition of GDGT-2, GDGT-3, and GDGT-4 with increasing temperature represents, as per cren, changes in their cyclopentane ring stereochemistry, in order to maintain membrane homeostasis.
Whereas the core lipid compositions of the "Ca. Nitrosotenuis uzonensis" cultures grown at 46°C and 50°C were most similar to each other, the cultures grown at 37°C and 46°C were the most similar in terms of polar headgroup composition (Table 3 and Fig.  3). A similar observation was made previously by others (33), who noted that the core lipid and polar headgroup distributions in Thaumarchaeota are affected by different factors. With increasing growth temperature, the main change in the "Ca. Nitrosotenuis uzonensis" polar headgroup composition was that two of the smaller headgroups (MH and DH) increased in relative abundance, while one of the largest headgroups (TH2) decreased. However, when we calculated the average number of sugars per GDGT, we found no significant difference between the different growth temperatures (Table 3). To examine polar headgroup adaptions further, we calculated for each growth temperature the ratio of glycolipids to phospholipids, which was higher at 50°C (4.2 Ϯ 2) than at 37°C and 46°C (2.2 Ϯ 2 and 2.0 Ϯ 0.3, respectively). Studies that describe the effect of temperature on archaeal polar headgroup composition are limited (for a review, see reference 70). The temperature-driven polar headgroup adaption reported for three strains of the Thaumarchaeota species Nitrosopumilus maritimus (71) (also from the NP order but a marine mesophile) was different from that seen in this study: all three N. maritimus strains generally decreased the relative percentage of MH lipids as the temperature increased, while the percentages of DH and HPH generally increased, and hence, in contrast to our findings, the ratio of glycolipids to phosphoglycolipids decreased as the temperature increased. However, it should be noted that the growth temperature range (18°C to 35°C) in the N. maritimus study was much lower. The results for "Ca. Nitrosotenuis uzonensis" are also different from those described previously for the Euryarchaeota species Thermoplasma acidophilum (thermophilic and acidophilic) (72), which was found to adapt to higher temperatures (and to lower pHs) by increasing the number of sugars in the polar headgroups. However, similar to our findings for "Ca. Nitrosotenuis uzonensis," the ratio of glycolipids to phosphoglycolipids in T. acidophilum increased at higher temperatures. The decrease in phosphoglycolipids relative to glycolipids may relate to adaptions such as decreased proton permeability of the membrane (72)(73)(74) or could relate to stress adaption during which P-containing lipids are replaced with non-P-containing lipids in order to utilize the P for other essential cell processes. Replacement of phospholipids with nonphospholipids as a response to nutrient limitation or other stresses has been previously observed in bacteria (75), algae (76)(77)(78), and archaea (50,74).

What determines thaumarchaeotal lipid composition?
We compared the core lipid composition of "Ca. Nitrosotenuis uzonensis" with the core lipid compositions of other thaumarchaeotal species reported in the literature (data used are listed in Table  S1 in the supplemental material). To produce this simplified data set, the isomers of the GDGTs were grouped together, with the exception of cren and cren=, which were treated separately, while hydroxy-GDGTs were not included. First, we used principalcomponent analysis (PCA) (Fig. 4) to examine this data set. The first two principal components accounted for 32 and 31%, respectively, of the variability in the core lipid composition. GDGT-0 to GDGT-3 were negatively loaded on the first principal component, while GDGT-4 and cren= were positively loaded. cren was negatively loaded on the second principal component. The majority of the NP order members were grouped across the two negative quadrants of the first principal component, while the majority of the