Document [original]

Phytoplankton exudates provide full nutrition to a

subset of accompanying heterotrophic bacteria via

carbon, nitrogen and phosphorus allocation

Falk Eigemann ,

1,2

*Eyal Rahav,

Hans-Peter Grossart,

Dikla Aharonovich,

Daniel Sher ,

Angela Vogts

and Maren Voss

Leibniz-Institute for Baltic Sea Research Warnemünde,

Rostock, Germany.

Water Quality Engineering, Technical University of

Berlin, Berlin, Germany.

Israel Oceanographic and Limnological Research,

Haifa, Israel.

Leibniz-Institute of Freshwater Ecology and Inland

Fisheries, Berlin, Germany.

Leon H. Charney School of Marine Sciences, University

Haifa, Haifa, Israel.

Summary

Marine bacteria rely on phytoplankton exudates as

carbon sources (DOCp). Yet, it is unclear to what

extent phytoplankton exudates also provide nutrients

such as phytoplankton-derived N and P (DONp,

DOPp). We address these questions by mesocosm

exudate addition experiments with spent media from

the ubiquitous pico-cyanobacterium Prochlorococcus

to bacterial communities in contrasting ecosystems

in the Eastern Mediterranean –a coastal and an

open-ocean, oligotrophic station with and without

on-top additions of inorganic nutrients. Inorganic

nutrient addition did not lower the incorporation of

exudate DONp, nor did it reduce alkaline phospha-

tase activity, suggesting that bacterial communities

are able to exclusively cover their nitrogen and phos-

phorus demands with organic forms provided by

phytoplankton exudates. Approximately half of the

cells in each ecosystem took up detectable amounts

of Prochlorococcus-derived C and N, yet based on

16S rRNA sequencing different bacterial genera were

responsible for the observed exudate utilization pat-

terns. In the coastal community, several phylotypes

of Aureimarina,Psychrosphaera and Glaciecola

responded positively to the addition of phytoplankton

exudates, whereas phylotypes of Pseudoalteromonas

increased and dominated the open-ocean communi-

ties. Together, our results strongly indicate that phy-

toplankton exudates provide coastal and open-ocean

bacterial communities with organic carbon, nitrogen

and phosphorus, and that phytoplankton exudate

serve a full-ﬂedged meal for the accompanying bac-

terial community in the nutrient-poor eastern

Mediterranean.

Introduction

Approximately 50% of the global primary production is

executed by phytoplankton (Field et al., 1998) with >100

Pg oceanically ﬁxed carbon per year (Huang

et al., 2021), and generally 50% of this photosynthetically

ﬁxed carbon is then consumed by oceanic heterotrophic

bacteria (Azam and Malfatti, 2007). In addition to utilizing

photosynthetically derived organic carbon, heterotrophic

bacteria also recycle nutrients, such as nitrogen and

phosphorus (N and P), and the recycling of these and

other micro- and macro-nutrients may directly impact

phytoplankton dynamics (Amin et al., 2012; Moore

et al., 2013; Buchan et al., 2014; Amin et al., 2015).

Thus, oceanic bacteria are key players in biogeochemical

cycles with global importance (Azam and Malfatti, 2007;

Amin et al., 2015; York, 2018), and interactions between

phytoplankton and bacteria are crucial for carbon and

nutrient ﬂuxes through aquatic food webs (Cole, 1982;

Amin et al., 2015; Christie-Oleza et al., 2017; Seymour

et al., 2017; Mühlenbruch et al., 2018). Dissolved organic

material (DOM) is released by phytoplankton (DOMp) via

passive leakage, active release, as well as through lysis

products of dead cells (Grossart, 1999; Thornton, 2014;

Christie-Oleza et al., 2017), whereby the type of release

may deﬁne its composition (Livanou et al., 2017).

Besides dissolved organic carbon (DOCp), phytoplankton

exudates contain organic nitrogen [DONp, e.g. amino

acids, peptides and proteins (Beliaev et al., 2014; Roth-

Rosenberg et al., 2021a)] and phosphorus [DOPp,

Received 26 July, 2021; accepted 3 February, 2022. *For correspon-

dence. E-mail [email protected]; Tel. +49-30-31425058; Fax

+49-30-31479621.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and

reproduction in any medium, provided the original work is properly cited.

Environmental Microbiology (2022) 24(5), 2467–2483 doi:10.1111/1462-2920.15933

e.g. DNA and RNA (Roth-Rosenberg et al., 2021a)].

These organic nutrient forms exuded by phytoplankton

serve the accompanying bacterial community as N

(Karlson et al., 2015) and P (Riemann et al., 2009)

source, because inorganic forms of both nutrients are

scarce in most oceanic environments (Moore et al., 2013;

Saito et al., 2014; Liefer et al., 2019). However, the

importance of such organic, phytoplankton-derived nutri-

ents relative to inorganic nutrient sources for heterotro-

phic bacteria is still unclear. Also, the role of inorganic

nutrients in the utilization of phytoplankton-derived

dissolved organic carbon (DOCp) is not consistent, as

inorganic nutrients may (Carlson et al., 2004;

Thornton, 2014) or may not (Carlson and Ducklow, 1996)

fuel the DOCp utilization by the bacterial community

under nutrient limiting conditions, depending on so far

unknown factors.

As suppliers of various carbon and nutrient sources,

phytoplankton drive bacterial community dynamics

(Rooney-Varga et al., 2005), and different phytoplankton

release different types of DOMp (Mühlenbruch

et al., 2018). This coupling between phytoplankton and

heterotrophic bacteria, however, might be less pro-

nounced in coastal environments compared to open-

ocean sites, due to higher amounts of allochthonous

material at coastal areas (Mor

an et al., 2002a), although

DOCp is preferred over other carbon sources

(Guillemette et al., 2016). Some bacteria are adapted to

DOMp derived from speciﬁc phytoplankton species

(Grossart et al., 2007; Sarmento and Gasol, 2012; Beier

et al., 2015), phytoplankton source communities (Carlson

et al., 2004) or growth-phases of the phytoplankton

(Becker et al., 2019), and thus act as specialists. Others,

however, apply generalist strategies and are more

affected by DOM concentrations than by its composition

(Sarmento et al., 2016; Becker et al., 2019). The ques-

tions of what percentage of the total bacteria are active in

exudate utilization as well as which exudate compounds

are used by them remained open in preceding studies.

In this study, we addressed the following questions:

(i) Does phytoplankton-derived dissolved organic mate-

rial (DOMp) serve as nitrogen and phosphorus source

for the accompanying bacterial community? (ii) Do bac-

terial cells selectively incorporate exudate-derived

organic carbon or nitrogen or do they incorporate both

at a constant ratio? (iii) Which fraction of the total bacte-

rial community is active in DOMp utilization? (iv) Which

speciﬁc bacterial taxa utilize DOMp, (v) How is the bac-

terial DOCp utilization affected by inorganic nutrients?

and (vi) Do bacterial communities from contrasting envi-

ronments (coastal and open-ocean) show consistent

patterns and magnitudes of reactions as responses to

exudate and inorganic nutrient additions? We explore

these questions in the Eastern Mediterranean at a

coastal and an open ocean station. The pelagic Eastern

Mediterranean is ultra-oligotrophic, comparable with

major ocean gyres (Hazan et al., 2018;Reich

et al., 2021), whereas conditions closer to the coast are

somewhat richer in nutrients (Sisma-Ventura and

Rahav, 2019). As a DOM source, we used spent media

from Prochlorococcus strain MIT9312, labelled with

and

N. The cyanobacterium Prochlorococcus is the

most abundant phototrophic organism on Earth, with an

annual global mean abundance of 2.9 10

cells

(Flombaum et al., 2013), and dominates phytoplankton

biomass in many oligotrophic oceans despite its small

size (Partensky et al., 1999). Prochlorococcus alone

was suggested to exude as much as 75% of the daily

photosynthetic organic carbon production in oligotrophic

environments (Ribalet et al., 2015), resulting in feeding

up to 40% of the total bacterial production

(BP) (Bertilsson et al., 2005; Biller et al., 2015). In the

Eastern Mediterranean, the community composition of

the free-living heterotrophic bacteria is weakly but signif-

icantly correlated with the presence of divinyl

Chlorophyll A, a diagnostic pigment of Prochlorococcus,

further supporting a potential link through DOM produc-

tion and uptake (Roth-Rosenberg et al., 2021b). The

speciﬁc strain used, MIT9312, was selected because it

is the most abundant ecotype globally (Johnson

et al., 2006), although not in the Mediterranean (Mella-

Flores et al., 2011), and has recently been shown to

exude large amounts of DOC (Roth-Rosenberg

et al., 2021a). To test for the above-raised questions,

we amended coastal and open-ocean bacterial commu-

nities with MIT9312 exudates with and without additions

of inorganic nutrients and analysed bacterial responses

via 16S rRNA amplicon sequencing, cell numbers, BP,

incorporation of DOCp and DONp, and alkaline phos-

phatase activity (APA).

Results

Contrasting conditions at the coastal versus open-ocean

sites

We performed two experiments –one at a coastal site

and one at an open-ocean station in the Eastern Mediter-

ranean Sea. At both sites, 25 μMCProchlorococcus

MIT9312 exudates were added to the natural bacterial

community with and without on top additions of inorganic

nutrients (for details see Experimental procedures and

Table 2). The two sampling sites exhibited distinctively

different chemical characteristics. The inorganic nutrient

concentrations at the coastal site under inﬂuence of a

storm event were 0.12 0.01 μMPO

, 3.00 0.63 μM

2+3

and 0.50 0.11 μMNH

. In contrast, the open

Environmental Microbiology,24, 2467–2483

2468 F. Eigemann et al.

ocean site was highly oligotrophic even though the sam-

ples were taken during winter when the water column

was relatively well mixed; 0.003 μMPO

, 0.22 μMNO

and 0.0025 μMNH

Cell numbers and cell-speciﬁc bacterial production

In the coastal community, after 24 h, cell numbers signiﬁ-

cantly decreased in the control +nutrients and exudate

without nutrients treatments (Fig. 1A), whereas in the

open-ocean community, both exudate treatments (with

and without nutrients) showed elevated cell numbers

(1.5-fold), which were, however, not signiﬁcant

(Fig. 1B). Thus, neither the inorganic nutrient nor the exu-

date additions resulted in a clear impact on cell numbers

after 24 h of incubation.

Although the cell-speciﬁc BP increased in all treat-

ments after 24 h, it differed signiﬁcantly from the control

only in the nutrient amendment for the coastal community

(Fig. 1C). In the open-ocean community the

exudate +nutrient treatment revealed increased cell-

speciﬁc BP (Fig. 1D).

