Enzyme specific isotope effects of the Nap and Nar nitrate reductases

Dissimilatory nitrate reduction (DNR) to nitrite is the first step in denitrification, the main process through which bioavailable nitrogen is removed from ecosystems. DNR fractionates the stable isotopes of nitrogen (14N, 15N) and oxygen (16O, 18O) and thus imparts an isotopic signature on residual pools of nitrate in many environments. Data on the relationship between the resulting isotopic pattern in oxygen versus nitrogen isotopes (18ε / 15ε) suggests systematic differences exist between marine and terrestrial ecosystems that are not fully understood. DNR can be catalyzed by both cytosolic (Nar) and periplasmic (Nap) nitrate reductases, and previous work has revealed differences in their 18ε / 15ε isotopic signatures. In this study, we thus examine the 18ε / 15ε of six different nitrate-reducing microorganisms that encode Nar, Nap or both enzymes, as well gene deletion mutants of the enzymes’ catalytic subunits (NarG and NapA) to test the hypothesis that enzymatic differences alone could explain the environmental observations. We find that the distribution of the 18ε / 15ε fractionation ratios of all examined nitrate reductases form two distinct, non-overlapping peaks centered around a 18ε / 15ε proportionality of 0.55 and a 18ε / 15ε proportionality of 0.91, respectively. All Nap reductases studied to date cluster around the lower proportionality (0.55) and none exceed a 18ε / 15ε proportionality of 0.68. Almost all Nar reductases, on the contrary, cluster tightly around the higher proportionality (0.91) with no values below a 18ε / 15ε proportionality of 0.84 with the notable exception of the Nar reductases from the genus Bacillus which fall around 0.62 and thus closely resemble the isotopic fingerprints of the Nap reductases. Our findings confirm the existence of two remarkably distinct isotopic end-members in the dissimilatory nitrate reductases that could indeed explain differences in coupled N and O isotope fractionation between marine and terrestrial systems, and almost but not fully match reductase phylogeny.


32
Nitrogen is an essential nutrient for life and consequently the availability of nitrogen is a vital 33 control on ecosystem productivity. Anthropogenic activity has severely altered the natural 34 balance of the nitrogen cycle. In particular, the use of the Haber-Bosch reaction to synthesize 35 fertilizers has resulted in excess amounts of nitrate and ammonium being introduced into 36 ecosystems 1,2 . Assessing the outcomes of excess nitrogen inputs into ecosystems requires a 37 mechanistic understanding of the competing processes that affect nitrogen cycling in the 38 environment. does not impart strong isotopic fractionation 9-13 , redox cycling of fixed nitrogen, especially the 45 isotopic fractionation associated with dissimilatory nitrate reduction to nitrite, controls the 46 isotopic composition of bioavailable nitrate in many environmental systems. Dissimilatory 47 nitrate reduction is the first step for two processes in the nitrogen cycle, denitrification to N2 and 48 dissimilatory nitrate reduction to ammonium (DNRA, also referred to as nitrate ammonification) 49 are transferred to NapB then NapA. Alternatively, a bacterium may express NapABCGH (solid 64 lines). Here NapH oxidizes menaquinol (MQH2) to menaquinone (MQ + ) and the electrons have 65 an additional transfer step from NapG to NapC, translocating two additional protons. The Nar 66 reductase uses NarI to oxidize UQH2 to UQ + and transfers electrons to NarH then NarG. NarK1 67 is a symporter that transports nitrate into the cytoplasm with a proton. NarK2 is an antiporter that 68 couples the import of nitrate to the export of nitrite.

70
The proportionality of N and O isotope fractionation ( 18 ε / 15 ε) associated with nitrate reduction 71 in marine ecosystems generally follows a proportionality of 0.9 to 1.0 [14][15][16][17][18][19] . In terrestrial 72 ecosystems, observational data with coupled N and O isotope measurements is more limited 73 (summarized in Fig. 2) but the existing data suggests that the 18 ε / 15 ε proportionality covers a 74 broader and generally lower range of values between 0.5 to 0.7 20-26 . To date, these systematic 75 differences in 18 ε / 15 ε proportionality are not fully understood and may indicate that we are 76 missing a key feature about how nitrogen cycling processes create the isotopic signatures of 77 nitrate observed in nature. Biogeochemical modelling and recent culturing work suggest that the 78 terrestrial observations of low 18 ε / 15 ε values could be the result of oxidative overprinting of the 79 isotopic signal of nitrate reduction by a combination of nitrate producing processes such as 80 anaerobic ammonium oxidation (annamox), nitrification, and enzymatic reversibility during 81 nitrate reduction 8,27 . However, an alternative hypothesis first proposed by Granger Table S2 for details) were aligned using ClustalOmega Multiple Sequence Alignment 52 . A 210 list of gene accession numbers is available in SI 4 Q 0 f 1 6 y 9 S R R J S 3 I t S w 3 g Z J S C R P Z b h c f V z P 6 z e 9 P 1 S K I G C y c o p k R G D I X 6 n Z 3 S 6 O B m J I L d C a l W N 2 8 N / L Y 1 4 o p H + x + q w u q b 7 9 L W Z Y z 4 z 6 9 c b Q T O Y F P 0 I w h l o k F n d 9 O u / o 0 T z I g O F X D J r u 2 G Q Y 6 9 k B g W X U N W i w k L O + B M b Q N d B x T K w v X I S a k V 3 H Z P Q V B v 3 F N I J + 3 q j Z J m 1 o y x 2 k x n D o X 2 v j c n P t G 6 B 6 X H P n Z c X C I p P j d J C U t R 0 / E M 0 E Q Y 4 y p E D j B s X A q d 8 y F x k 6 P 6 x 5 k I I 3 5 / 8 E d w d N M N W 8 + S 2 1 W i f z + J Y I t / I D t k j I T k i b X J F b k i H c G / D O / L a 3 g 9 / 2 z / 1 L / z L 6 a j v z X a 2 y J v y r / 8 C u g j K Y g = = < / l a t e x i t > Eq. 4 Eq. 6 Eq. 5

