Embryogenesis of flattened colonies implies the innovation required for the evolution of spheroidal colonies in volvocine green algae
Volvocine algae provide a suitable model for investigation of the evolution of multicellular organisms. Within this group, evolution of the body plan from flattened to spheroidal colonies is thought to have occurred independently in two different lineages, Volvocaceae and Astrephomene. Volvocacean species undergo inversion to form a spheroidal cell layer following successive cell divisions during embryogenesis. During inversion, the daughter protoplasts change their shape and develop acute chloroplast ends (opposite to basal bodies). By contrast, Astrephomene does not undergo inversion; rather, its daughter protoplasts rotate during successive cell divisions to form a spheroidal colony. However, the evolutionary pathways of these cellular events involved in the two tactics for formation of spheroidal colony are unclear, since the embryogenesis of extant volvocine genera with ancestral flattened colonies, such as Gonium and Tetrabaena, has not previously been investigated in detail.
We conducted time-lapse imaging by light microscopy and indirect immunofluorescence microscopy with staining of basal bodies, nuclei, and microtubules to observe embryogenesis in G. pectorale and T. socialis, which form 16-celled or 4-celled flattened colonies, respectively. In G. pectorale, a cup-shaped cell layer of the 16-celled embryo underwent gradual expansion after successive cell divisions, with the apical ends (position of basal bodies) of the square embryo’s peripheral protoplasts separated from each other. In T. socialis, on the other hand, there was no apparent expansion of the daughter protoplasts in 4-celled embryos after successive cell divisions, however the two pairs of diagonally opposed daughter protoplasts shifted slightly and flattened after hatching. Neither of these two species exhibited rotation of daughter protoplasts during successive cell divisions as in Astrephomene or the formation of acute chloroplast ends of daughter protoplasts as in volvocacean inversion.
The present results indicate that the ancestor of Astrephomene might have newly acquired the rotation of daughter protoplasts after it diverged from the ancestor of Gonium, while the ancestor of Volvocaceae might have newly acquired the formation of acute chloroplast ends to complete inversion after divergence from the ancestor of Goniaceae (Gonium and Astrephomene).
KeywordsMulticellularity Body plan Embryogenesis Volvocine green algae Gonium Tetrabaena
Differential interference contrast
The evolution of multicellular organisms from unicellular ancestors is one of the most important events in the history of life, known as evolutionary transitions in individuality . Such transitions to multicellularity have occurred independently at least 25 times, in many different eukaryotic and prokaryotic branches . In many lineages of multicellular organisms, including animals and land plants, the emergence of multicellularity was accompanied by the invention of embryogenesis, a developmental program for building a complicated body plan from a single cell . An investigation into the evolution of embryogenesis would provide some clues to understand the complex and divergent body plans found in multicellular organisms. However, many multicellular lineages lack extant intermediates between unicellular and multicellular organisms with complex body plans, which makes it difficult to resolve the first steps in the evolution of embryogenesis.
The volvocine algae has undergone parallel evolution of body plans, with two independent transitions from a flattened to spheroidal colony, in Volvocaceae and the goniacean Astrephomene (Fig. 1) [4, 5, 6]. Though the two lineages are alike in the morphology of their spheroidal colonies, they differ in their tactics for forming these colonies during embryogenesis: the embryos of volvocacean species undergo inversion, while the embryos of Astrephomene undergo rotation of daughter protoplasts (Additional file 6: Figure S1) .
Volvocacean species form spheroidal colonies by inversion, a unique morphogenetic event that occurs during embryogenesis (Additional file 6: Figure S1a). In volvocacean species, a mature reproductive cell initiates embryogenesis with successive cell divisions and forms a cup-shaped or hollow, spheroidal embryo composed of a single layer of daughter protoplasts laterally interconnected with one another by cytoplasmic bridges. Immediately following the successive divisions, the apical ends of daughter protoplasts (the side with basal bodies) are oriented toward the inside of the cell layer. The embryo then inverts its cell layer to form a spheroidal daughter colony with the apical ends of daughter protoplasts positioned on the outside. Developmental studies using V. carteri have revealed that the principal factors producing the force for folding the cell layer during inversion are the change in cell shape to flask-shaped with a stalk at the chloroplast ends (the opposite side from apical ends) of daughter protoplasts and the movement of protoplasts relative to cytoplasmic bridges (observed as relocation of cytoplasmic bridges to the stalks at chloroplast ends) [20, 21, 22]. The change in cell shape with the development of stalks or acute chloroplast ends and the location of cytoplasmic bridges at the chloroplast ends during inversion have also been reported in other Volvox species [23, 24, 25, 26] and in other genera in Volvocaceae [27, 28, 29, 30, 31], which suggests that the volvocacean genera use common cellular mechanisms for inversion.
