The influence of abrupt increases in seawater pCO2 on plankton productivity in the subtropical North Pacific Ocean

We conducted a series of experiments to examine short-term (2–5 days) effects of abrupt increases in the partial pressure of carbon dioxide (pCO2) in seawater on rates of primary and bacterial production at Station ALOHA (22°45’ N, 158° W) in the North Pacific Subtropical Gyre (NPSG). The majority of experiments (8 of 10 total) displayed no response in rates of primary production (measured by 14C-bicarbonate assimilation; 14C-PP) under elevated pCO2 (~1100 μatm) compared to ambient pCO2 (~387 μatm). In 2 of 10 experiments, rates of 14C-PP decreased significantly (~43%) under elevated pCO2 treatments relative to controls. Similarly, no significant differences between treatments were observed in 6 of 7 experiments where bacterial production was measured via incorporation of 3H-leucine (3H-Leu), while in 1 experiment, rates of 3H-Leu incorporation measured in the dark (3H-LeuDark) increased more than 2-fold under high pCO2 conditions. We also examined photoperiod-length, depth-dependent (0–125 m) responses in rates of 14C-PP and 3H-Leu incorporation to abrupt pCO2 increases (to ~750 μatm). In the majority of these depth-resolved experiments (4 of 5 total), rates of 14C-PP demonstrated no consistent response to elevated pCO2. In 2 of 5 depth-resolved experiments, rates of 3H-LeuDark incorporation were lower (10% to 15%) under elevated pCO2 compared to controls. Our results revealed that rates of 14C-PP and bacterial production in this persistently oligotrophic habitat generally demonstrated no or weak responses to abrupt changes in pCO2. We postulate that any effects caused by changes in pCO2 may be masked or outweighed by the role that nutrient availability and temperature play in controlling metabolism in this ecosystem.


Introduction
Human socioeconomic activities, specifically fossil fuel combustion, cement production, and changes in land use, have resulted in progressive increases in atmospheric and oceanic carbon dioxide (CO 2 ) inventories [1]. The ocean is a globally important net sink for CO 2 , and as such, increases in atmospheric CO 2 have raised seawater pCO 2 , with concomitant decreases in seawater pH [2][3][4][5]. However, studies examining the effects of changes in seawater carbonate chemistry on plankton productivity in open ocean ecosystems are relatively scarce. While an appropriate null hypothesis could be that ocean acidification may lead to no significant changes in microbial contributions to biogeochemical cycling [6], testing such a hypothesis demands rigorous experimental evidence. Previous results and observations suggest that, either as individual species or microbial assemblages, marine microbial physiology may be affected by increases in pCO 2 [7][8][9][10][11][12][13][14]. However, the reported signs and magnitudes of the effects vary [14]. Whether these changes in microbial physiology are large enough to impact ocean biogeochemical cycles remains an important unanswered question.
To date, relatively little is known about the capacity of phytoplankton to adapt or acclimate to changes in the seawater carbonate system, which are likely to have complex influences on ocean biology. Most contemporary lineages of phytoplankton evolved during periods in Earth's history when atmospheric and oceanic CO 2 inventories were considerably greater than today [15,16]. Indeed, for many algal species ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO), the enzyme that catalyzes the initial steps of carbon fixation, is less than half saturated at present day pCO 2 [17]. As a consequence, many algae, including cyanobacteria, appear to possess mechanisms for concentrating CO 2 [18]. Experiments examining the effects of elevated pCO 2 on natural phytoplankton communities have yielded enigmatic results; in several studies, rates of production have increased under elevated pCO 2 [9,14,19], although in other cases, no significant changes in rates of production have been observed [20,21]. There is also compelling evidence that the decline in carbonate ion concentrations that accompanies decreases in seawater pH can be detrimental to the growth of calcifying microorganisms [7], although this appears species-specific [22] and may not apply in the tropical oceans [23]. Additionally, decreases in seawater pH could affect other aspects of seawater chemistry, including altering availability of nitrogen substrates (i.e. decreasing ammonia relative to ammonium) and iron, with concomitant impacts on key processes in the marine nitrogen cycle [13,24].
