In vivo imaging of basement membrane movement:ECM patterning shapes Hydra polyps

Roland Aufschnaiter1,*,`, Evan A. Zamir2, Charles D. Little1, Suat Ozbek3, Sandra Munder4, Charles N. David4,Li Li1, Michael P. Sarras, Jr5 and Xiaoming Zhang1,`

1Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, Kansas 66160, USA2Georgia Institute of Technology, George W. Woodruff School of Mechanical Engineering, Georgia, 30332, USA3Center for Organismal Studies, Department of Molecular Evolution and Genomics, University of Heidelberg, 69120 Heidelberg, Germany4Department of Biology 2, Ludwig-Maximilians-Universitat Munchen, 82152 Martinsried, Germany5Department of Cell Biology and Anatomy, Rosalind Franklin University of Medicine and Science, North Chicago, Illinois, USA

*Present address: Max Plank Institute of Biochemistry, 82152 Martinsried, Germany`Authors for correspondence (aufschna@biochem.mpg.de; mzhang2@kumc.edu)

Accepted 7 July 2011Journal of Cell Science 124, 4027–4038� 2011. Published by The Company of Biologists Ltddoi: 10.1242/jcs.087239

SummaryGrowth and morphogenesis during embryonic development, asexual reproduction and regeneration require extensive remodeling of theextracellular matrix (ECM). We used the simple metazoan Hydra to examine the fate of ECM during tissue morphogenesis and asexualbudding. In growing Hydra, epithelial cells constantly move towards the extremities of the animal and into outgrowing buds. It is not

known, whether these tissue movements involve epithelial migration relative to the underlying matrix or whether cells and ECM aredisplaced as a composite structure. Furthermore, it is unclear, how the ECM is remodeled to adapt to the shape of developing buds andtentacles. To address these questions, we used a new in vivo labeling technique for Hydra collagen-1 and laminin, and tracked the fate of

ECM in all body regions of the animal. Our results reveal that Hydra ‘tissue movements’ are largely displacements of epithelial cellstogether with associated ECM. By contrast, during the evagination of buds and tentacles, extensive movement of epithelial cells relativeto the matrix is observed, together with local ECM remodeling. These findings provide new insights into the nature of growth and

morphogenesis in epithelial tissues.

Key words: Extracellular matrix, ECM, Hydra, Epithelium, Morphogenesis, Epithelial migration, Tissue remodeling

IntroductionTissues of multicellular animals are composed of cells mounted

on scaffolds of extracellular matrix (ECM). The first ECM is laid

down early in development and gives rise to basic epithelial

organization (Tyler, 2003). This usually happens before tissue

folding and morphogenesis (Ingber, 2006). Although we have

considerable understanding of how different cellular behaviours

contribute to the creation of shape in animal tissues, little is

known about the dynamic organisation of ECM during this

process.

Cell movements are one of the most important mechanisms

driving animal morphogenesis. Single cells or isolated groups of

cells can migrate based on invasive activity and cell–substrate

interactions, as in vertebrate neural crest cells, Hydra nematocyte

precursors, the lateral line in zebrafish, or Drosophila border

cells. Morphogenetic movements can also include large areas

of a tissue, such as vertebrate gastrulation and neurulation

or Hydra ‘tissue movements’. For most of these large-scale tissue

rearrangements, it is not yet clear whether cells move actively

or whether they are passively displaced by global tissue

deformations, which include cells and associated ECM. One

way to address this question is to track the fate of cells and at the

same time track changes in the associated ECM, as has been done

in vivo in the avian embryo (Benazeraf et al., 2010; Zamir et al.,

2006; Zamir et al., 2008).

Tissue of the freshwater polyp Hydra is organized as an

epithelial double layer; the outer epithelium is the ectoderm and

the inner the endoderm. These are separated by an intervening

ECM, which is called the mesoglea. The mesoglea has the

structural and molecular organisation of basement membranes in

higher animals and is synthesized by epithelial cells of both

layers (Epp et al., 1986; Sarras and Deutzmann, 2001). In

regularly fed polyps, epithelial cells in the body column

continually undergo cell division. This does not lead to an

increase in body size because, under steady state conditions,

tissue growth is balanced by tissue loss at the ends of the body

column and in developing buds. These tissue movements have

been systematically investigated with the help of various in

vivo markers for ectodermal and endodermal epithelial cells

(Campbell, 1967b; Campbell, 1973; Otto and Campbell, 1977a;

Shostak and Kankel, 1967; Shostak et al., 1965; Wittlieb et al.,

2006). However, the role of the mesoglea has remained unclear.

Two opposing views are possible: (1) the mesoglea is a stationary

structure that serves as a substratum for active epithelial motility

(Shostak and Globus, 1966; Shostak et al., 1965), or (2) tissue

movements are the result of continuous tissue expansion that

includes both epithelia and the mesoglea (Campbell, 1973;

Campbell, 1974).

