Introduction
Ocean circulation plays a strong role in global heat and moisture transport
(Rahmstorf, 2002) and is controlled in part by differences in temperature and
salinity, known as thermohaline circulation. Therefore, knowing these
parameters is important to reconstruct ocean circulation in the geological
past, which leads to a more robust understanding of our climate system. A
number of valuable proxies exist to reconstruct sea surface temperature, for
example, δ18Oforam (Emiliani, 1955), Mg / Ca
(Elderfield and Ganssen, 2000), TEX86 (Schouten et al., 2002),
U37K′ (Brassell et al., 1986), and long-chain diol index (Rampen et al., 2012).
However, there are currently very few proxies for reconstructing sea surface
salinity (SSS).
Evaporation, precipitation, continental runoff, and ice melt cause changes in
seawater salinity, thereby influencing ocean circulation. The isotopic ratios
of oxygen (δ18O) and hydrogen (δD) of water are strongly
tied to these environmental parameters (Craig and Gordon, 1965). Increasing
evaporation causes both enrichment in heavy isotopes (Clark and Fritz, 1997)
and an increase in surface water salinity. The resulting water vapor has a
depleted isotopic signature (Clark and Fritz, 1997) and the longer the water
stays in vapor form, the more depleted the isotopic signature of the vapor
becomes as relatively enriched water precipitates first. Therefore, a
depleted isotopic signature is found for most precipitation-fed rivers and
lakes (i.e., meteoric waters). As these waters drain into the ocean and mix
with seawater, the SSS is lowered, as is the water isotope value. This leads
to a strong linear correlation between δ18Owater values
and salinity in ocean water and therefore the δ18Owater
is a suitable proxy for sea surface salinity. However, the slope of the
correlation varies in space (ocean region) and time (Duplessy et al., 1993;
Mashiotta et al., 1999), severely complicating reconstructions of ancient
δ18Owater–S relationships and thus paleosalinity
reconstructions. Therefore, constraining the correlation between
δ18Owater values and S currently poses a challenge in
attempts to extract reliable paleosalinity estimates from inferred
δ18Owater values.
Over the last decade, culture studies have shown that the hydrogen isotopic
ratios of long-chain alkenones (δ2HC37, from here on referred
to as δDC37), biomarkers of Haptophyte algae from the
order Isochrysidales (Volkman et al., 1980), correlate with the hydrogen
isotopic ratios of the water in which the algae grow
(δDH2O) (Englebrecht and Sachs, 2005; Paul, 2002),
which in turn is correlated with salinity (Craig and Gordon, 1965). In
addition to the observed relationship between δDC37
and δDH2O values, biological hydrogen isotope
fractionation has been shown to decrease with increasing salinity, thereby
amplifying the salinity-to-seawater δD relationship of alkenones
grown in culture (Schouten et al., 2006; Wolhowe et al., 2009; M'Boule et
al., 2014; Chivall et al., 2014; Sachs et al., 2016). Therefore,
δDC37 has been proposed as an appropriate proxy for
reconstructing SSS (Englebrecht and Sachs, 2005; Schouten et al., 2006). For
example, δDC37 values measured on alkenones extracted
from Mediterranean sapropel S5 show similar trends to δ18O measured
on planktonic foraminifera and suggest a salinity decrease of 6 in the
eastern Mediterranean at the onset of sapropel formation (van der Meer et
al., 2007). δDC37 values from Panama Basin sediments
show changes in amount of runoff from the San Juan River, aligning well with
instrumental data and even tracking glacial to interglacial changes in
salinity (Pahnke et al., 2007). Salinity changes in the Agulhas Current
system were also recorded by changes in δDC37 values
during glacial terminations I and II and from the last glacial maximum into
the Holocene, which align with δ18Oforam values from
the same region (Simon et al., 2015; Petrick et al., 2015; Kasper et al.,
2014). Leduc et al. (2013) show a divergence in estimates of SSS between
proxies derived from δDC37 and planktonic
foraminifera (δ18Osw and Ba / Ca ratios) across the
Holocene in the Gulf of Guinea, which are attributed to differences in the
isotopic ratios of rainfall over the time period.
