# Source paper archive

- Source title: Stable isotope labeling reveals preferential ammonium acquisition by hydroponic lettuce under ambient and elevated CO2.
- Author(s): Fountain, L. L., Malbrough, S., Blancaflor, E. B., Wheeler, R. M., Massa, G. D.
- Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.16.767057v1
- PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.16.767057v1.full.pdf
- PDF SHA-256: 9aa7283849365db83b762f458600a39f250de129ded3167fd6dc16b6ef3595bc
- Converter: pymupdf4llm 0.0.27
- Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version.

---

&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


1
# Stable isotope labeling reveals preferential ammonium

2
# acquisition by hydroponic lettuce under ambient and 3 elevated CO 2 .

### 4 Luke L. Fountain &#91;1,2&#93;, Sebastien Malbrough &#91;3,4&#93;, Elison B. Blancaflor &#91;5,6&#93;, Raymond M. 5 Wheeler &#91;7&#93;, Gioia D. Massa &#91;6 &#93;


6 1 NASA Postdoctoral Program, Kennedy Space Center, FL, US

7 2 Department of Animal, Rural and Environmental Sciences, Nottingham Trent University, Nottingham,

8 UK


9 3 NASA OSTEM / Astrion Intern, Kennedy Space Center, FL, US


10 4 Department of Biological, Earth and Environmental Sciences, University of Southern Mississippi, MS,
11 USA


12 5 NASA, Biological and Physical Sciences Division, NASA Headquarters, DC, USA


13 6 NASA, Exploration Research and Technology Programs, Kennedy Space Center, FL, USA


14 7 NASA Retired

# 15 Abstract



16

17 Nitrogen (N) form acquired by plants reflects both physiological preference and environmentally
18 mediated availability. Hydroponic controlled-environment agriculture (CEA) removes many

19 constraints governing N availability yet remains predominantly reliant on nitrate (NO 3- )



19 constraints governing N availability yet remains predominantly reliant on nitrate (NO 3- )

20 nutrition. We used differential &#91;15&#93; N stable-isotope labeling to quantify ammonium (NH 4+ ) and



20 nutrition. We used differential N stable-isotope labeling to quantify ammonium (NH 4+ ) and

21 NO 3- acquisition by four lettuce ( _Lactuca sativa_ L.) cultivars grown hydroponically under
22 terrestrial ambient (410 ppm) and spaceflight-relevant elevated (3000 ppm) CO 2 . Plants received

+                       - 15

23 equivalent concentrations of NH 4 and NO 3, with either N form selectively enriched with N



+                       - 15

23 equivalent concentrations of NH 4 and NO 3, with either N form selectively enriched with N
24 during the final 24 h of growth. All cultivars exhibited significantly greater shoot &#91;15&#93; N enrichment

25 following &#91;15&#93; NH 4+ than 15 NO 3- labeling under both CO 2 environments, demonstrating consistent



+ and NO 3



+ than 15 NO 3



25 following NH 4 than NO 3- labeling under both CO 2 environments, demonstrating consistent

26 preferential NH 4+ acquisition. Preference magnitude varied among cultivars and broadly



26 preferential NH 4+ acquisition. Preference magnitude varied among cultivars and broadly
27 corresponded with biomass and total shoot N accumulation, driven predominantly by increased

28 NH 4+ -derived enrichment rather than reduced NO 3- -derived enrichment. Elevated CO 2 did not

29 significantly alter N-form preference, although strong NH 4+ preference persisted at 3000 ppm.



+ -derived enrichment rather than reduced NO 3



29 significantly alter N-form preference, although strong NH 4+ preference persisted at 3000 ppm.
30 These findings distinguish physiological acquisition preference from environmentally mediated
31 N availability and provide a basis for reconsidering nitrate-dominant nutrition in terrestrial and
32 space CEA, including integration with recycled reduced N.


33 **Summary statement:** Differential &#91;15&#93; N labeling revealed consistent preferential ammonium
34 acquisition across four hydroponically-grown lettuce cultivars under both ambient and
35 spaceflight-relevant elevated CO 2, challenging the assumption that nitrate-dominant nutrition
36 reflects lettuce N-form preference


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


37 **Keywords:** ammonium nutrition; bioregenerative life-support systems; controlled-environment
38 agriculture; elevated CO 2 ; hydroponics; nitrate nutrition; nitrogen acquisition; nitrogen-form
39 preference; stable isotope labeling


40

# 41 Introduction



42 Nitrogen (N) is an essential macronutrient required for plant growth and is acquired

43 predominantly as the inorganic ions ammonium (NH 4+ ) and nitrate (NO 3- ). Although plant



+
) and nitrate (NO 3



43 predominantly as the inorganic ions ammonium (NH 4 ) and nitrate (NO 3- ). Although plant
44 species are frequently characterized as preferentially acquiring one N form over another,
45 apparent N-form preference is strongly influenced by the environment in which N is supplied. In

46 soils, the relative availability of NH 4+ and NO 3- is governed by interacting biological and



+ and NO 3



46 soils, the relative availability of NH 4 and NO 3- is governed by interacting biological and
47 physicochemical processes, including mineralization, nitrification, denitrification, microbial
48 competition, ion mobility and retention, which are themselves influenced by environmental
49 conditions such as soil moisture, oxygen availability, temperature and pH (Britto &amp; Kronzucker,
50 2013; Chalk &amp; Smith, 2021). Consequently, the N form predominantly acquired by a plant under
51 field conditions may reflect the N form made most accessible by its environment as much as an
52 intrinsic physiological preference for that N source.



53 Evidence from natural ecosystems demonstrates the importance of this environmental control.
54 Across a tropical forest precipitation gradient, Houlton et al. (2007) observed a community-wide

55 shift from predominantly NO 3- acquisition in drier soils to NH 4+ acquisition in wetter soils,



55 shift from predominantly NO 3- acquisition in drier soils to NH 4+ acquisition in wetter soils,
56 closely corresponding with changes in the relative abundance of the two inorganic N pools.
57 Similarly, Wang and Macko (2011) demonstrated that plant N-form preference varied across an
58 aridity gradient and proposed that these differences were constrained by climatic conditions and

59 associated differences in the relative availability of NH 4+ and NO 3- . These observations illustrate




acquisition in drier soils to NH 4



+ and NO 3



59 associated differences in the relative availability of NH 4 and NO 3- . These observations illustrate
60 the difficulty of separating physiological N-form preference from environmentally mediated N

61 availability in soil-based systems. Indeed, NO 3- is highly mobile in soil and is commonly the



61 availability in soil-based systems. Indeed, NO 3- is highly mobile in soil and is commonly the

62 principal inorganic N source in well-aerated agricultural soils, whereas NH 4+ can dominate under



62 principal inorganic N source in well-aerated agricultural soils, whereas NH 4+ can dominate under
63 contrasting environmental conditions (Wang &amp; Macko, 2011). Thus, the widespread association

64 between crop production and NO 3- acquisition does not necessarily establish that NO 3- is



64 between crop production and NO 3- acquisition does not necessarily establish that NO 3- is

65 intrinsically the preferred N form when NH 4+ and NO 3- are equally available.



acquisition does not necessarily establish that NO 3



+ and NO 3



65 intrinsically the preferred N form when NH 4 and NO 3- are equally available.



