1
2
3
4
5
6
7
8
9
10
11
12
13
Supplementary Information
-
Lineage key
Figures SI Figure 1 to SI Figure 6
Tables SI T1 to SI T3
Detailed methods
SI reference list
Lineage key:
Lineage key used throughout SI.
14
15
Figures
Oxygen evolution
Oxygen consumption
1.50
Fold change to trait value in selection environment before CO2 enrichment
2.0
1.25
1.5
1.00
1.0
0.75
0.5
0.50
before
after
Chlorophyll content
1.2
before
after
Size
1.4
1.1
1.2
1.0
1.0
0.9
1.50
before
after
Lipid content
before
after
Selection environment
1.25
SA
FA
1.00
0.75
0.50
before
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
after
Figure 1A: The plastic responses of SA and FA evolved lineages to elevated CO 2 levels are
reversible before (left) and after (right) short term exposure to elevated pCO 2. After 7-14
generations at 1000µatm CO2, lineages were placed back in their respective environment, and a subset
of phenotypic traits was measured after two transfers (7-14 generations). Importantly, growth rates
were not impaired when lineages from either selection regimes were grown at 430µatm after short term
CO2 exposure ( Schaum & Collins, 2014). Likewise, there were no significant differences between
traits before and after short-term exposure to elevated CO2 levels. (ANOVA stability x before/after
F1,301 =0.2542 , p= 0.6145). SA in purple, FA in blue. All values are displayed as fold changes
compared to SA evolved lineages or FA evolved lineages prior to short term exposure to elevated CO 2 ,
depending on the selection environment lineages were evolved in. N=3 per lineage. Number of lineages
= 16. Dotted line indicates fold change 1 compared to lineages prior to short-term exposure to CO2 for
SA and FA respectively (i.e. no change). Units that fold changes were calculated from are: Oxygen
evolution and consumption in fmol/cell*hour. Chlorophyll content in fg/cell. Size as diameter in µm.
Lipid content is for polar lipids calculated from Nile Red Fluorescence units and in fg/cell.
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
SI Figure 1 B: Phenotypes of correlated responses of SH and FH evolved lineages.
SH displayed as purple, FH as blue.
When SH and FH evolved lineages were assayed at 430µatm CO2, growth rates were greatly reduced
in SH evolved lineages, but not in FH evolved lineages ( Schaum & Collins, 2014). Of the SH lineages
that did grow (albeit very slowly), oxygen evolution was reduced (Tukey post-hoc on omnibus
ANOVA testing for differences between SH and FH, p<0.05), but oxygen consumption was not, which
may partially explain the impeded growth rates. Size and chlorophyll content did not differ
significantly between SH and FH lineages, and were also not significantly different from FA (for FH
evolved lineages) or SA (for SH evolved lineagse). Note that SH lineages assayed at 430 ppm were
growing so poorly that no C:N samples could be obtained. Nile Red Fluorescence was below the
detection limit in SH and FH evolved lineages when assayed at 430µatm CO 2. Units that fold changes
were calculated from are as in SI Figure 1A.
47
48
49
50
51
52
53
54
55
56
SI Figure 2: Size selective gating of chlorophyll positive events. Side Scatter (SSC – proxy for
granularity) and Forward Scatter (FSC – proxy for size) gating of chlorophyll a positive cells (orange)
and chlorophyll negative cells (red) on a FACS CANTO. The SSC/FSC gate was established using
calibration beads of different sizes. The same gates were used to distinguish bacteria from algae in our
samples. A SYBR green stain had been used beforehand in order to distinguish debris from DNApositive particles, as Ostreococcus, with cells as small as 1µm is often in the same size range as debris.
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
SI Figure 3: In SH and FH evolved lineages the phenotype characteristic for plastic response of
SA and FA evolved lineages is not maintained in the long-term, i.e. the plastic response gradually
reverses for many traits.
For t0, t100, and t400, we display fold changes of mean trait values in FH evolved lineages compared
to FA evolved lineages, and fold changes of mean trait values in SH evolved lineages compared to SA
evolved lineages – the dotted line indicates a fold change of 1, i.e. no change. Traits are colour coded
and represented by individual symbols (see legend).
After 75-100 generations of evolution in SH or FH, lineages begin to show deviations from the plastic
response (compare to SI Figure 1A, where the plastic response is reversed almost completely after 7-14
generations) with significant differences between SH and FH evolved lineages for individual traits
(ANOVA F5,138 = 5.3269 , p<0.01). At t400 (400 generations into the experiment), mean trait values
are different between FH and SH, and differ from the plastic response as discussed in the main text.
