Content uploaded by Esther R Angert
Author content
All content in this area was uploaded by Esther R Angert on Jun 27, 2022
Content may be subject to copyright.
MICROBIOLOGY
A centimeter-long bacterium with DNA contained in
metabolically active, membrane-bound organelles
Jean-Marie Volland
1,2
*†, Silvina Gonzalez-Rizzo
3
†,OlivierGros
3,4
*†,TomášTyml
1,2
,
Natalia Ivanova
1
, Frederik Schulz
1
, Danielle Goudeau
1
, Nathalie H. Elisabeth
5
, Nandita Nath
1
,
Daniel Udwary
1
,RexR.Malmstrom
1
, Chantal Guidi-Rontani
6
, Susanne Bolte-Kluge
7
,
Karen M. Davies
5,8
‡, Maïtena R. Jean
3
, Jean-Louis Mansot
4
,NigelJ.Mouncey
1
, Esther R. Angert
9
,
Tanja Woyke
1,2,10
*, Shailesh V. Date
2,11,12
*
Cells of most bacterial species are around 2 micrometers in length, with some of the largest specimens
reaching 750 micrometers. Using fluorescence, x-ray, and electron microscopy in conjunction with
genome sequencing, we characterized Candidatus (Ca.) Thiomargarita magnifica, a bacterium that has
an average cell length greater than 9000 micrometers and is visible to the naked eye. These cells
grow orders of magnitude over theoretical limits for bacterial cell size, display unprecedented polyploidy
of more than half a million copies of a very large genome, and undergo a dimorphic life cycle
with asymmetric segregation of chromosomes into daughter cells. These features, along with
compartmentalization of genomic material and ribosomes in translationally active organelles bound
by bioenergetic membranes, indicate gain of complexity in the Thiomargarita lineage and challenge
traditional concepts of bacterial cells.
Bacteria and archaea are taxonomically
and metabolically the most diverse and
abundant organisms on Earth, but with
only a small fraction of them isolated
in culture, we remain grossly ignorant
of their biology (1). Although most model
bacteria and archaea are small, some remark-
ably large cells, referred to as giant bacteria,
are evident in atleast four phyla (2), and have
cellular sizes in the range of tens or even hun-
dreds of micrometers (3,4). Some exceptional
members of the sulfur-oxidizing gammapro-
teobacteria Thiomargarita namibiensis,for
instance, are known to reach up to 750 mm
(average size: 180 mm) (4–6). Such bacterial
giants raise the question of whether other
lineages of previously unidentified macro-
bacteria might exist.
Here,wedescribeasessilefilamentous
Thiomargarita species from a marine sulfidic
environment that is larger than all other known
giant bacteria by ~50-fold. Our multifaceted
imaging analyses revealed massive polyploidy
and a dimorphic developmental cycle in which
genome copies are asymmetrically segregated
into apparent dispersive daughter cells. We
show that centimeter-long Thiomargarita fila-
ments represent individual cells with genetic
material and ribosomes compartmentalized
into a metabolically active, membrane-bound
organelle. Sequencing and analysis of genomes
from five single cells revealed insights into dis-
tinct cell division and cell elongation mecha-
nisms. These cellular features likely allow the
organism to grow to an unusually large size
and circumvent some of the biophysical and
bioenergetic limitations on growth. In refer-
ence to its exceptional size, we propose to
name this species Thiomargarita magnifica,
which is hereafter referred to as Candidatus
(Ca.) Thiomargarita magnifica.
Ca. T. magnifica is a centimeter-long, single
bacterial cell
Some large sulfur bacteria (LSB) form very long
filaments that may reach several centimeters in
length but are composed of thousands of indi-
vidual cells that do not exceed 200 mm(7–10).
We observed seasonal “bouquets”of centimeter-
long white filamentous Thiomargarita cells
attached to sunken leaves of Rhizophora
mangle (fig. S1) in shallow tropical marine
mangroves from Guadeloupe, Lesser Antilles.
Thiomargarita spp. are sulfur-oxidizing gam-
maproteobacteria known to be morpholog-
ically diverse and to display polyphenism (11).
The morphology of the filaments observed
in Guadeloupe resembled those of sessile
Thiomargarita-like cells reported from deep-
sea methane seeps (12). They had a stalk-like
shape for most of their length and constricted
gradually toward the apical end, forming buds
(Fig.1,AtoE).Incontrasttorelativesthat
live buried in sediment, these filaments were
smooth in appearance and free of epibiotic
bacteria or any extracellular mucus matrix
(figs. S1 and S2) (11). Budding filaments had
an average length of 9.72 ± 4.25 mm, and
only the most apical constrictions closed
completely to form one to four separate, rod-
sha ped ce lls of 0.21 ± 0.05 mm. We also
noted that some filaments reached a length of
20.00 mm (Fig. 1A and figs. S1 and S3), much
larger than any previously described single-
celled prokaryote.