Numerous laboratory-based pCO 2 manipulation studies have examined the response of specific organisms to changes in seawater carbonate chemistry, where growth conditions are controlled and organisms are generally examined in isolation [12,25]. Other studies have examined natural planktonic communities, including work in the oligotrophic open ocean [8,9,20,21]. Recent mesocosm experiments conducted in nearshore waters found that elevated pCO 2 shifted the partitioning of carbon fixed via photosynthesis from the particulate to the dissolved phase [26] and increased bacterial growth [10]. These intriguing results highlight the need for experiments examining the effects of increased pCO 2 on plankton growth in the open ocean, where a major fraction of global productivity occurs [27] with a significant fraction of the production being partitioned into the dissolved phase which supports the microbial food web [28]. Increasing seawater pCO 2 could impact heterotrophic bacterial growth, through both direct changes to metabolic rates, for example alteration of enzymatic activities [29][30][31], or indirectly through changes in organic matter production or substrate lability [10,32]. Even small changes in rates of bacterial consumption of organic matter could have a large impact on carbon and nutrient cycling.
In this study, we conducted abrupt perturbations to the seawater carbonate system in the waters of the oligotrophic North Pacific Subtropical Gyre (NPSG) to artificially alter seawater pCO 2 to conditions projected for the surface ocean within the next 50 to 100 years [1]. During these experiments, we examined how such abrupt changes in pCO 2 influenced rates of primary and bacterial production in the near-surface ocean during a series of incubation experiments (2-5 days in duration). We also conducted a series of short-term (~12 hours; photoperiod) depth-resolved experiments to evaluate possible influences of increased pCO 2 on microbial production throughout the euphotic zone (0-125 m). Experiments were conducted in the open ocean of the NPSG, one of the largest biomes on the planet, and hence in an ecosystem that plays a major role in the global cycles of bioelements.

Experimental design
All seawater carbonate system manipulation experiments were performed on Hawaii Ocean Time-series (HOT) program cruises to Station ALOHA (22˚45' N 158˚W), the field site of the HOT program (between June 2010 and September 2012) or during two process cruises conducted in the vicinity of Station ALOHA (August 2010 and March 2011; Fig 1). Two types of carbonate system manipulation experiments were performed. The first kind of experiments (hereafter "bubbling") were performed as described in Böttjer et al. [21]. Briefly, near-surface (5-25 m) ocean seawater was collected near midnight using polyvinyl chloride sampling bottles attached to a conductivity-temperature-density (CTD) rosette, and subsampled from CTD rosette bottles under minimal light into acid-washed 20 L polycarbonate carboys fitted with sterile caps with ports for introducing and venting gases. After filling, carboys were placed into shaded (~50% surface irradiance) surface seawater-cooled incubators. Targeted pCO 2 levels were attained by bubbling control carboys with air (~387 μatm pCO 2 ) and treatment carboys with a mixture of air and CO 2 (targeting~750 or~1100 μatm pCO 2 ) for 6-8 hours (< 3 L min -1 ). Mixing and delivery of air or mixed air and CO 2 was regulated by use of mass flow controllers. During the initial 6-8 hours of bubbling, subsamples were collected regularly for measurements of seawater pH (see below for methods), and together with measurements of total alkalinity (TA) were used to estimate seawater pCO 2 using the 'seacarb' package [33] in the R statistical environment, with default settings for the carbonate dissociation constants [34][35][36][37]. Once the target pCO 2 was reached, the rate of bubbling was reduced ( 1.5 L min -1 ) for the duration of the experiment. Subsequent sampling was conducted before dawn at each time point, with the initial sample taken post equilibration considered the beginning of the experiment. Sampling was performed by applying positive pressure to the carboys and subsampling for measurements of TA, dissolved inorganic carbon (DIC), chlorophyll a, 14 C-based primary productivity ( 14 C-PP), and rates of 3 H-leucine ( 3 H-Leu) incorporation (as a proxy for bacterial production).
Additional CO 2 perturbation experiments were conducted at either ambient (~387 μatm) or elevated (~750 μatm) seawater pCO 2 to evaluate depth-dependent responses in 14 C-PP and 3 H-Leu incorporation to perturbation of the seawater carbonate system. For these experiments, samples were incubated in situ to simulate the vertical gradients in light and temperature representative of the depths from which samples were originally collected and rates of 14 C-PP and 3 H-Leu incorporation were measured. Seawater was collected before dawn from six euphotic zone depths (5,25,45,75,100, and 125 m) and subsampled from the CTD rosette bottles under minimal light into acid-washed 20 L polycarbonate carboys. These carboys were left untreated (controls) or amended with trace metal grade hydrochloric acid (43 mL of 0.1 N HCl) to increase pCO 2 to~750 μatm (elevated pCO 2 treatments) while minimizing potential changes to TA through additions of sodium bicarbonate (4 mmol). Once the carboys had been amended, seawater from each depth was subsampled into triplicate acid-cleaned 500 mL polycarbonate bottles and acid-cleaned 40 mL polycarbonate centrifuge tubes for subsequent measurements of 14 C-PP and 3 H-Leu incorporation, respectively. Following addition of radioactive substrates (see below), these bottles and tubes were affixed to a free-drifting array and incubated in situ at the depths of sample collection for the duration of the photoperiod (dawn to dusk).