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
SI Figure 4: In the ambient or plastic response, there are lineage-specific differences between SA
and FA evolved lineages. Likewise, lineages vary in their phenotypic evolution to SH or FH.
(ANOVA (all traits) F15 ,138 = 2.51, p<0.05) )
Mean trait values per lineage are represented by individual symbols (see key).
Error bars omitted for clarity (but see SI Figure 6, where lineages have been pooled according to
sampling depth). Blue slopes are linear models fitted for each pair of lineages in each environment.
Differences between lineages from stable and fluctuating environments are depicted for the ambient
response (SA and FA evolved lineages assayed at 430µatm CO 2), the plastic response (SA and FA
evolved lineages assayed at 1000µatm CO2), and the evolved response (SH and FH evolved lineages
assayed at 1000µatm CO2). Differences between lineages across selection regimes are shown as
differences in direction or steepness of slope. For example, oxygen evolution rates are on average not
different between SH and FH evolved lineages (see main text for statistics). On the level of the
individual lineage, however, there are significant differences (graphic representation direction of
slope), with rcc809, rcc789, rcc675, rcc501, rcc1662, rcc1645 and rcc1558 having higher oxygen
evolution rates after evolution in FH than after evolution in SH (all post hoc p<0.05), while rcc422,
rcc410, rcc343, rcc1114 and rcc1108 have significantly lower oxygen evolution rates in FH than in SH
(all post hoc p <0.05). All lineages have lower C:N ratios after evolution in FH than after evolution in
SH, but the amount by which C:N is lower (graphical representation: steepness of slope) differs
between lineages.
98
99
100
101
102
103
104
105
106
107
108
109
110
111
SI Figure 5: Since the phenotypic changes of SH and FH are not well correlated with each other
for most traits, evolutionary responses in SH cannot be used to predict evolutionary responses in
FH. For lineage legend see beginning of SI (or main manuscript Figure 3). Lineages are displayed as
means for 3 biological replicates; error bars are omitted for clarity. The blue line is a linear model fitted
to the means. The r2 value is only significant for Nile Red Fluorescence, but this might be an artifact of
lineages displaying either no or high Nile Red fluorescence after evolution in SH or FH. The units that
fold changes were calculated from are as in SI Figure 1.
112
6
8
12
6
9
Lipid
Size
Chlorophyll content
C:N ratio
R:P ratio
Oxygen consumption
Oxygen evolution
12
0.6
2.0
10
4
8
1.5
Trait Value
0.4
Isolation depth
4
surface
6
10-40 metres
>100 metres
1.0
5
2
4
0.2
2
3
0.5
113
114
115
116
117
118
119
120
121
122
123
124
SH
FH
SH
FH
SH
FH
SH
FH
0
0.0
0
0
0.0
0
0
SH
FH
SH
FH
SH
FH
Selection environment
SI Figure 6: There are significant within and between selection regime variations in evolved trait
values depending on where lineages of Ostreococcus had been sampled from originally
(ANOVA (all traits) F 2,153 = 3.88, p <0.05)
Isolation depths are colour coded with red = sampled at the surface, green = sampled at 10-40 metres,
blue = sampled at >100 metres. Bars are for mean trait values ± 1 SD. Oxygen evolution and
consumption are in fmol/cell*hour. Chlorophyll content is in fg/cell. Size is the average diameter in
µm. Lipid content is for polar lipids calculated from Nile Red Fluorescence units and in fg/cell.
0.4
PC2
0.2
0.0
SH evolved
FH evolved
SA evolved
FA evolved
Large symbols:
isolated 0-60 metres
Small symbols:
isolated >60 metres
-0.2
-0.2
125
126
127
128
129
130
131
132
133
134
135
136
137
0.0
0.2
0.4
PC1
SI Figure 7: Colour-coded version of Figure 3. Both environmental stability (PC1) and CO 2
elevation (PC2) drive phenotypic evolution of lineages. Large symbols lineages sampled from above
60 meters depth, small symbols: lineages sampled from below 60 meters depth. Red: SH evolved
lineages, blue: FH evolved lineages, black: SA evolved lineages, green: FA evolved lineages. Each
lineage is represented by a unique symbol. Together, fluctuation and elevation of pCO2 explain about
90% of the variance in evolved phenotypes, and lineages (within CO 2 treatments) cluster largely
according to sampling depth.
138
139
140
141
142
143
144
145
146
TABLES
SI Table 1: Sampling locations of Ostreococcus lineages used in this experiment.