To further characterize Ca. T. magnifica cells,
we highlighted membranes using osmium te-
troxide or the fluorescent dye FM 1-43X, and
visualized entire filaments in three dimensions
(3D) with hard x-ray tomography (n=4;Fig.1C
and movies S1 and S2) and confocal laser scan-
ning microscopy (CLSM) (n= 6; Fig. 1, C and
D, and movie S3). Filament sections (up to
850.6 mm in length) were visualized with trans-
mission electron microscopy (TEM) (n=15;
Fig. 1, E to G, and fig. S5). All techniques
consistently showed that each filament was
one continuous cell for nearly its entire length,
with no division septa, including the partial
constrictions toward the apical pole. Only the
most apical few buds were separated from the
filament by a closed constriction and these
represented daughter cells (Fig. 1, fig. S5, and
movies S1 to S4).
Similar to other LSB, Ca. T. magnifica cells
has a large central vacuole that reduces the
cytoplasmatic space. This may minimize growth
limitation due to the reliance on chemical diffu-
sion because bacteria lack an active intracellu-
lar transport system (2,4). In Ca.T.magnifica,
the central vacuole was continuous along the
whole filament and accounted for 73.2 ± 7.5%
(n= 4) of the total volume (Fig. 1, D and E; fig.
S5; and table S2). The cytoplasm was 3.34 ±
1.48 mm thick and was constrained to the pe-
riphery of the cell, so it was preserved from
such chemical diffusion limitations (Fig. 1, E
and F, and fig. S5) (4).
Within the cytoplasm, TEM revealed numer-
ous lucent vesicles 2.40 ± 1.03 mmindiameter,
which corresponded to the refractile gran-
ules observed with bright-field microscopy
and represented sulfur granules, as shown
by energy dispersive x-ray spectroscopy (Fig.
1F, figs. S5 and S13, and supplementary text).
The cytoplasm of Ca. T. magnifica appeared
to contain many electron-dense membrane-
bound compartments 1.31 ± 0.70 mmindiam-
eter (Fig. 1, F and G). Similar structures have
RESEARCH
Volland et al., Science 376, 1453–1458 (2022) 24 June 2022 1of6
1
Department of Energy Joint Genome Institute, Lawrence
Berkeley National Laboratory, Berkeley, CA, USA.
2
Laboratory
for Research in Complex Systems, Menlo Park, CA, USA.
3
Institut de Systématique, Evolution, Biodiversité, Université
des Antilles, Muséum National d'Histoire Naturelle, CNRS,
Sorbonne Université, EPHE, Campus de Fouillole, Pointe-à-
Pitre, France.
4
Centre Commun de Caractérisation des
Matériaux des Antilles et de la Guyane, Université des
Antilles, UFR des Sciences Exactes et Naturelles, Pointe-à-
Pitre, Guadeloupe, France.
5
Department of Energy Molecular
Biophysics and Integrated Bioimaging, Lawrence Berkeley
National Laboratory, Berkeley, CA, USA.
6
Institut de
Systématique, Evolution, Biodiversité CNRS UMR 7205,
Museum National d’Histoire Naturelle, Paris, France.
7
Sorbonne Universités, UPMC Univ. Paris 06, CNRS FRE3631,
Institut de Biologie Paris Seine, Paris, France.
8
Department
of Molecular and Cell Biology, University of California,
Berkeley, USA.
9
Cornell University, College of Agriculture and
Life Sciences, Department of Microbiology, Ithaca, NY, USA.
10
University of California Merced, School of Natural
Sciences, Merced, CA, USA.
11
University of California
San Francisco, San Francisco, CA, USA.
12
San Francisco
State University, San Francisco, CA, USA.
*Corresponding author. Email: jvolland@lbl.gov (J.-M.V.);
Olivier.Gros@univ-antilles.fr (O.G.); twoyke@lbl.gov (T.W.);
shailesh.date@lrc.systems (S.V.D.)
†These authors contributed equally to this work.
‡Present address: Electron Bio-Imaging Centre, Diamond
Light Source, Harwell Science and Innovation Campus,
Didcot, UK.
Downloaded from https://www.science.org at Cornell University on June 23, 2022
occasionally been observed in other LSB and
have been referred to as “blebs of cytoplasm,”
“putative endobionts,”“intracytoplasmic
structures which appear to contain nuclear
material,”or “membrane-enclosed cytoplas-
mic compartments with ribosomes and DNA
fibrils,”but their nature and function re-
mained elusive in these primarily ultrastruc-
tural studies (7,10,13). We hypothesized that
some of these membrane-bound compart-
ments within the cytoplasm may contain
dispersed genomic material because poly-
ploidy is evident in many giant bacteria
(2,14,15).
