2.1 Theoretical and Methodical Foundations of Cell Theory

Various ideas, first in the form of mythological narratives, on the origins and basis of life have existed since the dawn of humankind. Some of the first modern ideas on the substance of life and its developmental program being confined to a small piece of living organism came from Aristotle in the fourth century BC. On the basis of empirical experience of egg development and plant vegetative reproduction, he postulated “entelchy” as a driving principle that leads organisms toward fulfilling their form and potential (Welch and Clegg 2010). Until the seventeenth century, the developing science of biology only rarely sought the causes of live phenomena in the fine structure of organisms. Prominent trends such as French morphology and German Naturphilosophie looked for explanation of body plans and structures in abstract ideal forms toward which organisms are driven (Radl 1930); the approach was largely orthogonal to later and contemporary mechanistic views of life.

Idealistic concepts accompanied biology further, but a gradual shift toward empiricism and mechanistic tendencies in science appeared in the eighteenth century. The trend included revival of atomism, a theory that claims that properties of matter are given by the small indivisible particles it is composed of (Harris 2000). It is important to note that advances resulting in formulation of cell theory were not only led by technological improvements in microscopy, but also by a change in theoretical focus. Many scholars already had the idea that observed tissues were aggregates of more basic units, even before looking through microscopes (Harris 2000). Another important philosophical inspiration (quite distinct from common early analogies between cells and atoms or crystals, and much closer to the contemporary perception of cells) came from G.W. Leibniz (1646–1716). His idea established the often unrecognized basis of cell theory. In the idea of fully autonomous self-reproducing “monads,” developed in a critical discourse with the Cartesian mechanistic view of the universe, Leibniz stated that if living organisms were machines their parts would not merely be simple mechanical pieces of matter but smaller machines themselves. Importantly, the dynamics of monads is driven from the inside. This idea stimulated the concept of German philosopher Lorenz Oken (1779–1851) that all organisms are composed of “infusoria” and “Urbläschen” (primordial bubbles) as basic life units; this speculation directly preceded the works of the first empirical cell biologists (Canguilhem 2008; Harris 2000). However, it was only the invention and improvement of microscopes that enabled direct observation of the material basis and composition of organisms.

Based on early observations, the composition of tissues as fibers, globules, or twisted cylinders was postulated (Harris 2000). In the eighteenth century, Albrecht von Haller, inspired by atomism, speculated that fibers composed of strings of atoms were the basic structural elements of the body: “For the fiber is for the physiologist what the straight line for the geometrician, and from this fibre all shapes surely arise” (in Harris 2000). Robert Hooke was active in many fields of natural sciences in the second half of the seventeenth century and is considered to be the father of the term “cell” in biology. He used this term to describe the structures he saw with his simple microscope in slices of plant cork tissue because they resembled honeycomb cells (cellulae in Latin). At that time, cells were conceived as hollow and regarded as “avenues of communication, channels for conveyance of juices” (in Welch and Clegg 2010).

2.2 From Schleiden to Virchow: Formation of Cell Theory Tenets

More and more nineteenth century scientists were convinced that plant tissues were generally composed of cells, but Matthias Jakob Schleiden (1804–1881) made the first attempt to use cellular composition as a unifying explanatory principle in botany (Harris 2000). Schleiden wanted to establish botany on a firm ground as a more exact science, leaving behind the speculative tradition of German Naturphilosophie. Schleiden was mechanistically oriented and, like many of his contemporaries, inspired by Isaac Newton’s physics. He used crystal-like metaphors for conceptualizing the self-organization of organisms. He also emphasized inductive and empirical approaches, as well as the importance of following ontogeny (reflecting specification or differentiation of initial simpler general forms into more complex elaborated ones) in order to properly understand plant tissues. Schleiden’s efforts resulted in the formulation of a general rule that all plant tissues are composed of a single basic element, the polyhedral cell. He would subsequently call for “condemnation of every theory that explains processes in a plant otherwise than as combination of processes in individual cells” (in Radl 1930).

The cell wall as the boundary and structural element was still considered more important than the internal content, although the nucleus was already known and described. It was named in 1833 by Robert Brown who, however, did not recognize the general presence of nuclei in all cells. Such an opinion is understandable considering that the cell wall is morphologically the most conspicuous structure in differentiated plant cells and is often the functional determinant of the particular tissue. Moreover, the crucial importance of cell wall mechanics and its integration with the plant cell membrane and cytoplasmic core are currently well-recognized features of plant body organization. Schleiden was unclear about the ontogenic origin of cells; therefore, an extracellular protoplasm or sap played a role in his concept of cell formation. He postulated condensation of nuclei from this material and formation of cellular matter around them. Cells were formed from nuclei as growing vesicles until they touched each other (Harris 2000; in Radl 1930). Later in development, they mostly formed around nuclei inside other cells (Lombard 2014). Schleiden considered nuclei in mature cells dispensable and often reabsorbed (Harris 2000).

Because of the absence of distinct cell walls and difficulties in sample preparation, the cellular nature of animal bodies was less clear. Animal cells were studied, for example in developing embryos. However, the general empirical supposition postulated formation of the animal body from cells during early development, but not necessarily in its adult state. Henri Dutrochet (1776–1847) advocated a materialistic worldview and aimed to identify vital phenomena in animals and plants (Harris 2000). He claimed that both plant and animal tissues were composed of “vesicles” and “globules,” although he probably could not observe animal cells. Although Dutrochet’s morphological view of cells was largely erroneous, he was probably the first to perceive cells as basic physiological units of metabolic exchange with selective inflow of nutrients and outflow of waste. He also suggested the existence of the same underlying principles in animal and plant tissues: “[If] phenomena are tracked down to their origins, the differences are seen to disappear and an admirable uniformity of plan is revealed” (in Harris 2000).

The first claim of the widespread presence of “Kornchen” analogous to plant cells in animal tissues, backed by countless histological observations, was made by Bohemiam Jan Evangelista Purkyně/Purkinje (Harris 2000). He claimed that animal tissues were universally composed of cells, fibers, and fluids. Purkyně was also one of the first (following Dutrochet) to emphasize the functional significance of cells and facilitated the transition from “histomorphology” to “histophysiology” (Harris 2000), particularly through comprehensive studies of ciliary movements in several animal tissues. Unlike Schwann, who put most emphasis on the nucleus, Purkyně also focused on the active content of the cell, the “protoplasm” (see Sect. 2.3).

Theodor Schwann (1810–1882) got most credit for extending cell theory to animal tissue because he made stronger (although not always correct) claims than Purkyně (Harris 2000). Inspired by Schleiden’s conclusions, as well as the similarity between animal notochord cells and plant cells discovered by Schwann’s teacher Johannes Müller (Harris 2000), Schwann accumulated a vast number of examples of embryonic and adult animal tissues consisting of cells and claimed cellular origin as the unifying ontogenic principle for animals as well as plants (Radl 1930). Schwann was not certain about the exact origin of individual cells and postulated their origin either from homogenous life matter (possibly through first generating a nucleus) or from inside other cells, around their nuclei. According to Schwann, cells could thus originate inside or outside other cells (Harris 2000; Lombard 2014). Inspired by Schleiden, Schwann claimed that formation of cells from liquid via nuclei was a mechanistic crystallization-like process (Harris 2000).

Several different ideas about the mechanism of new cell generation coexisted and many scientists accepted that different mechanisms could work in different organisms and tissues (Harris 2000). Discovery of binary cell fission by Barthélemy Dumortier and Hugo von Mohl was of outstanding importance, although both admitted the plurality of mechanisms of cell formation. Franz Unger (1800–1870) was the first to oppose Schleiden’s aggregation/crystallization idea openly. He disregarded “cytoblasts” as source of cells and postulated that binary division was the most common mechanism of plant cell division (Harris 2000). Within a few years, sufficient empirical evidence had accumulated to abandon Schleiden’s concept of cell formation. Because of technical difficulties, it took much longer to accumulate precise observations of animal cell formation. Robert Remak (1815–1865) proposed the first explicit unifying theory of cell division in both plants and animals. Remak developed novel hardening agents that allowed him to carry out extensive studies of cell formation in many animal tissues. He concluded that extracellular formation of cells does not occur in animal tissues and that binary division is the universal mechanism of cell formation. Development is thus a sequence of binary divisions followed by morphological modifications; furthermore, the egg itself is a cell. Remak also proposed that the same rules governed cell division in both pathologic and embryonic tissues (in direct opposition to Müller’s theory of specific malignant tumor formation). Remak categorically opposed Schleiden and Schwann, particularly their analogies between cells and crystals: “It is hardly necessary to make special mention of the similarity or disparity of cells and crystals, for, in the light of the facts that I have discussed, the two structures offer no points of comparison” (in Harris 2000).