Incorporation of organic carbon and nitrogen

In order to verify that bacteria incorporate carbon and

nitrogen from phytoplankton-derived dissolved organic

material (DOMp), we investigated their uptake by means

of NanoSIMS measurements. Both coastal and open-

ocean bacterial communities showed incorporation of

labelled organic carbon and nitrogen derived from the

phytoplankton exudates, revealed by signiﬁcantly

increased

C and

N ratios (Fig. 2). As

expected, the coastal community at time 0 and all con-

trol treatments revealed values around the naturally

occurring ratios of 0.011 (

C) and 0.00367

(

N) (Fig. 2,not=0 samples were available for

the open-ocean location). Nutrient additions had neither

an effect on the DOCp (Fig. 2A and B) nor on the DONp

incorporation (Fig. 2C and D). We further tested the per-

centage of active cells in DOCp and DONp utilization

Fig. 1. Cell numbers (violin plots, top row) and bacterial production (box plots, bottom row) in the different treatment microcosm bottles. The let-

ters in the panels represent the outcomes of Tukey post hoc tests. Please note the different y-scales for the coastal and open-ocean communi-

ties. T0 samples are displayed in white, control samples in grey and exudate samples in blue.

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2469

(active cells deﬁned as cells with ratios above the 95%

percentile of pooled t0, control and control +nutrient

measurements, Supplement 1). After 24 h of incubation,

in the coastal bacterial community, 54% of the cells in

the pooled exudate and exudate +nutrients treatments

revealed increased

N values and 47% increased

values. In the open-ocean community, 39% and 51%

cells (pooled exudate and exudate +nutrients treat-

ment) with increased

N and

C values were found. In

the coastal as well as in the open-ocean bacterial com-

munity

Cand

N uptake were highly correlated

(Table 1). If the single treatments were considered sep-

arately, for both, the coastal and open-ocean communi-

ties both exudate treatments revealed signiﬁcant

correlations, but as expected, none of the other treat-

ments (Table 1). The slopes (

C=y-axis,

N=x-axis)

of the correlations in all exudate treatments were

steeper for the coastal community compared to the

open-ocean one (Table 1).

Cell-speciﬁc alkaline phosphatase activity

To test whether bacterial communities satisfy their phos-

phorus demand via DOMp, we analysed the activity of

the alkaline phosphatase enzyme (APA) in the different

treatments and environments. In both environments,

additions of phytoplankton exudates lowered the APA,

without any effect of on top additions of inorganic nutri-

ents. Nutrient additions in the controls also lowered the

APA, but not as strong as the exudate additions (Fig. 3).

In the coastal bacterial community, the control treatments

without nutrient additions revealed a signiﬁcantly higher

cell-speciﬁc APA compared to all other treatments, and in

the open-ocean community the exudate treatments had

Fig. 2.

C ratios (A, B) and

N ratios (C, D) of the coastal (A, C) and the open-ocean (B, D) bacterial community following 24 h incuba-

tions and in t0. T0 samples are displayed in white, control samples in grey and exudate samples in blue.

Environmental Microbiology,24, 2467–2483

2470 F. Eigemann et al.

not only lower APA compared to the control treatment,

but also lower APA compared to the t0 and

control +nutrient treatments (Fig. 3).

Dynamics in bacterial community composition

To answer how the active bacterial communities develop

in response to exudate additions, we analysed the diver-

sity and composition in the 16S rRNA derived dataset.

Since the amplicons were derived from RNA molecules,

these represent primarily the active members of the bac-

terial community. Shannon diversities did not show signif-

icant differences between the treatments in the coastal

communities (ANOVA, p=0.81), but in the open-ocean

community exudate treatments yielded lower Shannon

diversities (ANOVA, p=0.00, Fig. 4). If only t0 bacterial

communities were compared, both environments rev-

ealed slightly higher Bray–Curtis dissimilarities compared

to comparisons of t0 samples in the same environment

(average between the environments 0.76 0.06,

between coastal t0 samples 0.63 0.13, and between

open-ocean t0 samples 0.67 0.01, Supplement 2).

NMDS analyses of the bacterial communities showed

differences between experiments [analysis of similarities

(ANOSIM), p=0.001, R=0.44, Fig. 5], and in the

coastal community additionally between the free and

attached fractions (ANOSIM, p=0.001, R=0.45,

Fig. 5). However, the different treatments also clustered

signiﬁcantly differently (ANOSIM, p=0.02, R=0.11),

where strong differences occurred between samples with

added exudates and those without. The addition of nutri-

ents did not show any clear effects in either experiment

(Fig. 5). If only the 40 most abundant ASVs in both

experiments were considered and illustrated in

heatmaps, likewise strong community shifts were

observed between the treatments (Fig. 6A and B). LEfSe

(linear discriminant analysis effect size) analyses

suggested Fluviicola,Pseudofulvibacter,Syn-

echococcus and NS4 marine group being predominant

in t0 samples, Spongiispira predominant in the controls,

whereas the exudate treatments were dominated by

Glaciecola in the coastal environment [Supplement 3,

Fig. 6A, see also the 20 most abundant ASVs for every

single treatment (Supplement 4)]. In the open-ocean

community, Pseudoalteromonas was strikingly abundant

in the exudate treatments, whereas SAR11, SAR202 and

KI89a clade bacteria were predominant at t0 and in the

control samples (Fig. 6B, Supplements 4 and 5).

Discussion

In this study, we wanted to identify which fraction (and

which bacterial taxa) of the total bacterial community are

Table 1. Correlations between

C and

N incorporation derived from NanoSIMS analyses.

Experiment Treatment Linear function r

Exp. 1, coastal All treatments y=0.14x0.74 <0.0001

t0 y=0.13x0.03 0.63

Control y=0.07x0.01 0.39

Control +nutrients y=0.48x0.1 0.05

Exudates y=0.12x0.72 <0.0001

Exudates +nutrients y=0.15x0.48 <0.0001

Exp. 2, open-ocean All treatments y=0.06 0.91 <0.0001

Control y=0.002x0.03 0.77

Control +nutrients y=0.04x0.02 0.78

Exudates y=0.1x0.94 <0.0001

Exudates +nutrients y=0.06x0.95 <0.0001

The y-axis was deﬁned as =

C ratio, the x-axis as =

N ratio.

Fig. 3. Cell-speciﬁc alkaline phosphatase activity for the coastal (top

panel) and the open-ocean (bottom panel) bacterial community. The

letters in the panels represent the outcomes of Tukey post hoc tests.

T0 samples are displayed in white, control samples in grey and exu-

date samples in blue.

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2471

active in phytoplankton-derived dissolved organic mate-

rial (DOMp) utilization, whether DOMp serves as a sub-

stantial nitrogen and phosphorus source for the

accompanying bacterial community, and if so, whether

bacterial cells selectively incorporate carbon or nitrogen.

To test for general patterns as well as for the impact of

allochthonous material on utilization of DOMp, we

addressed these questions in two contrasting systems in

the Eastern Mediterranean: a coastal station after a storm

event with lots of suspended material and an oligotrophic

open-ocean station. Our experimental results strongly

indicate that DOMp serves as a substantial N and P

source for a subset of approximately 50% of the bacterial

community at the times and places sampled, and that the

active community members can fully satisfy their

demands with the phytoplankton-derived organic nutrient

sources. Thus, the Mediterranean Sea bacterial commu-

nities may compensate for inorganic nutrient limitations

(Krom et al., 2010; Krom et al., 2014) with organic forms

derived from phytoplankton exudates. Another raised

question was which factors limit heterotrophic production:

Are marine bacteria predominantly carbon (Christie-

Oleza et al., 2017) or nutrient (Carlson et al., 2004;

Fouilland et al., 2014) limited? To address this point, N

and P derived from the provided exudates seem to be

more important for the open-ocean community if com-

pared to the coastal community despite similar overall

reactions to DOMp additions in both environments.

Bacterial production and carbon incorporation following

phytoplankton exudate addition

Increased BP in both environments following exudate

additions (Fig. 1) suggests that exudate DOCp provide

carbon for both bacterial communities, which was con-

ﬁrmed by NanoSIMS measurements (Fig. 2). However,

as stated above, DOCp incorporation might be limited by

inorganic nutrients (Carlson et al., 2004; Fouilland

et al., 2014). For example, the conversion of carbon from

polysaccharides into bacterial biomass was enhanced by

inorganic nitrogen for de novo synthesis of cellular pro-

teins (Grossart et al., 2007; Piontek et al., 2011), and

additions of inorganic nitrogen and high nitrate concentra-

tions increased glucose assimilation by bacterioplankton

(Bianchi et al., 1998; Skoog et al., 2002). However, also

organic nutrients (e.g. dissolved free amino acids) were

already shown to increase BP rates in marine environ-

ments (Carlson and Ducklow, 1996). Our results suggest

that the nutrient demand for DOCp utilization in both

4.0

4.5

5.0

5.5

6.0

Shannon

4.0

4.5

5.0

5.5

6.0

control

control +

nutrients

exudates

exudates +

nutrients

Shannon

coastal community

open-ocean community

n.s.

a a ab b b

Fig. 4. Shannon diversity for the coastal (top panel) and the open-

ocean (bottom panel) bacterial community. The letters in the panels

represent the outcomes of Tukey post hoc tests. Please note that in

coastal communities size-fractionated (5 and 0.2 μm pore width)

samples were analysed, whereas in the open-ocean communities no

size-fractionation was performed (only 0.2 μmﬁlter pore width). T0

samples are displayed in white, control samples in grey and exudate

samples in blue.

MDS1

MDS2

fraction/experiment

coastal attached

coastal free

open-ocean no separation

sample group

control

control + nutrients

exudates

exudates + nutrients

coastal and open-ocean community Fig. 5. NMDS plot of the bacterial

communities (stress =0.15). The

symbol and colour key is given on the

right-hand side. T0 samples are dis-

played in white, control samples in

grey and exudate samples in blue.