We obtained oth95, rcc809 and rccc501 from the Plymouth Marine Laboratory (UK). All other strains
were supplied from the Roscoff Culture Collection through an ASSEMBLE grant. More information
on the strains used is available at http://www.sbroscoff.fr/Phyto/RCC/index.php?option=com_dbquery&Itemid=34.
We present a more detailed table as part of the SI to ( Schaum et al., 2013), but are also making the
information on sampling stations and sampling depth available here for easier access.
147
Lineage
148
149
Latitude
Longitude
Sampling station
Depth
First
[m]
isolated
oth95
43º 24'
3º 36'
Thau Lagoon
0m
1996
rcc1107
43º 3'
2º 59'
Mediterranean
0m
2006
rcc1108
42º 29'
3º 8'
Mediterranean
0m
2006
rcc1114
42º 48'
3º 1'
Mediterranean
0m
2005
rcc1558
42º 48'
3º 1'
Mediterranean
0m
2005
rcc1645
48º 46'
3º 56'
North Sea
10m
2007
rcc1662
50º 12'
0º19'
English Channel
10m
1995
rcc434
30º8’
10º3’
Atlantic Ocean
40m
1999
rcc410
29º 28'
34º 55'
Red Sea
100m
2000
rcc422
48º 37'
3º 51'
English Channel
0m
2001
rcc501
48º 46'
3º 1'
Mediterranean
10m
1995
rcc675
54º 11'
7º 54'
North Sea
0m
2001
rcc747
42º 35'
8º 49'
Atlantic Ocean
10m
1995
rcc789
not known
not known
Spanish Coast
0m
2001
rcc809
not known
not known
West Mediterranean
105m
1995
rcc810
21º 2'
31º 8'
Atlantic Ocean
120m
1991
150
151
152
153
154
155
156
157
158
159
160
161
162
SI Table 2: Carbonate chemistry table
Note that we also present this data as part of the SI to (Schaum & Collins, 2014). However, for easier
access, we are also including it in this paper.
Carbonate Chemistry: pH was measured routinely for all samples; DIC was measured for a random at
each transfer, and for all samples at the beginning and the end of the selection experiment. Data for
each time point in each selection regime was calculated from a minimum of six samples. Total
Alkalinity (TA) was measured at the beginning, 100 generations into the experiment and at the end of
the selection experiment. All other values were calculated using seacarb within R for 18 ºC and salinity
32. pH was determined for each sample at the end of each transfer. As samples were dilute with
maximum densities not exceeding ~ 5*104 cells per ml, pH did not drift more than 0.05 units.
Treatment
Control
HCO3-
CO32-
DIC
TA
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
8.001
1990.48
154.88
2161.94
2369.52
426.43
8.030
1900.95
158.09
2073.84
2290.35
430
466.76
8.013
1999.06
159.72
2174.99
2389.33
4
430
478.44
8.003
2003.79
156.56
2176.96
2386.41
5
430
477.79
8.003
2001.57
156.42
2174.59
2383.95
6
430
431.15
8.021
1880.38
152.99
2048.35
2258.28
7
430
367.72
8.101
1929.99
188.97
2131.74
2391.65
8
430
462.38
8.016
1996.29
160.78
2173.14
2389.16
9
430
455.04
8.023
1993.73
162.96
2172.50
2391.85
10
430
339.35
8.130
1904.46
199.39
2115.64
2391.56
11
430
455.43
8.022
1993.62
162.80
2172.24
2391.36
12
430
373.96
8.029
1664.34
138.19
1815.52
2013.85
13
430
343.32
8.111
1843.84
184.73
2040.49
2298.58
14
430
475.92
8.000
1980.08
153.69
2150.30
2356.58
15
430
484.24
8.001
2020.15
157.22
2194.19
2403.88
16
430
343.91
8.101
1805.47
176.82
1994.24
2242.84
17
430
458.27
8.018
1987.19
160.75
2163.86
2380.26
18
430
464.11
8.002
1939.24
151.16
2106.52
2310.94
19
430
446.81
8.013
1913.59
152.89
2081.99
2290.19
Week
pCO2
aimed for
pCO2
obtained
[ppm]
[ppm]
1
430
477.22
2
430
3
pH
20
430
448.00
8.001
1867.78
145.27
2028.61
2227.57
21
430
466.14
8.003
1952.80
152.62
2121.61
2327.56
22
430
474.71
8.001
1980.46
154.14
2151.09
2358.02
23
430
479.70
8.001
2001.17
155.74
2173.58
2381.94
24
430
465.74
8.012
1992.99
159.10