Rudolf Virchow (1821–1902), strongly inspired by Remak and spreading Remak’s ideas, consolidated cell theory with his famous statement “Omnis cellula e cellula,” reflecting the origin of existing cells from other cells and describing ontogeny as a gradual process of binary divisions from a fertilized egg to adult tissues. Classical cell theory thus stood on three major tenets:

  1. 1.

    All living organisms are composed of one or more cells.

  2. 2.

    The cell is the basic unit of structure and function in all organisms.

  3. 3.

    All cells arise from preexisting cells.

The history of discoveries leading to a unified picture of cell division (and the relationship between nucleus and cytoplasm during formation of new cells) is an excellent example of how the choice of methods and model system can influence the inferred theory. This aspect has always constrained experimental biology and is still relevant in our time. From the contemporary point of view (i.e., retrospective judgement), ideas about extracellular formation and crystallization around nuclei might seem obscure. However, one must acknowledge that many conclusions were based on observations of fixed tissues prone to artifacts and providing only a static view of underlying dynamic phenomena. The presence of open mitosis in both animals and plants (Sazer et al. 2014) made deciphering the relationships between “sap” (cytoplasm), nucleus, and cell division even more complicated until the nature of chromosomes was understood.

Moreover, some of the tissues used in the past as model systems are nowadays known as rather exceptional cases. Even original observations of plant tissues by Schleiden involved endosperm syncytium undergoing cellularization, which might have given him the wrong impression of cell formation (Harris 2000). Many early conclusions were also misled by mistaking starch grains (forming inside cells) for nuclei. On the other hand, Dumortier and von Mohl were able to make their outstanding discovery through observing an ideal model system for study of binary cell division—the filamentous alga Conferva (Draparnaldia by contemporary nomenclature) with cells dividing at the termini of filaments. Cartilage was repeatedly used as argument for the acellular origin of animal cells (Harris 2000). Developing embryos, which enabled direct observation of unfixed dividing cells in time, were the source both of support for a model of binary cell division and of erroneous judgment. Although many authors (working mostly with amphibian models) correctly interpreted the partitioning of egg as progressive cell division, French biologist Quatrefages de Bréau claimed in the middle of the nineteenth century that the development of gastropod embryos involves formation of cells within cells. Quatrefages de Bréau was probably driven by an attempt to support Schwann’s model. Even Dumortier, who discovered binary division in Conferva (Draparnaldia by contemporary nomenclature), acknowledged the possible formation of cells within cells and even formation of cells from acellular material after observing gastropod development (Harris 2000). Large yolky embryos with unequal cleavage were also source of confusion, as in the case of Carl Vogt who claimed that Alytes frog embryo furrowing was independent of formation of new cells.

In the light of incongruent fragmentary observations, it was honest of many contemporary scientists in the nineteenth century to acknowledge the plurality of animal cell formation mechanisms (Harris 2000). Strong universal claims required systematic comparison of many different tissues and improved techniques, as performed by Remak. Although he opposed ideas that involved intracellular formation of cells, he admitted that it was often not sloppiness of observation or ill judgment that lead to incorrect conclusions, but accidental choice of problematic material such as cartilage or muscle fiber. However, even Remak made an erroneous conclusion regarding nuclear division, possibly because of observation of static fixed specimens and a bias toward making an analogy between binary cell division and binary nuclear division. Karl Bogislaus Reichert (1811–1883) observed dissolution of nuclei during division of red blood cells, which he used as an argument for Schwann’s concept of de novo nuclei formation and against the concept of binary cell division. Remak claimed that he had failed to reproduce Reichert’s observation of nucleus dissolution in dividing red blood cells. Reichert’s ideas about cell formation were generally wrong but some of his observations were correct, whereas Remak’s ideas about cell formation were generally right but some of his observations were wrong. Remak occasionally observed nuclear dissolution but interpreted it as an artifact. Both Remak and Virchow supported a model of nuclear binary division that involved formation of grooves, constriction, and division of one nucleus into two. Some scientists advocated Remak’s and Virchow’s models, whereas others referred to nuclear dissolution (“Reichert’s doctrine”), often with interpretations close to Schwann’s original ideas about cell formation (Harris 2000).

2.3 Protoplasmic Concepts and Early Criticisms of Newly Established Cell Theory

Cell theory was popular with reductionists, who attempted to comprehend fundamental life phenomena by studying simple structural components. Technological improvements such as the oil immersion lens, Purkyně’s microtome technique (Harris 2000), and novel fixation and staining methods (McIntosh and Hays 2016) led to countless observations of cells and their contents in the nineteenth century. Criticism of cell theory also existed and, in extreme cases, many histological discoveries were accused of being staining and/or fixation artifacts. Skepticism over the universality of cell theory often cited the existence of cells without nuclei, multinuclear syncytia, and large amounts of extracellular material in adult tissues as evidence against cell theory. Nevertheless, all of these phenomena were ultimately understood as developmental products of cells. One of the last bitter arguments about the general validity of cell theory was over the nature of nervous tissue. “Reticulate theory” considered the nervous tissue as a continuous uninterrupted network, because of observation limits set by contemporary microscopes. Yet, cell theory envisaged nervous tissue as consisting of individual cells as in other tissues (the “neuronal doctrine”). Ramon y Cajal demonstrated the latter to be true by using a staining method that randomly marked only a few neurons within the tissue, clearly indicating discontinuity in the neuronal network (Radl 1930).

In addition to claims that cell theory cannot universally explain the functioning of organisms and that many observed structures might be fixation artifacts, cell theory was also repeatedly accused of being insufficient or even not relevant to understand the universal properties of life. Some of these incongruences were formulated in various forms of “protoplasmic theory,” which either complemented cell theory by closing a conceptual gap between the cell surface and cellular contents or competed with cell theory by completely shifting focus from cells as a mere building bricks to the living substance inside the cell. The term “protoplasm” was introduced by Jan Evangelista Purkyně/Purkinje in 1839, well before Hugo von Mohl and in a very similar sense (Janko and Štrbáňová 1988; Harris 2000; Zárský 2012; Liu 2016). Hugo von Mohl was critical of Schleiden’s and Schwann’s focus on understanding cells in terms of boundaries and building blocks and disliked analogies between cells and crystals. He redefined the cell’s function as more based on internal organization and formulated his protoplasmic theory in 1846 (Liu 2016).

Ferdinand Cohn proposed in 1850 that “plants and animals were analogous not only because of their construction from cells, but also, at a more fundamental level, by virtue of a common substance, protoplasm, filling the cavities of those cells” (Welch and Clegg 2010; Liu 2016). He thus connected von Mohl’s concept with the earlier idea of “sarcode,” a contractile substance proposed by Félix Dujardin to provide the life basis of unicellular eukaryotes (Liu 2016). The tendency to look for basic attributes of life (irritability, sensibility, contractility, reproduction, etc.) in the properties of protoplasm was not uncommon, and protoplasm itself was compared to an “elementary organism.” Anatomist Max Schultze suggested in the middle of the nineteenth century that the true basis of life would be found by studying protoplasm, not the cell (Welch and Clegg 2010) and redefined the cell as a “clump of protoplasm” around a nucleus (Liu 2016).