Although a relationship between δDC37,
δDH2O, and fractionation with salinity has been
observed in culture and some paleostudies show promising results, this
relationship is not always found in nature. Häggi et al. (2015) did not
find a significant relationship between δDC37 values
and salinity in suspended particulate matter from the Amazon Plume. In the
Chesapeake Bay estuary (USA), δDC37 values in
sediments relate to δDC37 values from suspended
particulate matter filters and δDH2O values, but
fractionation does not show a relationship with salinity (Schwab and Sachs,
2011). Nelson and Sachs (2014) also tested the use of
δDC37 in North American lakes covering a salinity
range of 10–133. In these North American lakes, there is a relationship
between δDH2O and δDC37, but no
trend between fractionation and salinity (Nelson and Sachs, 2014). These
environmental datasets suggest that there are other factors affecting
hydrogen isotope fractionation, which complicate the use of
δDC37 as a salinity proxy. Indeed, culture studies
have indicated that hydrogen isotope fractionation can be influenced by a
number of parameters, i.e., growth rate (Schouten et al., 2006; Wolhowe et
al., 2009; Sachs and Kawka, 2015), growth phase (Chivall et al., 2014),
species composition (M'Boule et al., 2014; Chivall et al., 2014), and
irradiance (van der Meer et al., 2015). When the hydrogen isotope ratios of
both the C37:3 and C37:2 alkenones are integrated (van der Meer et
al., 2013), the effect of temperature on hydrogen isotope fractionation has
been shown to be negligible on the δDC37 SSS proxy,
eliminating one impeding factor (Schouten et al., 2006). Growth rate and
irradiance have also been proven to influence total carbon isotope
fractionation of alkenones used as a pCO2 proxy (Pagani, 2014, and
references therein). Both of these factors are related and seem to play a
significant role for isotopic fractionation of alkenones, and the effects
remain to be completely understood.
The effect of alkalinity on hydrogen isotope ratios and fractionation has not
yet been tested. The effect of alkalinity on hydrogen isotope fractionation
is unknown because some of the culture experiments (Schouten et al., 2006;
M'Boule et al., 2014; Chivall et al., 2014) investigating hydrogen isotopes
from alkenones created media of different salinities by evaporation, which
changed alkalinity together with salinity in the culture media. In the
natural environment, precipitation and evaporation do not only influence
salinity but also affect the total alkalinity (AT) of the
surface ocean. In fact, a strong positive linear correlation between
AT and salinity is observed in surface ocean waters (Millero et
al., 1998; Lee et al., 2006), and, on top of that, large coccolithophore
blooms can bring about a significant decline in surface water AT
(Anning et al., 1996). Alkalinity is essentially the ability of water to
neutralize acid, which is linked to the amount of H+. H+ is readily
exchanged between extracellular and intracellular water; therefore, the
amount of H+ could potentially effect the hydrogen isotope composition
of intracellular water, which is a source of hydrogen for synthesis of
organic compounds. It is, therefore, crucial to decouple the effects of
salinity and alkalinity and assess how each effect hydrogen isotope
fractionation independently.
Furthermore, culture work has shown light intensity to have a strong effect
on αC37 at light intensities below
200 µmol photons m-2 s-1, but not above (van der Meer et al., 2015). However,
some of the culture studies that reported a strong correlation between
hydrogen isotope fractionation and salinity were performed at relatively low
light intensities (Wolhowe et al., 2009; M'Boule et al., 2014; Chivall et
al., 2014). Since algal blooms occur under high-light conditions in surface
waters across the globe (Nanninga and Tyrrell, 1996; Holligan et al., 1993),
and hydrogen isotope fractionation is less variable at high-light conditions
(van der Meer et al., 2015), the effect of salinity on hydrogen isotope
fractionation at high light intensity needs to be studied to better
understand the potential effect of salinity on alkenones synthesized in
nature. Here we addressed these two issues by using batch cultures of the
haptophyte algae Emiliania huxleyi in experiments in which alkalinity
was varied independently of salinity and where salinity was varied under high-light conditions.
Materials and methods
Media and culture conditions
Two separate batch culture experiments were conducted: (1) to assess whether
alkalinity affects hydrogen isotope fractionation between alkenones and
growth water (alkalinity–salinity experiment) and (2) to examine if the
fractionation–salinity relationship seen in previous culture experiments
still holds under high-light conditions (high-light experiment). A no-longer-calcifying strain of E. huxleyi, CCMP 1516, was used in these
batch cultures. Because the effects of alkalinity on hydrogen isotope
fractionation were being assessed, a non-calcifying strain was chosen to
avoid significant changes to the alkalinity of the media caused by the
organisms, changes that have previously been shown to occur during large
blooms of E. huxleyi (Holligan et al., 1993).
Media for all experiments were made using filtered North Sea water with added
nutrients, trace metals, and vitamins following the method for the F/2 medium
(Guillard and Ryther, 1962). The medium was diluted with ultrapure water to a
salinity of approximately 25 and NaCl was added to achieve higher salinities.
Salinity was measured using a VWR CO310 portable conductivity, salinity, and
temperature instrument.