Environmental Microbiology,24, 2467–2483

2472 F. Eigemann et al.

bacterial communities was satisﬁed with organic sources

derived from Prochlorococcus exudates because nutrient

additions on top of the exudates did not result in higher

cell-speciﬁc bacterial productivity nor in higher incorpora-

tion rates of

C-labelled carbon (Figs 1and 2, also see

the next paragraph). Despite similar bulk measurements

t0_a_0.2

t0_c_0.2

t0_b_0.2

control+nutrients_b_5

control_d_5

control_a_5

control+nutrients_a_5

control_b_5

t0_b_5

t0_c_5

t0_a_5

control_c_5

control_a_0.2

control_b_0.2

control+nutrients_a_0.2

control_c_0.2

control+nutrients_c_0.2

control+nutrients_b_0.2

control+nutrients_d_0.2

exudates+nutrients_b_5

exudates+nutrients_b_0.2

exudates_b_0.2

control+nutrients_d_5

control_d_0.2

exudates+nutrients_a_5

exudates+nutrients_a_0.2

exudates_c_0.2

Glaciecola1

Spongiispira1

NS9marinegroup1

KI89a clade1

Glaciecola2

Glaciecola3

Spongiispira2

Aureimarina1

Synechococcus CC9902_1

NS9marinegroup2

Glaciecola4

NS4marinegroup

Aureimarina2

Glaciecola5

Synechococcus CC9902_2

Glaciecola6

KI89a clade2

Spongiispira3

Aureimarina3

Glaciecola7

Aureimarina4

KI89a clade3

Pseudofulvibacter1

NS7marinegroup

SAR11 clade Ia1

Glaciecola8

NS9marinegroup3

Spongiispira

SynechococcusCC9902_3

Pseudofulvibacter2

Fluviicola

Pseudomonas

Psychrosphaera

Cryomorphaceae

NS2b marine group

KI89a clade4

Rhodobacteraceae

Glaciecola9

SAR11 clade Ia2

DEV007

Glaciecola10

NS9marinegroup4

coastal community

contro

ents_

ontro

_a_

ontro

ents_a_5

ontro

ates

nutr

ents_

ates

utr

ents_

ates_

ontro

_c_5

ontro

_a_

ontro

ents_a_

ontro

_c_0

ontro

ents_c_

ontro

ents_

ontro

ents_

ontro

ents_

ontro

ates

utr

ents_a_5

ates

utr

ents_a_

ates_c_0

t0_b

t0_c

control+nutrients_c

control+nutrients_a

control_a

control_c

t0_a

control_b

control+nutrients_b

exudates_a

exudates_b

exudates+nutrients_b

exudates+nutrients_a

exudates+nutrients_c

exudates_c

Pseudoalteromonas1

Pseudoalteromonas2

Pseudoalteromonas3

Pseudoalteromonas4

Pseudoalteromonas5

Pseudoalteromonas6

Pseudoalteromonas7

Pseudoalteromonas8

Pseudoalteromonas9

Pseudoalteromonas10

Pseudoalteromonas11

Pseudoalteromonas12

Pseudoalteromonas13

Pseudoalteromonas14

Pseudoalteromonas15

Pseudoalteromonas16

Pseudoalteromonas17

SAR11_cladeIa1

Pseudoalteromonas18

Pseudoalteromonas19

Pseudoalteromonas20

Pseudoalteromonas21

Pseudoalteromonas22

SAR11_cladeIa2

KI89a_clade1

Pseudoalteromonas23

Pseudoalteromonas24

Pseudoalteromonas25

AEGEAN_169_marine_group1

SAR11_cladeIa3

SAR202_clade1

Pseudoalteromonas26

Pseudoalteromonas27

KI89a_clade2

Synechococcus_CC9902

SAR202_clade2

SAR11_cladeIa_4

SAR202_clade3

Pseudoalteromonas28

SAR11_cladeIa_5

open-ocean community

ontro

ents_c

ontro

utr

ents_a

ontro

ents_

exu

ates_

exu

ates

nutr

ents_

exu

ates

nutr

ents_

ate

utr

ents_

ates_c

Fig. 6. Heat maps of the 40 most abundant

ASVs (normalized absolute read counts) in

the coastal (top) and open-ocean (bottom)

environment. Samples can be seen below

the heatmap, ASVs on the right side. If sev-

eral ASVs with the same taxonomy were

present, the taxonomy was numbered. Blue

colour indicates low read numbers and red

colour indicates high read numbers. The

letter behind the treatment names refer to

the replicate, the number behind this letter

indicates the ﬁlter pore width and thus the

fraction (only coastal experiment,

0.2 μm=free fraction, 5 μm=attached

fraction). Dendrograms on top are calcu-

lated based on Euclidian distances, the

ordering of ASVs refers to read counts (top

ASV: lowest read counts, bottom ASV:

highest read counts). T0 samples are dis-

played in white, control samples in grey

and exudate samples in blue.

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2473

in both environments (Supplement 6), we observed lower

cell-speciﬁc BP and higher

C ratios in the coastal

community if compared to the open-ocean community

(Figs 1and 2). The lower cell-speciﬁc BP in the coastal

community might partly be explained with the storm event

that introduced sediment and soil bacteria to the water

column. The re-suspension might have contributed signif-

icantly to our cell counts [at the start of the experiment

(t0), the coastal community revealed approximately ﬁve

times higher cell numbers compared to the open-ocean

community (Fig. 1A and B)], but not to the cell-speciﬁc

activity, lowering only the cell-speciﬁc BP. This assump-

tion is supported by Raveh et al.(

2015), where much

lower bacterial counts (510

cells ml

1

in January)

but comparable bacterial bulk productions were deter-

mined for a coastal area in the Eastern Mediterranean

Sea. The higher

C ratios in the coastal communi-

ties, on the other hand, appear surprising taking into

account the storm event with possible input of carbon

and contradict previous studies, where direct carbon cou-

pling between phytoplankton and bacteria was weak in

coastal waters due to substantial allochthonous carbon

sources (Mor

an et al., 2002b). Yet, our ﬁndings suggest

a preferential utilization of DOMp compared to other car-

bon sources, which ﬁts with observations from freshwater

bacterial communities that preferentially utilized

phytoplankton-derived carbon over allochthonous, terres-

trial one (Guillemette et al., 2016) and of coastal bacterial

communities in the Mediterranean satisfying more than

half of their carbon demand from phytoplankton exudates

(Fouilland et al., 2014). This can be explained by the fact

that pico-cyanobacteria (i.e. Prochlorococcus and Syn-

echococcus) exudates contain high fractions of low-

molecular-weight DOC, which is highly labile and can be

rapidly utilized by heterotrophic bacteria (Bertilsson

et al., 2005; Sharma et al., 2014), e.g. organic acids,

organo-halogens and isoprene (Shaw et al., 2003;

Bertilsson et al., 2005) as well as lipids, proteins and

fragments of DNA and RNA (Biller et al., 2015). The

strain Prochlorococcus MIT9312 used in our experiments

releases 90% of the ﬁxed carbon as DOCp (Roth-

Rosenberg et al., 2021a). Together, we did not ﬁnd indi-

cations that allochthonous carbon and nutrients at the

coastal site impacted the utilization of DOMp.

Incorporation of DONp and DOPp

Our NanoSIMS measurements, as well as APA analyses,

suggest that the coastal and open-ocean bacterial com-

munities cover their nitrogen and phosphorus demands

via DOMp (Figs 2and 3). This notion is in accordance

with previous studies, showing that pico-cyanobacteria

produce nitrogen and phosphorus-rich DOMp (Beliaev

et al., 2014; Christie-Oleza et al., 2017) that serves as

energy and also nutrient source for bacteria (Livanou

et al., 2017). For example, bacteria exposed to phyto-

plankton exudates upregulated transcription of nitrogen

and phosphorus utilization genes (McCarren et al., 2010),

and co-cultures of picocyanobacteria with several differ-

ent heterotrophic bacteria revealed an increased expres-

sion of genes for transporters of amino acids and

peptides (Beliaev et al., 2014). It has been previously

shown that Prochlorococcus MIT9312 exudes consider-

able amounts of organic nitrogen under laboratory condi-

tions (Roth-Rosenberg et al., 2021a), possibly in the form

of proteins, amino acids, DNA, RNA and nucleotides.

This could provide labile DONp to the bacterial commu-

nity (Sharma et al., 2014). In general, phytoplankton exu-

dates provide manifold organic nitrogen species to

heterotrophic bacteria, including urea, dissolved and free

amino acids, proteins, nucleic acids, amino sugars (Meon

and Kirchman, 2001; Berman and Bronk, 2003) and

methylated amines (Lidbury et al., 2015), and so on. In

co-culture and exudate addition experiments, especially

dissolved free amino acids and urea were used by bacte-

ria (Grossart and Simon, 2007; Bradley et al., 2010; Sar-

mento et al., 2013; Beliaev et al., 2014). Similar to

nitrogen, phytoplankton exudates may offer substantial

amounts of organic phosphorus in various forms

(Livanou et al., 2017), and our results indicate that both

bacterial communities appease their phosphorus demand

with the provided phytoplankton exudates (Fig. 3).

A previous modelling study has suggested that, under

conditions of N starvation, Prochlorococcus releases

P-containing molecules such as nucleobases and nucleo-

sides (Ofaim et al., 2021), although the magnitude of exu-

dation of DOP is not well constrained (Roth-Rosenberg

et al., 2021a). Increased APA in the control incubations

of both environments indicates phosphorus depletion in

these treatments, which is consistent with the low initial

phosphorus concentrations and our ‘batch-like’bottle

incubations. Indeed, we could show in the open-ocean

community a signiﬁcantly lowered APA after exudate

additions compared to the control +nutrient treatments,

indicating not only a complete phosphorus supply by the

exudates but a preferential utilization (Fig. 3).

However, organic N and P forms provided with the

phytoplankton exudates seem to have a higher impor-

tance for the open-ocean community than for the coastal

one: At time 0, cell-speciﬁc APA was approximately

seven times higher in the open-ocean community com-

pared to the coastal community (Fig. 3) [whereas bulk

analyses showed comparable outcomes in both environ-

ments (Supplement 7)], suggesting a severe P limitation

in the ﬁrst, which was completely diminished with the

addition of exudates. Likewise, correlations of C and N

uptake revealed steeper slopes (i.e. relatively higher C

Environmental Microbiology,24, 2467–2483

2474 F. Eigemann et al.

compared to N incorporation) in the coastal community,

indicating preferential uptake of N by the open-ocean

community (Table 1). The higher R

values of C to N cor-

relations furthermore suggest a more uniform behaviour

of the open-ocean community (Table 1). The ratios of

N incorporations derived from NanoSIMS analyses

in our experiments were in the same range as for phyto-

plankton associated bacteria in the Baltic Sea (Eigemann

et al., 2019), and different

Cto

N ratios in NanoSIMS

measurements of different treatments suggest selective

uptakes of either carbon or nitrogen.

In summary, our results strongly suggest that

phytoplankton-derived organic nitrogen as well as phos-

phorus is incorporated into bacterial biomass and provide

the primary nitrogen and phosphorus source for the

coastal as well as open-ocean bacterial communities.

Phytoplankton exudates seem to be a more important N

and P source for the open-ocean communities compared

to the coastal ones, where the latter may satisfy their N

and P demand partly with allochthonous sources. Our

experimental conditions, however, do not fully mimic nat-

ural conditions, for example, the added organic matter is

at relatively high concentration, and comes from a single

course (a single phytoplankton strain). Further research

is required in order to determine whether these patterns

also occur when considering naturally derived phyto-

plankton exudates in natural environments.