Nitrososphaerales (NS) order members were in the two positive quadrants of the first principal component, in the direction of cren=. The single Nitrosotaleales (NT) and Nitrosocaldales (NC) order members included in the data set were placed between the NP and NS orders (Fig. 4). There were three exceptions to this otherwise clear NP/NS separation. First, our three "Ca. Nitrosotenuis uzonensis" (NP order) cultures grown at different temperatures clustered between the NP and NS members. Second, a "Ca. Nitrososphaera gargensis" culture (moderately thermophilic; NS order), which was grown at 35°C (33), was placed within the NP order cluster. Third, "Ca. Nitrosocosmicus oleophilus" MY3 (NS order) (22) was placed within the NP order cluster. It should be noted that all members of the NS order that group closely together in the PCA are phylogenetically very closely related (Fig. S1). Without lipid analysis of further members of the genus Nitrosocosmicus, it is not possible to say whether "Ca. Nitrosocosmicus oleophilus" is an outlier or whether all Nitrosocosmicus members would group with the NP. Overall, the NP/NS cluster separation was driven by the fractional abundance of cren= (Fig. 4). Many of the NS members examined (e.g., "Ca. Nitrososphaera gargensis," Nitrososphaera viennensis, and "Ca. Nitrososphaera sp." strain JG1) contained a high percentage (14 to 29%) of cren= (33,36,40,54,63), while the majority of the NP members contained a lower fractional abundance (0 to 3%) of cren= (16,32,33,(37)(38)(39). This explains why the "Ca. Nitrosotenuis uzonensis" cultures did not cluster with the other NP order members, as all cells grown at the three different temperatures contained a relatively high fractional abundance of cren= (i.e., 5, 11, and 12%). It is commonly observed that the proportion of cren= is higher in thermophilic Thaumarchaeota than in mesophilic Thaumarchaeota (35,36,40), and our results suggest that this phenomenon is independent of order affiliation. The results of the PCA illustrate that while phylogeny seems to have a strong influence on GDGT distribution, environmental parameters like growth temperature can lead to inconsistencies between phylogenetic affiliation and GDGT composition, as exemplified by the moderate thermophile "Ca. Nitrosotenuis uzonensis" via its elevated amounts of cren= compared to other NP members. In this context, it should be kept in mind that factors not examined in this study have also been shown to have an effect on GDGT distribution (cf. lines linking points in Fig. 4), including O 2 concentration, pH, and salinity (52,71,79).
To further examine the relationship between cren= and temperature across the thaumarchaeotal orders, we calculated the cren=-to-cren ratio for all known thaumarchaeotal core lipid compositions reported in the literature (Table S1) and found a  Table S1 in the supplemental material) for "Ca. Nitrosotenuis uzonensis" and other species of the Thaumarchaeota for which lipid composition data have been reported in the literature. For this analysis, GDGT-0 to GDGT-4 were summed with their isomers, while crenarchaeol and cren= were included separately, and due to their trace abundance or absence, hydroxy-GDGTs, archaeols, GDDs, and GTGTs were not included. Data are from both this study and reports in the literature (both directly reported and estimated from figures) (n ϭ 56). The Nitrosopumilales (NP) order (group I.1a) members are in blue, Nitrososphaerales (NS) order (group I.1b) members are in green, the single Nitrosotaleales order (SAGMCG) member is in purple, and the single Nitrosocaldales order (HWCG) member is in orange. Experimental series are represented with filled circles and connected by a line (temperature [degrees Celsius] and/or O 2 concentration [percent] is shown in parentheses). Filled areas represent clusters based on GDGT composition, as discussed in the text. Unfilled circle with *, Nitrosopumilus maritimus SCM1; filled circle with **, Nitrosopumilus maritimus SCM1 (0.1, 1, 5, 10, and 21%; 30°C).