Growth of cultures 216
Growth rates are recorded in SI Table 1 Fig. 4). Together, our data suggest 18 ε / 15 ε differences can be 259 purely enzymatic, challenging the hypothesis that environmental 18 ε / 15 ε patterns require nitrite 260 re-oxidation from enzymatic reversibility, nitrification or anammox 27 . These observations for 18 ε 261 / 15 ε from nitrate reduction by the Nap reductase in PA14 ᐃnar are similar to all other available 262 observations from organisms that naturally have only this reductase, with 18 ε / 15 ε couplings of 263 0.63 and 0.51 observed in R. sphaeroides and S. gotlandica, respectively 7,31,32 (Fig. 4). filled points indicate cultures grown in LB.

272
As discussed above, the PA14 ᐃnap strain had an 18 ε / 15 ε proportionality of ~0.9 which is 273 consistent with previous reports from organisms that harbor only Nar (Fig. 4)  In contrast to all other data on Nar reductases, B. vireti and B. bataviensis have a significantly 295 lower 18 ε / 15 ε: 0.64 +/-0.04 and 0.61 +/-0.06, respectively (Table 1; Fig. 4). Although 18 ε / 15 ε 296 values between biological replicates covered a wider range than in other organisms, likely due to 297 analytical artifacts from nitrite build-up (see SI discussion on nitrite accumulation), Bacillus 18 ε / 298 15 ε values were robustly and consistently lower than all other Nar reductases (Fig. 4). Overall, 299 the Bacillus data indicate that it is possible for some Nar reductases to have distinct and lower 18 ε 300

Roles of Nap and Nar reductases 304
The Nar reductase is known as the primary respiratory reductase amongst denitrifying bacteria. 305 The nar operon is highly conserved, with narGHI present in every known Nar-bearing denitrifier  (Table 1; Fig. 3). This is a 323 midpoint value in comparison to the 18 ε / 15 ε proportionality measured in the PA14 ᐃnap and 324 PA14 ᐃnar strains and suggests that PA14 was using both nitrate reductases. The Nap reductase 325 for P. aeruginosa is used as a backup redox balancing mechanism, in particular under conditions 326 where electron acceptors are limiting 70 . When grown in LB, this strain exhibited a higher 18 ε / 327 15 ε proportionality of 0.97 +/-0.02 (Table 1). While LB broth is considered a rich medium, it is 328 actually carbon limited, with mostly amino acids available for uptake 71 . This would cause lower 329 C/N ratios in contrast to the MOPS minimal medium, in which we provide excess succinate as a 330 carbon source. Past research in E. coli has shown that the Nar reductase has a selective advantage 331 under low carbon and high nitrate concentrations, which is the case in our LB grown cultures 62 . 332 Furthermore, this effect does not occur in the PA14 ᐃnap strain, suggesting that this is not a 333 difference in how the Nar reductase performs in LB versus minimal medium, but a change in 334 expression pattern by PA14 to maximize energy conservation. 335 336