66 Controlled-environment agriculture (CEA), and hydroponic cultivation in particular, provides an
67 opportunity to disentangle physiological N acquisition from many of these environmental

68 constraints. In recirculating hydroponic systems, NH 4+ and NO 3- can be supplied directly to roots



+ and NO 3



68 constraints. In recirculating hydroponic systems, NH 4 and NO 3- can be supplied directly to roots
69 at defined concentrations without the soil processes that ordinarily determine their relative


                                                                                              70 availability. Nevertheless, modern hydroponic nutrient formulations remain predominantly NO 3 


                                                                                              70 availability. Nevertheless, modern hydroponic nutrient formulations remain predominantly NO 3 
71 based, with NH 4+ generally supplied as a comparatively small proportion of total N (Sonneveld



71 based, with NH 4+ generally supplied as a comparatively small proportion of total N (Sonneveld
72 &amp; Voogt, 2009; Voogt &amp; Bar-Yosef, 2019). This partly reflects the operational and physiological

73 challenges associated with higher NH 4+ supply, which can result in rapid rhizosphere



73 challenges associated with higher NH 4+ supply, which can result in rapid rhizosphere


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.



74 acidification, nutrient imbalances, and growth inhibition (Britto &amp; Kronzucker, 2013; Natarajan

75 Sakthivel et al., 2023). The predominance of NO 3- in hydroponic practice may therefore reflect



75 Sakthivel et al., 2023). The predominance of NO 3- in hydroponic practice may therefore reflect

76 the challenges associated with managing NH 4+ as well as assumptions inherited from soil-based



76 the challenges associated with managing NH 4+ as well as assumptions inherited from soil-based
77 crop production, rather than physiological acquisition preference alone. Determining which N
78 form plants preferentially acquire when both are supplied at equivalent concentrations and are
79 continuously available is consequently important for establishing whether nitrate-dominant
80 nutrition optimally aligns with crop N-acquisition physiology in engineered production systems.



81 There is a physiological basis to hypothesize that plants may preferentially acquire NH 4+ when

82 external availability of the two N forms is equivalent. Following uptake, NO 3- must first be



There is a physiological basis to hypothesize that plants may preferentially acquire NH 4



82 external availability of the two N forms is equivalent. Following uptake, NO 3- must first be

83 reduced to NO 2- and subsequently NH 4+ before N can be incorporated into organic compounds



83 reduced to NO 2- and subsequently NH 4+ before N can be incorporated into organic compounds

84 through the glutamine synthetase/glutamate synthase (GS/GOGAT) pathway. NH 4+, by contrast,




and subsequently NH 4



84 through the glutamine synthetase/glutamate synthase (GS/GOGAT) pathway. NH 4+, by contrast,
85 can enter this assimilation pathway directly, avoiding the additional energetic and reductant

86 requirements associated with NO 3- reduction (Bloom et al., 1992). Although NH 4+ nutrition



86 requirements associated with NO 3- reduction (Bloom et al., 1992). Although NH 4+ nutrition

87 imposes other physiological costs, and excessive NH 4+ concentrations can become detrimental




reduction (Bloom et al., 1992). Although NH 4



87 imposes other physiological costs, and excessive NH 4+ concentrations can become detrimental
88 (Britto &amp; Kronzucker, 2013), these limitations do not necessarily preclude preferential

89 acquisition of NH 4+ when both forms are present at non-toxic concentrations. The physiological



89 acquisition of NH 4+ when both forms are present at non-toxic concentrations. The physiological
90 preference expressed under controlled N availability may therefore differ from the apparent
91 preference inferred from N acquisition in soil.



92 Atmospheric CO 2 concentration may further modify this relationship. Elevated CO 2 has been

93 reported to inhibit shoot NO 3- assimilation in C 3 plants, and plant growth responses to CO 2
94 enrichment can consequently depend upon the N form supplied (Bloom et al., 2002; Bloom et
95 al., 2012). Environments with highly elevated CO 2 may therefore alter the relative physiological
96 advantages associated with acquisition of oxidized and reduced N. This interaction is particularly
97 relevant to space crop production, where plants may be cultivated under atmospheric CO 2
98 concentrations substantially greater than terrestrial ambient conditions. Concentrations of several
99 thousand ppm have historically occurred aboard the International Space Station (ISS), although
100 CO 2 concentrations vary spatially and temporally within the vehicle (Beard, 2020; Georgescu et
101 al., 2020). Elevated atmospheric CO � concentrations are likely to remain relevant to future
102 space crop production systems, owing to the challenges of managing crew-generated CO �
103 within closed habitats and the potential use of elevated CO � to enhance crop productivity
104 (Maiwald et al., 2020; Ewert et al., 2022). Determining whether elevated CO 2 alters
105 physiological N-form preference is therefore necessary to establish whether N-acquisition
106 strategies characterized under terrestrial conditions remain applicable to crops grown in
107 spaceflight-relevant environments.


108 Nitrogen form is also important at the crop-production system level. Space-based CEA
109 (SpaCEA) must provide sufficient nutrients to sustain crop productivity while minimizing
110 resource requirements. The integration of crop production with waste and nutrient recovery has
111 consequently been identified as an important step toward bioregenerative life-support systems


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.



112 (BLSS), with a recently proposed BLSS Readiness Level framework identifying nutrient-loop
113 closure and subsequent recycled fertigation as key stages in developing crops as operational
114 components of BLSS (Fountain et al., 2026). Human waste streams represent a potentially
115 important source of recoverable N. Urinary N is predominantly excreted as urea, which can be

116 hydrolyzed to NH 4+, whereas production of NO 3- from this recovered reduced N requires



+
, whereas production of NO 3



116 hydrolyzed to NH 4, whereas production of NO 3- from this recovered reduced N requires
117 subsequent oxidation through nitrification (Verbeelen et al., 2021). If candidate crops can utilize

118 a substantial proportion of their N directly as NH 4+, greater integration of crop production with



118 a substantial proportion of their N directly as NH 4+, greater integration of crop production with
119 waste-N recovery could potentially reduce the extent of nitrification required and associated
120 energy and infrastructure requirements before crop delivery and decrease reliance on N fertilizer

121 transported from Earth. Conversely, if plants intrinsically favor NO 3-, even when NH 4+ is



121 transported from Earth. Conversely, if plants intrinsically favor NO 3-, even when NH 4+ is

122 similarly available, conversion of recovered N to NO 3- may represent a necessary component of




, even when NH 4



122 similarly available, conversion of recovered N to NO 3- may represent a necessary component of
123 regenerative nutrient management. Establishing physiological N-form preference is therefore
124 important for rationally designing N-recycling strategies rather than assuming that nutrient forms
125 conventionally used in terrestrial agriculture are necessarily optimal for SpaCEA.