In contrast to Volvocaceae, Astrephomene does not employ inversion to construct a spheroidal daughter colony during embryogenesis (Additional file 6: Figure S1b) . During successive cell divisions in the embryogenesis of Astrephomene, daughter protoplasts rotate outward, with the apical ends of protoplasts moving from the anterior (the side with flagella of the mother cell and the direction of swimming of the daughter colony) to posterior region at the outer (apical) surface of the cell layer. After the successive cell divisions, the cell layer forms a nearly spheroidal shape with the apical ends of the daughter protoplasts oriented toward the outside. Therefore, no inversion occurs in the embryogenesis of Astrephomene.
How these two different mechanisms for making spheroidal colonies evolved in Volvocaceae and Astrephomene is currently unclear, since the embryogenesis of extant genera with ancestral flattened colonies, such as Gonium and Tetrabaena (Fig. 1), has not been investigated in detail from this point of view. Though embryogenesis has been observed in Gonium and Tetrabaena by light microscopy time-lapse imaging [32, 33, 34] and transmission electron microscopy [33, 34], it is currently unclear whether the rotation of daughter protoplasts seen in Astrephomene or the cell shape change of Volvocaceae occurs in these genera, as thus far their embryogenesis has only been observed at certain stages and from limited angles.
In this study, we focused on the formation of flattened colonies in G. pectorale and T. socialis to evaluate whether they employ cellular mechanisms that may have been co-opted to make spheroidal colonies in Volvocaceae and Astrephomene. We conducted light microscopy time-lapse imaging showing all successive stages of embryogenesis, as well as indirect immunofluorescence microscopy staining basal bodies, nuclei, and microtubules of both species to investigate their embryogenesis at cellular and subcellular levels. Our results define the similarities and differences in the developmental events that lead to the formation of flattened and spheroidal colonies, thus providing clues to the evolutionary pathway from flattened to spheroidal colonies.
Time-lapse imaging of embryogenesis in G. pectorale and T. socialis
Successive images of embryogenesis in G. pectorale from the anterior view (Additional file 1) and the anterior-lateral view (Additional file 2) were obtained and analyzed as movies. In G. pectorale, each cell in vegetative colonies performed four successive cell divisions to form 16-celled embryos (Additional file 6: Figure S2a–e). The cleavage patterns were as described in previous studies on G. pectorale [35, 36, 37]: the third cell division occurred nearly parallel to the first division to form an 8-celled embryo with a pattern of four zigzag rows of two protoplasts each (Additional file 6: Figure S2d), then the fourth cell division occurred nearly parallel to the second division to form a 16-celled embryo with four central protoplasts (A3, A3’, B3, and B3’ in Additional file 6: Figure S2e) and 12 peripheral protoplasts (Additional file 6: Figure S2e).
Additional file 1: Movie S1. Light microscopy time-lapse imaging of embryogenesis in G. pectorale from the anterior view. Scale bar: 5 μm, 900x speed. (AVI 2549 kb)
Additional file 2: Movie S2. Light microscopy time-lapse imaging of embryogenesis in G. pectorale from the anterior-lateral view. Scale bar: 5 μm, 900x speed. (AVI 2475 kb)
Additional file 3: Movie S3. Light microscopy video showing hatching of the daughter colony in G. pectorale. Scale bar: 10 μm, 1x speed. (MP4 8750 kb)
Successive images of embryogenesis in T. socialis from the anterior view (Additional file 4) and the anterior-lateral view (Additional file 5) were also obtained and analyzed as movies. The cleavage patterns in T. socialis (Additional file 6: Figure S2f–h) were identical to those previously observed . In the 4-celled embryo just after successive cell divisions, one pair of diagonally opposed daughter protoplasts (B and B′ in Additional file 6: Figure S2 h) were slightly attached to each other at the center of the embryo, while the other pair were separated from each other (A and A’ in Additional file 6: Figure S2 h). A similar daughter protoplast arrangement was also observed in the 4-celled embryos of G. pectorale (Additional file 6: Figure S2c) and has been observed in other species of volvocine algae .