Community responses to exudate additions

In the open-ocean communities, as a reaction to exudate

additions a decrease in Shannon diversity was obvious

(Fig. 4), which was caused by the dominance of a single

genus, namely, Pseudoalteromonas (Fig. 6, Supplement

6), whereas diversity in the coastal community remained

constant, but with different abundant genera in the differ-

ent treatments (Figs 4and 6). In general, exudate addi-

tions boosted copiotrophic members of the heterotrophic

bacterial community to the costs of oligotrophic genera

(Fig. 5, Supplement 4). However, one should keep in

mind that our analyses are based on rRNA, and thus only

the active part of the bacterial community is appropriately

reﬂected. Together, after exudate additions,

Pseudoalteromonas in the open-ocean communities, and

Glaciecola,Psychrosphaera and Aureimarina in the

coastal communities revealed strong positive responses,

whereas SAR11 showed negative responses to exudate

additions, especially in the open-ocean community

(Fig. 6, Supplements 3, 4, 5). Pseudoalteromonas and

Glaciecola belong to the order Alteromonadales which

signiﬁcantly contribute to carbon cycling in the surface

ocean (Pedler et al., 2014), possesses effective degrada-

tion systems for a wide array of polysaccharides (Gobet

et al., 2018), and showed positive responses to

phytoplankton exudates (Seymour et al., 2009; Taylor

and Cunliffe, 2017). The responsive bacteria in our

experiments possess a variety of carbohydrate-active

enzymes (CAZymes). For example, Glaciecola sp. 4H-

3-7YE-5 possesses the genomic possibility for the degra-

dation of several polysaccharides (Klippel et al., 2011),

which constitute major components of phytoplankton exu-

dates (Meon and Kirchman, 2001; Mühlenbruch

et al., 2018). Furthermore, cyanobacteria are known to

produce glycogen as a storage polysaccharide (Bertocchi

et al., 1990; Bhatnagar and Bhatnagar, 2019), and

Pseudoalteromonas,Psychrosphaera as well as

Glaciecola possess effective utilization systems for gly-

cogen (Lombard et al., 2014), and other polysaccharides

common in phytoplankton (Klippel et al., 2011; Pheng

et al., 2017; Gobet et al., 2018). The negative responses

of SAR11 related ASVs to exudate additions might be

partly attributed to the point in time when

Prochlorococcus spent medium was harvested, i.e. the

early stationary phase (see Experimental procedures

why this point in time was chosen). In co-culture experi-

ments between several Prochlorococcus strains and

SAR11 HTCC7211, stable long-term coexistence was

maintained if Prochlorococcus was kept in the log-phase,

but strong detrimental effects on SAR11 occurred when

Prochlorococcus entered the stationary phase (Becker

et al., 2019). These results were ascribed to growth

phase–dependent releases of metabolites, where log-

phase growing Prochlorococcus exudates fulﬁlled the

central carbon requirement of SAR11, but

Prochlorococcus metabolites from the stationary phase

caused a rapid decline in SAR11 cell numbers. Simulta-

neously, however, sympatric copiotroph bacteria in co-

cultures were boosted in cell numbers when

Prochlorococcus entered the stationary phase, highlight-

ing the higher functional and regulatory facilities of

copiotrophs compared to oligotrophs such as SAR11

(Becker et al., 2019). Despite clear responses of RNA-

based amplicon sequencing of speciﬁc genera to

Prochlorococcus exudates in our experiments, we can-

not relate these responses to speciﬁc demands of N, P,

or C, because the used methods (BP, APA and isotope

labelling) focused on bulk reactions. Nevertheless, we

could demonstrate the relative increase/decrease of spe-

ciﬁc genera, implying niche partitioning within the bulk

community.

We observed higher discrepancies and variability

between the control and the control +nutrient samples in

the attached fraction of the coastal community compared

to the free fraction and the open-ocean samples (Fig. 5).

The higher discrepancies between the control and

control +nutrient samples in the attached fraction might

be an indirect effect of the storm event with bacteria colo-

nizing introduced particles, whose utilization is fuelled by

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2475

the addition of inorganic nutrients, whereas in the open-

ocean the necessary carbon for a boost effect of inor-

ganic nutrients is just lacking. The higher variability in the

attached fraction might also be an indirect effect of the

storm event and may represent more chaotic communi-

ties attached to re-suspended and newly introduced parti-

cles. However, we can only speculate on this because

we did not perform analyses conﬁrming the above-raised

hypothesis. Despite considerable overlap in the bacterial

communities at t0 in coastal and open-ocean environ-

ments (Supplement 2), different responders to exudate

additions appeared at the contrasting sites (Fig. 5, Sup-

plements 3, 4, 5). Below, we list three possible explana-

tions for the different development after exudate

additions: First, the source communities differed although

several abundant members overlapped. Indeed,

Pseudoalteromonas, the main responder in the open-

ocean treatments, showed even higher relative read

abundances at t0 in the coastal community (mean

subsampled read-sums of Pseudoalteromonas ASVs in

t0 samples: ﬁve for open-ocean and 19 for coastal com-

munities, Supplement 8), but the most responsive genera

of the coastal community were completely lacking in the

open-ocean community (Psychrosphaera,Aureimarina,

Supplement 8) or present at low abundances

(Glaciecola, Supplement 8). This outcome suggests a

high importance of the bacterial source communities on

the effectiveness in DOC utilization as well as the ability

to use different DOC sources (Carlson et al., 2004;

Grossart et al., 2007). Second, environmental conditions

favour certain bacterial genera over others (Nemergut

et al., 2013;Wuet al., 2019), which may reﬂect the ability

to effectively utilize DOMp. Thus, under open-ocean con-

ditions paired with the additions of DOMp,

Pseudoalteromonas was the most successful genus that

outcompeted other genera, whereas under coastal condi-

tions it was outcompeted, despite its higher relative abun-

dances at t0. Our data suggest that other members in the

open-ocean community were not able to show such a

strong response in the 24 h of incubation, and

Alteromonadales are known as opportunists (Eilers

et al., 2000) and fast growers (Pedler et al., 2014). We

did not assess the effect of bacterivorous protists on

either of the two communities which may affect bacterial

abundances and communities after nutrient additions

(Lebaron et al., 2001). However, the overall effect of

grazers on bacterial community composition in Mediterra-

nean mesocosm experiments was not signiﬁcant (Baltar

et al., 2016), and it is unlikely that different grazing pres-

sures affected the differential outcomes of our experi-

ments (Baltar et al., 2016). Third, the dominant response

of the copiotrophic generalist Pseudoalteromonas

(Gammaproteobacteria) in the open-ocean environments

may be explained with drastic changes in relative DOM

concentrations, whereas in the coastal environment addi-

tions of DOM induced more specialists responses

(Sarmento et al., 2016). Absolute DOC concentrations

are an important factor for the uptake ability of heterotro-

phic bacteria, as only a few specialists were able to incor-

porate DOCp at low concentrations, but a broader range

of bacteria could use the same source of DOCp at high

concentrations (Sarmento et al., 2016). This might be

also true in our experiments, where the background DOC

concentrations probably have been much higher in the

coastal (especially after the storm event) compared to

the open-ocean environment.

Besides speciﬁc outcomes at the genus level, some

general patterns occur under phytoplankton bloom condi-

tions, with Flavobacteria, Alphaproteobacteria and

Gammaproteobacteria being dominant classes (Buchan

et al., 2014). This is partly reﬂected in our experiments

(Fig. 6), with especially Gammaproteobacteria

(Pseudoalteromonas,Psychrosphaera,Glaciecola) and

Flavobacteria (Aureimarina) showing strong positive

responses to exudate additions, whereas we did not

observe positive responses but rather relative declines of

Alphaproteobacteria (Supplement 4). This relative decline

of Alphaproteobacteria as response to phytoplankton

exudates partly contradicts previous studies, where espe-

cially the Roseobacter clade (class Alphaproteobacteria)

reveals numerous positive interactions with phytoplank-

ton (Romera-Castillo et al., 2011; Lidbury et al., 2015).

However, despite being capable of using a wide array of

substrates (Lidbury et al., 2015), Roseobacter did not

take up Prochlorococcus exudates in a similar experi-

ment (Sarmento and Gasol, 2012), emphasizing the

selective uptake of DOMp by different bacteria (Sarmento

et al., 2013).

A point not addressed in our experimental set-up is the

successive degradation from labile to recalcitrant DOM.

DOMp utilization by heterotrophic bacteria undergoes a

succession with different responsive organisms at differ-

ent times after DOMp pulses (Teeling et al., 2012;

Buchan et al., 2014; Teeling et al., 2016). Oligotrophic

organisms like SAR11 utilize highly labile low-molecular-

weight compounds, whereas copiotrophic bacteria such

as Alteromonadales utilize high-molecular-weight com-

pounds such as polysaccharides (Sharma et al., 2014).

With our experiments, we only reﬂect the community

response at a single time-point, i.e. 24 h after DOMp

addition, which may resemble daily changes in DOMp

concentrations accompanied with photosynthesis during

phytoplankton blooms. Furthermore, as mentioned

above, Prochlorococcus exudes high amounts of labile

DOMp, which can be rapidly utilized by heterotrophic

bacteria (Roth-Rosenberg et al., 2021a). Thus, the con-

centration of the added DOC as well as the incubation

time allow for a maximal comprehensive picture (for a

Environmental Microbiology,24, 2467–2483

2476 F. Eigemann et al.

single pulse), and thus should reﬂect responses of oligo-

trophic as well as copiotrophic community members at

environmental relevant concentrations (McCarren

et al., 2010; Seymour et al., 2010; Sarmento and

Gasol, 2012; Sharma et al., 2014; Beier et al., 2015; Sar-

mento et al., 2016). However, to fully comprehend the

dynamics of DOCp utilization and the community succes-

sion additional time-course studies are needed.