126 The same question is increasingly relevant to terrestrial CEA. Hydroponic production has
127 already decoupled crop nutrition from many of the soil processes that are primary determinants

128 of inorganic N availability, yet NO 3- -dominant nutrient strategies have largely persisted. At the



128 of inorganic N availability, yet NO 3- -dominant nutrient strategies have largely persisted. At the
129 same time, waste-derived nutrient streams such as anaerobic digestate, represent potential

130 sources of NH 4+ -rich N that could be integrated with hydroponic crop production (van Rooyen &amp;



130 sources of NH 4+ -rich N that could be integrated with hydroponic crop production (van Rooyen &amp;

131 Nicol, 2022). Understanding whether crops preferentially acquire NH 4+ when environmental



131 Nicol, 2022). Understanding whether crops preferentially acquire NH 4+ when environmental
132 constraints on its availability are removed could therefore inform both optimization of
133 conventional hydroponic nutrition and development of more circular CEA systems capable of
134 coupling nutrient recovery directly with crop production. Engineered cultivation environments
135 provide an opportunity not simply to reproduce nutrient regimes inherited from soil-based
136 agriculture, but to design nutrient supply around plant physiology and the resources locally
137 available to the production system.



Here, we used differential &#91;15&#93; N stable-isotope labeling to quantify the relative acquisition of NH 4



138 Here, we used differential N stable-isotope labeling to quantify the relative acquisition of NH 4+

139 and NO 3- by four lettuce ( _Lactuca sativa_ L.) cultivars grown in recirculating hydroponic systems.


+                              - +

140 Plants were supplied with NH 4 and NO 3 at equivalent concentrations while either the NH 4 or



+                              - +

140 Plants were supplied with NH 4 and NO 3 at equivalent concentrations while either the NH 4 or

141 NO 3- pool was selectively enriched with 15 N, allowing N-form acquisition to be compared under



+ and NO 3




at equivalent concentrations while either the NH 4



141 NO 3- pool was selectively enriched with 15 N, allowing N-form acquisition to be compared under
142 equivalent external supply. We further examined whether N-form preference was modified under
143 an elevated CO 2 concentration of 3000 ppm, selected as a spaceflight-relevant elevated
144 concentration within the range historically encountered aboard the ISS (Beard, 2020), relative to
145 terrestrial ambient CO 2 (410 ppm). We hypothesized that lettuce would preferentially acquire

146 NH 4+ when both N forms were equally supplied and that this preference would be enhanced



146 NH 4+ when both N forms were equally supplied and that this preference would be enhanced

147 under elevated CO 2 because of the additional reductant requirements associated with NO 3


147 under elevated CO 2 because of the additional reductant requirements associated with NO 3148 assimilation. By separating N-form acquisition from many of the soil processes that ordinarily
149 regulate N availability, this study aimed to determine whether nitrate-dominant nutrition reflects


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


150 the intrinsic N-acquisition physiology of lettuce and to establish a physiological basis for
151 reconsidering N management in terrestrial and SpaCEA.

# 152 Materials and Methods

## 153 Experimental design


154 The experiment was organized in a factorial design comprising four lettuce cultivars, two CO 2
155 concentrations, and two &#91;15&#93; N isotopic labeling treatments, resulting in 16 treatment combinations.
156 Eight plants were established per treatment combination in each experimental run, with five
157 plants targeted for individual isotopic analysis and the remaining plants maintained to
158 accommodate potential germination or establishment losses. The experiment was conducted in
159 three independent runs over time, with a target of 15 individual plants analyzed per treatment
160 combination across the three experimental runs. Where plant loss resulted in fewer than five
161 suitable biological replicates for a treatment combination in an earlier run, additional plants from
162 the corresponding treatment combination were analyzed in the final run to achieve the target
163 sample size.


164 Plants were maintained at their respective CO 2 concentrations throughout the experimental
165 period, whereas differential &#91;15&#93; N stable isotope labeling treatments were imposed only during the
166 final 24 h prior to harvest. At harvest, plant shoot biomass was recorded and shoot tissue was
167 collected from each analyzed plant for determination of total N and &#91;15&#93; N abundance by isotope168 ratio mass spectrometry (IRMS).


169

## 170 Growth conditions and plant material


171 Experiments were conducted in Percival (Perry, Iowa, USA) walk-in controlled-environment
172 growth chambers under either 410 or 3000 ppm CO � . The 410 ppm treatment was maintained in
173 a growth chamber equipped with active CO 2 scrubbing, which was required because background
174 CO 2 concentrations within the facility frequently exceeded terrestrial ambient concentrations.
175 The 3000 ppm treatment was maintained in a separate growth chamber. Consequently, CO 2
176 treatments remained assigned to the same respective chambers across experimental runs, and
177 growth chamber and CO 2 concentration were confounded. Effects associated with CO 2
178 concentration were therefore interpreted with appropriate caution. Chambers were maintained at
179 a constant temperature of 23 °C and 70% relative humidity. Plants were illuminated using
180 OSRAM Phytofy® LED fixtures at an average photosynthetic photon flux density (PPFD) of
181 300 µmol m � ² s � ¹ with a spectral composition of 49% red, 17% green, and 34% blue light and a
182 16-h photoperiod.


183 Four lettuce ( _Lactuca sativa_ L.) cultivars were used in this study: ‘Dragoon’ (Johnny’s Selected
184 Seeds, Winslow, ME, USA), ‘Outredgeous’ (Johnny’s Selected Seeds), ‘Waldmann’s Dark


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


185 Green’ (Johnny’s Selected Seeds), and ‘Black Seeded Simpson’ (Burpee, Warminster, PA,
186 USA). Three seeds were sown directly into 2.5 × 2.5 × 5 cm horticultural foam cubes
187 (Sublime®, BVB Substrates, the Netherlands) positioned within the hydroponic growth system
188 and remained in the system from sowing until harvest at 22 days after planting (DAP). A
189 humidity dome was maintained over emerging seedlings for the first 10 DAP.


190 Plants were grown in custom two-tier nutrient film technique (NFT) hydroponic systems. Each
191 tier comprised four 5-ft (1.5 m) NFT channels (CropKing Inc., Ohio, USA), with eight plant
192 positions per channel and a total capacity of 32 plants per tier. Each tier contained eight plants of
193 each lettuce cultivar, with individual plants assigned to positions in a randomized layout within
194 each tier for each experimental run. Nutrient solution was supplied from 20-L opaque reservoirs
195 and pumped to the NFT channels using a submersible pump at an average flow rate of
196 approximately 450 mL min � ¹. Channels were mounted on transparent acrylic supports of
197 progressively differing heights to provide a downward slope, with nutrient solution draining
198 from the channels through 2.5-inch (6.35 cm)-diameter PVC piping and returning to the
199 reservoir. Nutrient solution was continuously recirculated throughout the experimental period.