Additional file 4: Movie S4. Light microscopy time-lapse imaging of embryogenesis in T. socialis from the anterior view. Scale bar: 5 μm, 1,800x speed. (AVI 2090 kb)
Additional file 5: Movie S5. Light microscopy time-lapse imaging of embryogenesis in T. socialis from the anterior-lateral view. Scale bar: 5 μm, 1,800x speed. (AVI 1941 kb)
Indirect immunofluorescence microscopy of basal bodies, nuclei, and microtubules
The staining of basal bodies allowed visualization of the expansion of the apical surface of the concave or cup-shaped 16-celled embryo in G. pectorale (Fig. 4). During successive cell divisions, the basal bodies of daughter protoplasts were located within the central region of the cup-shaped embryos and did not change their positions much (Fig. 4a, b and c). This situation of the basal bodies during successive divisions is consistent with the absence of rotation of daughter protoplasts observed in time-lapse imaging (Fig. 2a, b and c). After successive cell divisions, the basal bodies of the 12 peripheral protoplasts positioned in the center of the cup-shaped embryo moved from the center to the periphery in accordance with expansion of the cell layer of embryos (Fig. 4d). The basal bodies of each protoplast were positioned more peripherally than the longitudinal axis of each daughter protoplast (Fig. 4d, e). After daughter colonies hatched from the mother cell wall, the basal bodies of their peripheral cells were further separated from one another and pointed outward (Fig. 4e), but they did not move toward the posterior region as in successive cell divisions of Astrephomene .
In contrast to the apparent movement of the basal bodies in G. pectorale, the basal bodies of daughter protoplasts did not show obvious movement during embryogenesis in T. socialis (Fig. 5). During successive cell divisions, the basal bodies of daughter protoplasts were positioned in the anterior face of the embryo (Fig. 5a, b and c). After successive cell divisions, the basal bodies of a pair of diagonally opposed daughter protoplasts (corresponding to A and A’ in Additional file 6: Figure S2 h) shifted slightly toward the anterior relative to those in the other pair (corresponding to B and B′ in Additional file 6: Figure S2 h; Fig. 5d). No significant movement of the basal bodies was observed in developing embryos within the cell wall, except for a slight shifting parallel to the anterior-posterior axis of the embryo. After daughter colonies hatched out from the mother cell wall, we observed no shifting of their cells, and the basal bodies of four cells were arranged to form a square shape in the same plane (Fig. 5e), which suggests that flattening and extension of the cell layer occurred after hatching.
The morphological changes in the cell layer after successive cell divisions during embryogenesis in G. pectorale and T. socialis described above were verified quantitatively by measuring the distance between basal bodies relative to the diameter of the embryo/daughter colony (Additional file 6: Figures S3 and S4). In G. pectorale, the ratio of the distance between basal bodies of a diagonally opposed pair of peripheral protoplasts to the diameter of the 16-celled embryo/daughter colony increased significantly after the formation of flagella (Additional file 6: Figure S3a, b). It further increased after the daughter colony hatched from the mother cell wall (Additional file 6: Figure S3c, d). In T. socialis, on the other hand, the ratio of the distance between the basal bodies of two pairs of diagonally opposed protoplasts/cells to the diameter of the 4-celled embryo/daughter colony did not increase following the formation of flagella before hatching (Additional file 6: Figure S4a, b, d, e), but increased after hatching in both pairs of diagonally opposed cells (Additional file 6: Figure S4c, f, g).