The alkalinity–salinity experiments consisted of batch cultures with a
salinity range of 26–42 and constant AT of 2.44 mM and batch
cultures of salinity 34 and AT values between 1.44 and 4.6 mM. For
batches in which alkalinity was changed, pH was kept constant (7.9 ± 0.07).
NaHCO3 and Na2CO3 were added to bring the medium to an
AT of 2.44 mM, an average value for open ocean waters, which
typically fall between 2.1 and 2.5 mM in the modern day ocean (Ilyina et al.,
2009; Takahashi et al., 1981). Concentrated HCl was added to reduce
alkalinity of the medium to 1.44 mM, and bubbling with air for 48 h allowed
for equilibration of CO2 with the atmosphere following the method of
Keul et al. (2013). To increase alkalinity of the medium, NaHCO3 and
Na2CO3 were added to achieve AT values of 3.3 and 4.6 mM,
respectively. Alkalinity was determined using titration with 0.1 M HCl and
calculated using Gran plots (Gran, 1952; Johansson et al., 1983; Hansson and
Jagner, 1973). Temperature was a constant 15 ∘C and light intensity
was consistently kept at 75 µmol photons m-2 s-1 using
cool white fluorescent light, with a light : dark cycle of 16:8 h.
The high light experiment was performed at five different salinities, from 25
to 35, under a light intensity of
600 µmol photons m-2 s-1 using cool white fluorescent
light, with a light:dark cycle of 16:8 h, a constant temperature of
18.5 ∘C, and a constant alkalinity. All batch culture experiments
were performed in triplicate. Cultures were transferred to a new medium five
times prior to starting the experiment to remove any possible memory effects
from the original stock culture and adapt the algae to the desired
experimental conditions. An Accuri C6 flow cytometer was used to count cell
concentrations daily to calculate growth rate over the length of the
experiment. Growth rate was calculated as the slope of the linear fit of the
natural logarithm of cell density (ln[cell density]) in the exponential part
of the growth curve. Cells were harvested during exponential growth when cell
abundance reached > 106 cells mL-1 to prevent
effects of shading or reduced nutrient content of the medium by the
haptophytes (10–12 days). Cultures (600 mL for the alkalinity–salinity
experiment and 150 mL for the high-light experiment) were filtered over
0.7 µm GF/F filters to collect organic material and the medium was
subsequently collected following filtration to determine δD of the
growth water.
Water isotope analysis
Hydrogen isotopic ratios of the medium (δDH2O) were
measured on water collected prior to the experiment and after the experiment
concluded. δDH2O was measured using elemental-analysis
thermal-conversion isotope-ratio-monitoring mass spectrometry (EA/TC/irMS)
(see Schouten et al., 2006). An amount of 1 µL of sample water was
injected at least 10 times during a single analytical run.
δDH2O values were corrected to an in-house North Sea
(5 ‰) standard, which was calibrated against VSMOW and VSLAP.
Growth parameters and hydrogen isotope ratios of alkenones from
alkalinity–salinity and high-light batch cultures of Emiliania huxleyi CCMP 1516.
Salinity
Temperature
Irradiance
Growth rate
δDH2O
SD
δDH2O
SD
δDC37
SD
αC37
Error
AT mM
AT mM
pH
pH
(∘C)
(µmol photons
(d-1)
(‰ vs.
(‰ vs.