200 Plants were supplied with 1× Yoshida nutrient solution, prepared as described by Yoshida _et_
201 _al._ (1976), which was selected because nitrogen is supplied exclusively as NH � NO � and

202 therefore provides NH 4+ and NO 3- at an equimolar ratio. The composition of the nutrient solution



+ and NO 3



202 therefore provides NH 4 and NO 3- at an equimolar ratio. The composition of the nutrient solution
203 is provided in Table 1. Reservoirs were maintained at a working volume of 20 L throughout
204 plant growth prior to isotopic labeling. Nutrient solution was maintained without replacement for
205 the first 10 DAP and was completely replaced every 2 d thereafter to minimize changes in the

206 relative availability of NH 4+ and NO 3- resulting from plant nutrient uptake. On days when



+ and NO 3



206 relative availability of NH 4 and NO 3- resulting from plant nutrient uptake. On days when
207 complete solution replacement was not performed, reservoir volume was restored to 20 L with
208 fresh 1× Yoshida nutrient solution as required; supplementation was not required during the first
209 10 DAP. Nutrient solution pH was initially adjusted to 5.8 and subsequently readjusted to 5.8
210 daily using 1 M KOH to compensate for progressive solution acidification.



211 **Table 1:** Yoshida nutrient solution composition, adapted from Yoshida _et al._ (1976).


Macronutrients

N NH 4 NO 3 40

P NaH 2 PO 4     - 2H 2 O 10

K K 2 SO 4 40

Ca CaCl 2    - 2H 2 O 40

Mg MgSO 4 40

Micronutrients

Mn MnCl 2    - 4H 2 O 0.5

Mo (NH 4 ) 2 Mo 7 O 24    - 4H 2 O 0.05
B H 3 BO 3 0.2


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


Zn ZnSO 4     - 7H 2 O 0.01

Cu CuSO 4    - 5H 2 O 0.01

Fe FeCl 3     - 6H 2 O 2

212


213


15
## 214 N stable isotope labeling



215 At 21 days after planting (DAP), differential &#91;15&#93; N stable isotope labeling treatments were imposed

216 for the final 24 h of plant growth to independently quantify acquisition of NH 4+ -N and NO 3- -N.



+ -N and NO 3



216 for the final 24 h of plant growth to independently quantify acquisition of NH 4 -N and NO 3- -N.

15
217 Ammonium and nitrate acquisition were traced using &#91;15&#93; NH 4 NO 3 and NH 4 NO 3, respectively (98



15
217 Ammonium and nitrate acquisition were traced using &#91;15&#93; NH 4 NO 3 and NH 4 NO 3, respectively (98
218 atom% &#91;15&#93; N; Thermo Fisher Scientific). Labeled and unlabeled NH 4 NO 3 were combined to
219 provide a nominal 5 atom% &#91;15&#93; N enrichment of the respective labelled N pool, with the quantity
220 of labelled material adjusted to account for its 98 atom% &#91;15&#93; N enrichment. The remaining
221 NH 4 NO 3 was supplied in unlabeled form. Thus, the two labeling solutions contained equivalent

222 total N concentrations and NH 4+ :NO 3- ratios and differed only in whether the NH 4+ or NO 3- pool
223 was enriched with &#91;15&#93; N. All other components of the Yoshida nutrient solution remained
224 unchanged.



+ :NO 3



+ or NO 3



ratios and differed only in whether the NH 4



225 Fresh nutrient solution was prepared immediately prior to the labeling period. For each
226 experimental run, 20 L of each labeling solution were prepared and divided equally between the
227 two CO � treatment chambers, providing 10 L per NFT tier. Existing nutrient solution was
228 completely drained from each reservoir, and the reservoir was rinsed with deionized water prior
229 to addition of the labeled solution. Within each growth chamber, one NFT tier received
15 15
230 NH 4 NO 3 - labeled solution and the other received NH 4 NO 3 - labelled solution. Assignment of
231 isotopic labeling treatment to NFT tier was alternated between independent experimental runs to
232 minimize potential confounding of labeling treatment with tier-specific effects.


233 Reservoir working volume was reduced from 20 to 10 L during the labeling period to reduce the
234 quantity of &#91;15&#93; N-labeled material required while maintaining continuous operation of the NFT
235 system. Following addition of the labelled nutrient solution, pH was adjusted to 5.8 but was not
236 subsequently corrected during the 24-h labeling period. All other environmental conditions were
237 maintained as described above. Plants were exposed to the labeled nutrient solutions for exactly
238 24 h and harvested immediately following completion of the labeling period.

## 239 Plant harvest and sample preparation


240 Plants were harvested at 22 days after planting (DAP), immediately following completion of the
241 24-h &#91;15&#93; N labeling period. Shoots were excised at the surface of the foam cube and weighed
242 immediately to determine shoot fresh weight. Roots were discarded because extensive root
243 growth through the foam substrate prevented consistent recovery of intact root systems.


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


244 Individual shoots were placed into paper bags and oven-dried at 70 °C for 7 d. Following drying,
245 shoot dry weight was recorded. The entire dried shoot from each individual plant was then
246 transferred to a 50 mL centrifuge tube containing approximately 5 mL by volume of 5 mm glass
247 beads and homogenized to a fine powder using a Bead Ruptor Elite bead mill homogenizer
248 (OMNI International, Kennesaw, GA, USA). Samples were homogenized at 5 m s � ¹ for two 30
249 s cycles separated by a 30 s dwell period to minimize sample heating.


250 Following homogenization, an aliquot of each powdered shoot sample was transferred to a 1.5
251 mL microcentrifuge tube.

## 252 Isotope-ratio mass spectrometry and calculation of nitrogen metrics


253 Approximately 3 mg of homogenized shoot tissue from each plant was weighed into a tin
254 capsule and analyzed for total N concentration and &#91;15&#93; N enrichment at the University of Florida
255 Stable Isotope Mass Spectrometry (SIMS) Laboratory. Samples were analyzed using a Costech
256 ECS 4010 CHNS-O elemental analyzer coupled via a ConFlo IV interface to a Thermo Finnigan
257 Delta V Advantage isotope-ratio mass spectrometer operating in continuous-flow mode. Samples
258 underwent flash combustion at 1000 °C in an oxygen-rich atmosphere, followed by reduction at
259 650 °C and removal of water using a magnesium perchlorate trap. N 2 and CO 2 were
260 subsequently separated using a 3.0 m gas chromatography column maintained at 55 °C prior to
261 elemental and isotopic analysis. Tissue N concentration was determined as percentage N (%N)
262 by comparison with calibration standards. For &#91;15&#93; N-enriched samples, isotopic composition was
263 reported as atom percent &#91;15&#93; N (atom% &#91;15&#93; N). A 1.0 atom% &#91;15&#93; N ammonium sulfate standard and
264 natural-abundance USGS40 reference material were included during analysis.



Differential &#91;15&#93; N enrichment was used to assess relative acquisition of NH 4



+ and NO 3



265 Differential N enrichment was used to assess relative acquisition of NH 4 and NO 3- . Because

266 plants supplied with &#91;15&#93; NH 4 NO 3 and NH 415 NO 3 received equivalent concentrations of NH 4+ and



266 plants supplied with &#91;15&#93; NH 4 NO 3 and NH 415 NO 3 received equivalent concentrations of NH 4+ and

267 NO 3- and differed only in the isotopically labelled N form, greater shoot atom% 15 N following



267 NO 3- and differed only in the isotopically labelled N form, greater shoot atom% 15 N following
15 +
268 NH 4 labeling was interpreted as greater acquisition of ammonium-derived N, whereas greater

269 enrichment following &#91;15&#93; NO 3- labeling indicated greater acquisition of nitrate-derived N.