Conclusions

Short-term responses of coastal and open-ocean bacte-

rial communities to phytoplankton exudates addition with

and without inorganic nutrients revealed similar overall

bacterial response patterns, but different responders in

the coastal versus the open ocean communities as well

as a higher importance of N and P provided with the exu-

dates for the open-ocean community. The different

responders suggest environmental factors, such as the

ambient DOM concentrations or the initial bacterial com-

munities to determine DOMp utilization effectiveness of

the respective bacterial communities to some extent. Our

results strongly indicate that the allocated phytoplankton

exudates provide a major fraction of the bacterial commu-

nity (50%) with organic carbon, nitrogen and phospho-

rus as combined additions of exudates and inorganic

nutrients neither enhance cell-speciﬁc BP nor lowered

incorporation of DONp and cell-speciﬁc APA. Utilization

of DOCp seems not to be limited by inorganic nutrients,

because addition of inorganic nutrients did not elevate

the incorporation of DOCp into bacterial cells. Conse-

quently, phytoplankton exudates may function as a full-

ﬂedged meal for the accompanying bacterial communi-

ties, and can be used as energy and nutrient sources by

bacteria independent of the surrounding inorganic nutri-

ent concentrations. Prochlorococcus is a major compo-

nent of the global phytoplankton (Flombaum et al., 2013),

and its exudates likely substantially contribute to BP in

oligotrophic environments (Biller et al., 2015; Ribalet

et al., 2015). Hence, together with the naturally occurring

amount of added DOM (in bloom occasions) (Sharma

et al., 2014; Beier et al., 2015), our results may reﬂect

important patterns in marine environments. Together, our

study emphasizes the dependency of heterotrophic bac-

teria on phytoplankton exudates and illustrates that

abiotic factors may resign beyond biotic interactions for

marine heterotrophic bacteria. Therewith, our outcomes

further strengthen the importance of phytoplankton–

bacteria interactions in carbon, nutrient and mineral

cycling, and thus in functioning of marine ecosystems.

Experimental procedures

Experimental set-up

Prochlorococcus MIT9312 was grown in 2 L bottles

under constant light (20 μEm

2

1

)at22

C in Pro99

media where the NH

concentration was reduced from

800 to 100 μM, resulting in the cells entering stationary

stage due to N starvation (Grossowicz et al., 2017). For

several generations before harvest, 98% labelled

N-

was used as the sole N source, and the media was

amended with 1 mM of 98% labelled

C-HCO

as a C

source, resulting in a fully labelled culture. To obtain cell-

free exudates, batch cultures were harvested at the early

decline phase by centrifugation followed by ﬁltration

through a 0.22 μm polycarbonate ﬁlter. Early stationary

phase was chosen to minimize the carryover of

N-NH

(which should be depleted from the media, see below), to

increase the amount of released DOC (Roth-Rosenberg

et al., 2021a), and because spent media from this stage

may reﬂect natural DOMp composition better than DOMp

from exponentially growing cultures (Christie-Oleza

et al., 2017). We note that the DOM in the media is likely

a result of both exudation and cell mortality. The spent

media was maintained at 20C until use.

Using the

N and

C labelled spent media we per-

formed two incubation experiments in order to test the

bacterial response to DOMp in coastal (Exp 1) compared

to open-ocean (Exp 2) systems. Each of these experi-

ments included the spent media with and without inor-

ganic nutrient additions of 20 μMNH

, and 2 μMPO

,to

eliminate nutrient limitation in the 24 h of exposure in the

nutrient spiked treatments (Table 2). Exp 1 was carried

out in the dark on January 9, 2019, in 1 m

natural sea-

water ﬂow-through tanks to maintain ambient tempera-

ture at the Israel Oceanographic and Limnological

Research centre in Haifa, Israel. Each treatment con-

sisted of four biological replicates. The coastal site was a

5 m intake pipe during a winter storm with high waves

Table 2. Summary of the exudate and nutrient additions to surface Eastern Mediterranean Seawater in January 2019.

Ingredient/treatment Control Control +nutrients Exudates Exudates +nutrients

Prochlorococcus MIT9312 exudates –– 25 μMC 25μMC

Nutrients –20 μMNH

,2μMPO

–20 μMNH

,2μMPO

NNH

5nM 5nM ––

Values shown are the ﬁnal concentration.

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2477

and signiﬁcant turbulence. Additionally, heavy rainfalls

caused considerable land-to-sea run-offs. As a result, the

water was brownish in colour, which may have added

allochthonous material, as well as soil and sediment bac-

teria. Exp 2 was carried out on board of the R/V Mediter-

ranean Explorer on January 23, 2019, at station

THEMO-2 (Reich et al., 2021), in an on-board ﬂow-

through system, and each treatment consisted of three

biological replicates. For both experiments, 4.5 L Nalgene

bottles were ﬁlled with 3 L of 50 μm pre-ﬁltered seawater

with the respective DOCp and nutrient amendments

(Table 2). In order to receive a clear response of the bacte-

rial community in the range of naturally occurring concen-

trations of DOC, we chose the addition of 25 μMDOCp

(Seymour et al., 2010; Beier et al., 2015; Sarmento

et al., 2016). Likewise, the incubation time was set to 24 h,

in order to see a meaningful response of both, fast and

slow responsive bacteria to phytoplankton exudates at the

above-mentioned DOC concentration (McCarren et al.,

2010; Sarmento and Gasol, 2012;Sharmaet al., 2014;

Beier et al., 2015). The spent Prochlorococcus medium

contained 0.9 μMPO

and 0.16 μMNH

, resulting in a ﬁnal

concentration of 5 nM

in the exudate treatments

(32diluted). We accounted for these labelled, inorganic

leftovers in the control and control +nutrient treatments in

order to differentiate between uptake of DOMp and inorganic

nutrients by the bacterial communities in our NanoSIMS

measurements (Table 2).

Nutrient measurements

For measurements of inorganic nutrients, samples were

pre-ﬁltered through 0.2 μm pore width ﬁlters and collected

in acid-clean plastic vials. Dissolved nutrients in the cul-

tures and for Exp. 1 and 2 were then determined using a

three-channel segmented ﬂow auto-analyser system

(AA-3 Seal Analytical).

Cell numbers

Bacterial cell numbers were determined by ﬂow cyto-

metry. Brieﬂy, duplicate samples were ﬁxed with

cytometry-grade glutaraldehyde (0.125% ﬁnal concentra-

tion), ﬂash-frozen with liquid nitrogen and stored at

80C until analyses. For measurements, samples were

thawed in the dark at room temperature, stained with

SYBR Green I (Molecular Probes/Thermo Fisher) for

10 min at room temperature, vortexed, and run on a BD

FACSCanto™II Flow Cytometry Analyser Systems

(BD 146 Biosciences) with 2 μm diameter ﬂuorescent

beads (Polysciences, Warminster, PA, USA) as a size

and ﬂuorescence standard. Bacterial cells were detected

at Ex494nm/Em520nm (FITC channel) and by the size of

cell (forward scatter). Phytoplankton cells were identiﬁed

based on their cell chlorophyll (Ex482nm/Em676nm,

PerCP channel) and by the size of cell (forward scatter).

Flow rates were determined several times during each

running session by weighing tubes with double-distilled

water. Finally, the data were analysed with the free soft-

ware ‘Flowing Software’(https://bioscience.ﬁ/).

Bacterial production

The activity of heterotrophic production was measured

using the [4,5-

H]-leucine incorporation method (Simon

et al., 1990). To this end, triplicate 1.7 ml of seawater

samples were collected from each microcosm bottle and

incubated with a 7:1 mixture of ‘cold’leucine and ‘hot’

H-leucine respectively, at a ﬁnal concentration of 100 nmol

leucine L

1

(Perkin Elmer, speciﬁc activity 156 Ci mmol

1

)

for 4 h in the dark under ambient surface seawater temper-

ature (19C). Incubations were stopped by the addition of

100 μl ice-cold 100% trichloroacetic acid. Next, the sam-

ples were brieﬂy spun with a desk-centrifuge and 1 ml of

scintillation cocktail (Ultima-Gold) was added to each vial.

Disintegrations per minute were measured using a TRI-

CARB 2100 TR (Packard) liquid counter. A conversion fac-

tor of 1.5 kg C mol

1

per mol leucine was used with an iso-

tope dilution factor of 2 (Simon and Azam, 1989).

Alkaline phosphatase activity

APA was determined by the 4-methylumbelliferyl phos-

phate (MUF-P: Sigma M8168) method according to

Thingstad and Mantoura (2005). Substrate was added to

triplicate 1 ml water samples (ﬁnal concentration of

50 μM) and incubated in the dark at ambient temperature

for 4 h (same as BP). The increase in ﬂuorescence by

the cleaved 4-methylumbelliferone (MUF) was measured

at 365 nm excitation, 455 nm emissions (GloMax

-Multi

Detection System E9032) and calibrated against a MUF

standard (Sigma M1508).

NanoSIMS analyses

Before analyses, ﬁlter pieces were covered with approxi-

mately 30 nm gold in a sputter coater (Cressington108

auto-sputter coater). SIMS imaging was performed as

described in Eigemann et al.(

2019) using a NanoSIMS

50 L instrument (Cameca, France). The scanning parame-

ters were 512 512 px for areas of 20–30 μm, with a dwell

time of 250 μs per pixel and a primary beam of 1 pA. All

NanoSIMS measurements were analysed with the Matlab

based program look@nanosims (Polerecky et al., 2012).

Brieﬂy, the 60 measured planes were checked for inconsis-

tencies, all usable planes accumulated, regions of interest

(i.e. bacterial cells) deﬁned based on

Nmasspictures

and

Caswellas

N ratios calculated from the

Environmental Microbiology,24, 2467–2483

2478 F. Eigemann et al.

ion signals for each region of interest. For analyses of each

measurement, ﬁrst the means of background measure-

ments were determined (i.e. regions on the ﬁlter without

bacterial cells), and this mean was factorized for theoretical

background values (0.11 for

C and 0.00367 for

N). These factors were applied to all non-background

regions of interest in the same measurement. For each

treatment, measurements of different spots on the same ﬁl-

ter as well as replicate ﬁlters (two replicates for each treat-

ment) were pooled.

RNA extraction, DNA digestion, cDNA synthesis

For RNA extraction, approximately 1.5 L of each incuba-

tion bottle was ﬁltered successively onto 5 and 0.2 μm

pore width polycarbonate ﬁlters (Exp 1) or directly onto

0.2 μmﬁlters (Exp 2). Filters were stored in 1 ml lysis

buffer (40 mM EDTA, 50 mM Tris pH 8.3, 0.75 M

sucrose), ﬂash-frozen in liquid nitrogen and stored at

80C upon extraction, approximately 3 months after the

experiments. For extraction, ﬁlters were cut into half with

scissors and placed in Eppendorf tubes. Then, 1 ml TRI

Reagent was added (to one half of a ﬁlter), ﬁlters were

vortexed and incubated for 10 min at room temperature

on an orbital shaker with 55 rpm. Then, 200 μl of chloro-

form was added, and ﬁlters were vortexed again and

incubated for 15 min at room temperature. Following, the

tubes were centrifuged at 12 000gfor 12 min at 4C, the

supernatant was transferred to fresh tubes, centrifuged

again as described above, and the aqueous phase was

transferred to fresh tubes. Next, 1 ml ice-cold 100% etha-

nol was added, vortexed and incubated for 1 h on ice,

centrifuged at 12 000gfor 10 min, the supernatant was

discarded, 1 ml ice-cold 75% ethanol was added, tubes

were vortexed, and incubated for 10 min at 20C. Last,

tubes were centrifuged for 5 min at 7500g, the ethanol

removed, and the RNA pellet air-dried in a hood. The

resulting pellet was resolved in autoclaved DEPC-treated

water and quality controlled with a Nanodrop device.