(‰ vs. VSMOW)
initial
end
initial
end
m-2 s-1)
VSMOW) initial
VSMOW) end
Alkalinity and salinity experiment strain CCMP1516
26.0
15
75
0.82
-10.0
1.1
-8.8
1.8
-230.9
1.8
0.776
0.002
2.39
2.49
8
8.8
26.0
15
75
0.87
-10.0
1.1
-9.6
1.6
-231.1
1.4
0.776
0.001
2.39
2.46
8
8.7
26.1
15
75
0.85
-10.0
1.1
-8.5
2.0
-224.3
2.1
0.783
0.002
2.39
2.46
8
8.7
30.9
15
75
0.93
-10.8
1.7
-10.1
2.6
-209.1
0.7
0.799
0.001
2.38
2.42
7.9
7.8
31.1
15
75
0.93
-10.8
1.7
-10.7
0.9
-220.3
0.2
0.788
0.000
2.38
2.47
7.9
8.7
30.9
15
75
0.93
-10.8
1.7
-9.6
1.1
-221.6
1.8
0.786
0.002
2.38
2.46
7.9
8.6
36.5
15
75
0.86
-8.7
1.1
-8.9
1.8
-207.8
0.5
0.799
0.001
2.42
2.48
7.8
8.7
37.2
15
75
0.83
-8.7
1.1
-11.1
1.5
-209.3
1.6
0.799
0.002
2.42
2.49
7.8
8.8
36.4
15
75
0.86
-8.7
1.1
-8.9
1.4
-212.5
1.1
0.795
0.001
2.42
2.45
7.8
8.6
42.2
15
75
0.69
-7.4
1.6
-7.4
1.6
-183.1
2.9
0.823
0.003
2.38
2.46
7.9
8.6
42.2
15
75
0.65
-7.4
1.6
-8.6
0.9
-182.8
2.3
0.824
0.002
2.38
2.45
7.9
8.6
42.3
15
75
0.68
-7.4
1.6
-7.3
1.0
-184.4
0.2
0.822
0.000
2.38
2.45
7.9
8.7
35.4
15
75
0.7
3.6
1.2
3.5
1.6
-197.0
1.4
0.800
0.001
1.39
1.5
7.8
8
35.3
15
75
0.7
3.6
1.2
4.0
1.7
-201.8
0.2
0.795
0.000
1.39
1.45
7.8
8.5
35.0
15
75
0.69
3.6
1.2
5.6
2.1
-201.5
0.3
0.795
0.000
1.39
1.48
7.8
8.5
34.8
15
75
0.86
4.4
1.2
5.0
1.1
-198.7
2.2
0.798
0.002
2.32
2.42
7.9
8.6
34.6
15
75
0.87
4.4
1.2
3.8
1.0
-190.2
3.0
0.806
0.003
2.32
2.39
7.9
8.5
34.9
15
75
0.9
4.4
1.2
4.8
1.7
-194.5
2.1
0.802
0.002
2.32
2.39
7.9
8.6
34.7
15
75
0.82
2.3
1.1
3.8
1.2
-197.1
1.0
0.800
0.001
3.32
3.45
8
8.6
34.5
15
75
0.8
2.3
1.1
3.3
1.0
-199.8
0.7
0.798
0.001
3.32
3.52
8
8.8
34.5
15
75
0.83
2.3
1.1
5.1
1.1
-198.1
1.4
0.799
0.001
3.32
3.69
8
8.7
35.0
15
75
0.93
4.1
1.5
4.2
1.3
-191.6
0.7
0.805
0.001
4.58
4.56
7.9
8.5
34.8
15
75
0.89
4.1
1.5
1.7
1.5
-197.5
0.6
0.800
0.001
4.58
4.63
7.9
8.6
34.7
15
75
0.87
4.1
1.5
4.0
1.4
-197.0
0.8
0.800
0.001
4.58
4.63
7.9
8.5
High-light experiment strain CCMP1516
25.3
18.5
600
0.60
-9.9
1.0
-10.1
0.9
-214.3
3.4
0.794
0.003
25.3
18.5
600
0.62
-9.9
1.0
-9.8
1.0
-218.3
0.4
0.789
0.000
28.2
18.5
600
0.65
-9.3
1.2
-10.0
0.8
-212.1
1.3
0.796
0.000
28.2
18.5
600
0.62
-9.3
1.2
-9.2
1.0
-213.1
0.0
0.794
0.001
28.2
18.5
600
0.64
-9.3
1.2
-9.1
1.1
-213.2
0.6
0.794
0.001
30.4
18.5
600
0.63
-10.7
1.2
-9.9
0.7
-204.9
0.5
0.803
0.001
30.4
18.5
600
0.54
-10.7
1.2
-8.7
1.2
-205.6
0.8
0.801
0.001
30.4
18.5
600
0.63
-1.7
1.2
-9.7
1.0
-203.1
0.8
0.805
0.001
33.1
18.5
600
0.56
-9.0
1.5
-9.7
1.3
-199.6
0.7
0.808
0.003
33.1
18.5
600
0.59
-9.0
1.5
-8.3
0.9
-196.8
2.9
0.810
0.000
33.1
18.5
600
0.59
-9.0
1.5
-9.0
1.5
-201.1
0.0
0.806
0.000
35.9
18.5
600
0.60
-9.5
0.9
-9.9
0.7
-197.2
0.0
0.811
0.001
35.9
18.5
600
0.55
-9.5
0.9
-9.0
1.3
-195.1
0.7
0.812
0.002
35.9
18.5
600
0.56
-9.5
0.9
-8.4
1.3
-197.2
1.9
0.810
0.000
Linear regression equations for hydrogen isotope fractionation–salinity (αC37)
relationship for a compilation of culture experiments growing different strains of Emiliania huxleyi.