Measurements of TA, DIC, and pH
Seawater samples for DIC and TA were collected from each carboy at every time point to evaluate the stability of the carbonate system during bubbling. Samples for determination of carbon system components (TA, DIC, and pH) were collected and analyzed following HOT program protocols [5,39]. DIC and TA samples were collected from carboys into precombusted 300 mL borosilicate bottles. Care was taken to avoid introducing bubbles into the sample during filling, and bottles were allowed to overflow three times during filling. Once filled, each sample was immediately fixed with 100 μL of a saturated solution of mercuric chloride; bottles were capped with a grease seal, and stored in the dark for later analysis. DIC concentrations were determined coulometrically using a Versatile INstrument for the Determination of Total inorganic carbon and Titration Alkalinity 3S (VINDTA) system [40]. TA was determined using an automated, closed-cell potentiometric titration. The precision and accuracy of these measurements were validated by comparison to a certified seawater CO 2 reference sample [40], with accuracies of approximately ±3 μmol L -1 for TA and ±1 μmol L -1 for DIC. Seawater pH (measured at 25˚C) was analyzed using spectrophotometric detection of m-cresol purple with a precision of 0.001 [5,41].

Measurements of 14 C-PP
Rates of 14 C-PP were measured at each sampling time point during the bubbling experiments. At each pre-dawn sampling, seawater was subsampled from the carboys into acid cleaned 500 mL polycarbonate bottles, and each bottle was amended with~1.85 MBq 14 C-bicarbonate. The total radioactivity added to each sample bottle was determined post-incubation by subsampling 250 μL aliquots of seawater into scintillation vials containing 500 μL of β-phenylethylamine. Bottles were placed in shaded (~50% irradiance) surface seawater-cooled incubators for the duration of the photoperiod. After sunset, 100 mL from each sample bottle was filtered at low vacuum (<50 mm Hg) onto 25 mm diameter, 0.2 μm porosity polycarbonate membrane filters. The filters were then stored frozen in 20 mL scintillation vials until analysis at the shore-based laboratory. At the shore-based laboratory, filters were acidified by the addition of 1 mL of 2 N hydrochloric acid, and allowed to passively vent for at least 24 hours in a fume hood to remove all inorganic 14 C, followed by addition of 10 mL Ultima Gold LLT liquid scintillation cocktail. The resulting radioactivity was determined on a Perkin Elmer 2600 liquid scintillation counter.
Measurements of 14 C-PP from the depth-dependent experiments were conducted similarly, except that the samples were incubated in situ at the respective depths of collection on a freedrifting array for the duration of the photoperiod, and the incubations were terminated via sequential size fractionated filtration onto 10 μm, 2 μm, and 0.2 μm pore sized polycarbonate . This image was derived from MODIS Aqua data using the color index (CI) algorithm [38] with no flags applied. The locations of the three stations occupied in August 2010 are also indicated; Station ALOHA is depicted as a circle, while stations S1 and S2 are depicted by triangles. Depth profiles of chlorophyll a during the depth-dependent experiments are also shown (panel B). https://doi.org/10.1371/journal.pone.0193405.g001 Effects of elevated pCO 2 on plankton productivity in the NPSG PLOS ONE | https://doi.org/10.1371/journal.pone.0193405 April 25, 2018 filters (25 mm diameter). Filters were treated as previously described for subsequent determination of 14 C activity.

H-leucine incorporation measurements
We measured 3 H-Leu incorporation into plankton protein as a proxy measurement for bacterial production [42,43]. Rates of 3 H-Leu incorporation were measured following incubations conducted in both the light ( 3 H-Leu Light ) and in the dark (through use of black cloth bags; 3 H-Leu Dark [44]). From the bubbling experiments, 125 mL polyethylene amber bottles were subsampled from each carboy in the pre-dawn hours; 6 aliquots of 1.5 mL were then subsampled from each bottle into 2 mL microcentrifuge tubes (Axygen) containing 20 nmol L -1 3 Hleucine (final concentration). In addition, 1.5 mL of seawater was subsampled into a 2 mL microcentrifuge tube containing 20 nmol L -1 3 H-leucine (final concentration) and 100 μL of 100% (w/v) trichloroacetic acid (TCA); these samples served as time zero "blanks". Samples were incubated for 2 to 12 hours in the same surface seawater-cooled incubator described previously. To terminate incubations, 100 μL of 100% TCA was added to each microcentrifuge tube and tubes were frozen (-20˚C) for later processing.
