the paring of^15 N-nitrite with residual un-
labeled nitrite transferred with the inoculum
when inoculating a new culture. Dinitrogen
production byN. maritimus, or other AOA
isolates, has not previously been reported.
Furthermore, our results show that both ni-
trogen atoms in the N 2 originated from nitrite,
with none coming from ammonium. This
result was confirmed by incubations with
(^15) N-ammonium, where no immediate conver-
sion of^15 N-ammonium to^29 N 2 or^30 N 2 was de-
tected(Fig.3,AandB).Instead,^15 N-ammonium
was most likely converted to nitrite and diluted
into the large existing nitrite pool in this ex-
periment. In contrast, when^15 N-ammonium
was added to washed cultures with a small
nitrite pool (5 and 25mM), the^15 N-nitrite
produced from ammonia oxidation was fur-
ther converted to N 2 (Fig. 2B and fig. S12),
which is consistent with our experiments with
(^15) N-nitrite.
Thus far we have shown that NO is a likely
intermediate in oxygen production and that
N 2 and O 2 are produced byN. maritimus.
Rates of O 2 accumulation and N 2 produc-
tion from the different incubations shown
in Figs. 2 and 3 are summarized in Table 1.
These results are generally consistent with
NO dismutation as a source of both N 2 and
O 2 , where NO is produced from the reduc-
tion of nitrite. In incubations with added
(^15) N-nitrite, however, oxygen accumulation
exceeded N 2 production in the first 20 hours
(incubations 1 and 2 in Table 1; Fig. 3C and
figs. S13 and S14), demonstrating a decou-
pling of O 2 and N 2 productioninthisphase
of the experiment. As net rates of O 2 accu-
mulation may underestimate gross rates of
O 2 production, as explored above, and as
N 2 production in our experiments is a gross
production rate, there is a definite imbal-
ance between the production rates of O 2
and N 2. Such an imbalance would be in-
consistent with O 2 and N 2 production di-
rectly from NO dismutation. If this was
the case, the O 2 accumulation rates should
not exceed N 2 production rates, but the im-
balance we measure suggests that further
intermediate(s) must exist between NO and
N 2 and O 2 production. We suggest that N 2 O
may be such an intermediate, where 2NO→
N 2 O + 0.5O 2.
Indeed, in our incubations supplied with
(^15) N-nitrite, (^46) N
2 O accumulated before
(^30) N
2
production accelerated (Fig. 3D), and the
rates of N 2 production and N 2 O accumu-
lation taken together in the first 20 hours
match the O 2 accumulation rates within the
uncertainties (Table 1). Furthermore, the dis-
mutation of NO [aqueous (aq)] to O 2 (aq)
and N 2 O (aq) is thermodynamically favor-
able (standard free energy change,DG^0 ′=
−165 kJ/mol O 2 ,), and there are no abiotic
mechanisms known for this dismutation re-
action. The accumulation of N 2 O distinguishes
the metabolism ofN. maritimusfrom that in
the NC10 bacteriumMethylomirabilis oxyfera,
whose O 2 production does not involve N 2 O,
and for which traces of N 2 O have been at-
tributed to other community members in
the enrichment ( 20 ). However, to explain
the subsequent formation of N 2 from N 2 O,
an unknown N 2 O reductase would need to
be present in theN. maritimusgenome ( 21 ).
Although no nitrous oxide reductases out-
side the NosZ family have been conclusively
identified, their existence has been proposed
multiple times on the basis of the occurrence
of complete denitrification in organisms that
lack a NosZ enzyme ( 22 , 23 ).
N-nitrosating hybrid N 2 formation, in which
oneNatomfromNO 2 −and one from NH 4 +
(or an intermediate of ammonia oxidation)
combine to form N 2 O, has been proposed
as a possible source for N 2 O production in
AOA ( 24 ), but the isotopic signature of the
N 2 O produced in our experiments (^46 N 2 O)
does not support this pathway (the produc-
tion of N 2 O from^15 N-ammonia and^14 N-nitrite
would yield^45 N 2 O). Thus, the N 2 O forma-
tion pathway in AOA when environmental
oxygen is present seems to differ from the
one in our experiments ( 15 , 25 ). However,
thed^18 OvaluesofN 2 Oproducedinincu-
bation experiments with AOA enrichment
cultures support an enzymatic N 2 O produc-
tion pathway from nitrite in AOA ( 26 ). Fur-
thermore, abiotic hybrid formation of N 2 O
from NO and hydroxylamine is insignificant
at the NO concentrations measured in our
incubations (see supplementary text in the
supplementary materials).
In incubations supplied with^15 N-ammonium
and a small nitrite pool (Fig. 2), N 2 O accumu-
lated transiently as well (fig. S15). In these
incubations, N 2 production far exceeded O 2
accumulation, and no N 2 O accumulation would
be required for a mass balance. This is not
surprising, as high rates of ammonia oxida-
tion (Table 1 and Fig. 2) indicate that O 2
accumulation rates in these experiments far
underestimate gross rates of O 2 production,
as explored above. This does not mean that
N 2 O was not an intermediate in these exper-
iments, only that these incubations did not
demonstrate an initial imbalance between
O 2 and N 2 production.
As for NO dismutation in NC10 bacteria, O 2
production inN. maritimusmost likely has
NO as an intermediate and produces N 2. How-
ever, unlike for NC10 bacteria, our results sug-
gest that the pathway of O 2 production used
byN. maritimushas an extra intermediate
that may be N 2 O and that it leads to the tran-
sient accumulation of oxygen. A proposal for
the metabolic pathway associated with oxy-
gen production inN. maritimusis shown
in Fig. 3E. Although ammonia oxidation to
nitrite is accomplished by the O 2 produced
byN. maritimus, the conversion of nitrite
to N 2 requires reducing equivalents regard-
less of the O 2 production pathway. The re-
quired electrons can partly, but not fully, be
obtained from the ongoing oxidation of am-
monia. Another source of electrons could be
intra- or extracellular organic matter produced
during normal aerobic ammonia oxidation
( 27 ) or, in situ, by dissolved organics available
in the water column. Furthermore, to sus-
tain the coupling of oxygen production and
SCIENCEscience.org 7 JANUARY 2022•VOL 375 ISSUE 6576 99
Fig. 2. Ammonia oxidation to nitrite and N 2
during oxygen production byN. maritimus.
(A)^15 N- nitrite production from^15 N-ammonium
(50mM) for two sets of incubations of washed
N. maritimusculture. Incubation I contained
a^14 N-nitrite pool of 5mM, and incubation II had a
(^14) N-nitrite pool of 25mM. (^15) N-nitrite production
continued after supplied oxygen was consumed
(10 hours). (B) Total N 2 production in incubations
I and II. Results include^28 N 2 from^14 N-nitrite as
well as^30 N 2 and^29 N 2 from added^15 N-ammonium
that was converted to^15 N-nitrite and partly
captured in the small^14 N-nitrite pool before
further conversion to^30 N 2 and^29 N 2 (results in
fig.S12).(C) Oxygen accumulation in a subset of
exetainers. Exetainers A and B belong to incuba-
tion I, and exetainers C, D, and E belong to
incubation II. Error bars represent the standard
deviation of three replicates.
RESEARCH | REPORTS