Reference
Strain
αC37–salinity relationship
R2
Number
of points
Schouten et al. (2006)
E. huxleyi PML B92/11
αC37=0.0033S+0.6928
0.74
11
M'Boule et al. (2014)
E. huxleyi CCMP 1516
αC37=0.0021S+0.7401
0.80
20
Sachs et al. (2016)
E. huxleyi CCMP 374
αC37=0.0015S+0.7770
0.88
9
Alkalinity and salinity
E. huxleyi CCMP 1516
αC37=0.0026S+0.7098
0.86
24
High light
E. huxleyi CCMP 1516
αC37=0.0020S+0.7408
0.92
14
Alkenone analysis
Following filtration, filters were freeze-dried and extracted ultrasonically
five times for 10 min each time using dichloromethane / methanol (2:1)
to obtain total lipid extracts (TLEs). TLEs were then separated into three
fractions over Al2O3 column using hexane / DCM 9:1 (v:v) to
elute the apolar fraction, hexane / DCM 1:1 (v:v) to elute the ketone
(alkenone) fraction, and DCM / MeOH 1:1 (v:v) to elute the polar
fraction. Ketone fractions were run on a gas chromatograph coupled to a flame
ionization detector (GC-FID) to determine alkenone concentrations prior to
running on GC/TC/irMS to measure compound-specific hydrogen isotope ratios
(δDC37). Both the GC-FID and the GC/TC/irMS were
equipped with an Agilent CP-Sil 5 column (25 m × 0.32 mm internal
diameter; film thickness = 0.4 µm). GC temperature programs
were the same as discussed in M'Boule et al. (2014). The H3+ factor
was measured daily on the GC/TC/irMS prior to running samples; values ranged
between 2.8 and 2.9 ppm mV-1 for the alkalinity–salinity experiments
and 5.4 and 5.5 ppm mV-1 for the high light experiments. A Mix B
standard (supplied by A. Schimmelmann, Indiana University) was run to assess
machine accuracy on a daily basis and samples were only run when standard
deviation and error of the Mix B standard were less than 5 ‰.
Samples were measured in duplicate and squalane was co-injected with each
analytical run to monitor quality of runs; the average value for squalane
co-injected with high-light experiment samples was -164.8 ‰ with a
standard deviation of 2.2 and -163.4 ‰ with a standard deviation
of 2.7 when co-injected with the alkalinity–salinity experiment samples. All
C37 alkenone peaks were integrated as a single peak and values are thus
reported as the combined values of the C37:2 and C37:3 alkenones
(van der Meer et al., 2013). The isotopic fractionation of alkenones compared
to media is expressed as αC37 and calculated using the
equation
αC37=δDC37+1000δDH2O+1000.
Statistics
Analysis of covariance (ANCOVA) was applied to test if a significant
difference exists between equations of the linear regression models
representative of the αC37–salinity relationship between
this study and previous culture studies of E. huxleyi. All
statistical analyses were run in R using the R stats package.
Discussion
Impact of alkalinity and light
In the alkalinity–salinity experiment, αC37 values changed
from 0.776 to 0.824 over a salinity range of 26 to 42 and constant
alkalinity, and they remained constant at 0.799 ± 0.003 at alkalinities
ranging from 1.4 to 4.6 (Fig. 2c). This shows that alkalinity, in contrast
to salinity, does not affect hydrogen isotope fractionation of non-calcifying
E. huxleyi. We note, however, that this experiment was performed
with a no-longer-calcifying strain of E. huxleyi, and results might
be different when haptophytes calcify since calcification may be
impacted by alkalinity, which in turn could have consequences for other
intracellular processes. At constant alkalinity over a range of salinity, we
see a 2.6 ‰ change in fractionation per salinity unit,
confirming that salinity does indeed have an effect on hydrogen isotope
fractionation between alkenones and growth water (Schouten et al., 2006;
M'Boule et al., 2014; Chivall et al., 2014; Sachs et al., 2016).
Alkenones synthesized by haptophytes growing at different salinities under
high light (600 µmol photos m-2 s-1) show a strong
correlation between αC37 values and salinity. This
unambiguously shows that there is also a strong correlation between salinity
and hydrogen isotopic fractionation in alkenones at high light intensities,
as encountered in the surface layers of the ocean. Haptophytes are not believed
to be photoinhibited and primarily bloom at light intensities above 500 µmol photons m-2 s-1 (Nanning and Tyrell, 1996),
falling under the range of expected ocean surface light levels, which can reach over 1600 µmol photons m-2 s-1
(Frouin and Murakami, 2007).
Furthermore, E. huxleyi has been shown to adapt to different light
conditions by expressing different genes under high- and low-light conditions
(Rokitta et al., 2012), showing that growth is possible under different light
regimes.