For those experiments incubated in situ on the free-drifting array, water was subsampled from each of the control and treatment carboys into 40 mL polycarbonate centrifuge tubes and each tube was inoculated with 3 H-leucine to a final concentration of 20 nmol L -1 . Time zero blanks were immediately subsampled from each tube; for these samples, 1.5 mL of seawater was aliquoted into 2 mL microcentrifuge tubes containing 100 μL of 100% TCA. The 40 mL tubes were then affixed to the same free drifting array utilized for the 14 C-bicarbonate assimilation measurements and samples were incubated under ambient light and in the dark (by placing the tubes in a darkened cloth bag). The array was then deployed for the duration of the photoperiod. After sunset, the array was recovered, and triplicate 1.5 mL subsamples were removed from each of the polycarbonate tubes and aliquoted into 2 mL microcentrifuge tubes containing 100 μL of 100% TCA. The microcentrifuge tubes were frozen (-20˚C) for later processing, following the procedures described in Smith and Azam [45].

Contextual biogeochemical data, statistics, and data analysis
Seawater samples for contextual biogeochemical analyses were collected and analyzed according to HOT program protocols (http://hahana.soest.hawaii.edu/hot/methods/results.html). Measurements of fluorometric chlorophyll a concentrations were performed as in Letelier et al. [46]. Samples for analysis of nutrient concentrations were subsampled from the CTD rosette bottles into acid washed polyethylene bottles (125 or 500 mL) and stored upright at -20˚C. Combined concentrations of nitrate and nitrite (N+N) were analyzed using the high sensitivity chemiluminescent technique [47,48], while concentrations of soluble reactive phosphorus (SRP) were determined via the magnesium-induced co-precipitation (MAGIC) method [49].
Statistical analyses were performed using Matlab (Mathworks). Data that were not normally distributed were log 10 transformed prior to subsequent analyses. Statistical differences between rates of 14 C-PP and 3 H-Leu incorporation at different pCO 2 levels during our bubbling experiments were determined by two-way analysis of variance (ANOVA), where pCO 2 and time were the factors of variation. For the depth-resolved experiments, we assessed significance for individual depths and for the depth-integrated rates based on the mean and standard deviation of the rates of 14 C-PP and 3 H-Leu incorporation in the treatments and controls using Student's t-tests.

Results
We conducted a total of 10 shipboard pCO 2 manipulation experiments where seawater pCO 2 was altered by bubbling with CO 2 -air gas mixtures. We measured rates of 14 C-PP in all 10 of these bubbling experiments, and rates of 3 H-Leu incorporation were measured in 8 of these experiments. An additional 5 experiments were conducted to evaluate depth-dependent responses in rates of 14 C-PP and 3 H-Leu incorporation to elevated pCO 2 . Experiments were conducted in all four seasons, spanning the range of conditions typically observed at Station ALOHA (Table 1). With two exceptions, all experiments were conducted with seawater collected at Station ALOHA; two of the depth-resolved experiments (August 26, 2010 and August 28, 2010) were conducted at sampling sites to the northwest of Station ALOHA (termed S1: 24˚45' N 160˚45' W and S2: 25˚35' N 160˚32' W) where concentrations of chlorophyll a in near-surface waters were elevated relative to Station ALOHA (Table 1; Fig 1).
For all bubbling experiments, initial concentrations of N+N and SRP were consistently below 10 nmol L -1 and 150 nmol L -1 , respectively, consistent with HOT program measurements of these nutrients (Table 1; Fig 2). Rates of particulate 14 C-PP at the beginning of the experiments ranged between 0.17 to 0.93 μmol C L -1 d -1 , with the higher rates measured at those stations to the northwest of ALOHA where chlorophyll a concentrations were elevated (Table 1). Seawater pCO 2 in the near-surface waters at the time the experiments were conducted ranged between 351 μatm to 419 μatm, consistent with HOT program observations at Station ALOHA [5,21], while sea surface temperatures ranged between~24 and 26˚C ( Table 1).