2168.27
2381.97
25
430
460.06
8.001
1919.23
149.36
2084.57
2287.24
26
430
447.83
8.001
1867.88
145.34
2028.78
2227.84
27
430
471.33
8.002
1969.50
153.53
2139.40
2345.94
28
430
447.92
8.003
1876.06
146.59
2038.21
2238.76
29
430
456.82
8.003
1913.27
149.49
2078.63
2281.76
30
430
460.06
8.001
1919.22
149.36
2084.56
2287.22
31
430
459.08
8.003
1922.72
150.23
2088.89
2292.68
Average control
430.00
444.05
8.022
1933.27
158.06
2106.75
2321.58
Standard
deviation
0.00
42.81
0.04
74.78
13.08
77.72
80.96
pCO2
aimed for
pCO2
obtained
pH
HCO3-
CO32-
DIC
TA
[ppm]
[ppm]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
1
430
470.00
8.002
1972.40
153.64
2142.45
2349.03
2
600
653.00
7.871
2121.57
122.31
2267.73
2419.75
3
1000
1007.00
7.720
2199.34
89.61
2323.93
2417.83
4
1000
1022.00
7.715
2207.61
88.96
2332.07
2424.41
5
1000
1002.00
7.739
2129.94
90.58
2252.98
2351.94
6
1000
1000.00
7.728
2225.02
92.36
2352.12
2449.68
7
1000
1200.00
7.652
2240.13
78.01
2359.82
2430.17
8
1000
1100.00
7.692
2249.81
85.84
2373.87
2458.49
9
1000
1034.00
7.712
2216.02
88.60
2340.54
2431.82
10
1000
1012.00
7.728
2248.94
93.23
2377.33
2475.31
11
1000
1002.00
7.725
2214.04
91.26
2340.11
2436.26
Treatment
Elevated
Week
12
1000
1100.00
7.689
2236.51
84.83
2359.55
2442.96
13
1000
1000.00
7.744
2307.99
99.37
2442.10
2548.03
14
1000
1090.00
7.701
2051.63
79.98
2165.71
2249.31
15
1000
1108.00
7.679
2202.56
81.68
2322.73
2401.96
16
1000
1171.00
7.668
2265.83
81.79
2388.30
2464.57
17
1000
1131.00
7.698
2059.74
79.81
2173.99
2256.87
18
1000
1111.00
7.684
2230.14
83.51
2352.24
2433.52
19
1000
1201.00
7.646
2212.42
76.03
2330.17
2398.09
20
1000
980.00
7.730
2188.66
91.19
2313.89
2411.11
21
1000
1000.00
7.731
2238.60
93.49
2366.82
2465.73
22
1000
1200.00
7.650
2231.19
77.39
2350.26
2419.86
23
1000
1090.00
7.693
2236.22
85.59
2359.67
2444.49
24
1000
1120.00
7.690
2279.72
86.57
2405.19
2489.68
25
1000
1200.00
7.639
2175.83
73.60
2291.12
2356.14
26
1000
1015.00
7.725
2240.35
92.25
2367.86
2464.50
27
1000
1002.00
7.728
2229.56
92.55
2356.91
2454.60
28
1000
800.00
7.808
2139.88
106.78
2274.45
2400.57
29
1000
1100.00
7.690
2238.85
85.01
2362.07
2445.69
30
1000
1100.00
7.696
2052.95
79.14
2166.60
2248.58
31
1000
980.00
7.743
2011.07
86.32
2127.76
2224.90
32
1000
1000.00
7.724
2018.72
82.97
2133.52
2224.24
Average
elevated
969.6875
1031.28
7.720
2183.54
89.82
2308.56
2402.81
Standard
deviation
5.44
87.11
0.07
84.88
14.73
84.29
79.75
pCO2
aimed for
pCO2
obtained
pH
HCO3
CO3
DIC
TA
[ppm]
[ppm]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
450
465.42
1979.44
197.94
157.05
2152.66
Treatment
control
fluctuating
Week
1
8.010
control
fluctuating
2
460
487.69
8.008
2063.84
162.93
2243.71
2459.83
control
fluctuating
3
480
475.00
8.011
2026.63
161.31
2204.44
2419.85
control
fluctuating
4
400
453.25
8.001
1888.26
146.75
2050.76
2250.97
control
fluctuating
5
570
598.76
7.891
1936.42
116.83
2074.05
2225.79
control
fluctuating
6
450
472.11
8.001
1968.50
153.12
2138.01
2343.99
control
fluctuating
7
450
474.50
8.000
1974.23
153.24
2143.95
2349.83
control
fluctuating
8
460
469.98
8.005
1978.41
155.37
2150.10
2358.96
control
fluctuating
9
430
433.00
8.014
1859.74
149.01
2023.79
2228.76
control
fluctuating
10
500
472.85
8.003
1979.89
154.65
2150.97
2358.69
control
fluctuating
11
400
430.00
8.049
2003.37
174.12
2192.43
2427.57
control
fluctuating
12
630
650.16
7.904
2166.34
134.66
2323.58
2492.84
control
fluctuating
13
520
532.70
7.940
1930.48
130.51
2079.50
2252.98
control
fluctuating
14
440
452.38
8.029
2013.17
167.13
2196.02
2420.56
control
fluctuating