269 enrichment following NO 3- labeling indicated greater acquisition of nitrate-derived N.



270 Total shoot N content was calculated for each individual plant from shoot dry mass and tissue N
271 concentration determined by elemental analysis as:


Total shoot N �g��shoot dry mass �g� x &#91;%�&#93;

100

## 272 Statistical analysis


273 Statistical analyses were performed in R using linear mixed-effects models fitted with
274 the lmerTest package. Individual plants represented biological observations; however, plants
275 within a common NFT tier shared the same recirculating nutrient reservoir and were therefore
276 treated as subsamples of a common experimental unit. NFT tier, uniquely defined within
277 experimental run and growth chamber, was included as a random intercept to account for this


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


278 hierarchical structure. Experimental run was included as a fixed blocking factor because the
279 experiment was conducted in three independent runs over time. Type III tests of fixed effects
280 were performed using Satterthwaite approximations for denominator degrees of freedom. Model
281 assumptions were assessed by visual inspection of residual-versus-fitted and normal Q–Q plots.
282 Statistical significance was assessed at α = 0.05.



283 Shoot &#91;15&#93; N enrichment (atom% &#91;15&#93; N) was analyzed separately for each lettuce cultivar using

+ 15                           
284 models containing isotopic label ( &#91;15&#93; NH 4 or NO 3 ), CO 2 concentration (410 or 3000 ppm), and



+ 15
or NO 3



284 models containing isotopic label ( NH 4 or NO 3- ), CO 2 concentration (410 or 3000 ppm), and
285 their interaction as fixed effects, with experimental run included as a fixed block and NFT tier as
286 a random intercept. Estimated marginal means were calculated using the emmeans package.
287 Planned pairwise comparisons were performed between isotopic labels within each CO 2
288 concentration and between CO 2 concentrations within each isotopic label, with _P_ -values adjusted
289 using the Holm method. Model-estimated marginal means and 95% confidence intervals were
290 used to summarize treatment responses. To visualize the magnitude and direction of N-form
291 preference, a preference index was calculated from the model-estimated marginal means as
15 15 +
292 N NO3 − N NH4, with negative values indicating greater acquisition of NH 4 -derived N and

293 positive values indicating greater acquisition of NO 3- -derived N. Standard errors for the



293 positive values indicating greater acquisition of NO 3- -derived N. Standard errors for the
294 preference index were obtained from the corresponding model-estimated contrast. Statistical
295 inference concerning N-form acquisition was based on the underlying mixed-effects models and
296 planned contrasts rather than significance testing of the preference index as a separate response

297 variable.



298 Shoot fresh weight and total shoot N content were analyzed using linear mixed-effects models
299 containing lettuce cultivar, CO 2 concentration, and their interaction as fixed effects. Isotopic
300 label and experimental run were included as additive fixed effects, and NFT tier was included as
301 a random intercept. Isotopic label was retained in these models to account for assignment of the
302 isotopic treatments, although the two labeling solutions were chemically equivalent apart from
303 the isotopically enriched N form. Where appropriate, estimated marginal means were compared
304 using Tukey-adjusted pairwise comparisons. For total shoot N, additional planned comparisons
305 between CO 2 concentrations within each cultivar were performed using Holm adjustment.
306 Figures show individual plant observations and/or treatment distributions together with model307 derived summaries or multiple-comparison groupings, as appropriate.

# 308 Results


309


310 **Differential** **&#91;15&#93;** **N labeling reveals consistent ammonium preference across lettuce cultivars**


311 Shoot &#91;15&#93; N enrichment differed significantly between isotopic labeling treatments in all four
312 lettuce cultivars, indicating differential acquisition of the two inorganic N forms. The effect of
313 isotopic label was significant for Dragoon, Outredgeous, Waldmann’s Dark Green, and Black


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


314 Seeded Simpson. By contrast, neither CO 2 concentration, nor the Label × CO 2 interaction were
315 significant for any cultivar (Fig. 1).



Within each CO 2 environment, shoot &#91;15&#93; N enrichment was consistently greater following &#91;15&#93; NH 4



316 Within each CO 2 environment, shoot N enrichment was consistently greater following NH 4+

                  - 15 + 15                  
317 than &#91;15&#93; NO 3 labeling. At 410 ppm CO 2, the model-estimated NH 4 − NO 3 differences were



- 15
labeling. At 410 ppm CO 2, the model-estimated NH 4



+ 15
− NO 3



                  -                  
317 than NO 3 labeling. At 410 ppm CO 2, the model-estimated NH 4 − NO 3 differences were
318 0.182 atom% for Dragoon, 0.383 atom% for Outredgeous, 0.493 atom% for Waldmann’s Dark
319 Green, and 0.718 atom% for Black Seeded Simpson. The same pattern was retained at 3000 ppm
320 CO 2, with corresponding differences of 0.195, 0.443, 0.587, and 0.703 atom%, respectively in all
321 cases (Fig. 1). No within-label comparison between 410 and 3000 ppm CO 2 was significant after
322 Holm adjustment.



323 The model-derived N-form preference index was negative for all cultivars under both CO 2

324 conditions, confirming greater acquisition of NH 4+ -derived than NO 3- -derived N (Fig. 2). The



+ -derived than NO 3



324 conditions, confirming greater acquisition of NH 4 -derived than NO 3- -derived N (Fig. 2). The
325 magnitude of ammonium preference increased in the order Dragoon &lt; Outredgeous &lt;
326 Waldmann’s Dark Green &lt; Black Seeded Simpson. Preference indices at 410 and 3000 ppm
327 CO 2, respectively, were −0.182 and −0.195 for Dragoon, −0.383 and −0.443 for Outredgeous,

328 −0.493 and −0.587 for Waldmann’s Dark Green, and −0.718 and −0.703 for Black Seeded

329 Simpson. Although Dragoon, Outredgeous and Waldmann’s Dark Green showed numerically
330 greater ammonium preference at 3000 ppm, the absence of significant Label × CO 2 interactions
331 indicated that CO 2 concentration did not significantly alter N-form preference.



332


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


333



334 **Figure 1: Differential acquisition of ammonium- and nitrate- derived nitrogen by lettuce under contrasting CO** **2**

+ 15                                                                  
335 **environments.** Shoot &#91;15&#93; N enrichment of four lettuce ( _Lactuca sativa_ L.) cultivars following differential &#91;15&#93; NH 4 or NO 3 labeling
336 under 410 and 3000 ppm CO 2 . Plants were exposed to isotopically labelled nutrient solutions for the final 24 h of growth. Small
337 points represent individual plants and larger points with error bars represent estimated marginal means ± 95% confidence
338 intervals derived from linear mixed-effects models fitted separately for each cultivar. Models included isotopic label, CO 2
339 concentration, and their interaction as fixed effects, experimental run as a fixed block, and NFT tier as a random intercept. Solid

340 brackets indicate significant pairwise differences between &#91;15&#93; NH 4+ - and 15 NO 3- - labeled treatments within a CO 2 concentration
341 following Holm adjustment. Only significant comparisons are shown. P &lt; 0.05 (*), P &lt; 0.01 (**), and P &lt; 0.001 (***).