Concentrations of chlorophyll a and rates of 14 C-PP and 3 H-Leu incorporation under elevated pCO 2
In 8 out of 10 bubbling experiments rates of 14 C-PP in the elevated pCO 2 treatments were not significantly different than rates measured in the controls (two-way ANOVA; p>0.05; Table 2; Fig 3). In the remaining 2 bubbling experiments (April 2011 and September 2012) rates of 14 C-PP in the controls were significantly greater than rates measured in the enhanced pCO 2 treatments (two-way ANOVA, p<0.05; Table 2); notably, concentrations of N+N and SRP in both of these experiments were elevated relative to other experiments (Table 1). There were no  In addition, during most experiments (7 of 10) concentrations of chlorophyll a in the elevated pCO 2 treatments were not significantly different than in the controls (two-way ANOVA; p>0.05; Table 2). In a single experiment (September 2010), chlorophyll a concentrations in the elevated pCO 2 treatments were greater than those in the controls. In contrast, rates of 14 C-PP and concentrations of chlorophyll a were greater in controls relative to the elevated pCO 2 treatments in two of the experiments (April 2011 and September 2012; two-way ANOVA; p<0.05; Table 2; Fig 3). We also normalized our measured rates of 14 C-PP to concentrations of chlorophyll a, and in 9 of 10 experiments there was no significant difference between controls and elevated pCO 2 treatments (two-way ANOVA; p>0.05; S1 Fig). In the experiment conducted in March 2011, chlorophyll a normalized rates of 14 C-PP were greater in elevated pCO 2 treatments than in controls (two-way ANOVA; p<0.05; S1 Fig).
We also examined possible responses in rates of 3 H-Leu incorporation during the seawater carbonate system manipulation experiments (Table 2). In total, rates of 3 H-Leu Dark incorporation were determined in 7 of the bubbling experiments, with coincident measurements of rates of 3 H-Leu Light incorporation in 6 of these 7 experiments (Table 2). In the enhanced pCO 2 treatments rates of 3 H-Leu Dark incorporation were similar to those measured in the controls, ranging between 7 and 41 pmol Leu L -1 h -1 , with measurements at subsequent time points ranging from 4 to 98 pmol Leu L -1 h -1 (Fig 3). Rates of 3 H-Leu Light incorporation in the controls ranged between 9 and 61 pmol Leu L -1 h -1 at the beginning of the experiments, and between 21 and 84 pmol Leu L -1 h -1 at subsequent time points. In the enhanced pCO 2 treatments, rates of 3 H-Leu Light incorporation at the beginning of the experiments ranged from 15 to 65 pmol Leu L -1 h -1 , and from 17 to 99 pmol Leu L -1 h -1 at subsequent time points. In 5 out of the 7 experiments where rates were measured, 3 H-Leu Dark incorporation rates increased significantly over time in both the controls and treatments (two-way ANOVA, p<0.05; Fig 3). Similarly, in 4 of the 6 experiments in which 3 H-Leu Light incorporation was measured rates increased significantly over the duration of the experiment in both the controls and treatments (two-way ANOVA, p<0.05). However, in the majority of experiments there were no significant differences in the enhanced pCO 2 treatments relative to the controls (twoway ANOVA, p>0.05; Table 2). In a single experiment (August 2010) rates of 3 H-Leu Dark incorporation in the pCO 2 treatments (750 μatm) were significantly greater than the controls (two-way ANOVA, p<0.05; Table 2; Fig 3). The median values of the percent differences ([CO 2 treatments-ambient controls]/controls) in rates of 3 H-Leu Dark and 3 H-Leu Light incorporation across all time points were 2% (mean 19%, standard deviation 78%) and 1% (mean 22%, standard deviation 87%) respectively ( Table 2). The resulting differences (%) were significantly different from zero in less than 20% of experimental time points (13% and 15% for 3

Depth-dependent responses in 14 C-PP and 3 H-Leu incorporation to elevated pCO 2
In addition to conducting pCO 2 perturbation experiments where near-surface ocean water was bubbled continuously for up to 5 days, we also conducted 5 experiments where we examined short-term, daytime (dawn to dusk), depth-dependent responses in rates of 14 C-PP and 3 H-Leu incorporation to perturbations in seawater pCO 2 . For these experiments, seawater pCO 2 at 6 discrete depths in the upper ocean was perturbed through the addition of acid (and bicarbonate to maintain constant alkalinity) and incubated in situ on a free-drifting array. For samples in the upper euphotic zone (<45 m), the pCO 2 derived from measurements of DIC and TA was withiñ 20% of the target pCO 2 (750 μatm), while in the lower euphotic zone (>75 m) the derived pCO 2 values were uniformly greater (by 2-52%) than the target pCO 2 (S2 Fig). This was likely due to a combination of depth-dependent natural increases in pCO 2 and the greater variability of the seawater carbonate system in the lower euphotic zone compared to surface waters [50].