DNA and ribosomes within a membrane-bound
bacterial organelle
Although bacteria were once presumed to be
uncompartmentalized “bags of enzymes,”re-
cent studies have revealed the presence of or-
ganelles with functions as diverse as anaerobic
ammonium oxidation, photosynthesis, and
magnetic orientation (16). No bacteria or
archaea are known to unambiguously segre-
gate their genetic material in the manner of
eukaryotes. There is, however, some evidence
of membrane-bound nucleoids in one mem-
ber of the Atribacteria, which has a compart-
ment containing DNA that occupies most of
the cell’svolume(17). Plancotomycetes such
as Gemmata obscuriglobus also have cytosolic
Volland et al., Science 376, 1453–1458 (2022) 24 June 2022 2of6
Fig. 1. Morphology and ultrastructure of
Ca. T. magnifica. (A) Size comparison of selected
bacterial (green) and eukaryotic (blue) model
systems on a log scale. (B) Light microscopy
montage of the upper half of a Ca. T. magnifica
cell, with a broken basal part revealing a tube-like
morphology due to the large central vacuole
and numerous spherical intracellular sulfur
granules (a tardigrade is shown for scale). (C)3D
rendering of segmented cells from hard x-ray
tomography (movies S1, S2, and S6) and CLSM
(movie S3) putatively at various stages of the
developmental cycle. From left to right, 3D
rendered cells D,B,F,G,andD(table S2). Note
that the smallest stage corresponds to the
cell Dterminal segment and was added to the left
for visualization purposes. (D)CLSMobservation
of cell K(table S2) after fluorescent labeling
of membranes with FM 1-43X showing the
continuity of the cell from the basal pole to the
first complete constriction at the apical end.
(E) TEM montage of the apical constriction
of a cell, with the cytoplasm constrained to
the periphery. (F) Higher magnification of the
area marked in (E), with sulfur granules and
pepins at various stages of development.
(G) Higher magnification of the area marked
in (F) showing two pepins (arrowheads).
S, sulfur granule; V, vacuole.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Cornell University on June 23, 2022
compartments with DNA, and recent work
has shown that these compartments repre-
sent deep invaginations of a Gram-negative
cytoplasmic membrane rather than a closed,
membrane-bound organelle (16).
4′,6-Diamidino-2-phenylindole (DAPI) stain-
ing revealed that the DNA in Ca.T.magnifica
cells was concentrated in the membrane-bound
compartments(Fig.2,HtoJ,andfig.S6).Al-
though we did not observe any connection of
these compartments to the cell envelope, their
mechanism of formation, which may include
cell membrane invagination, is yet to be studied.
They also harbored electron-dense structures
10 to 20 nm in size, similar to signatures of
Volland et al., Science 376, 1453–1458 (2022) 24 June 2022 3of6
Fig. 2. Characterization of the pepin organelles
by FISH and correlative TEM, as well as
membrane and DNA staining, immunohisto-
chemistry, and BONCAT. (Ato D) FISH of pepins
(arrows) in the cytoplasm of Ca. T. magnifica
(class gammaproteobacteria). Pepins are labeled
with the general bacterial probe EUB labeled with
Alexa Fluor 488 [(A), green], the gammaproteo-
bacteria-specific probe Gam42a labeled with
Cy3 [(B), yellow], and the Thiomargarita-specific
probe Thm482 labeled with Cy5 [(C), red] and with
DAPI [(D), blue] (see the supplementary text
for details). (E) TEM of a serial thin section
consecutive to the semithin section used for FISH.
The FISH- and DAPI-positive pepins appear as
electron-dense organelles under TEM. (Fand
G) Pepins from (E) under higher magnification.
Pepins are delimited by a membrane (arrowheads)
and contain numerous ribosomes that appear as
small, electron-dense granules throughout the
sections of the pepins. (Hto J) Fluorescent
labeling of membranes using FM 1-43X (H) and of
DNA using DAPI (I) and overlay (J) on a cross
section of a cell. The pepins labeled with DAPI are
also labeled with the dye FM 1-43X, confirming
the presence of a membrane. (Kto M)Visualization
of ATP synthase localization using immunohisto-
chemistry. Anti–ATP synthase antibodies label (K)
reveals the presence of bioenergetic membranes
around pepins [labeled with DAPI in (L)] and
throughout the cytoplasm. (N) 3D visualization of
a central portion of a cell after DAPI staining
(blue) showing the multitude of DNA clusters
spread throughout the cytoplasm (cell M; table S2
and movie S5). (Oto Q) Translational activity
revealed by BONCAT showing active protein
biosynthesis within an entire cell, including hot-
spots at constriction sites and pepins [enlarged
in (P) and (Q), respectively].
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Cornell University on June 23, 2022
ribosomes (Fig. 2, F and G, and fig. S6, G to K).
Fluorescence in situ hybridization (FISH) with
probes specifically targeting ribosomal RNA
sequences of Thiomargarita confirmed that
ribosomes were indeed present and were con-
centrated in these membrane-bound structures
(Fig. 2, A to G, and figs. S7 to S9) and spread
throughout the entire cell, including the apical
buds (fig. S9). This compartmentalization of DNA
and ribosomes is reminiscent of the genomic
nuclear compartmentalization in eukaryotes. By
analogy with pips, the numerous small seeds in
fruits such as watermelon or kiwi, we propose to
name this bacterial organelle a pepin (singular
pepin,pluralpepins:fromvulgarLatinpép,an
expressive creation used to express smallness).
Cytoplasmic localization of bioenergetic
membranes and translational activity
Adenosine triphosphate (ATP), the principal
and universal energy currency of cells, is pro-
duced by ATP synthase, a molecular machine
embedded in bioenergetic membranes and
powered by proton motive force. Bacterial and
archaeal ATP synthases are often observed
to be localized to the cell envelope, in contrast
to eukaryotes, in which ATP is generated by
mitochondria. This cell envelope localization
potentially constrains bacterial cell size be-
cause of the surface-to-volume ratio required
to satisfy energy needs; a theoretical maxi-
mum cell size of ~10
−14
m
3
has been estimated
recently (18). Giant Thiomargarita cells are
often cited as being exceptions to such ener-
getic constraints and are thought to rely on
their cell surface for ATP production (19–21).