The slope of the α–salinity correlation, or fractionation response
per unit salinity, is statistically similar (p < 0.05) for both
the alkalinity–salinity and the high-light experiments, based on ANCOVA
between the linear regression models fit to each dataset. There is a weak
negative correlation of growth rate (μ) with fractionation for both the
alkalinity–salinity and high-light experiments,
αC37 = -0.0692μ+0.8557 (R2=0.25, n=24, p < 0.05) and αC37 = -0.1257μ+0.8776 (R2=0.35, n=14, p < 0.05), respectively
(Fig. 3), which aligns with findings of Sachs and Kawka (2015), who report a
negative correlation between growth rate and fractionation, albeit a stronger
relationship. Growth rate is also negatively correlated with salinity in both
experiments (Table 1, Fig. 3), which is consistent with earlier work of
Schouten et al. (2006). However, our results show a direct effect of salinity
on both growth rate and fractionation, suggesting the correlation between
growth rate and fractionation might be largely indirect.
Comparison with previous studies
We performed a statistical comparison using ANCOVA between the different
αC37–salinity relationships for previous E. huxleyi cultivation experiments (Schouten et al., 2006; M'Boule et al.,
2014; Sachs et al., 2016; Table 2) and our experiments. Sachs et al. (2016)
report δD values for individual alkenones; thus, we used a weighted
mean average of the δDC37:3 and
δDC37:2 values to compare with other results
reporting integrated δDC37 values. The slopes of the
αC37–salinity relationships are not statistically
different from each other (p > 0.05), with the exception of
three comparisons: Sachs et al. (2016) was statistically different
(p ≤ 0.05) from Schouten et al. (2006), M'Boule et al. (2014), and
the alkalinity–salinity experiment (Table 2). A possible explanation for the
statistical difference between the αC37–salinity
relationship of Sachs et al. (2016) and the other three experiments could be
due to the fact that the Sachs et al. (2016) experiment was conducted using
chemostats with a controlled growth rate, whereas, the other experiments were
batch culture experiments in which growth rate varied. Growth rate has been
shown to effect hydrogen isotope fractionation of alkenones (Schouten et al.,
2006; Wolhowe et al., 2009; Sachs and Kawka et al., 2015), and, therefore,
could account for the difference between reported fractionation responses to
salinity. Although the slopes are statistically similar
(p > 0.05), different strains used in the experiments, and
the individual experiments themselves (i.e., conducted by different labs
using different techniques) do likely play a large role in the observed
differences between slopes. Furthermore, the intercepts of the regression
models applied to the αC37–salinity relationships for
the E. huxleyi culture data are all significantly different
(p ≤ 0.05), i.e., the absolute fractionation differs between the
different studies, except for the relationship reported by M'Boule et
al. (2014) and our high-light experiment. These differences in intercept may
be explained by a number of potential factors. One explanation could be due
to the different strains of E. huxleyi used in the experiments, as
each strain would respond in a similar fashion to salinity changes, but
fractionate to a different extent, similar to differences seen between
species (Schouten et al., 2006; Chivall et al., 2014; M'Boule et al., 2014).
This could be due to differences in fractionation between intra- and
extracellular sources of hydrogen or differences in lipid synthesis rates.
Another explanation for part of the discrepancies in intercepts could be
analytical differences between laboratories, i.e., small offsets in measured
absolute values of C37 alkenones.
With the exception of the high-light experiment, all other culture
experiments with E. huxleyi being discussed here were grown at light
intensities between 50 and 300 µmol photons m-2 s-1
(Schouten et al., 2006; M'Boule et al., 2014; Sachs et al., 2016;
alkalinity–salinity experiment). The fact that the strong
αC37–salinity response is also identified in E. huxleyi grown at high light conditions is important for understanding the
influence of light and depth effects (i.e., van der Meer et al., 2015;
Wolhowe et al., 2015) on the αC37 of alkenones preserved
in the sedimentary record. Van der Meer et al. (2015) suggested that at light
intensities above 200 µmol photons m-2 s-1,
αC37 responds differently to changes in light intensity
than below, with a larger reported range in fractionation values at light
intensities below 200 µmol photons m-2 s-1. Wolhowe et
al. (2015) also show this trend in aC37 measured on alkenones
in suspended particulate matter from the Gulf of California and eastern
tropical North Pacific. Krumhardt et al. (2016) indicate that although
haptophyte-indicative pigments were high below surface water layers in the
subtropical North Atlantic, they were also abundant in the upper 30 m of the
water column, especially during spring. Based on these findings and the
U37K′ core-top calibration, we can be confident that alkenones
preserved in the sediments largely reflect surface water temperatures during
the time of the year that haptophytes are known to bloom (Müller et al.,
1998). Furthermore, haptophytes are thought primarily to bloom at light
intensities above 500 µmol photons m-2 s-1 in the
surface ocean (Nanninga and Tyrrell, 1996), leading to the conclusion that
the light and depth effects discussed previously (van der Meer et al., 2015;
Wolhowe et al., 2015) might not have such a large effect on the
αC37–salinity response as previously believed.