Rates of 14 C-PP were measured from size fractionated water samples (>10 μm, 2-10 μm, and 0.2-2 μm) from all six depths from both the controls and pCO 2 -perturbed treatments ( Table 3; Fig 4). Overall, rates of 14 C-PP in all of the size fractions were greatest at stations S1 and S2, where concentrations of chlorophyll a were also elevated (Fig 1). Rates of 14 C-PP in the >10 μm size fraction at these two stations ranged from 0.4 to 0.5 μmol C L -1 d -1 , approximately an order of magnitude greater than rates observed at ALOHA (Fig 4). The resulting depth-integrated upper euphotic zone (0-45 m) rates of 14 C-PP in the >10 μm size fraction ranged between 1.0 and 17.7 mmol C m -2 d -1 , with average rates at S1 and S2 (August 26 and 28, 2010)~11-fold greater than at ALOHA (Table 3). Similarly, rates measured at S1 and S2 were elevated in the 2-10 μm and 0.2-2 μm size fractions, with depth-integrated (0-45 m) rates at these stations ranging from 3.1 to 6.8 and 8.8 to 13.2 mmol C m -2 d -1 , respectively (Table 3) compared to 1.6 to 3.2 mmol C m -2 d -1 and 3.8 to 6.0 mmol C m -2 d -1 , respectively at Station ALOHA. In the lower euphotic zone (75-125 m), rates of 14 C-PP in the two larger size fractions at Station ALOHA were 2-to 5-fold lower than in the upper euphotic zone, with rates in the 0.2-2 μm size fraction in the lower euphotic zone as much as 2.5-fold lower than the upper ocean (Table 3).
Consistent with results from the bubbling experiments, overall abrupt increases in pCO 2 had little or no effect on rates of 14 C-PP in these depth-resolved experiments. For at least one of the depths examined in 3 of the depth-resolved experiments, rates of 14 C-PP in the controls were greater than the pCO 2 -elevated treatments (t-Test; p<0.05; Fig 5). However, in 1 of the 5 experiments (occurring in August 2010) rates of 14 C-PP in the pCO 2 -elevated treatments were greater than in the controls for both the >10 μm and 2-10 μm size fractions at a single depth (25 m; t-Test; p<0.005 and p<0.05, respectively; Fig 5). There were no consistent differences in depth-integrated rates of 14 C-PP in the upper euphotic zone between controls and elevated pCO 2 treatments (0-45 m; Table 3; t-Test; p>0.05). In 4 of the 5 depth-resolved experiments there were no consistent differences in rates of 14 C-PP between controls and pCO 2 elevated treatments in the lower euphotic zone. In one of the experiments (August 2010) conducted at Station ALOHA, rates of 14 C-PP in the lower euphotic zone (75-125 m) were significantly greater in the >10 μm and 2-10 μm size fractions in the pCO 2 perturbed treatments relative to the controls (t-Test; p<0.05; Table 3).
Similar to rates of 14 C-PP, rates of 3 H-Leu Dark and 3 H-Leu Light incorporation were greater at stations S1 and S2 than at ALOHA (one-way ANOVA; p<0.0001 for both; Fig 6). The resulting depth-integrated (0-125 m) rates of 3 H-Leu Dark and 3 H-Leu Light incorporation were significantly greater in controls than in elevated pCO 2 treatments in 2 of 5 and 5 of 5 depthresolved experiments, respectively (t-Test; p<0.05; Table 4). In the upper euphotic zone (0-45 Table 3. Depth-integrated rates of 14   Depth-integrated rates of 14  Effects of elevated pCO 2 on plankton productivity in the NPSG m), rates of 3 H-Leu Dark and 3 H-Leu Light incorporation were significantly lower in the elevated pCO 2 treatments than in the controls in 2 of 5, and 3 of 5 experiments, respectively (t-Test; p<0.05; Table 4). In 4 of the 5 experiments, rates of 3 H-Leu Light incorporation in the lower euphotic zone (75-125 m) were significantly lower in the elevated pCO 2 treatments than the controls (t-Test; p<0.05; Table 4).