We used immunohistochemistry to assess the
localization of ATP synthase in Ca. T. magnifica.
We observed the distribution of ATP synthases
around pepins and throughout the complex
membrane network of the entire cytoplasm,
but they were absent from the outer cell en-
velope (Fig. 2, K to M, and fig. S20).
To examine the localization of activity
throughout the whole cell, we used bioor-
thogonal noncanonical amino acid tagging
(BONCAT) to detect protein biosynthesis (22).
Live fil ament s in cubated with a clickable
amino acid analog were labeled throughout
(Fig. 2O and fig. S10). Consistent with the
detection of ribosomes by FISH and TEM,
BONCAT showed protein biosynthesis activ-
ity in small, round-shaped areas that were
similar to pepins in size and localization, but
not all pepins appeared to be labeled (Fig.
2Q and fig. S10D). Labeled hotspots were also
observed at the site of constriction in the
apical part of cells, suggesting higher trans-
lational activity or concentration of newly
synthesized proteins in these areas (Fig. 2P).
The establishment of stable continuous lab-
oratory cultures of Ca. T. magnifica will likely be
necessary to undertake detailed st ud ie s of the
formation, biochemistry, and functions of pepins.
Based on these data, the metabolically active
biovolume of Ca. T. magnifica (excluding the
central vacuole) is two orders of magnitude
above the predicted maximum mentioned
above (table S2). This does not appear to con-
tradict previous studies; indeed, such models
have excluded bacteria with structural adap-
tations such as endomembrane systems and
slower growth rates, which allow much larger
cell volumes (18). Whereas most bacteria have
doubling times ranging from minutes to hours,
Ca. T. magnifica may be similar to other
Thiomargarita species, which require up
to 2 weeks to produce daughter cells (5). The
predicted maximum volume of bacteria as-
sumes binary fission as a division mode, but
Ca. T. magnifica does not have to double its
volume to produce a daughter cell because
only a small portion of the apex constricts
and detaches from the mother cell.
A highly polyploid cell with a
large genome
All previously described giant bacteria are
polyploid (2,3,14),meaningthattheircells
contain large numbers of genome copies
ranging from tens to tens of thousands that
are dispersed throughout the cell, supporting
the local need for molecular machineries and
overall cellular growth (15,23,24). Polyploidy
has been shown to decrease the selective pres-
sure on genes, allowing intracellular gene dup-
lication, reassortment, and divergence, and to
lead to extreme intracellular genetic diversity
in some LSB (25). Conversely, polyploidy may
allow balancing of genome copies through
homologous recombination and support a
high level of genome conservation (26). Ca.
T. magnifica, like all bacterial giants, appeared
to be polyploid; counts of DAPI-stained DNA
clusters on three CLSM 3D datasets suggested
an average of 36,880 ± 7956 genome copies per
millimeter of filament (737,598 ± 159,115 for a
fully grown 2-cm cell; see table S2, Fig. 2N,
movie S5, and details in supplementary text).
With its number of genome copies being one
order of magnitude above that of other giant
bacteria (2,24), Ca. T. magnifica accounts for
the highest estimated number of genome
copies within a single cell. Understanding the
mechanisms underlying the regulation of such
a large number of genome copies will require
additional work.
To genomically characterize Ca. T. magnifica,
we amplified, sequenced, and assembled the
DNA of five individual cells collected from a
single sunken leaf (tables S3 and S4). All five
draft genomes were highly similar to each
other, with an average nucleotide identity
>99.5% (table S5). Although extreme intra-
cellular genetic diversity has been shown in
some LSB (25), our variant analysis of DNA se-
quences recovered from a single cell revealed a
largely homogenous genome population (1.22 ±
0.18 single-nucleotide polymorphisms/100 kbp;
table S6) (27), which is similar to other poly-
ploid bacteria (26,28). The Ca. T. magnifica
genome assemblies were estimated to be nearly
complete at 91.0 to 93.7%, with total sequence
lengths between 11.5 and 12.2 Mb. This value
is twice as large as the only other sequenced
Thiomargarita species, Ca. T. nelsonii (29,30),
and at the upper range of bacterial genome
sizes; bacterial genomes are on average 4.21 ±
1.77 Mb (fig. S11). The Ca.T.magnificagenomes
contained up to 11,788 genes (only half with a
functional annotation; table S4), more than
three times the median gene count of pro-
karyotes (3935 genes) (31). For comparison
with eukaryotic organisms, Ca.T.magnifica
has a genome as large as the baker’s yeast
Saccharomyces cerevisiae (12.1 Mb) and contains
more genes than the model fungus Aspergillus
nidulans (≈9500 genes).