During bud formation an initially flat area of the body wall

in the lower gastric region evaginates and forms a small new

Research Article 4027

Journ

alof

Cell

Scie

nce

animal. Although it has been shown that bud outgrowth is basedon movement of epithelial cells from the mother polyp (Otto andCampbell, 1977a) and involves lateral intercalation of epithelial

cells (Philipp et al., 2009), little is known about how the ECM inthe bud is formed. Elevated levels of mesoglea synthesis in thebud (Hausman and Burnett, 1971; Zhang et al., 2007; Zhang et al.,

2002) indicate the need for new ECM material, but it remainsunknown whether the mesoglea, like the epithelial cells, isrecruited from the parent, and how the mesoglea is reorganised

during bud outgrowth while maintaining a functional epithelialbasement membrane at the same time.

To address these questions we used a method to label majorcomponents of Hydra mesoglea in the living animal. Themethod was originally developed for the avian embryo and

uses fluorescently tagged primary antibodies to target ECMcomponents (Czirok et al., 2004; Rongish et al., 1998). Theantibodies were microinjected into the mesoglea where they

bound stably to their epitopes without disturbing physiologicalfunctions. Grafting of labeled tissue fragments into unlabeledanimals and precise local injections of antibody allowed us to

track the fate of mesoglea in live animals. Our findings show thatthe mesoglea is a surprisingly dynamic structure. Only in thehead region did it remain stationary. In the adjacent body columnand tentacles it was continuously displaced toward the ends of the

animal. During bud evagination, it was stretched and remodeledto create the bud morphology. The dynamic displacement ofthe mesoglea along the body column and the tentacles largely

overlapped with the movement of epithelial cells. These tissuemovements, which might superficially appear as active cellularmigrations, are largely passive tissue displacements. At sites of

tissue evagination, however, the mesoglea was dramaticallyremodeled and epithelial cells moved relative to the mesoglea.

Our results, together with findings obtained in vertebratesystems (Czirok et al., 2004; Davidson et al., 2008; Larsen et al.,2006; Zamir et al., 2006; Zamir et al., 2008), suggest that our

understanding of tissue movements and morphogenesis needs tobe re-evaluated with respect to the role of ECM remodeling.

ResultsIn vivo labeling of Hydra ECM with non-interferingantibodies

We succeeded in labeling two Hydra ECM components,collagen-1 and laminin, by injecting fluorescently taggedmonoclonal antibodies into the mesoglea of living animals. To

test the specificity of the labeling, animals were injected withboth antibodies at the same time, fixed after 24 hours, andexamined with confocal microscopy (Fig. 1A–D). Antibodiesagainst Hydra laminin localized to the subepithelial basal lamina

(Fig. 1A,B,D), whereas those against Hydra collagen-1 localizedto the central interstitial matrix of the mesoglea (Fig. 1A,C,D).This is consistent with data obtained from previous studies

(Sarras et al., 1993; Shimizu et al., 2008). Unbound antibodywas taken up by epithelial cells and accumulated in terminallysosomes (Fig. 1A, arrowheads) where it could be clearly

distinguished from labeled mesoglea.

Budding rate and bud development were not different betweenlabeled and control polyps. Head and foot regeneration proceededat the same rate as in unlabeled animals. Finally, labeled polyps

could not be distinguished from control animals by morphology,contraction–elongation behaviour or feeding behaviour. Takentogether, in vivo labeling of mesoglea components in Hydra is

specific and does not interfere with major developmental and

physiological processes.

Hydra mesoglea is displaced along the body column ofthe animal

Small grafts containing labeled mesoglea were displaced down

the body column over the course of several weeks (Fig. 2A–C).

Both collagen-1 and laminin labels overlapped completely during

these displacements and also during tissue evagination (n512,

data not shown), indicating that basal lamina and interstitial

matrix function as a unit. Since label stability and the label to

background ratio were better in specimens labeled only with

collagen-1, we used it in all subsequent experiments.

The displacement speed of labeled grafts along the body

column varied considerably in different regions of the body

column (Fig. 2D). In the region from 0–9% body length, which

contains the hypostome, the mesoglea was stationary (see below).

Graft displacement was slow from 10–25% body length and

reached a peak between 25–50% body length. From 50–75%

body length, where the majority of budding events happen, the

displacement slowed again and reached its lowest level in the

peduncle region (75–100% body length). Previous studies have

shown that Hydra epithelial movements are dependent on cell

proliferation rates, which can be manipulated by different feeding

regimes (Bosch and David, 1984; Holstein et al., 1991; Otto and

Campbell, 1977b). Mesoglea displacement speed also depended

on feeding. Animals fed daily (Fig. 2D, purple bars) showed a

Fig. 1. Localization of injected antibodies to the three layers of Hydra

mesoglea. (A) Optical section through Hydra body wall showing ectoderm,

endoderm and the intervening mesoglea in differential interference contrast

(DIC) and immunofluorescent labeling. The mesoglea was labeled by

injection of monoclonal antibodies mAb52 raised against Hydra laminin

(green) and mAb39 raised against Hydra collagen-1 (red). Excess antibody

is taken up by ectodermal cells and sequestered in terminal lysosomes

(arrowheads). (B–D) Labeled mesoglea at higher magnification showing

laminin localization in the two subepithelial basal laminae (B,D) and

collagen-1 localization in the central interstitial matrix (C,D). Scale bars:

20 mm (A), 5 mm (D).