+ 15
or NO 3



+ - and 15 NO 3


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


342



343 **Figure 2: Nitrogen-form preference differs among lettuce cultivars.** Nitrogen-form preference of four lettuce ( _Lactuca_
344 _sativa_ L.) cultivars grown under 410 and 3000 ppm CO 2 . Preference indices were calculated from model-estimated marginal

345 means as shoot &#91;15&#93; N enrichment following &#91;15&#93; NO 3- labeling minus enrichment following 15 NH 4+ labeling. Negative values indicate

346 greater acquisition of NH 4+ -derived N, positive values indicate greater acquisition of NO 3- -derived N, and the dashed horizontal
347 line at zero represents equivalent acquisition of the two N forms. Points represent model-estimated preference indices and error
348 bars represent standard errors of the corresponding model-estimated contrasts. The preference index is presented as a summary
349 visualization; statistical inference concerning N-form acquisition was performed on the underlying &#91;15&#93; N enrichment data (Fig. 1).



- labeling minus enrichment following 15 NH 4



+
-derived N, positive values indicate greater acquisition of NO 3



350


351 **Shoot fresh biomass differs among cultivars but not CO** **2** **environments**


352 Shoot fresh weight differed significantly among lettuce cultivars, whereas there were no
353 significant effects of CO 2 concentration, isotopic label, or Cultivar × CO 2 interaction (Fig. 3).
354 Waldmann’s Dark Green and Black Seeded Simpson produced the greatest shoot biomass, with
355 estimated marginal means of 13.91 and 12.87 g plant � ¹, respectively, and did not differ
356 significantly from one another. Both cultivars had significantly greater shoot fresh biomass than
357 Dragoon (10.14 g plant � ¹) and Outredgeous (9.02 g plant � ¹), which did not differ significantly

358 from one another.


359


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


360


361 **Figure 3. Shoot fresh biomass differs among lettuce cultivars.** Shoot fresh weight of four lettuce ( _Lactuca sativa_ L.) cultivar **s**
362 following 22 d of hydroponic growth. Boxplots show the distribution of individual plant observations pooled across CO 2 an **d**
363 isotopic labeling treatments; horizontal lines indicate medians, boxes represent the interquartile range, and whiskers extend to 1. **5**
364 × the interquartile range. Triangles indicate arithmetic means. Data were analyzed using a linear mixed-effects model containin **g**
365 cultivar, CO 2 concentration, and their interaction as fixed effects, isotopic label and experimental run as additive fixed effects,
366 and NFT tier as a random intercept. Different letters indicate significant differences among cultivar estimated marginal mean **s**
367 following Tukey adjustment ( _P_ &lt; 0.05).


368 **Total shoot nitrogen content shows a cultivar-dependent response to CO** **2** **environment**


369 Total shoot N content differed significantly among cultivars and showed a significant Cultivar **×**
370 CO 2 interaction (Fig. 4). Neither the overall effect of CO 2 concentration, nor isotopic label wer **e**
371 significant.


372 The highest model-estimated total shoot N content occurred in Black Seeded Simpson at 41 **0**
373 ppm CO 2 (0.0526 g plant &#91;-1&#93; ), followed by Waldmann’s Dark Green at 410 ppm (0.0481 g plant &#91;-1&#93; **)**
374 and 3000 ppm (0.0458 g plant &#91;-1&#93; ) and Black Seeded Simpson at 3000 ppm (0.0445 g plant &#91;-1&#93; ).
375 Dragoon and Outredgeous generally accumulated less total shoot N. Despite the significan **t**
376 overall Cultivar × CO 2 interaction, Holm-adjusted comparisons between CO 2 concentration **s**
377 within individual cultivars were not significant. Black Seeded Simpson showed the larges **t**
378 numerical difference between CO 2 environments, with total shoot N 0.0081 g plant &#91;-1&#93; greater a **t**


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


379 410 than 3000 ppm, whereas the corresponding within-cultivar differences were smaller fo **r**
380 Dragoon, Outredgeous, and Waldmann’s Dark Green.


381


382 **Figure 4. Total shoot nitrogen content varies with lettuce cultivar and CO** � **environment.** Total shoot N content of fou **r**
383 lettuce ( _Lactuca sativa_ L.) cultivars grown under 410 or 3000 ppm CO 2 . Total shoot N was calculated from shoot dry mass an **d**
384 tissue N concentration determined by elemental analysis. Boxplots show individual plant distributions; horizontal lines indicat **e**
385 medians, boxes represent the interquartile range, and whiskers extend to 1.5 × the interquartile range. Triangles indicat **e**
386 arithmetic means. Data were analyzed using a linear mixed-effects model containing cultivar, CO 2 concentration, and thei **r**
387 interaction as fixed effects, isotopic label and experimental run as additive fixed effects, and NFT tier as a random intercept.
388 Different letters indicate significant differences among cultivar × CO 2 estimated marginal means following Tukey adjustmen **t**
389 ( _P_ &lt; 0.05).


390


391


392

# 393 Discussion



394 The present study demonstrates a consistent preference for ammonium-derived N across fou **r**


+                    
395 lettuce cultivars when NH 4 and NO 3 were supplied simultaneously at equivalent concentrations.



+ and NO 3



395 lettuce cultivars when NH 4 and NO 3- were supplied simultaneously at equivalent concentrations



.


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.



Shoot &#91;15&#93; N enrichment was significantly greater following &#91;15&#93; NH 4



+ than 15 NO 3



396 Shoot N enrichment was significantly greater following NH 4 than NO 3- labeling in every

397 cultivar and under both CO 2 environments. This contrasts with the nitrate-dominant nutrition

398 conventionally employed in hydroponic crop production (Sonneveld &amp; Voogt, 2009; Voogt &amp;

399 Bar-Yosef, 2019) but is physiologically plausible. NO 3- must be reduced to NO 2- and



399 Bar-Yosef, 2019) but is physiologically plausible. NO 3- must be reduced to NO 2- and

400 subsequently NH 4+ before entering the GS/GOGAT assimilation pathway, whereas absorbed




- must be reduced to NO 2



400 subsequently NH 4+ before entering the GS/GOGAT assimilation pathway, whereas absorbed

401 NH 4+ can be assimilated directly, avoiding the additional energetic and reductant requirements



401 NH 4+ can be assimilated directly, avoiding the additional energetic and reductant requirements

402 associated with NO 3- reduction (Bloom et al., 1992). Although NH 4+ nutrition carries additional



402 associated with NO 3- reduction (Bloom et al., 1992). Although NH 4+ nutrition carries additional
403 physiological costs associated with ion homeostasis, carbon-skeleton provision and pH
404 regulation (Britto &amp; Kronzucker, 2013), these results demonstrate preferential NH ��
405 acquisition when both inorganic N forms are similarly supplied.




reduction (Bloom et al., 1992). Although NH 4



406 This distinction between physiological acquisition and environmental availability is central to
407 interpreting N-form preference. Studies in natural systems demonstrate that apparent preference
408 can shift with environmentally mediated N availability, both across communities and within
409 species (Houlton et al., 2007; Wang &amp; Macko, 2011), while the concept of fixed N-form
410 preference has been questioned because differential acquisition in soil can reflect N
411 transformations and differences in availability rather than plant discrimination alone (Chalk &amp;

412 Smith, 2021). Hydroponics controls many of these environmental determinants, allowing NH 4+



412 Smith, 2021). Hydroponics controls many of these environmental determinants, allowing NH 4+

413 and NO 3- to be supplied directly and at defined concentrations. Our differential-labeling approach



413 and NO 3- to be supplied directly and at defined concentrations. Our differential-labeling approach
414 further minimized differences in external supply, yet all four cultivars preferentially acquired

415 NH 4+ . Consequently, the widespread use of NO 3- as the dominant N source in hydroponic crop
416 production should not necessarily be interpreted as evidence that it represents the intrinsically
417 preferred N form of lettuce.