Lipid Composition of "Ca. Nitrosotenuis uzonensis" Applied and Environmental Microbiology significant correlation with growth temperature (Spearman r ϭ 0.72; n ϭ 56; P ϭ Ͻ0.001). However, there appears to be a "tipping point" in the cren=-to-cren ratio at 35°C (Fig. 5). In the temperature range of 4°C to 35°C, the cren=-to-cren ratio was on average 0.02 Ϯ 0.03, while in the range of 36°C and above, the ratio was 0.3 Ϯ 0.2. However, it should be noted that taxon sampling is still relatively skewed in this analysis, with many members of the genus Nitrosopumilus being well represented and many members of the genera Nitrosocosmicus, Nitrosotalea, and Nitrosocaldus still awaiting lipid composition analysis. As discussed above, cren= has a stereochemically different cyclopentane ring than that of cren (31), which could lead to the two isomers having different effects on the fluidity of a cell membrane. It is possible that this apparent tipping point represents a Thaumarchaeota-wide temperature above which the different stereochemistry of the cren= cyclopentane ring provides a beneficial effect to the membrane. The fact that all "Ca. Nitrosotenuis uzonensis" cultures were grown at temperatures above this 35°C tipping point would then explain why the ratio of cren= to cren did not change between these growth temperatures. A direct comparison of the intact polar lipid (IPL) composition of "Ca. Nitrosotenuis uzonensis" with those of other thaumarchaeotal species reported in the literature is less straightforward than for core lipids due to variability in analytical methods used between studies and the nonquantitative nature in which IPL data have often been reported. In Table 6, we summarize, in a qualitative manner, thaumarchaeotal IPL distributions reported in the literature from studies that utilized a normal-phase liquid chromatography-mass spectrometry (LCMS) method comparable to the one used in this study. The "Ca. Nitrosotenuis uzonensis" cultures grown at 37°C and 46°C contained HPH as the dominant polar headgroup, as has been previously reported for a range of NP species (Table 6) and also for a moderately thermophilic terrestrial NS member, "Ca. Nitrososphaera gargensis." Conversely, the "Ca. Nitrosotenuis uzonensis" cultures grown at 50°C were dominated by DH, a characteristic IPL reported in high abundance for the NS members "Ca. Nitrososphaera sp." JG1 and N. viennensis (40). Neither form of TH detected in "Ca. Nitrosotenuis uzonensis" was reported in other members of the NP (16,37,38), but they have been reported in the NS member N. viennensis (40). Previously, the lipid compositions of a range of cultured representatives of the four thaumarchaeotal orders were examined (33), and it was suggested that the core lipid composition reflects phylogenetic orders, while the polar headgroup composition reflects habitat (either terrestrial thermophiles, marine mesophiles, or soil mesophiles). Knowledge of the lipid composition of "Ca. Nitrosotenuis uzonensis," which, unlike the other mesophilic members of the NP order, is moderately thermophilic (55), further confirms that environmental niche or habitat is a driver of headgroup composition. The "Ca. Nitrosotenuis uzonensis" cultures are more similar in headgroup composition to the three terrestrial NS members, all of which were cultured at temperatures above  Table S1 in the supplemental material) in Thaumarchaeota grown at temperatures between 4°C and 35°C and between 36°C and 72°C. Data used to calculate the ratios were taken from this study and from the literature (Table S1).
35°C, than the five other NP members, all of which were cultured at temperatures below 35°C ( Table 6).
The "Ca. Nitrosotenuis uzonensis" cultures were not found to contain IPLs with a core hydroxy-GDGT (MH-OH or DH-OH), whereas these have been detected in all other members of the NP order examined to date (16,33,37,38). It should be noted that hydroxy-GDGTs were not included in the PCA (Table S1). Their absence in "Ca. Nitrosotenuis uzonensis" may mean that hydroxy-GDGTs are found only within specific clades of the NP order. However, hydroxy-GDGTs have also been associated with growth temperature: a decrease in temperature has been observed to lead to an increase in hydroxy-GDGTs in both thaumarchaeotal cultures (80) and environmental samples (80)(81)(82)(83). Hence, it is possible that their absence in "Ca. Nitrosotenuis uzonensis" relates to the high cultivation temperatures, reflecting their thermophilic nature.

MATERIALS AND METHODS
Culturing. Highly enriched "Ca. Nitrosotenuis uzonensis" cultures that contained no other archaea (55)  Fresh medium batches were allowed to equilibrate for 2 weeks to ensure reaching low levels of hydrogen peroxide forming during the preparation process. Four biological replicates were grown at 37°C, 46°C, and 50°C, in 250-ml Schott flasks, in the dark. All replicates went through two 10% transfers at the respective temperatures to dilute out lipids from the inoculum culture. The cultures were fed with 1 mM NH 4 Cl (final concentration), and depletion was monitored with Nessler's reagent. The final batches were refed multiple times, consumed 4 mM ammonium in total, and depleted ammonium before the biomass was harvested. This final biomass production step took about 8 weeks, ensuring that at the end, the vast majority of the cells produced were in stationary phase.