Journal of Cell Science 124 (23)4028

Journ

alof

Cell

Scie

nce

higher speed in all body regions compared with animals fed every

other day (Fig. 2D, blue bars).

Mesoglea increases in size in the upper body column,

decreases in size in the budding region and is degraded at

the basal disc

The size of labeled mesoglea grafts increased in the upper body

column (Fig. 3A–D). This increase could be observed starting

from ,10% body length until the label reached the budding zone.

In the budding region, labeled mesoglea grafts decreased in size

(Fig. 3E–F, arrowheads). Because the decrease in size of the

labeled graft depended on the number of budding events it was

involved in, it appeared to be due to recruitment of labeled

mesoglea into the evaginating buds (see below). However, among

a large number of grafts in our record (n.50), five moved

through the budding area without being directly involved in a

budding event. Nonetheless, these grafts also exhibited a decrease

in size ranging from 43% to 72% of their labeled area, suggesting

that label loss into buds is not the only reason for the decrease in

label size.

Increases and decreases in labeled mesoglea size (Fig. 3G)

paralleled variations in the displacement speed of mesoglea along

the body column (Fig. 2D). During displacement along the

peduncle, the size of labeled grafts did not change. Only when

labeled mesoglea reached the aboral end of the animal, did it

decrease in size and ultimately disappear (data not shown). In

some animals, we observed fragments of labeled mesoglea in

the lumen, suggesting that at least part of the mesoglea, which

reached the end of the animal is digested.

Mesoglea is displaced toward the tip of the tentacles, but

is stationary in the head region

Labeled mesoglea in the tentacles was displaced toward the tip of

the tentacle (Fig. 4A,B, star). The displacement took about 7 days

to move from the base of the tentacle to the tip (n520) as shown in

Fig. 4A,B. No differences in label displacement were observed

between animals fed daily compared with animals fed every other

day (data not shown). Once the labeled mesoglea reached the tip of

the tentacle, it shrank in size and disappeared (data not shown).

Similarly to the situation in the basal disc, we observed fluorescent

material in the endodermal cells of the tentacles, which suggests

internalization of degraded mesoglea by the cells.

Labeled mesoglea in the head region remained stationary

(Fig. 4A,B, st) and a sharp border could be observed between the

Fig. 2. Mesoglea displacement along the body column.

(A–C) A typical graft in the body column labeled for

collagen-1 (green) on day 1, day 9 and day 34 after grafting.

The labeled area is progressively displaced toward the

aboral end of the animal. (D) Mean label displacement

speed (x-axis) in different regions of the body column

(y-axis). Tip of the hypostome, body column position 0%;

basal disc, body column position 100%. Animals fed daily

(purple bars) and animals fed every other day (blue bars)

were examined. Each bar represents the mean value of

individual label velocities (n510–14) for each body region

at each feeding regime. Scale bar: 1 mm.

Hydra ECM dynamics in vivo 4029

Journ

alof

Cell

Scie

nce

stationary head mesoglea and the mesoglea, which was displaced

outward along the tentacle (Fig. 4B, arrowhead). The stationary

region included the hypostome of the animal and extended to

about 9% body column length: when we grafted heads of labeled

donor animals, cut off at 9% (or less) body length, onto gastric

columns of non-labeled acceptor animals (or vice versa), the

label showed no displacement (Fig. 4C). Grafts at body column

position lower than 9%, however, were subject to displacement

(Fig. 2D). Labeled mesoglea within the stationary head region

faded without visible lateral displacement, which suggests local

turnover.

Parental mesoglea is recruited and stretched alongevaginating buds

Fig. 5A–C shows a typical example of mesoglea recruitment

from the parental body column into a bud. At an early stage

of evagination, the labeled mesoglea formed a contiguous

fluorescent patch (Fig. 5A, green label). Twenty-four hours

later, the fluorescent signal was retained only at the very tip

(Fig. 5B, arrow) and at the base of the developing bud (Fig. 5B,

arrowheads). As a result of extensive stretching, intervening

mesoglea appeared to have lost a significant amount of its

immunofluorescence signal. The stretched state of bud mesoglea

was also clearly detectable in fixed animals immunostained for

collagen-1. Trans-mesogleal pores (Shimizu et al., 2008), which

are circular in the parent animal (Fig. 5E, arrowheads) became

elongated along the oral–aboral axis in the developing bud

(Fig. 5E, arrows). Furthermore, fixed and immunostained

animals showed higher collagen concentrations in the tip of the

bud (supplementary material Fig. S1; Fig. 6B), confirming our in

vivo observations, that this region was excluded from stretching.