What were the molecular or genetic changes leading to the developmental modifications underlying the evolution of these body plans? One of the molecular bases associated with the formation of acute chloroplast ends in Volvocaceae is the regulation of cortical microtubules. In flask-shaped cells during inversion in V. carteri, the elongated stalks at the chloroplast ends are lined with cortical microtubules . Moreover, colchicine (which prevents the assembly of microtubules) and cold treatment (which causes rapid disassembly of microtubules) have both been reported to block the formation of stalks, which suggests that the microtubules play an important role in the formation of stalks in flask-shaped cells . Similar patterns of microtubules lining the acute chloroplast ends during inversion have also been observed in V. tertius  and Eudorina elegans . By contrast, the indirect immunofluorescence microscopy used to stain the microtubules of embryos in G. pectorale and T. socialis in the present study did not demonstrate the distinctive pattern of cortical microtubules at the chloroplast ends; the cortical microtubules lined the entire cell bodies and concentrated around the apical ends after successive cell divisions (Figs. 4c, d and 5c, d). Such a difference in the patterning of cortical microtubules between volvocacean species and the two species with flattened colonies, G. pectorale and T. socialis, suggests that the ancestral genes involved in the remodeling or the maintenance of cortical microtubules, which have not previously been reported, have been co-opted or new genes have been acquired to form the acute chloroplast ends in the ancestor of the Volvocaceae.
Moreover, our results clearly demonstrate the details of the formation of flattened colonies in G. pectorale and T. socialis (Additional file 6: Figure S5). A recent study on embryogenesis in G. pectorale reported that a “partial inversion” occurred within the mother cell wall , while earlier studies reported that the flattening of the cell layer coincided with the breaking of the mother cell wall or hatching of the daughter colony [35, 38]. Our time-lapse analyses are consistent with both: the cup-shaped embryos formed after successive cell divisions expanded gradually within the mother cell wall (Fig. 2c, d and e, Additional file 2) and the further flattening of the cell layer occurred simultaneously with hatching (Additional file 3). Immunofluorescence microscopy staining the basal bodies during these stages also allowed clear visualization of these developmental processes (Fig. 4c, d and e; Additional file 6: Figure S3). Such a stepwise formation of flattened colonies was also observed in the embryogenesis of T. socialis: the slight shifting between the two pairs of diagonally opposed protoplasts occurred within the mother cell wall (Figs. 3c, d and e and 5c, d; Additional file 5) and the cell layer was flattened after hatching (Fig. 5e; Additional file 6: Figure S4).
These developmental processes in G. pectorale and T. socialis might be accompanied by the molecular bases which have been co-opted for the spheroidal colony formation in Volvocaceae and Astrephomene. During inversion in V. carteri, InvA, a type of kinesin localized in cytoplasmic bridges, interacts with cortical microtubules and moves cell bodies relative to cytoplasmic bridges, which results in the production of the driving force for folding the cell sheet . Previous studies using transmission electron microscopy did not observe such a relocation of cytoplasmic bridges in G. pectorale and T. socialis [33, 34]. However, the genome data of both species include putative homologs of InvA (GPECTOR_36g129, Acc. No. KXZ47277.1 in the G. pectorale genome  and Kinesin-like protein KIF27, partial, Acc. No. PNH03015.1 in the T. socialis genome , only a partial sequence), and these putative homologs might be involved in development, such as in expansion of the cell layer in G. pectorale or the shifting of protoplasts in T. socialis. Another two known genes in V. carteri, InvB and InvC, encode a GDP-mannose transporter and a glycosyltransferase, respectively, which are involved in enlargement of the gonidial vesicle of the extracellular matrix that provides sufficient space for inversion [40, 41]. The enlargement of the mother cell wall after successive cell divisions was observed in G. pectorale (Fig. 2c–e), and several genes encoding proteins similar to InvB and InvC also exist in the genome of G. pectorale  and T. socialis . Thus, the genes involved in the enlargement of the mother cell wall in the volvocacean ancestor with flattened colonies, like those in Gonium spp., might have been co-opted to support inversion in the extant volvocacean species. These hypotheses regarding the co-option of ancestral genes for the molecular mechanisms of spheroidal colony formation may be verified by further investigation using recently established techniques such as transformation of G. pectorale  and its application to T. socialis.