Some authors regarded the cell as a nonliving envelope and focused on studying protoplasm as the “naked state of living matter” (Welch and Clegg 2010). For example, E.B.Wilson did not claim protoplasm to be the only living element inside the cell: “Protoplasm deprived of nuclear matter has lost, wholly or in part, one of the most characteristic vital properties, namely, the power of synthetic metabolism, yet we still speak of it as ‘living’, because it may for a long time perform some of the other functions, manifesting irritability and contractility, and showing also definite coordination of movements” (as in the enucleated protozoan) (Wilson 1899). He also disregarded strong versions of reductionism that searched for a single basic element of life in “any single substance or structural element of the cell,” because “life in its full sense is the property of the cell-system as a whole rather than of any one of its separate elements.” His theory is thus not atomistic or reductionistic but puts a strong focus on the properties of protoplasm by claiming “that the continuous substance is the most constant and active element and that which forms the fundamental basis of the system, transforming itself into granules, drops, fibrillae or networks in accordance with varying physiological needs” (Wilson 1899). Yet, Wilson prophetically admitted that he could not achieve any clear general conclusion because the basis of all phenomena lies in the “invisible organization of a substance which seems to the eye homogenous.” He believed that “ultramicroscopic bodies,” molecules, groups of molecules, and micellae formed the basis of protoplasmic organization (Wilson 1899).

2.4 Discovery of Organelles: Increasing Appreciation of Cellular Content

Along with protoplasmic concepts involving the actions of micelles, drops, and tiny fibrillae, the presence of larger structures localized within cells was more and more recognized and emphasized, including the notion of smaller living units present inside cells, inspired by Leibnitz’s theory of spontaneity and hierarchy of monads (see Sect. 2.1). Franz Unger described moving structures in pollen cytoplasm as an “army of monads full of inner vitality, full of an inner self-determination that revealed itself in their movements” (in Harris 2000). Observations of large unicellular eukaryotes such as amoebae and ciliates further stimulated thoughts about subcellular structures with specialized functions, analogous to macroscopic bodies. In 1884, Karl August Mobius suggested the term “organulum” (little organ) for such structures because they form parts of one cell, whereas true organs of multicellular animals consist of many cells. The term was later transformed into “organelle” and its meaning was expanded to cover subcellular structures of both unicellular and multicellular organisms (Schuldiner and Schwappach 2013).

An important breakthrough was made by van Benden and Boveri at the end of the nineteenth century. They discovered the autonomous life cycle of the centrosome and concluded that the structure had a life of its own; Boveri described the centrosome as a special organ of cell division (Harris 2000). Whitman perceived the cell as a “colony of simpler units, nucleus, centrosome, and so on,” much as a higher organism is colony of cells (Whitman 1893). In 1882, Julius Sachs wrote that “chlorophyl bodies” (chloroplasts) behaved like autonomous organisms that divide to adjust their number to the size of growing leaves (Kutschera and Niklas 2005). In 1883, Andreas Schimper noticed the similarity between chloroplasts and cyanobacteria and proposed the symbiotic cyanobacterial origin of plastids (Taylor 1987). In 1890, Altmann postulated the universal presence of “bioblasts” (named “mitochondria” by German microbiologist Benda in 1898) and discovered that they had same staining properties as bacteria; he concluded that they were modified bacteria (Ernster and Schatz 1981; Kutschera and Niklas 2005).

This idea of the endosymbiotic origin of chloroplasts and xenobiotic origin of eukaryotic cells as an evolutionary amalgam of once-independent organisms was further elaborated by Konstantin Mereschkowsky between 1905 and 1920 (Taylor 1987; Kutschera and Niklas 2005), but was not generally accepted until its revival in the 1970s. With improved microscopes and staining methods, novel organelles were added to the nuclei, chloroplasts, and vacuoles known from earlier observations (Ernster and Schatz 1981). With the discovery of “ergatoplasm” (later named “endoplasmic reticulum”) in 1897 and the Golgi apparatus one year later, most large common components of the cell “inventory” were known by the end of the nineteenth century (Ernster and Schatz 1981).

2.5 Disputes over Cell Boundaries

For a living system, the existence and properties of a boundary to the outside world are as important as the properties of its internal composition. Yet, the presence and identity of a boundary between cells and the outside environment was not clear in the nineteenth century and (especially from the contemporary perspective) was largely neglected by proponents of both cell and protoplasmic points of view. Schwann assumed that surfaces/membranes always limit the mobility in/out of a cell, even if invisible, and this could be inferred from the Brownian motion of cell components, which do not escape the cell volume as delimited by the surface structure. Generally, however, comparison of the cell surfaces of plant cells (with walls) and animal cells were confusing and the terms “wall” and “membrane” were often used interchangeably. True membranes were impossible to detect with nineteenth century histology techniques. Thus, in the second half of the nineteenth century, little attention was paid to membranes and, if present, they were considered unessential secondary structures originating from hardening of the cell surface. Max Schulze, the proponent of protoplasmic theory, was also an eager opponent of the membrane concept (Lombard 2014). He postulated, in place of cells, small blebs of contractile protoplasm immiscible with water. Detected membranes were simply the result of protoplasm hardening caused by contact with the outside environment or an artifact of degeneration and the hallmark of dead cellular material.

The main support for the membrane concept came from osmotic studies. Hewson published experiments on the swelling and shrinking of blood cells as early as 1773. In the first half of the nineteenth century, Dutrochet explained plant turgescence by osmosis via a border with “chemical sieves” (Harris 2000; Lombard 2014). The first artificial membranes were created by precipitation of copper ferrocyanide (from potassium ferrocyanide and copper sulfate) and were thus named precipitation membranes. Together with the contemporary colloidal concept of cell interiors and ideas about cell membranes originating through surface hardening, the existence of artificial precipitation membranes fueled belief that the surface of colloidal protoplasm precipitates and forms an osmotic barrier. Overton’s pioneering experiments (published between 1895 and 1900) showed cell volume changes in more than 500 different solutions and allowed him to conclude that a barrier distinct from the plant cell wall must exist and is made of ether-soluble components (i.e., is hydrophobic). He suggested cholesterol and phospholipids as possible candidates. In combination with works on electrophysiology and microinjection experiments, acceptance of the plasma membrane as a real structure was established in the early twentieth century (Harris 2000; Lombard 2014).

2.6 Toward Cellular Determinants of Heredity

A clear picture of nuclear division formed only after the mitotic spindle and chromosomes were discovered and understood. Recurrent observations eventually led to the consensus that nuclei disassemble and reassemble during cell division. Strassburger proposed homology of plant and animal cell division before the end of the nineteenth century (Harris 2000). In the 1870s, details of cell division events were repeatedly observed and, in 1879, Walter Flemming coined the term “mitotic process” and described its basic chronology. Flemming also introduced the term “chromatin” and was the first to describe longitudinal division of chromosomes in both animal and plant cells. He was a sharp critic of the direct nuclear division concept advocated by Remak and Virchow, but at the same time fully acknowledged the continuity of nuclear material during cell division by expanding Virchow’s statement into “Omnis nucleus e nucleo.”

At that time, there was also a major effort to localize the material determinants of heredity. Many great biologists of the nineteenth century, even if not working with cells themselves, postulated such particles (Darwin postulated gemulae; Haeckel, plastiduls; Spencer, physiological units; de Vries, pangenes; Galton, strips, etc.) and thus stimulated the search for them (Radl 1930). Cumulative descriptive work helped characterize the progression of cell division and behavior of chromosomes in sufficient detail that biological interpretations and manipulative experiments were possible. As early as 1885, the concept of chromosomal loops as storage place for hereditary information was proposed by A. Weissmann (McIntosh and Hays 2016) and helped to explain the phenomena of meiosis and recombination (Harris 2000). The work of Theodor Boveri (1862–1915) not only definitively demonstrated chromosome function in heredity, but also shifted work from solely combination of observations and deduction to the introduction of manipulative experiments (Harris 2000). His experiments with sea urchin embryos involved polyspermy and manipulation of early embryo cleavage, resulting in blastomeres with unequal chromosome distribution. Boveri discovered that the fate of blastomeres correlated with introduced chromosomal abnormalities and deduced that different chromosomes carry different genetic loads. After the rediscovery of Mendel’s laws, Boveri was the first to point out the similarity between segregation of elements, as proposed by Mendel, and physical segregation of chromosomes (Harris 2000). The first concept of genes was purely phenomenological and did not necessarily ask for the material agent of heredity. Later, in the light of mechanistic trends, a material component responsible for transmission of genetic information was envisaged. Boveri proposed that the material basis of Mendel’s laws of inheritance lay in the properties of chromosomes and thus contributed to the development of molecular genetics in the twentieth century (Harris 2000).