Native mesoglea washed free of cells (Day and Lenhoff, 1981)

can be stretched mechanically and snaps back to resting length

when released from the deforming force (Hausman and Burnett,

1969; Wainwright et al., 1976), indicating a high inherent

elasticity (Sarras and Deutzmann, 2001). The conspicuous

stretched appearance and the lowered collagen content of

mesoglea in early bud stages (Fig. 5D,E, Fig. 6A,B) raised the

question of whether the ECM is simply in a state of elastic

stretching imposed by morphogenetic forces from the overlying

epithelia or whether the ECM of the early bud is already stably

deformed. To answer these questions, we selected intact early

buds (stage 3–5), isolated their mesoglea and compared the shape

of the bud before and after isolation (Fig. 5F,G). In all cases

(n510), the isolated mesoglea shortened about 40–60% in length

along the oral–aboral axis of the animal, suggesting that it

is under tension when integrated in tissue. Little change

perpendicular to this axis was observed (Fig. 5F,G). However,

in all observed specimens, the tube shape of the evaginated bud

was preserved, indicating that the mesoglea of early buds is

already stably deformed.

Biosynthesis of new collagens is not required for early

stages of bud evagination

Transcription of collagens and laminin is upregulated in buds

(supplementary material Fig. S2) (Deutzmann et al., 2000;

Fig. 3. Size changes in mesoglea grafts

moving down the body column. (A–C) A graft

labeled for collagen-1 (green) and ectodermal

epithelial cells (yellow) in the upper body

column at three different time points. The length

of the labeled graft increases from 6.6% body

length on day 1 to 21% body length on day 11.

(D) Length (y-axis) of seven individual grafts

labeled for collagen-1 during a time period of 2

weeks (x-axis). The initial length of all grafts is

between 5–7% body length and the initial graft

position between 10–20% body length on day 1.

The length of the labeled grafts was measured

every other day. (E,F) A graft labeled for

collagen-1 (green) before (E) and after moving

through the budding region (F). Note the

decrease in size of the labeled area (arrowheads).

A decrease in size was typical for all grafts

that moved through the budding region.

(G) Schematic drawing of labeled graft

expansion (1) in the upper body column and

labeled graft shrinkage (2) in the budding region

during graft displacement along the body

column. Scale bar: 500 mm.

Journal of Cell Science 124 (23)4030

Journ

alof

Cell

Scie

nce

Fowler et al., 2000; Zhang et al., 2007; Zhang et al., 2002),

indicating that the developing bud is a site of extensive mesogleal

biosynthesis. Nevertheless, owing to rapid bud outgrowth during

early stages, collagen levels drop compared with levels in the

parent (Fig. 6A). The lowest collagen levels were observed in 24

hour buds. By 48 hours, buds exhibited collagen levels only

slightly lower than those in the adult body column, and no

difference could be detected between parent and 72 hour buds.

Changes in the fluorescence intensity were more pronounced for

collagen-4 than for collagen-1 (Fig. 6A). The reason for this

effect remains unclear; however, the results clearly indicate that

the concentrations of both collagens drop during early bud stages

and they only return to original levels in later stages.

To investigate the importance of collagen synthesis for bud

evagination we applied the chemical inhibitor 2,2 Dipyridyl

(DP). DP is a competitive inhibitor of prolyl-hydroxylase, which

Fig. 4. Mesoglea fate in the tentacles and the

head region. (A,B) Labeled collagen-1 (green)

in the head region and the tentacle base 3 hours

after injection and 7 days after injection. The

label in the tentacle (star) moves to the tip of the

tentacle by day 7, the label in the head region

(st) remains stationary. (C) Position (y-axis) of

the border between labeled and unlabeled tissue

of seven individual head grafts over the course

of 11 and 13 days (x-axis). No displacement is

observed. Scale bar: 500 mm.

Fig. 5. Mesoglea remodeling during bud formation.

(A–C) A graft labeled for collagen-1 (green) and epithelial

cells (yellow) during bud formation. Tissue labeling

preceded bud evagination. At an early stage of evagination,

the labeled mesoglea forms a continuous fluorescent patch

(A). 24 hours later fluorescence is only visible at the tip

(B, arrow) and the base of the bud (B, arrowheads), whereas

fluorescence in the intervening region has disappeared. Note

the decrease in size of the labeled mesoglea, which remains

in the parent body column (A–C, arrowheads). (D) Fixed

stage-6 bud stained for collagen-1. (E) Detail from D

showing the border region between parent and bud. The

collagen-1 matrix in the bud has a clearly stretched

appearance: mesoglea pores, which are circular in the parent

animal (arrowheads) appear stretched along the oral–aboral

axis of the developing bud (arrows). (F) A living bud stage

4–5 labeled for collagen-1 (green). (G) Isolated mesoglea of

the same bud as in F. Some endodermal cell debris

(arrowhead) remains trapped in the ECM. Scale bars:

500 mm (C,G), 50 mm (E).