Mechanism for isotopic differences 337
Regardless of differences in gene regulation, the Nap and Nar reductases still catalyze the same 338 reaction and yet have different isotope effects. The active site of both reductases are similar, with 339 both containing a Mo-bis-MGD cofactor and iron sulfur cluster. 28,72 . One distinction is that the 340 Nar reductase's Mo center is coordinated by an aspartate residue, while the Nap reductase is 341 coordinated by a cysteine. Cysteine is a more reduced residue that may impact the redox 342 potential of the Mo center, affecting how nitrate is bound and reduced [73][74][75] . Studies indicate that 343 Nap generally has a higher affinity for nitrate than Nar 62,76-78 . Furthermore, the base of its 344 substrate channel is lined with positively charged amino acid residues that guides nitrate to the 345 active site 74,79 . In contrast, Nar has a substrate channel with negatively charged residues that 346 may impact the rate of nitrate binding overall 80 . Thus, it is possible that the root of isotopic 347 differences lies within the nitrate molecule's interaction with the active site of these enzymes. 348 349 Additionally, it has been proposed that nitrate binds to the catalytic site of Nap and Nar 350 differently. For the Nar reductase, the general mechanism for nitrate binding allows nitrate to Nar reductase may be subject to an intramolecular isotope effect. While the precise mechanism 356 of nitrate binding and reduction for both Nap and Nar are still uncertain, the Nap reductase's 357 high affinity for nitrate and its faster reduction mechanism may be key in understanding the 358 differences in 18 ε / 15 ε proportionality. Contrary to expectations, our results for the Bacillus 359 experiments indicate that a Nap-like isotopic signature with respect to 18 ε / 15 ε proportionality is 360 possible in a Nar-reductase. Future work on the structural differences between the Bacillus and 361 other Nar reductases may hold the key to uncovering the mechanistic basis for these isotopic 362 differences. 363

ε coupling in ecosystems 365
Our research shows that nitrate reduction by Nap reductases consistently produces 18 ε / 15 ε 366 proportionality values that are lower than those observed in marine ecosystems and may explain 367 the 18 ε / 15 ε signals observed in terrestrial ecosystems. The isotopic data sets collected for the 368 terrestrial data in Fig. 2 come from a diverse set of ecosystems ranging from soils to lakes to 369 riparian zones and groundwater runoff from agriculture (see SI for details). Soils in particular 370 can have a large range of redox gradients contained within a few centimeters and experience 371 drastic changes in moisture on short time scales, impacting oxygen availability 83 . 372 In comparison, marine systems operate at larger scales and experience less heterogeneity over 373 short spatial and temporal scales with dissimilatory nitrate reduction occurring predominately in 374 oxygen minimum zones (OMZ) and anoxic sediments [14][15][16][17]19 . The nar operon has a much 375 narrower regulatory range of permissible environmental conditions than the nap operon and, 376 unlike the latter, is always inhibited by the presence of O2 84-86 , which may explain the 377 predominance of nar-based nitrate reduction in stable low oxygen systems like OMZs 87,88 . It is 378 thus conceivable that the Nap reductase's multiple functions are more suitable for maintaining 379 bacterial homeostasis in terrestrial aquatic ecosystems that can fluctuate significantly over short 380 spatial and temporal timescales. 381 382 Though this hypothesis may appear at odds with the established assumption that the Nap 383 reductase is used less commonly than the Nar reductase, limited data is available on Nap versus 384  While similar studies specifically targeting Nap and Nar gene abundances have not been carried 389 out in marine ecosystems, at minimum this data indicates that the Nap reductase serves an 390 important role in nitrate reduction for bacteria and that its expression is comparable to Nar in 391 freshwater and terrestrial ecosystems. 392 Since the Nap reductase is not embedded in the cytosolic membrane, and thus not directly 394 involved in proton motive force (PMF) generation, it is frequently presumed to be rarely used for 395 respiration. This explains the common assumption that the isotopic signal of nitrate reduction in 396 ecosystems must stem mainly from the membrane bound cytosolic Nar reductase, as PMF 397 generation is essential for survival and growth 7,14,17,32,33 . However, the potential to perform 398 nitrate reduction with only a Nap reductase appears to be common place, and with the right 399 auxiliary genes present in the nap operon, can be just as efficient as the Nar reductase at 400 producing a proton motive force (PMF) (Fig. 1B)  Bacillus-like Nar enzymes in nature that may have lower 18 ε / 15 ε values. The regulation patterns 405 observed in the PA14 wild type strain in MOPS versus LB medium also emphasize the 406 importance of performing transcriptomics over metagenomics, as bacteria with both reductases 407 may switch between Nap and Nar depending on environmental constraints. This is particularly 408 important when considering processes such as DNRA which can use either NapA or NarG to 409 reduce nitrate. Though the Nap reductase is often implicated as the main reductase used during 410 DNRA, many species of bacteria appear to catalyze DNRA solely via the Nar reductase 43,93-95 . 411 The data presented in this study provides a clear indication that even closely related enzymes can 412 have very distinct isotopic signatures that may allow more comprehensive interpretations of 413 environmental data in the future. Colorado Boulder and a NASA Exobiology grant (80NSSC17K0667) to SHK. We would like to 418 thank Daniel Sigman for support and access to analytical instrumentation at Princeton 419 University. We are grateful to Emma Kast, Dario Marconi, Sergey Oleynik and other members 420 of the Sigman Lab for providing guidance and support throughout sample processing and 421 isotopic analysis. We also sincerely thank the Dietrich Lab for providing us with the P.