2.7 Cells in Tissues: Early Holistic and Reductionist Experimental Approaches

Since the early days of cell theory, many scientists have stressed that organisms are more than just an assembly of their parts, and that functional aspects of life should be studied in the context of the whole developing embryo/organism. Attitudes ranged from sharp criticism of cell doctrine as insufficient and misleading, through attempts to introduce novel organizing principles that would supplement and coordinate the action of cells, to a systematic attempt to understand developing embryos purely from the collective interactions of individual cells.

T.H. Huxley put forward a physiological interpretation of the cell in opposition to Schleiden’s and Schwann’s morphological concept. He claimed that “the cell-theory of Schleiden and Schwann” was not only “based upon erroneous conceptions of structure,” but it also led “to errors in physiology” (Richmond 2000). He particularly disliked that “cell doctrine” overstated the assumption of anatomic individuality of cells and felt that cells should be studied in their mutual relation in the context of development, because the entire life history of an organism is “dominated by development” (Richmond 2000). Whitman stated that “the fact that physiological unity is not broken by cell-boundaries is confirmed in so many ways that it must be accepted as one of the fundamental truths in biology” (Whitman 1893). Sachs advocated the organism-standpoint and considered the presence of cells, although a general phenomenon of life, to be of secondary importance and only one of the many manifestation of formative life forces (Whitman 1893). The idea of Sachs that growth and change of plant forms is primary and that planes of cell division are secondary and dependent on overall growth (Radl 1930) was also shared by de Bary, who coined the famous statement: “The plant forms cells, the cells do not form plants” (Thompson 1917).

Major attempts at causal analysis of embryonic development as a result of collective interaction of individual cells crystallized into the discipline of Entwicklungsmechanik (developmental mechanics in the sense of natural causation), enthusiastically advocated by Wilhelm Roux (Radl 1930; Sander 1991). Roux shifted focus from speculations based purely on descriptive observations to manipulative experiments in a quest for causal explanation of development by combination of individual acting forces (Priven and Alfonso-Goldfarb 2009; Sander 1991). Based on his experiments with amphibian embryos, Roux advocated a mosaic concept of development, stating that cells of the early embryo determine the position of later parts of the organism.

Other scientists proposed different concepts of development, largely because they used other model systems, such as cnidarians and early developing embryos that display an astonishing capacity for regeneration and a certain degree of invariance of morphogenesis with respect to the number of cells participating. Such experiments suggested that cells of the same lineage can have different fates and cells of different lineages the same fate, depending on the position they acquire within the embryo. Whitman claimed that “Comparative embryology reminds us at every turn that the organism dominates cell-formation, using for the same purpose one, several, or many cells, massing its material and directing its movements, and shaping its organs, as if cells did not exist, or as if they existed only in complete subordination to its will” (Whitman 1893). Some of the trends even resulted in the search for holistic principles that precede formation of cells and organize actions of cells across the whole developing organism.

Hans Driesch also attempted to break the continuous process of animal morphogenesis into its ultimate elements (first principles) at the outset of his career (Sander 1992a). In a visionary manner, he considered development to “start with a few ordered manifoldnesses,” which would gradually “create, by interactions, new manifoldnesses,” which “acting back upon the original ones (manifoldnesses) provoke new differences.” “With each response, a new cause is immediately provided, and a new specific reactivity for further specific responses.” (Sander 1992a). Parts of the developing embryo thus constitute a gradual conversion of states and receptivity to other stimuli. Governed by the nucleus, organogenetic chemicals are formed in the cytoplasm, which acts as intermediaries between external stimuli and the nucleus. A cascade of stimuli between cells and their partial activations drive development of the organism (Sander 1992a). Later in his life, Driesch became critical of overestimating the explanatory potential cell theory (Whitman 1893) and even revoked some of his original positions (Sander 1992b). Experiments with cnidarians, acrasid slime molds, plants, and echinoderm embryos (Markoš 2002; Sander 1992b) led him to search for fundamental laws determining the spatiotemporal coordinating system that leads cells into form (Priven and Alfonso-Goldfarb 2009; Sander 1993). Driesch advocated a mathematical and physical approach (Priven and Alfonso-Goldfarb 2009) but also wanted biology to be a science with autonomy and thus searched for organization principles, around which the undergoing chemical and physical phenomena are constituted (Priven and Alfonso-Goldfarb 2009). His conclusion that contemporary chemistry and physics were not sufficient to explain embryogenesis could in fact be extended until the 1970s, when cell research incorporated advances in cybernetics and genetics (Roth 2011). Driesch put strong emphasis on teleology in development (Sander 1992b) and unsuccessfully tried to formulate entelechy as a new collective physical quantity (Markoš 2002; Priven and Alfonso-Goldfarb 2009), specific for organisms, which might be analyzed using mathematical approaches (Priven and Alfonso-Goldfarb 2009).

Driesch’s attempt to uncover laws of organization typical for biology was further developed by Alexander Gurwitsch (Beloussov 1997; Markoš 2002). Gurwitsch studied developing shark brain, fungal fruiting bodies, and composite flowers and arrived at the general conclusion that the overall shape repeatedly develops in an exact manner despite fluctuations in the shape and growth rate of individual parts. He also thought that the outline of a part or a whole embryo can be formulated mathematically more precisely than the shape and arrangements of their internal components (Beloussov 1997). Looking for a supracellular principle that orders and coordinates cells over the embryo, and inspired by contemporary developments in physics, he formulated the concept of a “species-specific field” that organizes morphogenesis (Beloussov 1997; Markoš 2002). Cells produce the field that extends to and affects an extracellular space and, at the same time, the field acts back on the cells. Fields from cells form an aggregate field, which depends on the configuration of the multicellular whole and there is feedback between the field and its morphogenetic consequences (Markos 2002). The interdependence between cell properties and their coordinates of position within a developing organism should be precise and mathematically simple (Beloussov 1997). Gurwitsch even attempted to define the field in vectorial manner (as a geometric description, not in a strictly physical sense), where cells followed the vectors of the field (Markoš 2002).

By the 1930s, many crucial discoveries in experimental embryology had been accomplished. Many studies involved isolation and recombination of embryonic parts and mapping of the differentiation and inductive potential of the isolated parts of embryos and the effects of parts transplanted onto other embryos, including interspecific transplants (Oppenheimer 1966; Gilbert et al. 1996). Phenomena such as the inductive potential of neural folds and establishment of limb polarity were intensively studied. Hans Spemann reintroduced the term “field of organization” to describe the inductive properties of the amphibian dorsal blastopore (Gilbert et al. 1996), conceptually building upon Driesch’s concept of a “harmonious equipotential system.” The concept of a field was thus still vital and, in 1939, Paul Weiss postulated that field is the key organizing principle of embryology; developmental phenomena have field properties and components of fields are connected by a web of interactions (Gilbert et al. 1996). Field concepts in the 1930s experimental embryology were materialistic. Weiss claimed that field has physical existence and is bound by physical substrates from which morphogenesis arises and should be the object of research like any other physical phenomena. The morphogenetic field was supposed to become the basic paradigm of embryology in its attempt to discover the laws of morphogenesis (Gilbert et al. 1996).

2.8 Establishment of Molecular Biology

Details of the birth and early history of biochemistry are beyond the scope of this review. However, we mention several key discoveries and concepts because the paradigm and methodology elaborated by biochemists largely influenced the advent of modern cell biology, especially in the twentieth century. Although most German scientists studying cells focused on their structure and formation, the French naturalist Francois Vincent Raspail (1794–1878) was interested in the chemistry of cells. He analyzed the chemical composition of cells by adopting chemical combustion analysis for small samples (microburning) and developed staining procedures to detect starch, albumin, silica, mucin, sugar, chlorides, and iron. He also stressed that that the cell is itself a microlaboratory, carefully balancing catabolism and anabolism (Harris 2000). In 1833, Payen and Persoz purified a thermolabile fraction able to breakdown starch into sugar. Such “agents” were later named enzymes by Wilhelm Kuhne. In 1893, Eduard Buchner was able to replicate the whole yeast fermentation process by a cell-free extract. Thomas Burr Osborne systematically crystallized proteins and demonstrated a vast diversity of protein species (Kyne and Crowley 2016). In 1926, James Sumner managed to isolate and crystallize an enzyme (urease) for the first time. He redissolved urease from the crystal (thus free of any small compounds potentially co-purified from the cell) and showed its catalytic activity, also demonstrating the proteinaceous (and biopolymer) nature of enzymes (Quastel 1985; Kyne and Crowley 2016).