Hydra ECM dynamics in vivo 4031

Journ

alof

Cell

Scie

nce

catalyses a critical step required for formation of the collagen

triple helix (Kivirikko et al., 1989). DP interferes with collagen

biosynthesis in vertebrate and invertebrate models, including

Hydra (Kawano et al., 2005; Mizoguchi and Yasumasu, 1983;

Sarras et al., 1991b; Sarras et al., 1993). Two sets of experiments

were performed to test the effect of DP on bud formation. In the

first experiments, animals were treated with 0.05 mM DP starting

at the first visible sign of bud formation (stage 1–2). Bud

development was normal until the first tentacles appeared, but

later developmental stages were blocked or severely inhibited

(Fig. 6C,D). By 72 hours, control buds had formed a basal disk

and were ready to detach from the parent animal. In treated

animals, formation of the basal disk at the border to the parent

animal was severely inhibited (Fig. 6C,D): 76% of the buds did

not form a basal disk, 11% formed an incomplete disk and only

8% formed a complete basal disk (n583). In addition, tentacle

outgrowth was greatly reduced in treated animals. Whereas the

tentacles in control buds had grown out to an average length of

12.1 (±3.5) times their diameter at 72 hours, the length of

tentacles of treated animals was only 2.1 (±0.8) times their

diameter (n530 tentacles in 15 animals per measurement,

compare Fig. 6C and 6D). Washout experiments were

performed to determine whether the effects of the inhibitor

were reversible. Although tentacle outgrowth recommenced in all

treated animals, foot differentiation did not occur, and the

majority of inhibited buds did not detach from the parent animal

(not shown). Immunostaining of buds after 48 hours of treatment

revealed that collagen-4 (Fig. 6C,D) and collagen-1 (not shown)

were greatly reduced in comparison to controls. In a second set of

experiments, we used synchronized animals (Sudhop et al., 2004)

to determine the effects of DP on bud initiation. Budding animals

were synchronized by a weekly regime of feeding for 5 days and

starving for 2 days. After resumption of feeding, about 50% of

large, budless polyps started to form a bud by day 3. We pre-

incubated budless animals starting on day 2 with DP and

monitored budding events over the next 36 hours. Treated

animals initiated buds and we observed no difference in the

budding rate of treated versus untreated animals (0.1 mM DP,

56%, n554; 0.05 mM DP, 48%, n558; control 0.1% DMSO,

53%, n555). The effects of the drug on later bud development

were analogous to those described above. Hence we conclude

that DP does not affect the initial phase of tissue evagination

during budding.

Mesoglea is subject to different levels of proteolytic

digestion depending on the body region

Stable deformation of mesoglea, as occurs during bud formation,

requires remodeling of the collagen and laminin network. To

investigate mesoglea turnover at these sites, we prepared extracts of

appropriate tissue pieces from different body regions and assayed

their ability to degrade Hydra laminin. Laminin was chosen as a

marker for mesoglea integrity because western blotting for Hydra

laminin, in contrast to collagen-1, yielded a stronger and more

reproducible signal. Our antibody against Hydra laminin (mAb52)

produced a clear band at about 200 kDa. As shown in Fig. 7, Hydra

laminin was rapidly degraded by early bud extracts and basal disc

tissue. Body column extracts and head extracts also degraded

laminin, but at a slower rate. a-Tubulin, which was also present in all

tissue lysates, was not significantly degraded under the conditions

used. This suggests that the laminin degradation observed is not the

result of non-specific protease activity present in tissue lysates.

Thus, the body-part-specific differences in laminin proteolysis

appear to reflect the different levels of mesoglea remodeling

observed in our in vivo tracking experiments.

Mesoglea and overlying epithelial cells move at similarrates in most regions of the animal

The apparent similarity between previously reported cell

movement patterns and the mesogleal displacement patterns,

which we observed in the current study, suggested that epithelial

cells move in concert with their associated mesoglea. To address

this question directly, we labeled epithelial cells and the mesoglea

Fig. 6. The role of collagen synthesis during bud formation. (A) Relative protein concentration of collagen-1 and collagen-4 in different bud stages. The graph

shows average fluorescence intensities of collagen-1 and collagen-4 in 24 hour, 48 hour and 72 hour buds (n510 for each time point). (B) For the values in the

graph, the fluorescence intensity of immunocytochemically labeled specimens was measured along a line in the center of the bud (yellow lines) and compared with

the fluorescence intensity of the upper body column of the parent animal (blue line). (C) Control bud development at 24, 48 and 72 hours and a 48 hour control

bud stained for collagen-4. (D) A typical example of bud development under continuous treatment with 0.05 mM DP and a 48-hour-treated bud stained for

collagen-4. Scale bar: 500 mm.

Journal of Cell Science 124 (23)4032

Journ

alof

Cell

Scie

nce

in the same region and compared their displacement dynamics.