Potential mechanisms for salinity and light responses
As mentioned above, our results show that the response in hydrogen isotopic
fractionation of alkenones to salinity is statistically similar across
different E. huxleyi strains and under different growth conditions,
including low- and high-light conditions. How salinity affects hydrogen
isotope fractionation is still unknown, although several possible mechanisms
have been proposed (e.g., Maloney et al., 2016, and references therein). The
effect of salinity on hydrogen isotope fractionation seems to be a general
feature recorded in alkenones, fatty acids, sterols, phytene, and diploptene
produced by algal photoautotrophs (Heinzelmann et al., 2015; Schouten et al.,
2006; Sachse and Sachs, 2008; Sachs and Schwab, 2011; Nelson and Sachs,
2014). Nicotinamide adenine dinucleotide phosphate (NADPH) is associated with
large isotope fractionation values, larger than the fractionation between
extracellular and intracellular water, but both are used as sources of H for
synthesis of organic compounds (Schmidt et al., 2003). NADPH has been
proposed to supply around 50 % of the H eventually used for lipid
production in the bacterium Escherichia coli (Kazuki et al., 1980),
and it is presumed to be roughly the same for photosynthetic algae (Zhang et
al., 2009). The cell generates NADPH either photosynthetically or via the
oxidative portion of the pentose phosphate pathway (OPP pathway) (Schmidt et
al., 2003; Wamelink et al., 2008). NADPH derived via ferredoxin-NADP+
reductase (FNR) in photosystem 1 (photosynthetically derived) tends to be
depleted by ∼ 600 ‰ in D when compared to intracellular water
(Luo et al., 1991), whereas NADPH produced via the OPP pathway is also
depleted compared to intracellular water, but much less than
photosynthetically derived NADPH (Schmidt et al., 2003; Maloney et al.,
2016). Schmidt et al. (2003) suggested that a transfer of H from NADPH
generated as part of the OPP pathway causes depletion in D, which is further
enhanced during continued biosynthesis, meaning organic compounds containing
H largely derived from metabolically reduced NADPH are characterized by
depletion in D. However, this depletion is still less than what is observed
for photosynthetically derived NADPH. Up- or down-regulation of the OPP
pathway relative to other NADPH-generating pathways (FNR derived, for
instance) could, therefore, cause differences in the amount of D depletion of
organic compounds. Up-regulation of the pentose phosphate pathway observed in
the bacterium Vibrio sp. at high salinities led to an increase in
NADPH generated by the pathway for use in biosynthesis (Danevčič and
Stopar, 2011). Danevčič and Stopar (2011) also found that
intracellular production of L-proline, an osmoregulating amino acid,
increased. The advantage of up-regulating the OPP-derived NADPH would be tied
to this increase in L-proline, which helps continue growth and biosynthesis
at higher salinities in Vibrio sp. A similar mechanism could be
present in E. huxleyi, causing the metabolically reduced NADPH pool
to increase relative to other pools and possibly become a more important
source of NADPH for biosynthesis if the OPP pathway exists in the same
location as the site of alkenone synthesis. However, in photoautotrophic
organisms, the reduction of NADP+ to NADPH is also directly linked to
photosystem activity (FNR) and therefore light intensity (Allen, 2002).
Furthermore,
as previously mentioned, this initial reduction is characterized by a very
large fractionation step (Schmidt et al., 2003; Maloney et al., 2016; Luo et
al., 1991). Because of this, we would expect isotope ratios to change with
different light intensities. Indeed, van der Meer et al. (2015) showed this
to be the case for irradiances between 15 and
200 µmol photons m-2 s-1, but the effect of changing
light intensity on hydrogen isotope fractionation is much lower at light
intensities > 200 µmol photons m-2 s-1.