15
450
467.00
8.018
2022.34
163.38
2201.93
2420.57
control
fluctuating
16
400
446.56
8.017
1930.12
155.63
2101.26
2312.75
control
fluctuating
17
500
538.31
7.937
1937.87
130.14
2086.71
2259.28
control
fluctuating
18
500
510.00
7.995
2095.86
160.68
2274.26
2485.63
control
fluctuating
19
480
489.44
7.974
1916.67
140.02
2073.70
2262.40
control
fluctuating
20
430
440.00
8.026
1940.77
159.70
2115.76
2332.76
control
21
460
474.49
7.980
1886.58
139.94
2043.00
2232.99
fluctuating
control
fluctuating
22
400
428.99
8.025
1891.97
155.66
2062.53
2275.87
control
fluctuating
23
500
478.00
8.026
2109.71
173.71
2300.03
2529.76
control
fluctuating
24
560
566.21
7.945
2076.38
142.05
2238.10
2422.59
control
fluctuating
25
590
561.74
7.938
2025.97
136.31
2181.80
2359.82
control
fluctuating
26
420
420.00
8.051
1964.52
171.42
2150.53
2383.56
control
fluctuating
27
630
620.89
7.898
2044.17
125.55
2191.29
2351.86
control
fluctuating
28
480
498.34
7.961
1893.15
134.17
2044.63
2225.49
control
fluctuating
29
420
435.00
8.041
1988.89
169.64
2173.64
2402.96
control
fluctuating
30
460
460.00
7.997
1898.32
146.14
2060.44
2259.26
control
fluctuating
31
500
510.00
7.995
2095.91
160.68
2274.31
2485.69
control
fluctuating
32
480
497.96
7.961
1892.01
134.11
2043.41
2224.24
average control
fluctuating
478.125
490.96
7.989
1980.62
151.77
2085.80
2342.84
standard
deviation
62.14
96.85
0.04
176.73
17.13
361.56
97.54
pCO2
aimed for
pCO2
obtained
pH
HCO3-
CO32-
DIC
TA
[ppm]
[ppm]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
[µmol
kg-1]
Treatment
Week
elevated
fluctuating
1
430
456.46
8.069
1972.40
153.64
2142.45
2349.03
elevated
fluctuating
2
653
633.80
7.908
2121.57
122.31
2267.73
2419.75
elevated
fluctuating
3
870
947.19
7.789
2199.34
89.61
2323.93
2417.83
elevated
4
1200
1266.23
7.668
2207.61
88.96
2332.07
2424.41
fluctuating
elevated
fluctuating
5
1300
1158.52
7.630
2129.94
90.58
2252.98
2351.94
elevated
fluctuating
6
1115
1266.09
7.685
2225.02
92.36
2352.12
2449.68
elevated
fluctuating
7
936
968.23
7.768
2240.13
78.01
2359.82
2430.17
elevated
fluctuating
8
840
879.13
7.898
2249.81
85.84
2373.87
2458.49
elevated
fluctuating
9
1200
1199.95
7.652
2216.02
88.60
2340.54
2431.82
elevated
fluctuating
10
1300
1276.32
7.615
2248.94
93.23
2377.33
2475.31
elevated
fluctuating
11
718
725.85
7.851
2214.04
91.26
2340.11
2436.26
elevated
fluctuating
12
1250
1313.70
7.626
2236.51
84.83
2359.55
2442.96
elevated
fluctuating
13
1100
996.15
7.696
2307.99
99.37
2442.10
2548.03
elevated
fluctuating
14
938
909.66
7.761
2051.63
79.98
2165.71
2249.31
elevated
fluctuating
15
978
978.39
7.732
2202.56
81.68
2322.73
2401.96
elevated
fluctuating
16
1400
1298.18
7.600
2265.83
81.79
2388.30
2464.57
elevated
fluctuating
17
800
735.41
7.817
2059.74
79.81
2173.99
2256.87
elevated
fluctuating
18
1060
1093.57
7.699
2230.14
83.51
2352.24
2433.52
elevated
fluctuating
19
1200
1115.58
7.652
2212.42
76.03
2330.17
2398.09
elevated
fluctuating
20
1200
1223.44
7.662
2188.66
91.19
2313.89
2411.11
elevated
fluctuating
21
1300
1312.05
7.613
2238.60
93.49
2366.82
2465.73
elevated
fluctuating
22
741
761.69
7.861
2231.19
77.39
2350.26
2419.86
elevated
fluctuating
23
870
878.04
7.793
2236.22
85.59
2359.67
2444.49
163
164
elevated
fluctuating
24
1304
1255.67
7.633
2279.72
86.57
2405.19
2489.68
elevated
fluctuating
25
702
656.44
7.878
2175.83
73.60
2291.12
2356.14
elevated
fluctuating
26
1077
1002.24
7.716
2240.35
92.25
2367.86
2464.50
elevated
fluctuating
27
1200
1149.60
7.656
2229.56
92.55
2356.91
2454.60
elevated
fluctuating
28
746
777.68
7.843
2139.88
106.78
2274.45