+
. Consequently, the widespread use of NO 3



418 Elevated CO 2 was hypothesized to strengthen NH 4+ preference because elevated CO 2 can inhibit

419 shoot NO 3- assimilation in C 3 plants and alter responses to N form (Bloom et al., 2002; Bloom et
420 al., 2012). Contrary to this hypothesis, CO 2 concentration did not significantly alter N-form

421 preference. One possibility is that the strong NH 4+ preference already present at 410 ppm



Elevated CO 2 was hypothesized to strengthen NH 4



421 preference. One possibility is that the strong NH 4+ preference already present at 410 ppm

422 provided limited scope for a further shift. Alternatively, changes in shoot NO 3- assimilation may



422 provided limited scope for a further shift. Alternatively, changes in shoot NO 3- assimilation may
423 not translate directly into whole-plant acquisition when both N forms remain continuously

424 available. Importantly, NH 4+ preference persisted in all cultivars at 3000 ppm, a spaceflight


424 available. Importantly, NH 4+ preference persisted in all cultivars at 3000 ppm, a spaceflight425 relevant elevated CO 2 concentration within the range historically encountered aboard the ISS
426 (Beard, 2020; Georgescu et al., 2020). The CO 2 comparison should nevertheless be interpreted
427 cautiously because each concentration was maintained in a dedicated growth chamber,
428 preventing separation of CO 2 from persistent chamber effects. The strongest inference therefore
429 lies in the replicated differential-label comparisons within each environment: lettuce

430 preferentially acquired NH 4+ under both ambient and highly CO 2 -enriched conditions.



430 preferentially acquired NH 4+ under both ambient and highly CO 2 -enriched conditions.



431 The magnitude of NH 4+ preference varied substantially among cultivars and broadly
432 corresponded with biomass and total shoot N accumulation. Dragoon exhibited the weakest
433 preference and comparatively low biomass and shoot N, whereas Black Seeded Simpson
434 exhibited the strongest preference alongside greater biomass and N accumulation. Importantly,


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.



this gradient was driven predominantly by increasing &#91;15&#93; N enrichment following &#91;15&#93; NH 4



435 this gradient was driven predominantly by increasing N enrichment following NH 4+ labeling

436 rather than decreasing &#91;15&#93; NO 3- enrichment. A simple effect of plant size would be expected to



436 rather than decreasing NO 3- enrichment. A simple effect of plant size would be expected to
437 influence acquisition from both labelled pools, while atom% &#91;15&#93; N represents proportional isotopic
438 composition rather than total tracer mass. The data therefore are consistent with a comparatively


                                        - +

439 stable contribution from NO 3 acquisition accompanied by increasing incorporation of NH 4 



                                        - +

439 stable contribution from NO 3 acquisition accompanied by increasing incorporation of NH 4 440 derived N as plant growth and N demand increase. Shoot atom% &#91;15&#93; N does not provide an
441 absolute uptake flux, so this mechanism cannot be demonstrated conclusively; however, it
442 suggests a testable hypothesis that rapidly growing or highly N-demanding genotypes may

443 increasingly satisfy additional N demand through NH 4+ acquisition when both forms are




acquisition accompanied by increasing incorporation of NH 4



443 increasingly satisfy additional N demand through NH 4+ acquisition when both forms are

444 available.



445 These findings have implications for nutrient management in space crop production, where crop
446 performance must ultimately be considered alongside resource use and integration with
447 regenerative life-support systems. Nutrient-loop closure and recycled fertigation have been
448 identified as important steps toward operational BLSS (Fountain et al., 2026). Increasing the

449 proportion of N supplied as NH 4+ could better align nutrient provision with lettuce acquisition



449 proportion of N supplied as NH 4+ could better align nutrient provision with lettuce acquisition
450 physiology, although whether this improves yield or N-use efficiency requires direct testing

451 across optimized NH 4+ :NO 3- ratios. Recent work showing improved hydroponic lettuce



+ :NO 3



451 across optimized NH 4 :NO 3- ratios. Recent work showing improved hydroponic lettuce
452 performance when NH �� comprised 20% of supplied N further supports optimization of crop453 specific N-form ratios rather than assuming nitrate-dominant nutrition is optimal (Xu et al.,

454 2025).



455 N-form preference may become particularly important when crop production is coupled with

456 nutrient recovery. Urinary N is predominantly excreted as urea and converted to NH 4+ following



456 nutrient recovery. Urinary N is predominantly excreted as urea and converted to NH 4+ following

457 hydrolysis, whereas producing NO 3- from this recovered N requires nitrification (Verbeelen et



457 hydrolysis, whereas producing NO 3- from this recovered N requires nitrification (Verbeelen et

458 al., 2021). Greater direct utilization of recovered NH 4+ could therefore potentially reduce the



458 al., 2021). Greater direct utilization of recovered NH 4+ could therefore potentially reduce the
459 extent of processing required before crop delivery and help integrate crop production with
460 nutrient-loop closure (Verbeelen et al., 2021; Fountain et al., 2026). Similar opportunities exist in
461 terrestrial CEA, where ammonium-rich recovered nutrient streams such as anaerobic-digestion
462 digestate could be integrated with hydroponic crop production (van Rooyen &amp; Nicol, 2022).

463 Preferential NH 4+ acquisition should not, however, be interpreted as evidence that unrestricted or



463 Preferential NH 4+ acquisition should not, however, be interpreted as evidence that unrestricted or

464 sole-source NH 4+ nutrition is optimal. Excessive NH �� can impair growth and drive rapid root


464 sole-source NH 4+ nutrition is optimal. Excessive NH �� can impair growth and drive rapid root465 zone acidification (Britto &amp; Kronzucker, 2013; Natarajan Sakthivel et al., 2023), making

466 optimization of NH 4+ :NO 3- supply and pH management essential.



+ :NO 3



466 optimization of NH 4 :NO 3- supply and pH management essential.