Extraction. Freeze-dried biomass lipids were extracted using a modified Bligh-Dyer procedure (84). Briefly, the biomass was treated ultrasonically three times for 10 min with a solvent mixture of methanol (MeOH), dichloromethane (DCM), and phosphate buffer (2:1:0.8, vol/vol/vol). After sonication, the combined supernatants were phase separated by adding additional DCM and buffer to a final solvent ratio of 1:1:0.9 (vol/vol/vol). The organic phase containing the intact polar lipids (IPLs) was collected, and the aqueous phase was reextracted three times with DCM. Finally, the combined extract was dried under a stream of N 2 gas. h Hydroxy moiety on the GDGT core and not the polar headgroup (40).
In order to remove the headgroups from the IPLs and to obtain the remaining core lipids, the Bligh-Dyer extract was hydrolyzed with 5% (vol/vol) HCl-MeOH by refluxing (3 h). The hydrolysate was neutralized with KOH to pH 7/8, extracted with DCM, and dried over Na 2 SO 4 .
Core lipid analysis. The hydrolyzed Bligh-Dyer extracts were analyzed using high-performance liquid chromatography/atmospheric-pressure chemical ionization mass spectrometry (HPLC/APCI-MS) on an Agilent 1100/Hewlett Packard 1100 MSD instrument equipped with automatic injector and HP-Chemstation software according to methods described previously (85), with the following modifications. Separation was achieved in normal phase with two Prevail Cyano columns in series (150 mm by 2.1 mm; 3 m) with a starting eluent of hexane-propanol (99.5:0.5, vol/vol) and a flow rate of 0.2 ml min Ϫ1 . This remained isocratic for 5 min, and thereafter, there was a linear gradient to 1.8% propanol at 45 min. The injection volume was 10 l.
The ratios and calculations that were carried out on the core lipid data are as follows: Intact polar lipid analysis. The Bligh-Dyer extracts were directly analyzed for IPLs. Extracts were redissolved in a mixture of hexane-2-propanol-water (72:27:1, vol/vol/vol) at a concentration of 10 mg ml Ϫ1 . IPL extracts were analyzed by HPLC-ion trap mass spectrometry (ITMS) according to methods described previously (86), with modifications as described previously (87). The analysis was performed on an Agilent 1200 series LC instrument (Agilent, San Jose, CA), equipped with a thermostated autoinjector and column oven, coupled to an LTQ XL linear ion trap with an Ion Max source and an electrospray ionization (ESI) probe (Thermo Scientific, Waltham, MA). Separation was achieved on a LiChrospher diol column (250 by 2.1 mm, 5-m particles; Alltech) maintained at 30°C. The following elution program was used with a flow rate of 0.2 ml min Ϫ1 : 100% eluent A for 1 min, followed by a linear gradient to 66% eluent A-34% eluent B in 17 min, maintained for 12 min, followed by a linear gradient to 35% eluent A-65% eluent B in 15 min (where eluent A is hexane-2-propanol-formic acid-14. . The lipid extract was analyzed by an MS routine where a positive-ion scan (m/z 1,000 to 2,000) was followed by a data-dependent MS 2 experiment where the base peak of the mass spectrum was fragmented (normalized collision energy [NCE] of 25, isolation width of 5.0, and activation Q of 0.175). IPLs were examined in terms of their MS peak area response. Thus, the relative abundance of the peak area does not necessarily reflect the actual relative abundance of the different IPLs; however, this method allows for comparison between the strains analyzed in this study. The peak areas were determined from extracted ion chromatograms of the [M ϩ NH 4 ] ϩ ion for each individual IPL species.
Phylogenetic analyses. An amino acid alignment of 34 universal marker genes was extracted from previously reported ammonia-oxidizing archaeon (AOA) genomes using CheckM (88). A maximum likelihood tree was constructed using IQTREE multicore version 1.6.2 (89) with 1,000 ultrafast bootstraps (90) under the best-fit model LG_F_R4, determined using ModelFinder (91) (where LG ϭ general amino acid exchange matrix [92], F ϭ empirical amino acid frequencies from the data, and R4 ϭ rate heterogeneity calculated under the FreeRate model [93,94] with four categories).

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/AEM .01332-19.