2400.57
elevated
fluctuating
29
1500
1450.15
7.571
2238.85
85.01
2362.07
2445.69
elevated
fluctuating
30
870
843.87
7.813
2052.95
79.14
2166.60
2248.58
elevated
fluctuating
31
1100
1098.26
7.697
2011.07
86.32
2127.76
2224.90
elevated
fluctuating
32
560
770.52
7.898
2018.72
82.97
2133.52
2224.24
average elevated
fluctuating
985.46
1012.44
7.74
2183.54
89.82
2308.56
2402.81
standard
deviation
262.41
244.48
0.12
86.24
14.96
85.64
81.02
165
166
167
168
169
170
171
172
173
174
175
176
177
178
SI Table 3:
Summary of comparisons. SH evolved lineages are different from FH evolved lineages (i.e.
environmental stability affects evolution of phenotypes). SH evolved lineages are also different from
SA evolved lineages, and FH evolved lineages differ from FA evolved lineages. Further, the plastic
response is not maintained in either selection environment (SA plastic response differs from SH
evolved phenotypes, and FA plastic response differs from FH evolved phenotypes) – we use the
difference from the plastic response to estimate how much evolution there has been. For untits or traits
see SI Figure 1.
* significant (p<0.05), ** highly significant (p<0.01). NS not significant. Compare main manuscript
Figure 2.
Comparison
O2
O2
R:P
Size
ChloroNile Red
Consumption
Evolution
Ratio
phyll a
C:N
Ratio
Trait value SH
evolved compared to
trait value FH
evolved
NS
*
lower
NS
**
higher
**
higher
NS
**
higher
Evolved response SH
compared to ambient
response SA
*
higher
**
lower
*
lower
NS
NS
**
higher
**
higher
Evolved response FH
compared to ambient
response FA
*
higher
NS
NS
**
lower
*
lower
**
higher
**
lower
Evolved response SH
compared to plastic
response SA
**
lower
**
lower
**
lower
NS
NS
**
higher
*
lower
Evolved response FH
compared to plastic
response FA
NS
NS
NS
**
lower
**
lower
**
higher
**
lower
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
METHODS
Additional detailed Methods for measuring phenotypic traits and rationale for chosing
fluctuations
Rationale for adding a fluctuating environment component:
Little is known about how evolving in regimes of environmental fluctuation that select for plasticity
affects key traits in primary producers, or whether the phenotypes of populations evolved in stable
environments are good predictors of phenotypes evolved in fluctuating environments. For example,
populations evolved in fluctuating environments could simply evolve the same mean trait values as
those in stable environments, with the plastic populations having an increase in the variance in trait
value. In this case, it is the average environment that is important in determining the average
phenotype, and experiments carried out in stable environments are probably good predictors of average
phenotypes even in fluctuating environments (though there is some indication variation in phenotypes
is important in determining community composition, see Menden-Deuer & Rowlett (2014)). In
contrast, evolution in a fluctuating environment may select for qualitatively different phenotypes, such
that the average environment plays a limited role in determining the average phenotype, and
experiments carried out in stable environments may be poor predictors of phenotypic evolution in less
stable environments. Experiments carried out in stable environments may be poor predictors of
phenotypic evolution in less stable environments when evolution in a fluctuating environment selects
for qualitatively different phenotypes, such that the average environment plays a limited role in
determining the average phenotype.