Epithelial cells move in both directions along the body column: up

toward the tip of the tentacles and down toward the basal disc

(Campbell, 1967b). The dividing line between upward and

downward movement, the ‘stationary region’, is located in the

upper body column. In our experiments, the stationary region was

located in the upper body column at ,5–10% body length in the

endoderm and at ,13–18% body length in the ectoderm (Fig. 8E),

which is consistent with previous reports (Campbell, 1967b;

Campbell, 1973; Otto and Campbell, 1977b). In the body column

below ,18% body length, all three layers, mesoglea, endoderm

and ectoderm, moved towards the basal disc. Furthermore,

longitudinal expansion of all three tissue elements was observed

concomitant with body column motion (Fig. 8A,B,A9,B9).

Mesogleal and endodermal markers exhibited virtually complete

motion pattern overlap, whereas the ectoderm moved downward at

a slightly lower speed (Fig. 8A,B,A9,B9, compare also with

Fig. 3A–C). The ectodermal cell sheet thus resembled a ‘sleeve’

being pushed up relative to the mesoglea and endoderm (Fig. 8B9).

In the vicinity of the budding region and in the outgrowing bud

itself, mesoglea and the overlying cells of both layers moved at

different rates (see below). In the lower peduncle, both epithelial

Fig. 7. Mesoglea turnover in different body regions estimated as laminin

stability in tissue lysates. Isolated mesoglea samples containing laminin were

incubated in tissue lysates from different body regions of the animal as

indicated. For head lysate, only the head without tentacle tissue was used

(compare Fig. 4C, st). Bud lysate was taken from bud stages 3–5. Samples

taken at different time points were analyzed by western blotting using

antibodies against laminin and a-tubulin. The a-tubulin is not significantly

degraded in any of the lysate samples; the a-tubulin blot shown corresponds

to the basal disc lysate.

Fig. 8. Epithelial cell and mesoglea dynamics in the body column and tentacles. Representative examples of labeled tissue displacement in individual animals.

(A,B) A graft labeled for collagen-1 (green), ectodermal epithelial cells (bright yellow) and endodermal epithelial cells (orange) in the upper body column on day

1 and day 8. (A9,B9) Magnified views. (A0,B0) Schematic views. The perspective of A is about 180˚rotated from B, as indicated by the position of cell cluster 2 in

A0 and B0. (C,D) Labeled patch of collagen-1 (green) and ectodermal epithelial cells (yellow) in a tentacle on day 0 and day 4. Cells and ECM retain a 100%

overlap. (E) Results of mesoglea and cell sheet movement illustrated schematically based on observations in multiple animals (n510 in head and tentacle base,

n510 in tentacle, n530 in the upper body column, n550 in budding area and bud, n515 in the lower peduncle region). Regions where epithelial cells and

mesoglea move together (blue) are shown. In the upper body column, the ectodermal layer moves slightly slower than the endodermal layer and the mesoglea

(asterisk). The stationary region for ectodermal epithelial cells is located at a body column position of about 13–18% (yellow double-headed arrow) and for

endodermal epithelial cells at a body column position of about 5–10% (orange double-headed arrow). Scale bars: 1 mm (B), 500 mm (B,D).

Hydra ECM dynamics in vivo 4033

Journ

alof

Cell

Scie

nce

layers and the adjacent mesoglea were displaced as a composite

structure until they reached the aboral end of the animal (data not

shown). In the tentacles, only ectodermal cells and the associated

mesoglea were examined. Both tissue elements moved to the tip of

the tentacle at identical rates (Fig. 8C,D), which contrasts with the

situation at the tentacle base (see below). The tissue displacement

velocities of epithelial cells and mesoglea observed at different

body column positions and in the tentacles are in good agreement

with previous measurements on Hyda vulgaris (Otto and

Campbell, 1977b) and are about 30% slower than similar

measurements on Hydra littoralis (Campbell, 1967b) and Hydra

viridissima (Campbell, 1973).

Mesoglea and epithelial cells move at different rates during

tentacle and bud formation

Epithelial cells of the tentacle originate in the body column

(Fig. 3A–C) (Campbell, 1967b; Hobmayer et al., 1990),

whereas the new tentacle mesoglea is made at the base of the

tentacle (Fig. 4A,B). Consequently, when cells moved into the

tentacles, they appeared to leave the stationary mesoglea in

the head region and moved onto newly forming mesoglea at the

tentacle base (Fig. 9A,B). Once in the tentacle, ectodermal cells

were displaced together with their underlying mesoglea

(Fig. 8C,D). A very similar situation could also be observed

during budding. Numerous cells of both epithelial layers moved

from the parent body column into developing buds (Otto

and Campbell, 1977a), whereas the amount of recruited

mesoglea remained relatively small (Fig. 9C–E, Fig. 4D,E;

supplementary material Fig. S3B–E). The area of recruited cells

had a diameter approximately 3–4 times larger than the area of

recruited ECM (Fig. 9F, n55).

DiscussionThe new experimental approach of labeling ECM in living

systems (Davidson et al., 2008; Larsen et al., 2006; Ohashi et al.,

1999; Sivakumar et al., 2006) provides a promising tool to assess

the role of the ECM during tissue growth and morphogenesis. At

the whole organism level, such an analysis has so far only been

done on the vertebrate avian embryo (Czirok et al., 2004; Zamir

et al., 2006; Zamir et al., 2005; Zamir et al., 2008). These studies

Fig. 9. Epithelial cell and mesoglea dynamics

at the tentacle base and during budding.