Using light microscopy time-lapse imaging and staining cellular components with indirect immunofluorescence microscopy, we described the formation of flattened colonies in G. pectorale and T. socialis in detail at cellular and subcellular levels. Compared to species with spheroidal colonies, G. pectorale and T. socialis showed simpler embryogenesis, without the rotation of daughter protoplasts during successive cell divisions and the formation of acute chloroplast ends after successive cell divisions, implying that these mechanisms were independently acquired by the distinct ancestors of species that form spheroidal colonies. Although some molecular mechanisms of inversion have been investigated in Volvox, the molecular or genetic bases underlying the rotation of daughter protoplasts in Astrephomene are currently unknown. Further analyses of these cellular mechanisms in Volvocaceae and Astrephomene and comparative analyses with Gonium and Tetrabaena would unveil the molecular and genetic bases, and the modification of these, which brought about the evolution of the body plan from flattened to spheroidal colonies.
Strains and culture conditions
We used T. socialis strain NIES-571  to observe embryogenesis in Tetrabaena. As the Gonium strains used in recent genome  and morphological studies [34, 44] frequently show abnormal morphology in vegetative colonies and no formation of completely flattened 16-celled colonies (data not shown), we established several new strains of Gonium. We used one of these new strains, G. pectorale strain 2017–0423-IsgGo1, in the present study (Additional file 6: Figure S6). The culture was grown in 10 mL of SVM medium  in screw-capped tubes (18 × 150 mm) and maintained at 25 °C on a 12-h light/12-h dark schedule under cool-white fluorescent lamps at an intensity of 50–90 μmol·m− 2·s− 1. Species identification was based on morphological and molecular data (Additional file 6: Figures S6 and S7).
For developmental observation, G. pectorale strain 2017–0423-IsgGo1 was grown synchronously in silicon-capped 500-mL Erlenmeyer flasks containing 250 mL SVM medium with aeration at 32 °C on a 16-h light/8-h dark schedule under cool-white fluorescent lamps at an intensity of 140–180 μmol·m− 2·s− 1, and T. socialis strain NIES-571 was grown synchronously in silicon-capped 500 mL Erlenmeyer flasks containing 250 mL SVM medium with aeration at 20 °C on a 12-h light/12-h dark schedule under cool-white fluorescent lamps at an intensity of 240–280 μmol·m− 2·s− 1.
Light microscopy time-lapse imaging
Embryogenesis of G. pectorale strain 2017–0423-IsgGo1 and T. socialis strain NIES-571 was observed by time-lapse light microscopy based on methods reported previously [7, 34] with some modifications. As the polyethyleneimine coating on coverslips hampered the normal development of G. pectorale and T. socialis embryos (data not shown), we used coverslips without coating. Culture containing fully mature vegetative colonies of G. pectorale or T. socialis was placed on slides, and then the coverslips, surrounded by thin banks of petroleum jelly, were placed on top of the culture. In this method, cells and embryos are attached to the coverslip with weak adhesion of the flagella or cell walls. Preparations were observed under a BX-53 microscope (Olympus, Tokyo, Japan) equipped with Nomarski interference optics. Photomicrographs were obtained using DP Controller 1.2.1108 (Olympus) at 1-min (G. pectorale) or 2-min (T. socialis) intervals. This difference in timing was implemented in response to the observation that T. socialis embryos ceased embryogenesis when photomicrographs were taken at 1-min intervals, likely because they are sensitive to light and/or heat, so that frequent exposure to light under light microscopy prevented T. socialis embryos from proceeding with their development. As the embryos of G. pectorale and T. socialis moved and rotated during observation, positions and angles of images were adjusted manually using Adobe Photoshop CC (Adobe Systems Inc., San Jose, CA, USA). The sequence of adjusted images was analyzed and processed using ImageJ 1.50b (National Institutes of Health, Bethesda, MD, USA). Embryogenesis was observed at least three times in each species and showed almost identical patterns in each observation.