Remaining DNA was digested using the Turbo DNA free

kit (Invitrogen) using the manufacturer’s instructions, and

successful digestions were tested by PCRs using primers

com1f and com2rph (Schwieger and Tebbe, 1998), with

initial denaturation at 94C for 3 min, 30 cycles of dena-

turation at 94C for 1 min, annealing for 1 min at 50C

and elongation for 90 s at 72C, ﬁnalized by elongation at

72C for 4 min. RNA was transcribed into cDNA using

MultiScribe reverse transcriptase following the manufac-

turer’s instructions (Invitrogen).

Sequencing

Complementary DNA was PCR ampliﬁed with primers

515F and 926R (Walters et al., 2016) targeting the V4

and V5 variable regions of the microbial small subunit

ribosomal RNA gene using a two-stage ‘targeted

amplicon sequencing’protocol (Naqib et al., 2018).

Primers were modiﬁed to include linker sequences at the

50ends (i.e. so-called ‘common sequences’or CS1 and

CS2 on forward and reverse primers respectively). First

stage PCR ampliﬁcations were performed in 10 μl reac-

tions in 96-well plates, using the MyTaq HS 2

mastermix (BioLine, Taunton, MA, USA). PCR conditions

were 95C for 5 min, followed by 28 cycles of 95C for

3000,50

C for 6000 and 72C for 9000. Subsequently, a sec-

ond PCR ampliﬁcation was performed in 10 μl reactions

in 96-well plates. A mastermix for the entire plate was

made using the MyTaq HS 2mastermix. Each well

received a separate primer pair with a unique 10-base

barcode, obtained from the Access Array Barcode Library

for Illumina (Fluidigm, South San Francisco, CA, USA;

Item# 100-4876). These Access Array primers contained

the CS1 and CS2 linkers at the 30ends of the oligonucle-

otides. Cycling conditions were as follows: 95C for

5 min, followed by 8 cycles of 95C for 3000,60

C for 3000

and 72C for 3000.Aﬁnal, 7-min elongation step was per-

formed at 72C. Samples were pooled in equal volume

using an EpMotion5075 liquid handling robot (Eppendorf,

Hamburg, Germany). The pooled library was puriﬁed

using an AMPure XP cleanup protocol (0.6, vol./vol.;

Agencourt, Beckman-Coulter) to remove fragments

smaller than 300 bp. The pooled libraries, with a 20%

phiX spike-in, were loaded onto an Illumina MiniSeq mid-

output ﬂow cell (2 150 paired-end reads). Based on the

distribution of reads per barcode, the amplicons (before

puriﬁcation) were re-pooled to generate a more balanced

distribution of reads. The re-pooled library was puriﬁed

using AMPure XP cleanup, as described above. Next,

the re-pooled libraries, with a 15% phiX spike-in, were

loaded onto a MiSeq v3 ﬂow cell and sequenced using

an Illumina MiSeq sequencer. Fluidigm sequencing

primers, targeting the CS1 and CS2 linker regions, were

used to initiate sequencing. Library preparation, pooling

and MiniSeq sequencing were performed at the Univer-

sity of Illinois at Chicago Sequencing Core (UICSQC),

Research Resources Center (RRC), University of Illinois

at Chicago (UIC). All forward and backward sequence

reads were deposited at the European Nucleotide

Archive under the accession number PRJEB44710.

Sequence analyses

All sequences were analysed using the Dada2 (Callahan

et al., 2016) pipeline and the software packages R

(R Development Team, 2020) and RStudio (RStudio

Team, 2020). Brieﬂy, forward and backward primers were

trimmed, forward and backward reads truncated after

quality inspections to 280 and 210 bases respectively,

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2479

and after merging of forward and backward sequences, a

consensus length only between 404 and 417 bases was

accepted. For taxonomic assignment, Silva database ver-

sion 138 (Quast et al., 2013) was used. All chloroplasts,

mitochondria, archaea, eukaryotes and amplicon

sequence variants (ASVs) without any taxonomic afﬁlia-

tion were discarded from downstream analyses. The

complete ASV table with absolute read counts,

sequences, sequence lengths as well as metadata for all

samples is accessible as Supplement 9. After inspections

of rarefaction curves, all samples with <2509 reads were

discarded from further analyses, and all remaining sam-

ples subsampled to 2509 reads.

Statistical analyses

Bacterial communities were analysed with Shannon

diversities, non-metric-multidimensional-scaling (NMDS)

and heatmaps of the most abundant ASVs. NMDS was

performed using the ‘metaMDS’command and the Bray

distance in the ‘vegan’package (Oksanen et al., 2013)

in order to analyse differences between the communities

based on the subsampled absolute ASV tables. An

ANOSIM, which tests for signiﬁcant differences between

communities, was conducted using the ‘vegan’pack-

age. Heatmaps were calculated separately for both

experiments using the 40 most abundant ASVs (i.e. the

ASVs with the highest sum of subsampled reads for all

samples in the same experiment) and the heatmap

3 package (Zhao et al., 2014). Homogeneities of vari-

ances were tested by Bartlett or Levene’s (Shannon

diversity) tests. Differences between the different treat-

ments in terms of Shannon diversity, incorporation of

C and

N, cell numbers and BP were tested by ANO-

VAs with subsequent Tukey post hoc tests if homogene-

ity of variances was given. If homogeneity was not

given, Kruskal–Wallis tests were calculated, with subse-

quent Tukey–Nemenyi post hoc tests. For all analyses,

signiﬁcance was assumed for pvalues <0.05. To test for

speciﬁc ASVs associated with treatments, LDA effect

size (LEfSe) analyses (Segata et al., 2011)wereexe-

cuted with the online tool https://huttenhower.sph.

harvard.edu/galaxy. For the open-ocean community,

treatments were assigned as class without subclass,

whereas for the open-ocean community, treatments

were assigned as class, and the fraction (free and

attached) assigned as subclass. For this multi-class

analysis, a ‘one-against-all’strategy was applied. All

analyses, except LEfSe were performed with R

(R Development Team, 2020) and RStudio (RStudio

Team, 2020), and all graphics were executed with the

ggplot2 package (Ginestet, 2011), and reﬁned with the

freeware Inkscape (https://inkscape.org).

Acknowledgements

We thank two reviewers for valuable input, the captain and

crew of the R/V Mediterranean Explorer (EcoOcean) for help

at sea, Mike Krom, Anat Tsemel and Tal Ben-Ezra for the

inorganic nutrient analysis, and Stefan Green (DNA Services

Facility at the University of Illinois at Chicago) for the

amplicon sequencing. We also thank Tom Reich, Dalit Roth-

Rosenberg, Tal Luzzatto-Knaan, Noam Nago and Natalia

Belkin for excellent help with the experiments, and Annett

Grüttmüller for NanoSIMS measurements. This work was

supported by the Human Frontier Science Program (HFSP)

through the grant number RGB 0020/2016 (DS, MV and

HPG), by the National Science Foundation –United

States-Israel Binational Science Foundation Program in

Oceanography (grant number 1635070/2016532 to DS) and

by the Israel Ministry of Science and Technology (grant num-

ber 3-17404 to DS). The experiment at THEMO-2 was per-

formed as part of the SoMMoS (Southeastern Mediterranean

Monthly cruise Series) project, with ship-time funded by the

Leon H. Charney School of Marine Sciences with help from

EcoOcean and IOLR. The NanoSIMS at the Leibniz Institute

for Baltic Sea research in Warnemuende (IOW) was funded

by the German Federal Ministry of Education and Research

(BMBF), grant identiﬁer 03F0626A.

References

Amin, S.A., Hmelo, L.R., van Tol, H.M., Durham, B.P.,

Carlson, L.T., Heal, K.R., et al. (2015) Interaction and sig-

nalling between a cosmopolitan phytoplankton and associ-

ated bacteria. Nature 522:98–101.

Amin, S.A., Parker, M.S., and Armbrust, E.V. (2012) Interac-

tions between diatoms and bacteria. Microbiol Mol Biol

Rev 76: 667–684.

Azam, F., and Malfatti, F. (2007) Microbial structuring of

marine ecosystems. Nat Rev Microbiol 5: 782–791.

Baltar, F., Palovaara, J., Unrein, F., Catala, P., Horˇn

ak, K.,

Šimek, K., et al. (2016) Marine bacterial community struc-

ture resilience to changes in protist predation under phyto-

plankton bloom conditions. ISME J 10: 568–581.

Becker, J.W., Hogle, S.L., Rosendo, K., and Chisholm, S.W.

(2019) Co-culture and biogeography of Prochlorococcus

and SAR11. ISME J 13: 1506–1519.

Beier, S., Rivers, A.R., Moran, M.A., and Obernosterer, I.

(2015) The transcriptional response of prokaryotes to

phytoplankton-derived dissolved organic matter in seawa-

ter. Environ Microbiol 17: 3466–3480.

Beliaev, A.S., Romine, M.F., Serres, M., Bernstein, H.C.,

Linggi, B.E., Markillie, L.M., et al. (2014) Inference of inter-

actions in cyanobacterial-heterotrophic co-cultures via

transcriptome sequencing. ISME J 8: 2243–2255.

Berman, T., and Bronk, D.A. (2003) Dissolved organic nitro-

gen: a dynamic participant in aquatic ecosystems. Aquat

Microb Ecol 31: 279–305.

Bertilsson, S., Berglund, O., Pullin, M., and Chisholm, S.

(2005) Release of dissolved organic matter by

Prochlorococcus. Vie Et Milieu-Life Environ 55: 225–231.

Bertocchi, C., Navarini, L., Cesàro, A., and Anastasio, M.

(1990) Polysaccharides from cyanobacteria. Carbohydr

Polym 12: 127–153.

Environmental Microbiology,24, 2467–2483

2480 F. Eigemann et al.

Bhatnagar, M., and Bhatnagar, A. (2019) Diversity of poly-

saccharides in cyanobacteria. In Microbial Diversity in

Ecosystem Sustainability and Biotechnological Applica-

tions: Volume 1 Microbial Diversity in Normal & Extreme

Environments, Satyanarayana, T., Johri, B.N., and Das, S.

K. (eds). Springer Singapore: Singapore, pp. 447–496.

Bianchi, A., Van Wambeke, F., and Garcin, J. (1998) Bacte-

rial utilization of glucose in the water column from eutro-

phic to oligotrophic pelagic areas in the eastern North

Atlantic Ocean. J Mar Syst 14:45–55.