(A) Head region of an animal with labeled

mesoglea (green) extending from the head

region into the tentacle and a group of labeled

ectodermal epithelial cells (yellow, circled in

red) between two tentacles at day 0.

(A9) Schematic. (B) On day 2, the group of

epithelial cells (circled in red) have moved from

the stationary head region into the tentacle. Note

the gap with strongly reduced green fluorescence

at the proximal end of the tentacle (arrowheads)

indicating mesoglea expansion. (B9) Schematic.

(C–E) A graft labeled for collagen-1 (green) and

epithelial cells (yellow) in the body column over

the course of three consecutive days before

(C) and during (D,E) budding. Cells move

relative to their underlying substrate into the

newly developing animal (D,E). Note that only a

small part of the labeled, parental mesoglea is

recruited into the bud (arrowheads). (F) Fate

map of epithelial cells and ECM of the parent

body column recruited into a bud. The schematic

shows the area of epithelial cells recruited to the

bud (yellow) compared with the area of ECM

recruited to the bud (green). The fate map was

constructed based on an idealized body column

diameter (Otto and Campbell, 1977a). Scale

bars: 500 mm.

Journal of Cell Science 124 (23)4034

Journ

alof

Cell

Scie

nce

revealed that the ECM is unexpectedly dynamic, and contributes

significantly to overall tissue movements during gastrulation and

neurulation. Here, we have investigated the behaviour of the

basement membrane system in a simple invertebrate animal, the

cnidarian Hydra. The epithelium of Hydra is one of the earliest

examples of true epithelial organisation in metazoan evolution

(Tyler, 2003). The surprisingly dynamic nature of Hydra ECM

observed in our experiments – together with similar observations

in the distantly related vertebrate embryo – suggests that ECM in

the tissues of all multicellular animals has a more dynamic nature

than so far assumed.

The role of mesoglea during Hydra tissue movements

Although the mesoglea appears to be a stable structure and can

be isolated intact free of cells from Hydra, our experiments in

vivo with patches of labeled mesoglea indicate that it is in

fact a dynamic structure that undergoes constant remodeling

(Fig. 10A). Mesoglea is produced in the upper body column and

displaced to the basal disk, where it is degraded. Similarly,

mesoglea is produced at the tentacle base and displaced to the tip

of the tentacle, where it is degraded. This pattern of ECM flow

from central body regions to the periphery of the animal is similar

to the flow of epithelial cells, and overlaps with it to a significant

degree (Fig. 8E). Thus, not only epithelial sheets but also the

intervening mesoglea is subjected to a constant centrifugal tissue

renewal. These observations reject the assumption that the

general nature of Hydra tissue movements represents active

epithelial migration relative to a stationary mesoglea (Burnett and

Hausman, 1969; Shostak and Globus, 1966; Shostak et al., 1965).Rather, they support the view of Campbell (Campbell, 1973;

Campbell, 1974) that the apparent ‘movement’ of epithelial cellsrelative to the morphology of the animal can be understood ascontinuous ‘outward’ expansion of the whole tissue.

Despite these general similarities in the dynamic behaviour

of epithelial cells and the mesoglea, there is not a simple

relationship between cells and matrix. At least five different

situations can be distinguished: (1) Mesoglea in the head region

is very stable and appears to change neither in size nor in

position. The overlying epithelial cells, however, undergo

continuous proliferation and slow movement toward the base of

tentacles and the tip of the hypostome (Campbell, 1967b; Dubel,

1989). (2) Mesoglea in the upper gastric region grows in length

along the oral aboral axis essentially at the same rate as the

overlying epithelial sheet expands due to cell proliferation

(Fig. 8A,B). (3) Mesoglea at the base of tentacles expands

extensively to accommodate new epithelial cells, which flow in

from the head region (Fig. 9A,B). (4) Similarly, mesoglea of the

evaginating bud is subjected to extensive stretching and thinning

to accommodate large numbers of epithelial cells, which flow in

from the parent body column (Fig. 9C–E). (5) In the distal

tentacles, labeled mesoglea patches remain constant in size and

move together with overlying epithelial cells distally along the

tentacle (Fig. 8C,D). Similarly, patches of stained mesoglea in

the peduncle change little in size while moving down the body

column together with overlying epithelial cells.