Analysis of the Ca.T.magnificagenome
revealed a large set of genes for sulfur oxida-
tion and carbon fixation, suggesting chemo-
autotrophy, which is consistent with other
evidence for thioautotrophy (figs. S12 to S14
and supplementary text). Like its sister lineage
Ca. T. nelsonii, Ca. T. magnifica encoded a wide
range of metabolic capabilities with one no-
table difference: It lacked nearly all genes in-
volved in dissimilatory and assimilatory nitrate
reduction and denitrification except for Nar and
Nap nitrate reductases (fig. S12). This absence
suggests that nitrate can solely be used as
an electron acceptor to be reduced to nitrite,
which is not further reduced (29,30) (see the
supplementary text for extended genome
analysis). The absence of epibiotic bacteria
maybeexplainedbythehighnumberofgenes
encoding secondary metabolism. With 25.9%
of sequences dedicated to biosynthetic gene
clusters (Fig. 3A), the genome encoded dozens
of modular nonribosomal peptide synthetase
and polyketide synthase systems, hinting at
numerous secondary metabolism pathways
(similar to the Actinobacteria; see table S7)
that are indicative of antibiotic or bioactive
compound production (32).
Giant bacteria have been shown to repurpose
their cell division machinery as an adaptation
to extreme cell size (33). The Ca. T. magnifica
genome also holds clues for its unusual cell
morphology in the form of an atypical com-
plement of cell division and elongation genes.
Many genes that encode core cell division pro-
teins, including the core components of Z ring
assembly and regulation, FtsA, ZipA, and FtsE-
FtsX,werelacking(Fig.3BandtableS8).
By contrast, genes that encode the cyto-
skeletal protein FtsZ, which is part of the
well-conserved dcw (“division and cell wall”)
operon and the core component of the Z
ring, were conserved. Proteins ZapA, ZapB,
and ZapD, which interact with FtsZ and reg-
ulate Z ring assembly, were likewise conse rv ed
Volland et al., Science 376, 1453–1458 (2022) 24 June 2022 4of6
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Cornell University on June 23, 2022
(34).The entire set of genes that encode late
divisome proteins, including peptidoglycan
polymerases FtsI and FtsW, as well as FtsQ,
FtsL, FtsB, and FtsK, was absent from all Ca.
Thiomargarita genomes (Fig. 3B, fig. S19,
and table S8). In contrast to the conspicuous
lack of cell division genes, a complete set of
genes encoding cell elongation proteins was
present, three of which, mreD,rodZ, and the
peptidoglycan transpeptidase mrdA,have
undergone recent dupl icati on s, wi th bo th
copies located next t o eac h oth er on the
chromosome (Fig. 3B, figs. S16 to S18, and
table S8) (34). It is possible that an increased
number of cell elongation genes, coupled
with the lack of key cell division genes, may be
responsible for producing the unusually
long filaments of Ca. T. magnifica (see the
supplementary text).
Dimorphic developmental cycle of
Ca. T. magnifica
Laboratory observations of live Ca. T. magnifica
revealed eventual apical bud detachment from
the filament and release into the environment,
likely representing a dispersive stage of the
developmental cycle (Fig. 1C; fig. S1, B to F;
and supplementary text). We observed dozens
of cells at all intermediary stages, from the
smallest attached cells resembling terminal
segments recently settled to the largest fila-
ments with apical constrictions (fig. S1 and
movies S1 to S3 and S6). Such a dimorphic life
cycle resembles the aquatic single-celled model
system Caulobacter crescentus, as well as the
multicellular segmented filamentous bacteria,
albeit at a different scale, in which stalked
“parent”cells produce free-living “daughter”
cells (35,36). Because of this asymmetrical
division mode, only a small fraction of the ge-
nome copies present within pepins in the most
apical bud (≈1%) were transmitted to the
daughter cell (fig. S9). Like the polyploid giant
bacterium Epulopiscium spp., Ca. T. magnifica
apparently transmits only a subset of its ge-
nomes, so-called “germline genomes,”to the
offspring (14,24). If terminal buds are indeed
daughter cells, then such a developmental
cycle may have evolved to enhance disper-
sion similar to the fruiting bodies of the so-
cial myxobacteria or to the aerial hyphae of
Streptomyces spp. (37). This apparent life cycle
is also somewhat analogous to the sulfur-
oxidizing giant ciliate symbiosis, Zoothamnium
niveum (38), possibly representing a case
of convergent evolution of developmental
cyc le a cross domains (see the supplementary
text). As with other aspects of Ca. T. mag-
nifica biology, detailed investigations of cell
division and its regulation will require the
establishment of stable laboratory cultures,
and considering the cell size, spatial omics
approaches within a single cell might be
tractable.
Concluding remarks
Confirmation bias related to viral size pre-
vented the discovery of giant viruses for more
than a century, and their ubiquity is only now
being recognized (39,40). The discovery of
Ca. T. magnifica suggests that large and more
complex bacteria may be hiding in plain sight.
Volland et al., Science 376, 1453–1458 (2022) 24 June 2022 5of6
Fig. 3. Genome analysis and proposed model for the subcellular organization of Ca. T. magnifica.
(A) Genome phylogenetic tree with added information about genome quality [red: low quality, orange:
medium quality, and yellow: high quality (46)], estimated level of completeness, assembly size, coding
sequence (CDS) count, and percentage of sequence dedicated to biosynthetic gene clusters (BGCs).