Discussion
The overarching goals of this study were to examine whether abrupt changes to the ocean carbonate system would impact organic matter productivity and bacterial growth in the NPSG. Our experiments were not designed to investigate adaptations at the gene, species, or community level. To address these objectives, two types of experiments were conducted: 1) Manipulation of the near-surface (5-25 m) seawater carbonate system by gentle bubbling with air or a mixture of air and CO 2 and subsequent daily measurements of 14 C-PP and 3 H-Leu incorporation over 2 to 5 day incubation periods; and 2) Perturbation of the seawater carbonate Effects of elevated pCO 2 on plankton productivity in the NPSG system through the addition of acid (and bicarbonate to keep TA unchanged) at different depths throughout the euphotic zone, examining subsequent depth-dependent responses in rates of 14 C-PP and 3 H-Leu incorporation during in situ incubations lasting over the course of a photoperiod (~12 hours). We detected no consistent changes in rates of either 14 C-PP or 3 H-Leu incorporation in response to elevated pCO 2 over the course of our bubbling experiments. This lack of a consistent effect of enhanced pCO 2 on either 14 C-PP or 3 H-Leu incorporation suggests that the contemporary microbial assemblages in this region of the NPSG appear relatively resilient to rapid increases in seawater pCO 2 . Such observations are in agreement with results from other studies conducted in oligotrophic ocean ecosystems [20,21]. However, several studies have reported small to moderate increases in rates of 14 C-PP [7,51] and bacterial production [31] under elevated pCO 2 in more eutrophic nearshore ecosystems.
Similar to the lack of response to increased pCO 2 observed in the bubbling experiments, rates of production in the depth-resolved experiments also demonstrated no significant or consistent response to increases in seawater pCO 2 . Intriguingly, in one of our experiments (conducted in August 2010), rates of 14 C-PP by larger phytoplankton (>2 μm) in the lower euphotic zone demonstrated greater rates of production under elevated pCO 2 . These dimly lit waters and larger phytoplankton size classes account for a relatively small fraction of the euphotic zone productivity in the NPSG [52], so the resulting stimulation by pCO 2 resulted in no significant change in the depth-integrated (0-125 m) productivity from this experiment. Although the changes were modest (~20% and~50%, for 2-10 μm and >10 μm, respectively), the apparent stimulation of productivity by larger phytoplankton may reflect ecological adaptations of phytoplankton in these dimly lit waters. The lower euphotic zone of the NPSG is dynamic with respect to changes in the seawater carbonate system, as a result of the large vertical gradient in DIC concentrations [5] together with the greater influence of mesoscale variability in the lower euphotic zone compared to surface waters [50]. Hence, the observed response to rapid perturbation in pCO 2 could reflect an adaptive response by phytoplankton communities to abrupt changes in the seawater carbonate system in the lower euphotic zone. In addition, this observation could reflect carbon limitation of phytoplankton growing in the  lower euphotic zone during summer months. Net production of oxygen (O 2 ) in the sub-mixed layer waters of the NPSG results in accumulation of dissolved O 2 throughout the spring, with supersaturating concentrations through the summer and early fall [53]. We speculate that the enhanced rates of 14 C-PP during this single experiment may reflect alleviation of CO 2 limitation of the larger phytoplankton growing in waters with elevated O 2 : CO 2 ratios, where competitive binding of O 2 by RuBisCO could decrease photosynthetic efficiency and increase photorespiration [17,54].
We also sought to examine the sensitivity of bacterial production to abrupt increases in seawater pCO 2 during our depth-resolved experiments. While we observed no consistent response in rates of 3 H-Leu Light or 3 H-Leu Dark incorporation to the pCO 2 treatments during the bubbling experiments, rates of 3 H-Leu incorporation in our depth-resolved experiments were frequently sensitive to changes in pCO 2 . In all 5 of the depth-resolved experiments, euphotic zone (0-125 m) rates of 3 H-Leu Light incorporation were always significantly lower in the enhanced pCO 2 treatments than in controls. In contrast, rates of 3 H-Leu Dark incorporation did not vary in a consistent manner, with rates of 3 H-Leu Dark incorporation greater in the controls than the pCO 2 treatments in 2 of 5 experiments. We incubated samples in both the light and dark to evaluate how elevated pCO 2 might alter the known photostimulation of 3 H-Leu incorporation previously reported in the euphotic zone of the NPSG [44,55]. Based on flow cytometric sorting of picoplankton populations, Björkman et al. [56] determined that , and full euphotic zone (0-125 m). Significant differences between rates in controls and treatments indicated by letters (two-sample t-Test), no significant difference indicated by "ns".