Biller, S.J., Berube, P.M., Lindell, D., and Chisholm, S.W.

(2015) Prochlorococcus: the structure and function of col-

lective diversity. Nat Rev Microbiol 13:13–27.

Bradley, P.B., Sanderson, M.P., Nejstgaard, J.C., Sazhin, A.

F., Frischer, M.E., Killberg-Thoreson, L.M., et al. (2010)

Nitrogen uptake by phytoplankton and bacteria during an

induced Phaeocystis pouchetii bloom, measured using

size fractionation and ﬂow cytometric sorting. Aquat

Microb Ecol 61:89–104.

Buchan, A., LeCleir, G.R., Gulvik, C.A., and Gonzalez, J.M.

(2014) Master recyclers: features and functions of bacteria

associated with phytoplankton blooms. Nat Rev Microbiol

12: 686–698.

Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W.,

Johnson, A.J.A., and Holmes, S.P. (2016) DADA2: high-

resolution sample inference from Illumina amplicon data.

Nat Methods 13: 581–583.

Carlson, C.A., and Ducklow, H.W. (1996) Growth of bacter-

ioplankton and consumption of dissolved organic carbon

in the Sargasso Sea. Aquat Microb Ecol 10:69–85.

Carlson, C.A., Giovannoni, S.J., Hansell, D.A., Goldberg, S.

J., Parsons, R., and Vergin, K. (2004) Interactions among

dissolved organic carbon, microbial processes, and com-

munity structure in the mesopelagic zone of the northwest-

ern Sargasso Sea. Limnol Oceanogr 49: 1073–1083.

Christie-Oleza, J.A., Sousoni, D., Lloyd, M., Armengaud, J.,

and Scanlan, D.J. (2017) Nutrient recycling facilitates

long-term stability of marine microbial phototroph-

heterotroph interactions. Nat Microbiol 2: 17100.

Cole, J.J. (1982) Interactions between bacteria and algae in

aquatic ecosystems. Annu Rev Ecol Syst 13: 291–314.

Eigemann, F., Vogts, A., Voss, M., Zoccarato, L., and

Schulz-Vogt, H. (2019) Distinctive tasks of different cyano-

bacteria and associated bacteria in carbon as well as

nitrogen ﬁxation and cycling in a late stage Baltic Sea

bloom. PLoS One 14: e0223294.

Eilers, H., Pernthaler, J., Glöckner, F.O., and Amann, R.

(2000) Culturability and in situ abundance of pelagic bac-

teria from the North Sea. Appl Environ Microbiol 66:

3044–3051.

Field, C.B., Behrenfeld, M.J., Randerson, J.T., and

Falkowski, P. (1998) Primary production of the biosphere:

integrating terrestrial and oceanic components. Science

281: 237–240.

Flombaum, P., Gallegos, J.L., Gordillo, R.A., Rinc

on, J.,

Zabala, L.L., Jiao, N., et al. (2013) Present and future

global distributions of the marine Cyanobacteria

Prochlorococcus and Synechococcus. Proc Natl Acad Sci

USA110: 9824–9829.

Fouilland, E., Tolosa, I., Bonnet, D., Bouvier, C., Bouvier, T.,

Bouvy, M., et al. (2014) Bacterial carbon dependence on

freshly produced phytoplankton exudates under different

nutrient availability and grazing pressure conditions in

coastal marine waters. FEMS Microbiol Ecol 87: 757–769.

Ginestet, C. (2011) ggplot2: elegant graphics for data analy-

sis. J R Stat Soc A Stat Soc 174: 245–246.

Gobet, A., Barbeyron, T., Matard-Mann, M., Magdelenat, G.,

Vallenet, D., Duchaud, E., and Michel, G. (2018) Evolu-

tionary evidence of algal polysaccharide degradation

acquisition by Pseudoalteromonas carrageenovora 9(T) to

adapt to macroalgal niches. Front Microbiol 9: 2740.

Grossart, H.-P. (1999) Interactions between marine bacteria

and axenic diatoms (Cylindrotheca fusiformis, Nitzschia

laevis, and Thalassiosira weissﬂogii) incubated under vari-

ous conditions in the lab. Aquat Microb Ecol 19:1–11.

Grossart, H.-P., and Simon, M. (2007) Interactions of plank-

tonic algae and bacteria: effects on algal growth and organic

matter dynamics. Aquat Microb Ecol 47: 163–176.

Grossart, H.P., Tang, K.W., Kiorboe, T., and Ploug, H.

(2007) Comparison of cell-speciﬁc activity between free-

living and attached bacteria using isolates and natural

assemblages. FEMS Microbiol Lett 266: 194–200.

Grossowicz, M., Roth-Rosenberg, D., Aharonovich, D.,

Silverman, J., Follows, M.J., and Sher, D. (2017)

Prochlorococcus in the lab and in silico: the importance of

representing exudation. Limnol Oceanogr 62: 818–835.

Guillemette, F., Leigh McCallister, S., and del Giorgio, P.A.

(2016) Selective consumption and metabolic allocation of

terrestrial and algal carbon determine allochthony in lake

bacteria. ISME J 10: 1373–1382.

Hazan, O., Silverman, J., Sisma-Ventura, G., Ozer, T.,

Gertman, I., Shoham-Frider, E., et al. (2018) Mesopelagic

prokaryotes alter surface phytoplankton production during

simulated deep mixing experiments in eastern Mediterra-

nean Sea waters. Front Mar Sci 5:1.

Huang, Y., Nicholson, D., Huang, B., and Cassar, N. (2021)

Global estimates of marine gross primary production

based on machine learning upscaling of ﬁeld observa-

tions. Global Biogeochem Cycles 35: e2020GB006718.

https://doi.org/10.1029/2020GB006718

Johnson, Z.I., Zinser, E.R., Coe, A., McNulty, N.P.,

Woodward, E.M.S., and Chisholm, S.W. (2006) Niche par-

titioning among Prochlorococcus ecotypes along ocean-

scale environmental gradients. Science 311: 1737–1740.

Karlson, A.M., Duberg, J., Motwani, N.H., Hogfors, H.,

Klawonn, I., Ploug, H., et al. (2015) Nitrogen ﬁxation by

cyanobacteria stimulates production in Baltic food webs.

Ambio 44: 413–426.

Klippel, B., Lochner, A., Bruce, D.C., Davenport, K.W.,

Detter, C., Goodwin, L.A., et al. (2011) Complete genome

sequence of the marine cellulose- and xylan-degrading

bacterium Glaciecola sp. strain 4H-3-7+YE-5. J Bacteriol

193: 4547–4548.

Krom, M., Kress, N., Berman-Frank, I., and Rahav, E. (2014)

Past, present and future patterns in the nutrient chemistry

of the eastern Mediterranean. In The Mediterranean Sea:

Its History and Present Challenges, Goffredo, S., and

Dubinsky, Z. (eds). Dordrecht: Springer Netherlands,

pp. 49–68.

Krom, M.D., Emeis, K.C., and Van Cappellen, P. (2010) Why

is the eastern Mediterranean phosphorus limited? Prog

Oceanogr 85: 236–244.

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2481

Lebaron, P., Servais, P., Troussellier, M., Courties, C.,

Muyzer, G., Bernard, L., et al. (2001) Microbial community

dynamics in Mediterranean nutrient-enriched seawater

mesocosms: changes in abundances, activity and compo-

sition. FEMS Microbiol Ecol 34: 255–266.

Lidbury, I.D., Murrell, J.C., and Chen, Y. (2015)

Trimethylamine and trimethylamine N-oxide are supple-

mentary energy sources for a marine heterotrophic bacte-

rium: implications for marine carbon and nitrogen cycling.

ISME J 9: 760–769.

Liefer, J.D., Garg, A., Fyfe, M.H., Irwin, A.J., Benner, I.,

Brown, C.M., et al. (2019) The macromolecular basis of

phytoplankton C:N:P under nitrogen starvation. Front

Microbiol 10: 763.

Livanou, E., Lagaria, A., Psarra, S., and Lika, K. (2017) Dis-

solved organic matter release by phytoplankton in the con-

text of the Dynamic Energy Budget theory.

Biogeosciences 42:1–33.

Lombard, V., Golaconda Ramulu, H., Drula, E., Coutinho, P.

M., and Henrissat, B. (2014) The carbohydrate-active

enzymes database (CAZy) in 2013. Nucleic Acids Res

42: D490–D495.

McCarren, J., Becker, J.W., Repeta, D.J., Shi, Y., Young, C.

R., Malmstrom, R.R., et al. (2010) Microbial community

transcriptomes reveal microbes and metabolic pathways

associated with dissolved organic matter turnover in the

sea. Proc Natl Acad Sci U S A 107: 16420–16427.

Mella-Flores, D., Mazard, S., Humily, F., Partensky, F.,

Mahé, F., Bariat, L., et al. (2011) Is the distribution of

Prochlorococcus and Synechococcus ecotypes in the

Mediterranean Sea affected by global warming? Bio-

geosciences 8: 2785–2804.

Meon, B., and Kirchman, D.L. (2001) Dynamics and molecu-

lar composition of dissolved organic material during exper-

imental phytoplankton blooms. Mar Chem 75: 185–199.

Moore, C.M., Mills, M.M., Arrigo, K.R., Berman-Frank, I.,

Bopp, L., Boyd, P.W., et al. (2013) Processes and pat-

terns of oceanic nutrient limitation. Nat Geosci 6:

701–710.

Mor

an, X.A., Estrada, M., Gasol, J.M., and Pedr

os-Ali

o, C.

(2002a) Dissolved primary production and the strength of

phytoplankton- bacterioplankton coupling in contrasting

marine regions. Microb Ecol 44: 217–223.

Mor

an, X.A.G., Josep, M.G., Pedros-Alio, C., and Marta, E.

(2002b) Partitioning of phytoplanktonic organic carbon

production and bacterial production along a coastal-

offshore gradient in the NE Atlantic during different hydro-

graphic regimes. Aquat Microb Ecol 29: 239–252.

Mühlenbruch, M., Grossart, H.P., Eigemann, F., and

Voss, M. (2018) Mini-review: phytoplankton-derived poly-

saccharides in the marine environment and their interac-

tions with heterotrophic bacteria. Environ Microbiol 20:

2671–2685.

Naqib, A., Poggi, S., Wang, W., Hyde, M., Kunstman, K.,

and Green, S.J. (2018) Making and sequencing heavily

multiplexed, high-throughput 16S ribosomal RNA gene

amplicon libraries using a ﬂexible, two-stage PCR proto-

col. Methods Mol Biol 1783: 149–169.