Pattern 1 corresponds to “complete gene cluster for cell division of model bacteria.”Pattern 2
corresponds to “mreD,mrdA,androdZ genes are duplicated.”(B) Gene neighborhoods centered on
the ddl,mreB,androdZ genes showing the incomplete set of divisome genes (lack of ftsQ and ftsA)
in both Thiomargarita species, as well as the duplication of elongasome genes (mreD,mrdA,androdZ)in
Ca. T. magnifica. Note that the Beggiatoa sp. PS, Achromatium sp. WMS3, and Ca. Thiomargarita sp.
Thio36 draft genomes were too fragmented and thus are not included here. (C) Light microscopy
image and model proposed for the subcellular organization in Ca. T. magnifica showing how the pepin
organelles might develop into other cellular compartments, resulting in an increase of surface area of
the bioenergetic membranes.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Cornell University on June 23, 2022
Investigating the biology, energy metabolism,
and the formation, nature, and role of pepins
will take us a step closer to understanding the
evolution of biological complexity.
Although cells of most bacteria and archaea
are ~2 mm, eukaryotic cells are usually be-
tween 10 and 20 mm, with some of the largest
single-cell eukaryotes reaching 3 to 4 cm
(41). Several theoretical frameworks explain
the restriction of bacteria and archaea to mi-
croscopic sizes, including: (i) the lack of active
intracellular transport and the reliance on
chemical diffusion, which is efficient only along
micrometer distances (4); (ii) a predicted maxi-
mum cell volume constraining the number of
needed ribosomes should the cell grow larger
(21); or even (iii) a decrease in energy effi-
ciency due to mismatched surface area to
volume ratio when considering placement of
membrane-bound ATP synthases (18,20).
These frameworks all suggest that with in-
creasing size, the physiological or metabolic
needs of a bacterial cell grow faster than the
cell’s capacity to sustain it and should reach a
limit. The next largest prokaryote known after
Ca. T. magnifica, Ca. T. nelsonii, has a meta-
bolically active biovolume of 1.05 × 10
−14
m
3
(excluding the central vacuole), close to the
predicted maximum due to ribosome limita-
tions, 1.39 × 10
−15
m
3
, and to the bioenergetic
membrane limitation of 10
−14
m
3
(18,21).
Our precise 3D measurements on a 4.27-mm
Ca. T. magnifica cell revealed a cytoplasm
biovolume several orders of magnitude above
that limit (5.91 × 10
−12
m
3
; table S2). It is
possible that changes in spatial organization
of cellular components, such as DNA and
ribosome compartmentalization and rear-
rangement of the bioenergetic membrane
system, may allow Ca. T. magnifica to overcome
many such limitations (Fig. 3C).
Distributed in at least 23 phyla are 19 known
types of bacterial organelles, of which only
seven are membrane bound (16,42). Cyano-
bacteria can form multicellular, centimeter-long
filaments and are capable of cell differentia-
tion (43). Planctomycetes have special energy
transduction organelles called anammoxosomes
and a compartmentalized cell, and some are
even capable of phagocytosis (16,44). The
social Myxobacteria have large genomes and
a complex developmental cycle, and are ca-
pable of moving and feeding cooperatively
in predatory groups (45). Through its gi-
gantic cell size, its large genome, and its
dimorphic life cycle, but most importantly
through its compartmentalization of genetic
material in membrane-bound pepins, Ca. T.
magnifica adds to the list of bacteria that
have evolved a high level of morphological
complexity. Because it segregates its genet-
ic material in membrane-bound organelles,
Ca. T. magnifica challenges our concept of a
bacterial cell.
REFERENCES AND NOTES
1. A. D. Steen et al., ISME J. 13, 3126–3130 (2019).
2. D. Io nescu, M. Bi zic, “Giant bacteria,”in eLS (Wiley,
2019); https://doi.org/10.1002/9780470015902.
a0020371.pub2.
3. P. A. Levin, E. R. Angert, Cold Spring Harb. Perspect. Biol. 7,
a019216 (2015).
4. H. N. Schulz, B. B. Jorgensen, Annu. Rev. Microbiol. 55,
105–137 (2001).
5. H. N. Schulz et al., Science 284, 493–495 (1999).
6. E. R. Angert, K. D. Clements, N. R. Pace, Nature 362, 239–241
(1993).
7. D. C. Nelson, C. O. Wirsen, H. W. Jannasch, Appl. Environ.
Microbiol. 55, 2909–2917 (1989).
8. M. Huettel, S. Forster, S. Kloser, H. Fossing, Appl. Environ.
Microbiol. 62, 1863–1872 (1996).
9. M. R. Jean et al., PLOS ONE 10, e0117832 (2015).
10. J . M. Larkin , M. C. Henk, Microsc. Res. Tech. 33,23–31
(1996).
11. V. Salman, J. V. Bailey, A. Teske, Antonie van Leeuwenhoek
104, 169–186 (2013).
12. J. V. Bailey et al., ISME J. 5, 1926–1935 (2011).
13. S. Maier, H. Volker, M. Beese, V. A. Gallardo, Can. J. Microbiol.
36, 438–448 (1990).
14. E. R . Ang ert , Annu.Rev.Microbiol.66,197–212
(2012).