467 More broadly, these results demonstrate that nitrate-dominant nutrition does not necessarily
468 reflect intrinsic N-form acquisition preference. Engineered hydroponic systems provide an
469 opportunity to approach N nutrition differently from soil-based agriculture: rather than allowing
470 environmental processes to determine the N forms available to roots, nutrient supply can be

471 designed around crop acquisition physiology. The consistent NH 4+ preference observed across
472 four lettuce cultivars provides a physiological basis for reconsidering N-form provision in CEA


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


473 and for integrating crop nutrient requirements with locally available and recovered N resources
474 on Earth and beyond.

# 475 References


476 Beard, B. L. (2020). Characterization of how CO 2 level may impact crew performance related to
477 the HSIA risk. NASA/TM–20205011433. NASA Ames Research Center.

478 Bloom, A. J., Sukrapanna, S. S., &amp; Warner, R. L. (1992). Root respiration associated with
479 ammonium and nitrate absorption and assimilation by barley. _Plant Physiology_, 99,
480 1294–1301.
481 Bloom, A. J., Smart, D. R., Nguyen, D. T., &amp; Searles, P. S. (2002). Nitrogen assimilation and
482 growth of wheat under elevated carbon dioxide. _Proceedings of the National Academy of_
483 _Sciences of the United States of America_, 99, 1730–1735.
484 Bloom, A. J., Rubio Asensio, J. S., Randall, L., Rachmilevitch, S., Cousins, A. B., &amp; Carlisle, E.
485 A. (2012). CO 2 enrichment inhibits shoot nitrate assimilation in C 3 but not C 4 plants and
486 slows growth under nitrate in C � plants. _Ecology_, 93, 355–367.
487 Britto, D. T., &amp; Kronzucker, H. J. (2013). Ecological significance and complexity of N-source
488 preference in plants. _Annals of Botany_, 112, 957–963.
489 Chalk, P. M., &amp; Smith, C. J. (2021). On inorganic N uptake by vascular plants: Can &#91;15&#93; N tracer
490 techniques resolve the NH �� versus NO �� “preference” conundrum? _European_
491 _Journal of Soil Science_, 72, 1762–1779.
492 Ewert, M. K., Chen, T. T. &amp; Powell, C. D. (Eds.). (2022). Life Support Baseline Values and Ass
493 umptions Document(NASA/TP-2015494 218570/REV2). National Aeronautics and Space Administration, Johnson Space Center.
495 https://ntrs.nasa.gov/citations/20210024855
496 Fountain, L.L., Gilliham, M., Amitrano, C., Arouna, N., Barker, R.J., Bohmer, M., Braun, M.,
497 Brereton, N.J.B., Brocato, R.L., Bunchek, J.M., Canaday, E.L.J., Caplin, N., Castano, P.,
498 Decourteix, M., Del Bianco, M., De Micco, V., Doherty, C.J., Escobar, C., Franke, M.F.,
499 Fuentes, S., Gilroy, S., Harrison, L., Hasenstein, K.H., Hauslage, J., Herranz, R., Iyer500 Pascuzzi, A., Izuma, D.S., Junya, K., Kiss, J.Z., Legue, V., Lloyd, J.P.B., Massa, G.D.,
501 Maffei, M.E., Meyers, A.D., Perera, I.Y., Poulet, L., Roychoudry, S., Sena, G., Shippen,
502 D.E., Stoochnoff, J., Takahashi, H., Wyatt, S.E., &amp; Blancaflor, E.B. (2025). Expanding
503 frontiers: harnessing plant biology for space exploration and planetary sustainability. _New_
504 _Phytologist_, 249(2), 656–669.
505 Georgescu, M.-R., Meslem, A., &amp; N ă stase, I. (2020). Accumulation and spatial distribution of
506 CO � in the astronaut's crew quarters on the International Space Station. _Building and_
507 _Environment_, 185, 107278.
508 Houlton, B. Z., Sigman, D. M., Schuur, E. A. G., &amp; Hedin, L. O. (2007). A climate-driven switch
509 in plant nitrogen acquisition within tropical forest communities. _Proceedings of the_
510 _National Academy of Sciences of the United States of America_, 104, 8902–8906.
511 Maiwald, V., Vrakking, V., Zabel, P., Schubert, D., Waclavicek, R., Dorn, M., Fiore, L., Imhof,
512 B., Rousek, T., Rossetti, V. &amp; Zeidler, C. (2020). From ice to space: A greenhouse design
513 for Moon or Mars based on a prototype deployed in Antarctica. CEAS Space Journal, 12,
514 507–526.
515 Natarajan Sakthivel, Song, J., &amp; Jeong, B. R. (2023). Ammonium phytotoxicity and tolerance:
516 An insight into ammonium nutrition to improve crop productivity. _Agronomy_, 13, 1487.


&#91;bioRxiv preprint doi: https://doi.org/10.64898/2026.11.16.767057; this version posted November 17, 2026. The copyright holder for this preprint&#93;(https://doi.org/10.64898/2026.11.16.767057)

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.


517 Sonneveld, C., &amp; Voogt, W. (2009). _Plant Nutrition of Greenhouse Crops_ . Springer, Dordrecht.
518 van Rooyen, I. L., &amp; Nicol, W. (2022). Nitrogen management in nitrification-hydroponic
519 systems by utilizing their pH characteristics. _Environmental Technology &amp; Innovation_,
520 26, 102360.
521 Verbeelen, T., Leys, N., Ganigué, R., &amp; Mastroleo, F. (2021). Development of nitrogen
522 recycling strategies for bioregenerative life support systems in space. _Frontiers in_
523 _Microbiology_, 12, 700810.
524 Voogt, W., &amp; Bar-Yosef, B. (2019). Water and nutrient management and crops response to
525 nutrient solution recycling in soilless growing systems in greenhouses. In M. Raviv, J. H.
526 Lieth, &amp; A. Bar-Tal (Eds.), _Soilless Culture: Theory and Practice_ (2nd ed., pp. 425–507).
527 Elsevier.
528 Wang, L., &amp; Macko, S. A. (2011). Constrained preferences in nitrogen uptake across plant
529 species and environments. _Plant, Cell &amp; Environment_, 34, 525–534.
530 Xu, J., Che, G., Yan, J., Luo, Y., Good, A. G., Yuan, T., Lei, Z., Zhang, Z., &amp; Shizumi, K.
531 (2025). Optimizing nitrogen form ratio for enhanced nutrient uptake and lettuce
532 performance in hydroponic systems. _Bioresource Technology Reports_, 32, 102384.


533
# Acknowledgements


534 L.L.F. was supported by an appointment to the NASA Postdoctoral Program at NASA Kennedy
535 Space Center, administered by Oak Ridge Associated Universities under contract with NASA.
536 The authors thank the Kennedy Space Center LASSO-II team for assistance with experimental
537 activities and the University of Florida Stable Isotope Mass Spectrometry Laboratory for isotope538 ratio mass spectrometry analyses.

# 539 Author Contributions


540 L.L.F., E.B.B., R.M.W., and G.D.M. conceived and designed the experiments. L.L.F. performed
541 the experiments. L.L.F. and S.M. analyzed the data. L.L.F. drafted the manuscript with
542 contributions from S.M. E.B.B., R.M.W., and G.D.M. critically reviewed and revised the
543 manuscript. All authors approved the final manuscript.