The initial approach of biochemistry was thus orthogonal to that of microscopy. The properties of life would be studied outside of the organismal context, irrespective of the structural principles in the intact body. The aim was to replicate life or life-like processes in an isolated system with a minimal set of components and thus isolate the underlying substances in order to understand the ongoing properties and changes of matter. Parts of the “protoplasmic” concept were dropped or overshadowed by the advent of classical biochemistry, which focused on isolated molecules in buffered water solutions of simple composition (Kyne and Crowley 2016). The simplified “bag of enzymes in solution” perception of cell content, where molecules randomly encounter each other and follow the law of mass action, was criticized at the outset of the science of biochemistry. It was suggested that catalytic agents act as part of an integral and dynamic proteinaceous network in the cell. However, the original focus of early biochemistry on enzymes as catalytic agents provided a unified mechanistic tool set for characterizing subsets of cellular components and phenomena (Welch and Clegg 2010; Kyne and Crowley 2016). Molecular biology is currently understood as based on molecular genetics, but before the ability to modify genetic information was acquired, it was biochemistry that established the first true molecular-level reductionist description of some life processes.

Synthesis of Mendelian and chromosomal heredity theories in the early twentieth century put genes into the spatial context of location on chromosomes and stimulated institutionalization of genetics as a discipline. As a result of the successful reductionist approach and the immediate economic impact on breeding, there was a common tendency to put genetics into the center of a mechanistic biology framework (Gayon 2016). For example, developmental genetics arose as an alternative program that competed with established experimental embryology (instead of being proposed as a complementary approach). Both the concept of gene used by geneticists and the concept of field used by embryologists were abstract and both were considered to have a physical basis, although understood only vaguely. At that time, genes were still considered to be associated with the action of proteins, possibly enzymes (Oppenheimer 1966; Gilbert et al. 1996). Genocentric tendencies were thus evident in biology at least two decades before the tenets of molecular biology were consolidated. The concept of field as an organizing principle was eventually abandoned, largely because biochemical techniques to examine field phenomena in detail were not available, whereas techniques for study of gene expression in model systems gradually appeared (Gilbert et al. 1996). Despite continuous attempts to interpret life in a holistic framework or perspective, reductionist approaches prevailed in biology as a pragmatic framework for finding mechanistic explanations of complex phenomena.

Genetics, biochemistry, and biophysics developed independently for some time, but started to converge after the 1930s. Key experiments on genetic regulation of Neurospora biochemistry in the 1940s showed that each step in a metabolic pathway is controlled by a single gene and this led to the “one gene–one enzyme hypothesis,” which suggested that each gene acts directly as an enzyme or determines the specificity of an enzyme (Gayon 2016). This further stimulated perception of the gene as a central unit of biological function and much of the attention turned to the relationship between nucleic acid and protein macromolecules and the search for the molecular basis of heredity. Introduction of novel techniques such as X-ray crystallography and ultracentrifugation helped to turn the focus from colloidal theories to biopolymers and their structures.

Recapitulating the great endeavors of twentieth century molecular biology is beyond the scope of this review and is thoroughly described elsewhere (e.g., Rheinberger 2010). Most importantly, the material basis of hereditary information in the form of nucleotide sequences of nucleic acids was discovered and the genetic code solved, uncovering the relationship between a gene sequence and the protein macromolecule it encodes. Discoveries of the basic principles of molecular biology further stimulated the search for genes responsible for all sorts of processes in living organisms.

With basic metabolic pathways mapped, biochemists became interested in the regulation of metabolism. After the pioneering research of Jacques Monod (1910–1974) on the regulation of biochemical pathways and gene expression (Pardee and Reddy 2003), the concepts of positive feedback, negative feedback, allosteric regulation, cooperativity, induction of enzymes, control by repression, nonlinear regulation, cross-inhibition, and boolean integration of regulatory processes became the standard vocabulary of molecular biology (Monod 1972; Pardee and Reddy 2003). Parallels between molecular biology and cybernetics were thus grounded (Monod 1972), although ideas about cell signaling and gene expression at the time were rooted in biochemistry and simple cybernetic relations. Newly developed tools shifted the focus onto study of individual genes and their protein products or simple signaling, genetic, and biochemical pathways. It was understood that other components such as extracellular matrix (ECM) components and membrane lipid composition also play important roles (Monod 1972) but, because of technological difficulties, they were neglected in comparison with research performed on DNA and proteins. These molecules were understood to be localized inside cells but more focus was put on understanding their function at a molecular level than on their cellular functions in terms of structural organization of the cells.

2.9 Biological Membranes in the Twentieth Century: From Discovery of Lipid Bilayers to the Fluid Mosaic Model

Despite initial neglect of the cell barrier in the nineteenth century, the nature of biological membranes became an important topic in twentieth century cell biology. In 1925, Gorter and Grendel performed a pioneering experiment addressing the structural nature of the plasma membrane. They picked erythrocytes, cells devoid of internal membranes, as the model system and showed that the ratio of monolayer area formed from extracted lipids and erythrocyte surface area was 2:1, suggesting the bilayer nature of the plasma membrane (Lombard 2014). It is noteworthy that the experiment was criticized for several shortcomings, including neglecting the protein components of the plasma membrane and wrong calculation of erythrocyte surface. It is now believed that several experimental errors reciprocally cancelled each other, leading to the correct conclusion. However, the validity of this early model can only be appreciated in the light of much later experiments. Regardless of the criticism, the immediate impact of the lipid bilayer hypothesis was to open discussion on the molecular nature of membrane structure. Trends based on Traube precipitation membranes and Overton lipid membranes were both popular. In terms of molecule permeability prediction, a crucial component of the former was pore size and of the latter, hydrophobicity. The unifying theories assumed membranes to be lipid layers interrupted by pores. The mixed roles of lipids and proteins in the function of membranes were acknowledged, but their relative contribution was a controversial issue (Lombard 2014).

In addition to the iconic character of the search for molecular heredity determinants and solving the differential role of proteins and nucleic acids in the nucleus, another key question in twentieth century cell biology was the nature of protein and lipid interplay in the functioning of biological membranes. Various models involved mixtures of lipid and protein fractions within or between postulated layers of the membrane. Interestingly, one of the concepts dominating membrane research for decades was the “paucimolecular model,” which postulated a lipid layer sandwiched between two protein layers. The model was based on measurement of surface tension between echinoderm/teloostei cells and an oil layer, as well as the structure of myelinized axons. The surface tension experiments were soon criticized for using triacylglycerol instead of native membrane phospholipids, and for using myelinized axons as representative model for a general cell membrane. Nevertheless, the concept became popular for a long time and early low quality electron microscopy (EM) images were interpreted as supporting the paucimolecular membrane model. As in many other cases, a well-intended set of experiments and choice of model system led to wrong assumptions that persisted for decades (Lombard 2014).

Mosaic models of the plasma membrane were also popular. Speculations involving fat-like parts and protoplasmic-like parts, a mixture of sieve-like and solvent elements, were supported by permeability experiments at the beginning of the twentieth century. Permeability experiments also suggested that “pore” diameter could change according to the hydration of the pore, pH, metabolic activity, and cell type but the molecular mechanisms of membrane properties were unclear. Even the breakthrough experiments of Hodgkin and Huxley on membrane excitability (1952) were phenomenological and the mechanism of differential membrane permeability toward Na+ and K+ ions was not known (Lombard 2014). Because hydrated Na+ ions are larger than hydrated K+ ions, selective protein agents facilitating Na+ transport were difficult to imagine. Lipid-based carriers specific for Na+ were postulated. Furthermore, several arguments against the lipid nature of plasma membranes were based on its high water permeability. These conundrums were eventually solved in the context of a delicate structure of the potassium channel and the late discovery of aquaporins, membrane proteins that facilitate water permeability.