Prochlorococcus incorporation of 3 H-Leu in the light was a major factor controlling this photostimulation. Given the large contribution of Prochlorococcus to rates of 14 C-PP (39% ±20%) and 3 H-Leu incorporation (20% in the dark and 60% in the light) at Station ALOHA [56,57], our results suggest Prochlorococcus growth may be relatively insensitive to, or perhaps negatively affected by, abrupt increases in pCO 2 . These results are consistent with previous findings in culture that suggest that while Synechococcus growth responds to elevated pCO 2 , Prochlorococcus growth appears largely insensitive to variations in pCO 2 [11]. Several previous studies have reported increased rates of 3 H-Leu Dark incorporation under elevated pCO 2 [10], suggesting a shift in the partitioning of primary production from the particulate to the dissolved pool [26,58,59], with subsequent increased growth by heterotrophic bacteria on this newly available DOM [31,60]. However, other studies that have specifically measured rates of dissolved organic carbon production under elevated pCO 2 have reported inconsistent responses [61,62]. In our experiments, based on both bubbling and depthresolved experiments, rates of 3 H-Leu Dark incorporation were most often unchanged under conditions of elevated pCO 2 . Similarly, in the near-surface waters (represented by our bubbling experiments) rates of 3 H-Leu Light incorporation were unaffected by increases in pCO 2 , but in the deeper regions of the euphotic zone (75-125 m), rates of 3 H-Leu Light incorporation were significantly lower under elevated pCO 2 treatments relative to the controls in 4 out of 5 experiments.
In general, we found that abrupt increases in pCO 2 have little or no influence on rates of 14 C-PP and 3 H-Leu Light or 3 H-Leu Dark incorporation at Station ALOHA. On a single occasion, we did observe apparent stimulation of 14 C-PP by larger phytoplankton dwelling in the lower euphotic zone, an observation we hypothesize could reflect seasonally-dependent carbon limitation by phytoplankton growing in these dimly lit waters. However, the majority of our experiments suggest that contemporary microbial growth in the euphotic zone at Station ALOHA is relatively resilient to abrupt increases in pCO 2 . Such results are somewhat surprising given the low temporal fluctuations in seawater pCO 2 this habitat experiences; however, we suspect that in this persistently oligotrophic environment, both rates of 14 C-PP and 3 H-Leu incorporation are strongly controlled by the availability of growth-limiting substrates, whether in the form of inorganic nutrients or in the form of labile dissolved organic carbon. In particular, decadalscale (2006-2016) rates of 14 C-PP (at 25 m) at Station ALOHA varied more than threefold (from 0.28 to 0.98 μmol C L -1 d -1 ), approximately equivalent to the largest variation in rates of 14 C-PP we measured between control and treatment rates during our experiments. Consequently, even large perturbations to the carbonate system appear to have only a weak influence on microbial growth in this ecosystem. Additionally, the short division times and large population sizes of open ocean phytoplankton may provide some capacity to adapt to or evolve in response to anthropogenic changes to the ocean carbonate system [63]. In contrast, it is likely that increasing ocean temperatures will exert a relatively stronger influence on microbial metabolism. Temperature effects could manifest directly, for example by changing microbial metabolic rates or growth efficiencies [64,65] or indirectly; through increased vertical stratification with concomitant reduction in nutrient supply and expansion of the oligotrophic gyres [66]. The combination of temperature-driven increases in respiration and decreased nutrient supply to the euphotic zone would likely decrease rates of net community production, with decreases in the amount of organic carbon available for upper trophic levels and export to the deep ocean. Hence, based on our observations together with those from previous reports, responses in planktonic metabolism to elevated pCO 2 appear variable and likely depend on the types of organisms present and the environmental conditions under which they grow [61,67,68]. It remains an open question whether our findings reflect physiological flexibility by the resident microbial community in acclimating to changes in the carbonate system, or whether the growth of these organisms is so tightly regulated by resource availability that any influence due to variations in the carbonate system are obscured by these other controlling factors. This question could be addressed by carrying out similar perturbation experiments that examine whether microbial growth responds to elevated pCO 2 coincident with alterations in the availability of growth-limiting nutrients.