15. S. Schorn et al., FEMS Microbiol. Ecol. 96,fiz200
(2020).
16. C. Gr eening, T. L ithgow, Nat. Rev. Microbiol. 18,677–689
(2020).
17. T. Katayama et al., Nat. Commun. 11 , 6381 (2020).
18. P . E. Schavemaker, S. A. Muñoz-Gómez, The role of
mitochondrial energetics in the origin and diversification
of eukaryotes. bioRxiv 465364 [Preprint] (2021);
https://doi.org/10.1101/2021.10.23.465364.
19. N. Lane, Curr. Biol. 30, R471–R476 (2020).
20. N. Lane, W. M artin, Nature 467,929–934
(2010).
21. C. P. Kempes, L. Wang, J. P. Amend, J. Doyle, T. Hoehler,
ISME J. 10,2145–2157 (2016).
22. R. Hatzenpichler et al., Environ. Microbiol. 16, 2568–2590
(2014).
23.J.E.Mendell,K.D.Clements,J.H.Choat,
E. R. Angert, Proc. Natl. Acad. Sci. U.S.A. 105,6730–6734
(2008).
24. E. R. An gert, Geno me Biol. Evol. 13 , evab037
(2021).
25. D. Ionescu, M. Bizic-Ionescu, N. De Maio, H. Cypionka,
H. P. Grossart, Nat. Commun. 8, 455 (2017).
26. V. Salman-Carvalho, E. Fadeev, S. B. Joye, A. Teske,
Front. Microbiol. 7, 1173 (2016).
27. B. J. Sh apiro, Curr. Opin. Microbiol. 31,116–123
(2016).
28. T. Woyke et al., PLOS ONE 5, e10314 (2010).
29. B. E. Flood et al., Front. Microbiol. 7, 603 (2016).
30. M. Winkel et al., Front. Microbiol. 7, 964 (2016).
31. K. Chiyomaru, K. Takemoto, R. Soc. Open Sci. 7, 191859
(2020).
32. L. Katz, R. H. Baltz, J. Ind. Microbiol. Biotechnol. 43, 155–176
(2016).
33. D. A. Miller, G. Suen, K. D. Clements, E. R. Angert, BMC
Genomics 13, 265 (2012).
34. A. J. F. Egan, J. Errington, W. Vollmer, Nat. Rev. Microbiol. 18,
446–460 (2020).
35. S. A. Woldemeskel, E. D. Goley, Trends Microbiol. 25, 673–687
(2017).
36. E. A. Hutchison, D. A. Miller, E. R. Angert, Microbiol. Spectr. 2,
2.5.22 (2014).
37. D. Claessen, D. E. Rozen, O. P. Kuipers, L. Søgaard-Andersen,
G. P. van Wezel, Nat. Rev. Microbiol. 12, 115–124
(2014).
38. M. Bright, S. Espada-Hinojosa, I. Lagkouvardos, J. M. Volland,
Front. Microbiol. 5, 145 (2014).
39. C. Abergel, M. Legendre, J. M. Claverie, FEMS Microbiol. Rev.
39, 779–796 (2015).
40. F. Schulz et al., Nature 578, 432–436 (2020).
41. J. Dumais, K. Serikawa, D. F. Mandoli, J. Plant Growth Regul. 19,
253–264 (2000).
42. C. A. Kerfeld, C. Aussignargues, J. Zarzycki, F. Cai,
M. Sutter, Nat. Rev. Microbiol. 16,277–290 (2018).
43.E.Flores,A.Herrero,Nat. Rev. Microbiol. 8,39–50 (2010).
44. T. Shiratori, S. Suzuki, Y. Kakizawa, K. I. Ishida, Nat. Commun.
10, 5529 (2019).
45. J. Muñoz-Dorado, F. J. Marcos-Torres, E. García-Bravo,
A.Moraleda-Muñoz,J.Pérez,Front. Microbio l. 7,781
(2016).
46. R. M. Bowe rs et al., Nat. Biotechnol. 35,725–731 ( 2017).
ACKNO WLED GMEN TS
We are thankful to the following centers where the electron
microscopy analyses were performed: the Centre Commun de
Caractérisation des Matériaux des Antilles et de la Guyane in
Guadeloupe, F.W.I., which is supported by The European Regional
Development Fund, the Regional Council of Guadeloupe, and the
French Research Department; the Electron Microscope Lab
(EML) of theUniversity of CaliforniaBerkeley; the ElectronMicroscopy
Resource in Donner at LBNL, Berkeley; and the FEI Eindhoven
Center. We are particularly grateful to D. Jorgens at EML for advice
and assistance in electron microscopy sample preparation and
data collection. The x-ray tomography data were acquired at
the Stanford Nano Shared Facilities at Stanford University (Palo
Alto, CA), and we are particularly grateful to A. Vailionis for his
technical support during hard x-ray tomography scan acquisitions.
The confocal microscopy observations were performed at the
Advanced Microscopy Facility at LBNL (Berkeley, CA). Preliminary
confocal microscopy observations were acquired at the IBPS
Imaging Facility, which is supported by “Conseil Regional d’Ile-de-
France.”We thank S. Volland for his help with 3D rendering
animations and H. Maughan for copyediting this manuscript.