The fluid mosaic model dominated the membrane field in the 1970s. It was compatible with most contemporary experiments and predicted future observations; the model remained basically unaltered for next few decades. One of its main advantages over several competing models was compatibility with the thermodynamics of protein–lipid and lipid–lipid binding within membranes, largely based on hydrophobic interactions (Singer 2004; Lombard 2014). The general focus on proteins was fostered by tools developed for molecular biology, resulting in membrane proteins being the primary target of research looking for molecular agents of particular membrane functions. Lipids were considered to be passive structural elements that mostly ensured fluidity of proteins within the membrane. Such an idea is still advocated in many textbooks.

2.10 Insights into Cell Ultrastructure and Organelle Origin in the Twentieth Century

The classical descriptive endeavor of cell theory continued during the twentieth century with the disciplines of histology and cytology. The methodological barrier of microscopy was broken in the 1930s by the introduction of electron microscopy. In combination with novel fixation, sectioning, and staining techniques, it became possible to image subcellular structures with the precision of tens of nanometers. First EM images of mitochondria immediately revealed the presence of a double membrane with inner membrane folds, named cristae (Ernster and Schatz 1981). In 1953, EM helped rediscover the endoplasmic reticulum (Schuldiner and Schwappach 2013). EM not only served as a tool for discovering novel details of subcellular structures, but also brought independent confirmation of conclusions on some older conundrums or questions. For example, several competing models of plasma membrane structure existed and Fischer still opposed membrane theory in 1921, arguing that membranes were invisible even when boundaries of cells were visible (see Sect. 2.9). EM eventually confirmed the presence of a plasma membrane lipid bilayer even in bacterial cells, where its presence had been debated for a long time (Lombard 2014). The generally accepted neuronal theory was also unequivocally confirmed by visualizing the synaptic cleft, a small space between neighboring neural cells. The high spatial resolution enabled detection of novel fine branching structures connecting other cellular components (Welch and Clegg 2010). This microtrabecular network was considered the “basic solid component of cytoplasm,” but was also deemed a fixation artifact by many opponents. The concept of solid/liquid phases and heterogeneity of cytoplasm thus became hot topic for some time but then disappeared, only to come back in recent years (Welch and Clegg 2010).

The idea of symbiogenesis (introduced by Mereschkowsky) as the appearance of evolutionary novelties, including novel cell organelles, was revived by Lynn Margulis in the 1970s (Taylor 1987; Chapman and Margulis 1998; Kutschera and Niklas 2005). Margulis also propagated the concept of serial endosymbiosis, stating that modern eukaryotic cells originated by multiple successive symbiogenetic events of once independent organisms (Taylor 1987), and the idea that symbiogenetic events were a common driving force in eukaryotic speciation (Kutschera and Niklas 2005). With employment of molecular biology techniques, support for the endosymbiotic origin of mitochondria and plastids soon accumulated and the paradigm of eukaryotic cell evolution shifted from gradual accumulation of changes as the only mechanism to the possibility of abrupt acquisition of organelles (Taylor 1987). Revival of the symbiogenetic organelle concept and the idea of the eukaryotic cell as a product of cellular fusion between Archea and Eubacteria (Kutschera and Niklas 2005) points to the crucial role of cooperative processes in the evolution of life and to the fact that the evolution of cells could not be fully understood as a simple progressive, incremental process but involved singularities with crucial macroevolutionary impact.

2.11 Formation of Modern Cell Biology and Methodical Trends in Twenty-First Century Cell Biology

Whereas nineteenth century biology had to decide which of the big theories were correct, late twentieth century cell biology was marked by the trend to put together the discoveries of genetics, molecular biology, biochemistry, and cytology into a congruent whole. Top-down (more and more detailed observation of tissue ultrastructure) and bottom-up (examining the properties of smallest functional components in the form of molecules and their relationships) approaches were eventually used together as a common tool set of a unified scientific field. Many processes were attributed to specific genes and their protein products. Proteins were successfully mapped into biochemical, signaling, and gene regulatory pathways. With the help of cell fractionation techniques and EM, combined with antibody staining, it became possible to map biochemical pathways and protein activities to specific subcellular compartments (Schuldiner and Schwappach 2013). The ability to maintain, grow, and manipulate cells outside organisms (a relatively simple task for plant cells), together with the expansion of live cell imaging techniques, especially discovery of genetically encoded fluorescent proteins, led to countless observations of dynamic processes in living cells. Cells have always been perceived as dynamic entities, but the new techniques allowed observation of molecular processes in vivo with the proper spatial and temporal context.

Emphasis has gradually shifted from the role of individual genes to how the actions of individual components within the cell collectively contribute to a particular process. This trend does not negate the earlier discoveries of twentieth century molecular biology in any sense, but demonstrates the importance of studying molecular components within live cells, taking into account structural and dynamic properties of the cellular environment. The cell has thus re-emerged as both a biological and an interpretational platform, connecting molecular mechanisms with macroscopic phenomena.

Several technological trends are typical for cell biology in this new millennium. First, improved techniques now allow cellular components and processes to be followed with greater and greater precision. The resolution of fluorescence microscopes is increasing in time and space, beyond the limitation imposed by the diffraction barrier (Wollman et al. 2015). The classical resolution limit of light microscopy has been surmounted by combination of fluorescence technologies and specialized fluorophore excitation methods. These techniques, along with sophisticated computer analyses, allow almost angstrom (Å) resolution in specific cases (Zeng and Xi 2016). Structural analyses of large macromolecular machines such as the ribosome (Yusupova and Yusupov 2017) and nuclear pore (Beck and Hurt 2017) are not uncommon. Fast tools for intracellular manipulation, such as optical tweezers (Ritchie and Woodside 2015), optogenetically activated proteins (Toettcher et al. 2011), and small photoactivated molecules (Hoglinger et al. 2014), now supplement traditional genetic and pharmacological tools.

Some of the new techniques are helping to bridge traditional approaches. For example, correlative light and electron microscopy enables live cell imaging. High resolution EM data can be acquired for a specific part of the cell after rapid freezing of the sample at a chosen time point (Kobayashi et al. 2016). During imaging mass spectrometry, specific regions of a cell/tissue are separately analyzed by mass spectrometry, which is thus enriched with spatial information (Asano et al. 2016). Analyses of protein structural properties, previously obtained by in vitro measurements, can be performed within the cellular environment in some cases (Schwamborn et al. 2016). Another dominant trend of contemporary cell biology is increasing experimental throughput with the help of automatized data acquisition and processing. Such tendencies were largely introduced for sequencing of whole genomes and transcriptomes but “omics” approaches are becoming widespread in connection with most techniques, including fluorescence microscopy (Mattiazzi Usaj et al. 2016), cell sorting (Warkiani et al. 2015), electron microscopy (Eberle et al. 2015), and structural biology (Grabowski et al. 2016).

2.12 Modular Cell Biology

It has become evident that, although some simple cellular functions are executed by a single molecular component (potassium transport through the plasma membrane via a membrane channel, metabolite conversion by a specialized enzyme), most cellular functions (growth regulation, cell differentiation, chemotaxis) arise from the interactions of many components (Hartwell et al. 1999). After decades of characterizing individual cell components and trends for their total catalogization, focus is now shifting from identifying individual parts to understanding their relationships, spatiotemporal associations, and collective behavior. Systems biology approaches rely on combining high-throughput data generated by various omics and quantitative computational analyses to generate new integrated insights into how individual parts produce emergent phenomena. Precise definition and methodology of systems biology is not unified and often elusive (Simpson 2016), but the main emphasis is on deducing the properties of interaction networks governing cellular processes. Ongoing debate exists about the need to change perception and scientific language if we are to understand cellular functions.

The concept of “modular biology” (closely linked with the concept of synthetic biology) is based on the realization that omics approaches alone are unable to uncover and understand the “design” or “engineering” (in a functional sense) principles of living organisms (Hartwell et al. 1999). Yuri Lazebnik has called for a new formalized language that is better suited to comprehend modules in living systems (Lazebnik 2002). Inspired by Hartwell et al. (1999), he uses the putative example of an effort to understand the functions of a radio and repair it using the methodology of molecular biology: dissecting the functioning system into a pile of random smaller parts or describing the effects of their removal (as in classical developmental genetics). Such an approach would undoubtedly lead to identification of a few components that are crucial for functioning, and replacement of which would repair the radio if those components had been damaged. However, this procedure is futile if the individual components are functional but not tuned properly. Similarly, the quest of the pharmaceutical industry to find “miracle drugs” by identifying “critical molecular targets” does not often work because the malfunction may be the result of improper “tuning” of the whole system rather than damage to the critical molecular target.