Indirect immunofluorescence microscopy
To verify the behavior of daughter protoplasts during embryogenesis in G. pectorale and T. socialis, immunostaining of basal bodies and microtubules was initially performed as previously described . However, we found that the cortical microtubules of G. pectorale and T. socialis were fragile and broke when subjected to that protocol, and thus we introduced the following modifications. The samples were placed in fixing solution-A (3.7% formaldehyde [Sigma-Aldrich, St. Louis, MO, USA], 0.1% Triton X-100 [Sigma-Aldrich], and 1 mM DTT [Nacalai Tesque Inc., Kyoto, Japan] in phosphate-buffered saline [PBS]) for 1 min and then into fixing solution-B (3.7% formaldehyde and 1 mM DTT in PBS; i.e., fixing solution-A without Triton X-100) for 14 min, instead of the original protocol of 15 min in fixing solution-A. Chlorophyll and other pigments in the samples were extracted in − 20 °C methanol for 1 min, instead of extracting solution  for 60 min. The samples were then subjected to two primary antibodies and two secondary antibodies, each for 1 h at 37 °C. The primary antibodies were a rat anti-tubulin α antibody (clone YL1/2, Bio-Rad Laboratories, Inc., Hercules, CA, USA) and a rabbit anti-CrSAS-6 antibody  diluted 1:500 with blocking buffer (0.44% gelatin [Sigma-Aldrich], 0.05% NaN3, and 1% bovine serum albumin [Sigma-Aldrich] in PBS). The secondary antibodies were an Alexa Fluor 568-conjugated goat anti-rat IgG (Invitrogen) and an Alexa Fluor 488-conjugated goat anti-rabbit IgG (Invitrogen) diluted 1:500 with the blocking buffer described above. The samples were placed on a drop of SlowFade Gold Antifade Mountant (Thermo Fisher Scientific Inc., Waltham, MA, USA) mixed with 0.5 μL of 0.1 mg/mL DAPI (Sigma-Aldrich). Preparations were observed under a BX-60 microscope (Olympus) equipped with Nomarski interference optics, a mercury lamp, and filter sets with DP Controller 1.2.1108 (Olympus). Differential interference contrast and fluorescence images were merged using Adobe Photoshop CC (Adobe Systems Inc.). For quantitative analysis of the shapes of embryos in successive developmental stages, we measured the diameter of the embryo/daughter colony in both species, as well as the distance between basal bodies of a diagonally opposed pair of peripheral protoplasts/cells in G. pectorale or a diagonally opposed pair of protoplasts/cells in T. socialis, using ImageJ 1.50b (National Institutes of Health).
We would like to thank Dr. Kaoru Kawafune and Dr. Takumi Murakami for providing soil samples to establish the new strain of G. pectorale. We would like to thank Dr. Masafumi Hirono and Dr. Akira Noga for providing anti-CrSAS-6 antibody and helpful suggestion.
SY and HN conceived and designed the experiments. SY performed the experiments and analyzed the data. SY and HN wrote the manuscript. All authors read and approved the manuscript.
This work was supported by Grants-in-Aid for JSPS Research Fellow (grant number 17 J03439 to SY) and Scientific Research (grant number 16H02518 to HN) from MEXT/JSPS KAKENHI. The funding bodies had no role in the design of the study and collection, analysis, and interpretation of data or in writing the manuscript.
Ethics approval and consent to participate
Consent for publication
Author HN is a member of the BMC Evolutionary Biology Editorial Board. The authors declare that they have no competing interests.
- 16.Featherston J, Arakaki Y, Hanschen ER, Ferris PJ, Michod RE, Olson BJSC, et al. The 4-celled Tetrabaena socialis nuclear genome reveals the essential components for genetic control of cell number at the origin of multicellularity in the volvocine lineage. Mol Biol Evol. 2018;35(4):855–70.PubMedCrossRefGoogle Scholar
- 24.Kelland JL. Inversion in Volvox (Chlorophyceae). J Phycol. 1977;13(4):373–8.Google Scholar
- 27.Gottlieb B, Goldstein ME. Colony development in Eudorina elegans (Chlorophyta, Volvocales). J Phycol. 1977;13(4):358–64.Google Scholar
- 36.Gerisch G. Die Zelldifferenzierung bei Pleodorina californica Shaw und die Organisation der Phytomonadinenkolonien. Arch Für Protistenkd. 1959;104:292–358.Google Scholar
- 38.Hartmann M. Über die Veränderung der Koloniebildung von Eudorina elegans und Gonium pectorale unter dem Einfluß äußerer Bedingungen. IV. Mitt. der Untersuchungen über die Morphologie und Physiologie des Formwechsels der Phytomonadinen (Volvocales). Arch Für Protistenkd. 1924;49:375–95.Google Scholar
- 43.Kasai F, Kawachi M, Erata M, Mori F, Yumoto K, Sato M, et al. NIES-collection list of strains, 8th edition. Jpn J Phycol Sôrui. 2009;57(1, Supplement):1–350.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.