Nemergut, D.R., Schmidt, S.K., Fukami, T., O’Neill, S.P.,

Bilinski, T.M., Stanish, L.F., et al. (2013) Patterns and

processes of microbial community assembly. Microbiol

Mol Biol Rev 77: 342–356.

Ofaim, S., Sulheim, S., Almaas, E., Sher, D., and Segrè, D.

(2021) Dynamic allocation of carbon storage and nutrient-

dependent exudation in a revised genome-scale model of

Prochlorococcus. Front Genet 12: 586293.

Oksanen, J., Kindt, R., Legendre, P., O’Hara, B.,

Simpson, G.L., Solymos, P., et al. (2013). vegan: Commu-

nity ecology package, R package version 2(0). http://cran.

r-project.org/,http://vegan.r-forge.r-project.org/.

Partensky, F., Hess, W.R., and Vaulot, D. (1999)

Prochlorococcus, a marine photosynthetic prokaryote of

global signiﬁcance. Microbiol Mol Biol Rev 63: 106–127.

Pedler, B.E., Aluwihare, L.I., and Azam, F. (2014) Single

bacterial strain capable of signiﬁcant contribution to car-

bon cycling in the surface ocean. Proc Natl Acad Sci U S

A111: 7202–7207.

Pheng, S., Ayyadurai, N., Park, A.Y., and Kim, S.G. (2017)

Psychrosphaera aquimarina sp. nov., a marine bacterium

isolated from seawater collected from Asan Bay, Republic

of Korea. Int J Syst Evol Microbiol 67: 4820–4824.

Piontek, J., Handel, N., De Bodt, C., Harlay, J., Chou, L.,

and Engel, A. (2011) The utilization of polysaccharides by

heterotrophic bacterioplankton in the Bay of Biscay (North

Atlantic Ocean). J Plankton Res 33: 1719–1735.

Polerecky, L., Adam, B., Milucka, J., Musat, N., Vagner, T.,

and Kuypers, M.M. (2012) Look@NanoSIMS –a tool for

the analysis of nanoSIMS data in environmental microbiol-

ogy. Environ Microbiol 14: 1009–1023.

Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T.,

Yarza, P., et al. (2013) The SILVA ribosomal RNA gene

database project: improved data processing and web-

based tools. Nucleic Acids Res 41: D590–D596.

R Development Team. (2020) R: A Language and Environ-

ment for Statistical Computing. Vienna, Austria: R Foun-

dation for Statistical Computing. https://www.R-project.

org/.

Raveh, O., David, N., Rilov, G., and Rahav, E. (2015) The

temporal dynamics of coastal phytoplankton and bacter-

ioplankton in the eastern Mediterranean Sea. PLoS One

10: e0140690.

Reich, T., Ben-Ezra, T., Belkin, N., Tsemel, A.,

Aharonovich, D., Roth-Rosenberg, D., et al. (2021) Sea-

sonal dynamics of phytoplankton and bacterioplankton at

the ultra-oligotrophic southeastern Mediterranean Sea.

bioRxiv.

Ribalet, F., Swalwell, J., Clayton, S., Jiménez, V., Sudek, S.,

Lin, Y., et al. (2015) Light-driven synchrony of

Prochlorococcus growth and mortality in the subtropical

Paciﬁc gyre. Proc Natl Acad Sci U S A 112: 8008–8012.

Riemann, L., Holmfeldt, K., and Titelman, J. (2009) Impor-

tance of viral lysis and dissolved DNA for bacterioplankton

activity in a P-limited estuary, northern Baltic Sea. Microb

Ecol 57: 286–294.

Romera-Castillo, C., Sarmento, H., Alvarez-Salgado, X.A.,

Gasol, J.M., and Marrasé, C. (2011) Net production and

consumption of ﬂuorescent colored dissolved organic mat-

ter by natural bacterial assemblages growing on marine

phytoplankton exudates. Appl Environ Microbiol 77:

7490–7498.

Environmental Microbiology,24, 2467–2483

2482 F. Eigemann et al.

Rooney-Varga, J.N., Giewat, M.W., Savin, M.C., Sood, S.,

LeGresley, M., and Martin, J.L. (2005) Links between phy-

toplankton and bacterial community dynamics in a coastal

marine environment. Microb Ecol 49: 163–175.

Roth-Rosenberg, D., Aharonovich, D., Omta, A.W.,

Follows, M.J., and Sher, D. (2021a) Dynamic macromo-

lecular composition and high exudation rates in

Prochlorococcus. Limnol Oceanogr 66: 1759–1773.

Roth-Rosenberg, D., Haber, M., Goldford, J., Lalzar, M.,

Aharonovich, D., Al-Ashhab, A., et al. (2021b) Particle-

associated and free-living bacterial communities in an oli-

gotrophic sea are affected by different environmental fac-

tors. Environ Microbiol 23: 4295–4308.

RStudio Team. (2020) RStudio: Integrated Development for R.

Boston, MA. URL: RStudio, PBC. https://www.rstudio.com/.

Saito, M.A., McIlvin, M.R., Moran, D.M., Goepfert, T.J.,

DiTullio, G.R., Post, A.F., and Lamborg, C.H. (2014) Multi-

ple nutrient stresses at intersecting Paciﬁc Ocean biomes

detected by protein biomarkers. Science 345: 1173–1177.

Sarmento, H., and Gasol, J.M. (2012) Use of phytoplankton-

derived dissolved organic carbon by different types of

bacterioplankton. Environ Microbiol 14: 2348–2360.

Sarmento, H., Morana, C., and Gasol, J.M. (2016) Bacter-

ioplankton niche partitioning in the use of phytoplankton-

derived dissolved organic carbon: quantity is more impor-

tant than quality. ISME J 10: 2582–2592.

Sarmento, H., Romera-Castillo, C., Lindh, M., Pinhassi, J.,

Sala, M.M., Gasol, J.M., et al. (2013) Phytoplankton

species-speciﬁc release of dissolved free amino acids and

their selective consumption by bacteria. Limnol Oceanogr

58: 1123–1135.

Schwieger, F., and Tebbe, C.C. (1998) A new approach to

utilize PCR-single-strand-conformation polymorphism for

16S rRNA gene-based microbial community analysis.

Appl Environ Microbiol 64: 4870–4876.

Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L.,

Garrett, W.S., and Huttenhower, C. (2011) Metagenomic

biomarker discovery and explanation. Genome Biol

12: R60.

Seymour, J.R., Ahmed, T., Durham, W.M., and Stocker, R.

(2010) Chemotactic response of marine bacteria to the

extracellular products of Synechococcus and

Prochlorococcus. Aquat Microb Ecol 59: 161–168.

Seymour, J.R., Ahmed, T., and Stocker, R. (2009) Bacterial

chemotaxis towards the extracellular products of the toxic

phytoplankton Heterosigma akashiwo. J Plankton Res 31:

1557–1561.

Seymour, J.R., Amin, S.A., Raina, J.B., and Stocker, R.

(2017) Zooming in on the phycosphere: the ecological

interface for phytoplankton-bacteria relationships. Nat

Microbiol 2: 17065.

Sharma, A.K., Becker, J.W., Ottesen, E.A., Bryant, J.A.,

Duhamel, S., Karl, D.M., et al. (2014) Distinct dissolved

organic matter sources induce rapid transcriptional

responses in coexisting populations of Prochlorococcus,

Pelagibacter and the OM60 clade. Environ Microbiol 16:

2815–2830.

Shaw, S.L., Chisholm, S.W., and Prinn, R.G. (2003) Iso-

prene production by Prochlorococcus, a marine

cyanobacterium, and other phytoplankton. Mar Chem 80:

227–245. https://doi.org/10.1016/S0304-4203(02)00101-9

Simon, M., Alldredge, A.L., and Azam, F. (1990) Bacterial

carbon dynamics on marine snow. Mar Ecol Prog Ser 65:

205–211.

Simon, M., and Azam, F. (1989) Protein content and protein

synthesis rates of planktonic marine bacteria. Mar Ecol

Prog Ser 51: 201–213.

Sisma-Ventura, G., and Rahav, E. (2019) DOP stimulates

heterotrophic bacterial production in the oligotrophic

southeastern Mediterranean coastal waters. Front

Microbiol 10: 1913.

Skoog, A., Whitehead, K., Sperling, F., and Junge, K. (2002)

Microbial glucose uptake and growth along a horizontal

nutrient gradient in the North Paciﬁc. Limnol Oceanogr 47:

1676–1683.

Taylor, J.D., and Cunliffe, M. (2017) Coastal bacter-

ioplankton community response to diatom-derived poly-

saccharide microgels. Environ Microbiol Rep 9: 151–157.

Teeling, H., Fuchs, B.M., Becher, D., Klockow, C.,

Gardebrecht, A., Bennke, C.M., et al. (2012) Substrate-

controlled succession of marine bacterioplankton

populations induced by a phytoplankton bloom. Science

336: 608–611.

Teeling, H., Fuchs, B.M., Bennke, C.M., Kruger, K.,

Chafee, M., Kappelmann, L., et al. (2016) Recurring pat-

terns in bacterioplankton dynamics during coastal spring

algae blooms. Elife 5: e11888.

Thingstad, T.F., and Mantoura, R.F.C. (2005) Titrating

excess nitrogen content of phosphorous-deﬁcient eastern

Mediterranean surface water using alkaline phosphatase

activity as a bio-indicator. Limnol Oceanogr Methods 3:

94–100.

Thornton, D.C.O. (2014) Dissolved organic matter (DOM)

release by phytoplankton in the contemporary and future

ocean. Eur J Phycol 49:20–46.

Walters, W., Hyde, E.R., Berg-Lyons, D., Ackermann, G.,

Humphrey, G., Parada, A., et al. (2016) Improved bacterial

16S rRNA gene (V4 and V4-5) and fungal internal tran-

scribed spacer marker gene primers for microbial commu-

nity surveys. mSystems 1: e00009-15.

Wu, X., Wang, Y., Zhu, Y., Tian, H., Qin, X., Cui, C., et al.

(2019) Variability in the response of bacterial community

assembly to environmental selection and biotic factors

depends on the immigrated bacteria, as revealed by a soil

microcosm experiment. mSystems 4: e00496-00419.

York, A. (2018) Marine biogeochemical cycles in a changing

world. Nat Rev Microbiol 16: 259–259.

Zhao, S., Guo, Y., Sheng, Q., and Shyr, Y. (2014)

Heatmap3: an improved heatmap package with more

powerful and convenient features. BMC Bioinformatics

15: P16. https://doi.org/10.1186/1471-2105-15-S10-P16

Supporting Information

Additional Supporting Information may be found in the online

version of this article at the publisher’s web-site:

Appendix S1: Supporting Information.

Environmental Microbiology,24, 2467–2483

Phytoplankton exudates provide full nutrition to bacteria 2483