Oxygen evolution and consumption rates :
Samples were acclimated to the respective assay pCO2 for two transfers (a total of ~ 14 generations).
This way, we are more likely to rule out measuring instantaneous stress responses that are not specific
to changes in pCO2 levels. The protocols for pre-acclimation and acclimation were the same. For
oxygen evolution and respiration measurements, 15ml of each sample were gently spun down to
increase cell density. We found that spinning down for 15 minutes at 1500g yielded better results than
filtrations. Pellets were then suspended in 2ml of medium of the appropriate pCO 2 and oxygen
production rates in the light and oxygen consumption rates in the dark measured in a Clark-type
oxygen electrode illuminated at 160 - 180 µmol photons m-2 s-1, with stirring, at 18 ± 0.8 ºC. Prior to
measurements, oxygen levels in the cultures were reduced to 15% by bubbling with N 2 and N2/ CO2.
For oxygen evolution measurements, changes in oxygen concentration were recorded for 6 minutes in
the light. For respiration, oxygen concentration was measured for another 6 minutes in the dark.
Oxygen consumption by the electrode was determined at the beginning of each measurement. To avoid
bias due to circadian rhythms in Ostreococcus, photosynthesis and respiration measurements were
always carried out in a 4-8 h window after the beginning of the light period. Furthermore, we estimated
bacterial background respiration by using a cell sorter to obtain medium that only contained bacteria,
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
but no Ostreococcus. The sample was then treated the same way as a ‘community’ sample containing
239
Carbonate chemistry:
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
both Ostreococcus and associated bacteria, to estimate bacterial background respiration.
Cell size, chlorophyll content, and Nile Red fluorescence.
Chlorophyll content, lipid content, and relative cell size were determined using a FACS CANTO (for
general protocols see (Petersen et al., 2012; Marie et al., 2005)). The red fluorescence channel was
used as a proxy for total chlorophyll content and chlorophyll content per cell was calculated using
chlorophyll calibration beads as standards (Bangs Laboratories, Inc.) as well as through comparing to
samples of known chlorophyll a content, where chlorophyll had been extracted according to (HolmHansen & Riemann, 1978) The same channel was used to determine polar lipids after a Nile Red stain
(la Jara et al.; Guzmán et al., 2009). Samples (200µl) were measured twice, once with and once
without the dye, in order to account for the overlap of Nile Red and chlorophyll auto fluorescence. As
the cultures were not axenic, we accounted for possible bacterial contribution to the Nile Red signal by
setting the gates so that only chlorophyll positive events within an SSC/FSC gate specific to
Ostreococcus were considered. Cell size was inferred from the FSC (forward scatter) and comparison
to 1µm, 2µm and 5µm calibration beads. Flow cytometry data was also compared to coulter counter
measurements for random samples (an average assay including technical replication consists of more
than 800 samples, making coulter counter measurements too low-throughput as the main means of
determining cell size).
The desired CO2 levels in the seawater was established by bubbling the required amount of
artificial seawater for at least 24 hours with the appropriate pCO 2, in an incubator that was set to the
same pCO2 as the water was bubbled with. Flasks were inoculated quickly and transferred back to the
incubator set to the respective selection or assay pCO2.
Seawater carbonate chemistry was calculated from pH and alkalinity using the CO2sys software
(1998). DIC was measured colourimetrically (Stoll et al., 2001) and Total alkalinity (TA) was inferred
from linear Gran-titration plots (Dickson, 1981). DIC and pH samples for all lineages were taken at
the beginning and at the end of the selection experiment. Additionally, we measured pH and either
alkalinity or DIC before medium was used for a transfer. pH was measured routinely for all samples at
the beginning and at the end of a transfer; DIC was measured for a random set of samples at each
transfer, and for all samples at the beginning and the end of the selection experiment. Data for each
time point in each selection regime was calculated from a minimum of six samples. Total Alkalinity
(TA) was measured at the beginning, 100 generations into the experiment and at the end of the
selection experiment. All other values were calculated using seacarb within R for 18 ºC and salinity 32.
We did not measure TA and DIC for all samples at each transfer, as we had at least 192 samples in
ongoing culture at any point during the selection experiment and taking samples at the end of each
transfer only would have amounted to at least 20 000 DIC and TA samples. This would have been
outwith the scope of this experiment.