On the other hand, the formal language of electronics (with components such as triggers and amplifiers) used by engineers provides direct insight into processes that the components are wired to perform. The analogy is not entirely fair because engineers have designed artifacts from first principles and formulated suitable language on the way, whereas the reverse-engineering approach of molecular biology meets systems that have evolved on their own for billions of years in complex environments. Nevertheless, biologists could learn more from taking an engineering perspective. Even the original models of gene expression regulation were inspired by Boolean logic, and many modern machines are now complex enough to foster further dialogue between biology and engineering, at least in the realm of signal transmission, processing, and interpretation (Csete and Doyle 2002). The concepts of amplification, adaptation (short and long term), robustness, insulation, attractors, bistability, waves and oscillations, memory switches, filtering, pattern recognition, discrimination of time series, hysteresis, complex logic gate operations, error correction, and coincidence detection should become staple parts of cell biology vocabulary. Cellular modules reflecting these concepts, rather than individual molecules, are of primary interest in understanding collective cell phenomena (Hartwell et al. 1999; Klipp and Liebermeister 2006; Lim et al. 2013; Mast et al. 2014). Novel bioinformatic methods can be used to search for similar network motifs, and it can be experimentally tested whether similar motifs play the same role in different contexts (Lim et al. 2013). The general functions of positive feedback (bistability, memory, switch-like behavior) and negative feedback (noise resistance, input-induced steady state) have been known for a long time (Lim et al. 2013).

The list of common motifs and architectures associated with specific functions in cells is now being expanded. For example, coherent feedforward loops often act as persistence detectors, which switch “on” only when the input persists for minimal amount of time (Lim et al. 2013). If the set of solutions for a particular problem is small enough, more analogies between artificial systems and cells should be possible to find and a table of frequent motifs with their functions established (Lim et al. 2013). There are even calls for verification of these rules by building minimal biological processing networks, with the use of a “synthetic biology” as the ultimate proof of understanding (Mast et al. 2014). However, it should be emphasized that networks and their motifs in living systems have their own specificities, because they often evolved to play multiple roles and work in unstable environments (Klipp and Liebermeister 2006). Yet, many modern artifacts are not dominated by minimal function but by modular buildup, which ensures robustness and further evolvability, so more similarities with evolving living systems could be discovered in the future (Csete and Doyle 2002). The languages of modular cell biology and molecular cell biology are complementary, because the same functional motifs studied by modular biology can be implemented by many different molecular agents: “Cell biology is in transition from a science that was preoccupied with assigning functions to individual proteins or genes, to one that is now trying to cope with the complex sets of molecules that interact to form functional modules” (Hartwell et al. 1999).

2.13 Cells in Tissues: Molecular and Modular Mechanisms of Morphogenesis

Contemporary biology is again realizing the importance of an old wisdom that multicellular animals and plants are not composed of cells in a brick-like manner, but that tissues form specialized domains by cell growth, division, and differentiation. In addition to focusing on individual cell activity in this process, the dynamic integrated whole of the organism that produces and controls cells should be considered. As in cell biology, attempts have been made to understand multicellular developing systems in terms of the information processing networks of signaling pathways and gene expression regulation (Davidson 2010). It is also understood that, along with regulatory modules embodied in protein–protein interactions and gene promoter structures, the dynamic shape of tissue needs to be taken in account. For example, gradients of signaling molecules are dynamically reshaped by changes in tissue shape (Bollenbach and Heisenberg 2015). Therefore, each specific type of cell within an organism can be fully understood only within the context of its specific position within a tissue and its function. Bottom-up molecular and modular approaches must be complemented by top-down concepts that take into account the structure of developing tissues (Levin 2012).

Understanding both the modular and interconnected nature of living systems has allowed revival of the supracellular concept of field in developmental biology and its re-formulation in a framework compatible with molecular biology (Gilbert et al. 1996; Levin 2012). Such modular fields, displaying both autonomy and hierarchy and interacting with each other, have been proposed as mediators between genotype and phenotype in both ontogeny and evolution. Unlike some early field concepts, these fields are based on genetically defined interactions between cells. Their hierarchy and establishment are influenced by genetic information, but the field concept allows a shift of focus to the supracellular level of organization (Gilbert et al. 1996).

For a long time, the ECM was considered a passive material that filled the space between cells (Rozario and DeSimone 2010). Now it is understood as a dynamic repository of signaling molecules. The ECM can inhibit or facilitate signal spreading (Yan and Lin 2009; Rozario and DeSimone 2010), as well as store the morphogens and release them upon proteolytic degradation or stimulation by additional signals (Rozario and DeSimone 2010). Moving cells reorganize the structure and position of ECM and ECM tracks the drive direction of cell migration (Rozario and DeSimone 2010). The actions of cells and ECM are thus bidirectional and complementary. More than a century after Roux defined a program of developmental mechanics, mechanical concepts are becoming the hallmark of mainstream developmental biology.

A program ridiculed by early developmental geneticists for not having achieved any mechanical understanding (Gilbert et al. 1996) now works fully within the framework of molecular biology. Developmental biology can also focus on mechanical aspects of development as a result of technological advances such as optical tweezers (Le et al. 2016), laser ablation of selected cells within tissue (Polacheck and Chen 2016), and atomic force microscopy to measure quantitatively the mechanical properties of cell/ECM surfaces at microscale resolution (Alcaraz et al. 2017). An increasing number of studies have demonstrated how the mechanical signaling within interconnected cellular–ECM nets strongly regulates growth, gene expression, and differentiation (Heisenberg and Bellaïche 2013), including mechanical aspects of regulation of cellular invasivity in normal development and in cancer establishment (Parekh and Weaver 2016).

2.14 Insights into Cytoplasm Structure in the Twenty-First Century

Together with the established tradition of associating cellular processes with membrane-bound organelles, attempts to comprehend the structure and properties of cytoplasm have reemerged 100 years after the decline of protoplasmic concepts, as nicely expressed in a quotation by T. Mitchison (2010): “Nothing epitomizes the mystery of life more than the spatial organization and dynamics of the cytoplasm.”

The aqueous phase of the cytoplasm is not a bag of freely diffusing enzymes, as often wrongly perceived in the light of classical biochemistry, but is crowded with macromolecules. Diffusive transport and partitioning of macromolecules and organelles in cytoplasm is highly restricted by steric hindrance and by unexpected binding interactions (Luby-Phelps 2013). High viscosity and crowding are thought to play major roles in the mobility of cytoplasmic components. Mobility measurements by modern techniques indeed show behavior different from mere passive diffusion. Oddly, small proteins often move faster than inert molecules (Ross 2016). Weak interactions with surrounding cytoplasmic components possibly enhance their mobility. Recent advances have accumulated sufficient evidence for the existence of membraneless or “naked” compartments in the cytoplasm. Such compartments are formed by multivalent weak interactions between low complexity repeat domains and/or distorted hydrophobic domains (Luby-Phelps 2013; Uversky 2017). Self-interaction of domains ensures phase separation of the components from the rest of the cytoplasm. Upon formation of such a compartment by polyvalent interacting proteins, monovalent interacting partners can enter the compartment and concentrate there.

Membraneless droplets could play a role in concentrating components of a cellular pathway without the need for a membrane barrier or other cage. Individual droplets of the same kind can split and coalesce, and components are constantly exchanged with the soluble pool (Weber and Brangwynne 2012). These structures thus possess a high level of internal dynamics and are characterized by liquid-like behavior, such as dripping, fusion, wetting, and the ability to become reversibly deformed when encountering a physical barrier (Uversky 2017). Droplets of different kinds (each based on a different self-interaction domain) can coexist within the cytoplasm without mixing together. Many such compartments are ribonucleoprotein granules consisting of long multivalent RNA molecules and specific RNA-binding proteins (Weber and Brangwynne 2012). Formation of membraneless compartments is condition-dependent, reversible, and controlled, including by posttranslational modification (Uversky 2017). The environment of these compartments is even more crowded than the rest of the cytoplasm (Uversky 2017). The combination of phase separation and molecular crowding can even trap together proteins with extremely low copy number (Wolde and Mugler 2014). The effects of crowding on the dynamics of signaling pathways, gene regulation networks, and metabolic networks are still not well understood, but crowding alters the diffusion of proteins and the kinetics of biochemical reactions (due to entropic changes), often in nonlinear dependence on the concentrations of molecules involved (Wolde and Mugler 2014).