Furthermore, Sachs et al. (2017) showed a light effect on hydrogen isotope
fractionation of C14:0 fatty acid, but no effect of light was
observed on hydrogen fractionation of C16:0 and C16:1 fatty acids from the
diatom Thalassiosira pseudonana grown over a low light range from
6–47 µmol m-2 s-1. The lack of correlation with light
intensity for the longer fatty acids is explained by enzymatic reprocessing
that causes further hydrogen fractionation and overwrites the light effect
seen for the C14:0 fatty acid (Sachs et al., 2017). This effect could also
apply to alkenones since alkenone synthesis has been linked to fatty acid
biosynthesis (Volkman et al., 1980; Marlowe et al., 1984; Rontani et al.,
2006). However, at high-light conditions, where more photosynthetically
derived NADPH is expected to be available (e.g., our high-light experiment),
the same fractionation response to salinity is observed as at low-light
conditions (e.g., M'Boule et al., 2014; our alkalinity–salinity experiment),
where less photosynthetically derived NADPH is expected. This suggests that
light intensity does not directly have an effect on the predominance of
photosynthetically derived versus metabolically derived NADPH, or
enzymatically reprocessed NADPH used in biosynthesis, or that the
up-regulation of metabolically derived NADPH with increasing salinity exerts
a stronger control on hydrogen isotope fractionation than irradiance.
Another explanation for the observed significant correlation with salinity at
both high and low light intensity could be that the cell synthesizes
alkenones in a closed cell compartment, similar to the “coccolith
vesicle-reticular body” in which coccoliths are formed (Wilbur et al., 1963;
Sviben et al., 2016), where the amount of NADPH used for biosynthesis is
regulated and the fraction of NADPH derived from the OPP pathway into the closed
compartment increases with increasing salinity.
In addition to a higher abundance of NADPH generated by the OPP pathway at
higher salinities, cells could also produce more D-depleted compounds,
osmolytes for instance (Dickson et al., 1982; Sachs et al., 2016, and
references therein; Sachse et al., 2008; Danevčič and Stopar, 2011),
which would leave the intracellular NADPH pool more enriched and would
result in D enrichment of other biosynthetic products such as alkenones. The
production of DMSP, an osmolyte produced by marine microalgae, is not coupled
to light intensity (van Rijssel and Gieskes, 2002); therefore, osmolyte
production could be a major factor responsible for the salinity response
observed over a range of light intensities. An added complication could be
that cells excrete more isotopically depleted osmolytes at high salinities
than at low salinities (Demidchik et al., 2014), which could leave the
fraction of NADPH used for other organic compounds more isotopically enriched
at high salinities. These processes are however correlated with salinity if
NADPH plays a central role in hydrogen isotope fractionation, and the
reduction of NADP+ to NADPH is directly coupled to photosystem activity
and therefore light intensity. Furthermore, different slopes for
αC37–salinity are expected for cells grown at high-light and
low-light conditions, which is in contrast to what our results show.
Based on the compilation of E. huxleyi culture data, a significant
relationship between hydrogen isotope fractionation and salinity is observed.
However, we do not see a clear relationship between hydrogen isotope
fractionation and light intensity. Due to balancing between ATP and NADPH
production and consumption within the cell (Walker et al., 2014), NADPH
formation dominates at high light levels, whereas ATP synthesis dominates at
lower light levels (Beardall et al., 2003), leading to the idea that a larger
pool of photosynthetically derived NADPH inside the cell under high light
conditions would, in turn, cause differences in hydrogen isotope
fractionation during the synthesis of alkenones, but this is not what the
data show. Transhydrogenase enzyme activity can remove NADPH when in excess
by reducing NAD+ to NADH using NADPH (Kim and Gadd, 2008; Zhang et al.,
2009), which is associated with a large isotope fractionation effect of
between 800 and 3500 ‰ (Zhang et al., 2009), leaving a
relatively D-enriched pool of NADPH behind. At high-light conditions, an
excess of NADPH, and therefore increased transhydrogenase
activity, is expected, something also seen with increasing salinity (Danevčič and
Stopar, 2011). However, if transhydrogenase enzyme activity is responsible
for reducing the excess NADPH, we might expect to see a difference in isotope
values and a different fractionation response to salinity at high- and low-light conditions, which is not the case. There is, of course, the possibility
that the cell could use excess NADPH for other intracellular processes or
synthesis of compounds that are not being investigated or measured in our
experiments, which could be light intensity dependent as well, similar to
what was reported for different fatty acids by Sachs et al. (2017). This
would explain why having an abundance of NADPH at high light intensities does
not seem to affect hydrogen isotope fractionation of alkenones since this
abundance of photosynthetically derived NADPH is being used for other
processes or is counteracted by enzymatic activity (Sachs et al., 2017). A
better understanding of the hydrogen isotopic composition of different
relevant hydrogen pools and how the alkenone synthesis process works is
required for more accurate determinations of the mechanism responsible for
the strong salinity response to hydrogen isotope fractionation of alkenones.