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
Statistics:
ANOVA design:
ANOVAs were performed on mixed linear model output. Models were built as described in Pinheiro
and Bates (2000). We used the lmer and lme4 packages - which are based on similar algorithms but
differ in how easily random effects can be specified or p values obtained. Note that in lme degrees of
freedom are not calculated the same way as in other statistics software packages. We therefore
recalculated degrees of freedom manually. In spite of our balanced and non-hierarchal design, it was
necessary to use mixed models, because lineages between treatments were related to each other due to
being founded from clones, and to control for any ‘pseudo’ replication arising from batch-culturing the
same culture for many hundreds of generations.
In lme, when there was only one random effect and one fixed effect the model was specified as random
as below:
fit2.lme <- anova(lme(variable~fixed effect , random=~+1|random effect))
In lmer, when there were several random effects (for example, in order to include sampling origin), the
general code read:
lmer(response variable~ fixed effect1*fixed effect2 + (1|random effect1) + (1|random effect2) +
(1|random effect 3))
For example, a model testing the Ho that the interaction of assay and selection environment pCO2 do
not yield different phenotypes , might read like this (this also tests for differences between biological
and technical replicates)
m1 <-lmer( traitvalues ~ selected*measured +(1|lineage)+(1|lineage:selected), data = data)
m2<lmer(traitvalues~1+(1|Lineage)+(1|Lineage:Selection)+(1|Lineage:Selection:Biorep)+(1|Lineage:S
election:Biorep:Techrep), data = data)
Correction for multiple comparisons:
Even in a simple one-way ANOVA with K group means, there are K(K − 1)/2 pairwise comparisons.
Our ANOVAs -with many comparisons - were followed by a Holm-Bonferroni correction for multiple
comparisons to protect against inflation of Type 1 errors. Correcting at the level of the ANOVA, this
limits the amounts of comparisons carried out in the next step (post-hoc tests). The ‘post hoc’ values
reported in the manuscript were obtained from Tukey’s HSD tests, which also control for multiple
comparisons.
In addition to this we broke down omnibus ANOVAs (for example, an ANOVA assuming as H0 that
there was no difference between the means of populations in stable and fluctuating environments)
where p-values indicated a significant outcome, into several smaller ANOVAS testing for more
specific outcomes on the level of individual traits or lineages from different sampling origins.
We are happy to make R code available on request.
303
304
305
306
307
SI References
308
309
Guzmán HM, la Jara Valido de A, Duarte LC, Presmanes KF. (2009). Estimate by means of flow
cytometry of variation in composition of fatty acids from Tetraselmis suecica in response to culture
Dickson AG. (1981). An exact definition of total alkalinity and a procedure for the estimation of
alkalinity and total inorganic carbon from titration data. Deep Sea Research Part A Oceanographic
Research Papers 28:609–623.
310
conditions. Aquacult Int 18:189–199.
311
312
Holm-Hansen O, Riemann B. (1978). Chlorophyll a Determination: Improvements in Methodology.
Oikos 30:438.
313
314
la Jara de A, Mendoza H, Martel A, Molina C. R. Diaz (2003). Flow cytometric determination of lipid
content in a marine dinoflagellate, Crypthecodinium cohnii.
315
316
Marie D, Simon N, Vaulot D. (2005). Phytoplankton Cell Counting by Flow Cytometry. In:Algal
Culturing Techniques, Elsevier, pp. 253–267.
317
318
Menden-Deuer S, Rowlett J. (2014). Many ways to stay in the game: individual variability maintains
high biodiversity in planktonic microorganisms. J R Soc Interface 11:20140031–20140031.
319
320
Petersen TW, Brent Harrison C, Horner DN, van den Engh G. (2012). Flow cytometric characterization
of marine microbes. Methods 57:350–358.
321
322
Schaum CE, Collins S. (2014). Plasticity predicts evolution in a marine alga. Proc Biol Sci
281:20141486–20141486.
323
324
Schaum E, Rost B, Millar AJ, Collins S. (2013). Variation in plastic responses of a globally distributed
picoplankton species to ocean acidification. Nature Climate change 3:298–302.
325
326
Stoll MHC, Bakker K, Nobbe GH, Haese RR. (2001). Continuous-Flow Analysis of Dissolved
Inorganic Carbon Content in Seawater. Anal Chem 73:4111–4116.
327
328
329
330
(1998). CO2SYS—Program developed for the CO2 system calculations. Carbon Dioxide Inf. Anal.
Center. available at
http://scholar.google.com/scholar?q=related:Wnqve01Uz4MJ:scholar.google.com/&hl=en&num=30&
as_sdt=0,5.
331