Some of the ideas involving aqueous phase separation as a self-organizing mechanism trace back to 1899 or possibly earlier. E.B. Wilson proposed at the end of nineteenth century that non-membrane-bound compartments such as P-granules and Cajal bodies could be explained by the principles of colloid chemistry (Luby-Phelps 2013). Membraneless protein bodies of crystalline or quasicrystalline organization, probably formed by self-assembly, have also been known for some time. The shells of such compartments are permeable for small metabolites but otherwise keep the inside isolated from the rest of the cytoplasm (O’Connell et al. 2012). Most of these structures were discovered in bacterial cells, but examples from eukaryotes have also been described. In addition to the well-known polymerization of actin and tubulin into cytoskeletal fibers, some metabolic enzymes such as CTP synthase also tend to form fibers. Large-scale fluorescence microscopy screens revealed the localization of many supposedly cytoplasmic yeast proteins in fibers. The studies avoided overexpression artifacts and were supported by additional methods such as mass spectrometry for selected candidates (O’Connell et al. 2012). Packing of many proteins into as-yet uncharacterized structures is thus becoming evident.

Various roles for protein fibers and foci have been proposed, including efficient allosteric regulation, shielding of metabolic intermediates and their channeling into complex pathways, and storage of inactive proteins. Each of these functions has been demonstrated in particular cases but, for most proteins, the impact of assembly into aggregates is not known and the impact of the highly organized structure of the cytoplasm is currently not well documented or understood. However, it is clear that certain emergent physicochemical properties of the cell interior cannot be revealed by reductionist experiments with a few isolated components. A challenge for postreductionist biochemistry is to study biochemical phenomena far from chemical equilibrium and under physiologically relevant conditions (i.e., inside cells, in complex cell extracts, or in crowded solutions) (Kyne and Crowley 2016).

2.15 Lipid and Membrane Research in the Twenty-First Century

Although support for the widespread existence of membraneless compartments in cells is accumulating, modern research also demonstrates the vital role of biological membranes. In interplay with cytoplasmic components, membranes expand the mechanisms of cell compartmentalization and functional regulation with additional layers of complexity. Lipids, although previously overlooked as mere passive components of membranes, are now appreciated as crucial determinants of membrane properties at different scales and are a key research topic in modern cell biology (Mouritsen and Bagatolli 2015). Improved lipidomic analyses demonstrate that the diversity of lipids could match the diversity of protein species in a eukaryotic cell and that the catalogue of lipid diversity is still expanding (Saliba et al. 2015). One year after the formulation of the fluid mosaic model of plasma membranes, it was hypothesized that more stable domains exist within evenly mixed membranes (Sezgin et al. 2017). This “lipid-raft” hypothesis, based on biochemical extractions indicating stable sphingolipid and sterol-enriched compartments within membranes, was never fully accepted. However, the expanded computational, biophysical, and biochemical tool set, including molecular dynamic simulations and advanced spectroscopic methods (Sezgin and Schwille 2011, 2012; Gumí-Audenis et al. 2016; Sommer 2013), is leading to better understanding of membrane heterogeneity at different spatial and temporal levels. Like macromolecules in cytoplasm, membrane components show anomalous diffusion and undergo clustering (Honigmann and Pralle 2016). Transient self-organized domains driven by segregation of components are reported at scales from a few molecules to micrometers. Moreover, the cortical actin cytoskeleton obviously fine tunes the organization of microdomains, not only by acting as a boundary to membrane protein diffusion but also by influencing lipid organization and phase transition, which can be further facilitated or suppressed by actin (depending on other specific conditions) (Honigmann and Pralle 2016). The existence of a fine actin–spectrin network has been observed in red blood cells and recently demonstrated in neurons with the help of super-resolution microscopy (D’Este et al. 2016), indicating a general cellular phenomenon. Fast local rearrangements of the domains as a result of feedback between the local phosphoinositide composition and actin cytoskeleton are also possible (Honigmann and Pralle 2016). Like the cytoplasmic cortex, the ECM is believed to influence the mobility of membrane proteins, which has been demonstrated in the case of selective limiting of the mobility of plant plasma membrane proteins by the cell wall (Martinière et al. 2012). Differences in local lipid composition regulate the function of membrane proteins, and a substantial fraction of membrane lipids are bound to transmembrane proteins in the form of a hydrophobic solvation shell instead of being freely mobile within the bilayer (Poveda et al. 2017). The effects of lipid composition on the physical properties of a membrane are complex and difficult to predict. For example, cholesterol can increase or decrease local membrane fluidity depending on the other components (Schmid 2017).

Computational and experimental tools now allow assessment of the effect of specific compositions on membrane physical properties and protein structure in different situations (Poveda et al. 2017). Once cytoplasmic proteins are recruited to the membrane, the dimensionality of their mobility is reduced from three to two dimensions, increasing their effective concentration by orders of magnitude. Membranes thus serve as interaction platforms for proteins, which can be further fine-tuned by segregating interaction partners to specific microdomains (Honigmann and Pralle 2016; Stoeger et al. 2016). Membranes are now also understood to serve as tunable capacitors for integration and storage of information in the form of accumulation of specific signaling phospholipid species (Stoeger et al. 2016). Coincidence detection of more lipid species, or a specific lipid together with a protein interaction partner, regulates protein binding to the microdomains and membranes of different organelles (Saliba et al. 2015). Large-scale protein–protein interaction maps are now being complemented by high-throughput screens testing protein–membrane interactions and their dependence on the complex composition of the membrane and biophysical properties such as curvature (Saliba et al. 2015). The dynamic effects of lipid composition on cellular processes have been difficult to study, because membrane composition is subject to tight and fast regulation in the form of phospholipid headgroup modification, fatty acyl chain transfer, and movement of lipids between membrane leaflets (Sekereš et al. 2015). Furthermore, lipid transfer proteins in connection with membrane contact sites are being studied as regulated highways for lipid transport. Such a transport mechanism is possibly much faster than vesicular transport, previously considered to be the major agent of lipid movement between compartments (Jain and Holthuis 2017). Emerging technologies such as optogenetic activation of lipid-modifying enzymes (Idevall-Hagren and De Camilli 2015) and photoactivation of caged phospholipids (Hoglinger et al. 2014) now enable monitoring the effect of membrane composition changes on cellular processes at the physiological spatiotemporal scale.

2.16 Into the Unknown: The Future of Cell Biology

In addition to the increasing resolution and coverage of molecular measurements, discovery of some previously unknown fundamental components and mechanisms has been achieved. Discovery of RNA interference in the 1990s reshaped the perception of gene expression regulation and fostered growing interest in noncoding RNA species (Deniz and Erman 2016). There are also factors that probably have a large impact but are difficult to measure and factors whose existence we do not even suspect, the true “dark matter of cell biology” (Ross 2016). Examples of the former are the properties of intrinsically disordered proteins, small intracellular and intercellular DNA species, weak interactions impossible to detect using traditional biochemical methods, and intracellular distribution of ion species. The latter factors could be undiscovered protein–protein interaction motifs, exotic phases, undetected types of small molecules existing at low copy numbers, unknown posttranslational modifications, or new modes of collective behavior of biomolecules. With further improvement of available tools, it is possible that previously abandoned and possibly forgotten concepts in the framework of molecular biology will be revived, as happened with endosymbiotic theories and epigenetics. Cell biologists will continue to use the combination of top-down and bottom-up approaches. Detailed mechanistic characterization of individual components will be combined with large-scale systems level approaches, enabling identification of novel functional cellular modules. The future of cell biology (and of biology as a whole) also lies in capturing life processes simultaneously at different spatiotemporal scales and the integration of results into multiscale models, so that the relationship between the interactions of individual components and collective emergent phenomena can be understood.