The work (proposal: 10.46936/10.25585/60001074) conducted
by the US Department of Energy Joint Genome Institute
(https://ror.org/04xm1d337), a Department of Energy (DOE)
Office of Science User Facility, is supported by the Office of Science
of the DOE operated under contract no. DE-AC02-05CH11231.
The Nagoya permit TREL1820249A/50 (unique identifier ABSCH-
IRCC-FR-246822-1) can be consulted publicly online at:
https://absch.cbd.int/en/search. Funding: This work was
supported by the John Templeton Foundation (grant 60973 to
J.-M.V., S.V.D., and T.T.), the Gordon and Betty Moore
Foundation (grant GBMF7617 to J.-M.V., S.V.D., and T.T.),
DARPA (award no. HR001120036 to J.-M.V., S.V.D., and T.T.);
the DOE Office of Science (contract no. DE-AC02-05CH11231 to
T.W., F.S., T.T., J.-M.V., K.M.D., and N.H.E.), and Region
Guadeloupe (F.W.I. grant to M.R.J.). Author contributions:
Conceptualization: O.G., S.G.R., J.-M.V., T.W., S.V.D., F.S., R.R.M.,
N.I., T.T., N.H.E.; Data curation: N.I., F.S., J.-M.V., D.U., S.G.R.;
Formal analysis: J.-M.V., S.G.R., F.S., N.I., D.U., O.G.; Funding
acquisition: T.W., S.V.D., O.G., J.L.M.; Investigation: J.-M.V.,
S.G.R., O.G., T.T., F.S., D.U., D.G., N.N., N.I., C.G.R., S.B.K., N.H.E.,
M.R.J., J.L.M.; Methodology: O.G., T.T., J.-M.V., S.V.D., S.G.R.,
D.G., N.N., R.R.M., T.W., C.G.R., S.B.K.; Resources: K.M.D., T.W.,
R.R.M., O.G., J.L.M., N.H.E., N.J.M.; Supervision: O.G., T.W., S.V.D.;
Visualization: O.G., S.G.R., J.-M.V., F.S., N.I., D.U.; Writing –
original draft: J.-M.V.; Writing –review and editing: J.-M.V., S.G.R.,
O.G., T.W., S.V.D., E.A., N.I., R.R.M. Competing interests: S.V.D.
serves as the CEO of Sample Exchange. The remaining authors
declare no competing interests. Materials and data availability:
All data needed to evaluate the conclusions in this study are
present in the main manuscript or the supplementary materials.
ThecompleteassembliesaswellastheextractedCa.T.
magnifi ca draft ge nomes hav e been depo sited in IM G (https://
img.jgi.doe.gov/). The raw reads have been deposited in SRA
(https://www.ncbi.nlm.nih.gov/sra). Accession numbers are
provided in tables S3 and S4. License information: Copyright ©
2022 the authors, some rights reserved; exclusive licensee
American Association for the Advancement of Science. No claim to
original US government works. https://www.science.org/about/
science-licenses-journal-article-reuse
SUPPLEMENTARY MATERIALS
science.org/doi/10.1126/science.abb3634
Materials and Methods
Supplementary Text
Figs. S1 to S20
Tables S1 to S11
References (47–83)
Movies S1 to S6
MDAR Reproducibility Checklist
View/request a protocol for this paper from Bio-protocol.
Submitted 16 October 2021; resubmitted 26 March 2022
Accepted 9 May 2022
10.1126/science.abb3634
Volland et al., Science 376, 1453–1458 (2022) 24 June 2022 6of6
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Cornell University on June 23, 2022
Use of this article is subject to the Terms of service
Science (ISSN ) is published by the American Association for the Advancement of Science. 1200 New York Avenue NW, Washington, DC
20005. The title Science is a registered trademark of AAAS.
Copyright © 2022 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works
A centimeter-long bacterium with DNA contained in metabolically active,
membrane-bound organelles
Jean-Marie VollandSilvina Gonzalez-RizzoOlivier GrosTomáš TymlNatalia IvanovaFrederik SchulzDanielle
GoudeauNathalie H. ElisabethNandita NathDaniel UdwaryRex R. MalmstromChantal Guidi-RontaniSusanne Bolte-
KlugeKaren M. DaviesMaïtena R. JeanJean-Louis MansotNigel J. MounceyEsther R. AngertTanja WoykeShailesh V. Date
Science, 376 (6600), • DOI: 10.1126/science.abb3634
A magnificent megabacterium
We usually think of bacteria as microscopic isolated cells or colonies. Sampling a mangrove swamp, Volland et al.
found an unusually large, sulfur-oxidizing bacterium with a complex membrane organization and predicted life cycle
(see the Perspective by Levin). Using a range of microscopy techniques, the authors observed highly polyploid cells
with DNA and ribosomes compartmentalized within membranes. Single cells of the bacterium, dubbed Candidatus
Thiomargarita magnifica, although thin and tubular, stretched more than a centimeter in length. —MAF
View the article online
https://www.science.org/doi/10.1126/science.abb3634
Permissions
https://www.science.org/help/reprints-and-permissions
Downloaded from https://www.science.org at Cornell University on June 23, 2022