Fig. 10. Fate map of Hydra mesoglea and a model of

cell–matrix interactions during growth and

morphogenesis. (A) Mesoglea (a) and epithelial cell

(b) dynamics. Black arrows in a indicate the direction of

mesoglea movements; shaded areas indicate mesoglea

stretching and growth; white areas indicate stationary

mesoglea in the head and bud tip; stars mark sites of

mesoglea degradation. Yellow arrows in b indicate the

direction of epithelial cell movements; shaded area marks

the area of epithelial cell proliferation (Campbell, 1967a;

Holstein et al., 1991). (B) Mesoglea growth in the body

column is associated with cell proliferation. (C) Mesoglea

stretching (top) and cell movements (bottom) during steady

state tentacle evagination. Light green represents the area of

mesoglea expansion (double-headed arrow); dark green

represents the region of stable mesoglea; arrow indicates

the direction of cell movement. The relationship between

cells and substrate changes during these morphogenetic

movements: cells move over stable mesoglea (1); cells

move over expanding mesoglea (2); cells moved together

with mesoglea (3). (D) Expansion of the mesoglea (top) and

cell movements (bottom) during bud evagination. Labeling

analogous to that in C.

Hydra ECM dynamics in vivo 4035

Journ

alof

Cell

Scie

nce

These apparently diverse behaviours of cells and matrix indifferent body parts are consistent with the following principles.

Epithelial sheets can change shape easily by rearrangingindividual cells within the epithelial sheet, as can be observedduring their movement into buds and tentacles (Clarkson andWolpert, 1967; Philipp et al., 2009). Furthermore, epithelia are

able to detach from the mesoglea in situations where they moveover its surface. By comparison, the mesoglea is rigid andrequires biochemical remodeling to change shape, including

both softening to permit stretching of existing networks andsubsequent new synthesis to restore a mature mesoglea structure.Most interestingly, and in contrast to epithelial cells, the

mesoglea is made separately in every body part (Fig. 10Aa)and these ‘compartments’ of mesoglea do not mix (except for thebud mesoglea, which is derived from a small patch of parentalmesoglea and subsequently strengthened by newly synthesized

mesoglea components). These results suggest the need for localpatterning signals to achieve this body-part-specific mesogleafate. Because mesoglea is made by the overlying epithelial sheet,

this implies that the epithelial cells in different regions of theanimal secrete a different repertoire of modifying enzymes anddifferent amounts of new mesoglea components according to the

local need.

Mechanisms of mesoglea morphogenesis in differentbody regions

In view of the mechanical stability of mesoglea (Day andLenhoff, 1981; Hausman and Burnett, 1969; Wainwright et al.,1976), it is clear that changing mesoglea shape, as occurs during

bud formation, tentacle formation and more slowly duringbody column expansion, requires alterations to the network ofconstituent macromolecules. This might involve proteolytic

remodeling of the mesoglea. Results shown in Fig. 7demonstrate that enzymes that degrade laminin displaydifferent activities in different body parts of the animal, which

is consistent with the level of mesoglea shape change observedduring in vivo tracking experiments. Lysates from evaginatingbuds for instance, where high rates of mesoglea shape changeoccur, display much higher rates of laminin degradation

compared with head lysates, where the mesoglea is comparablystable. Matrix metalloproteinases (MMPs) have been shown tocleave ECM components in vertebrates (Page-McCaw et al.,

2007) and two such MMPs have been described in Hydra.HMMP is expressed in the distal end of tentacles and the basaldisc (Leontovich et al., 2000; Shimizu et al., 2008). It has been

shown to cleave collagen in vitro and might be involved in thedegradation of mesoglea at the tip of the tentacles and the basaldisc. MMP-A3 (Munder et al., 2010) is expressed in a narrow

region at the base of late bud stages and might function in theseparation of buds from parent animals. In addition to theseMMPs, there are at least five more MMPs of unknown functionpresent in the genome of Hydra (Chapman et al., 2010) (data not

shown). Local activity of these metalloproteinases during budevagination, at the base of tentacles and in the upper bodycolumn could, in principle, loosen the stable mesoglea structure

and permit shape changes.

Although enzymatic remodeling of mesoglea by matrixmodifying enzymes such as MMPs might help to make existing

ECM scaffolds more susceptible to shape change by ‘softening’rigid networks, the shape change itself must be brought aboutby physical forces created by the morphogenetic activities of

overlying cells. Mesoglea growth in the upper body column isassociated with epithelial expansion (Fig. 8), which happens as a

result of cell proliferation (Otto and Campbell, 1977b). Changingthe feeding rate affects both cell proliferation (Holstein et al.,1991; Otto and Campbell, 1977b) and mesoglea displacementrates, suggesting that mesoglea growth in this region is related to

epithelial cell proliferation. In the simplest case, cell proliferationand the resulting increase in epithelial area could directlygenerate the physical force necessary to stretch the underlying

mesoglea (Fig. 10B). Recent observations based on clones ofproliferating GFP-expressing epithelial cells in the body column(Wittlieb et al., 2006) (R.A., unpublished results) show that the

clones are aligned along the oral–aboral axis of the animal,supporting the hypothesis of cell proliferation as the driving forcefor mesoglea expansion along that axis. It remains unclear,however, why the mesoglea in other body regions, where

epithelia proliferate and expand, such as the head or the lowergastric region, does not change in size. A lower level ofproteolytic remodeling and the resulting higher mechanical

stability of mesoglea could account for this phenomenon.

By contrast, mesoglea expansion at the base of tentacles and inthe evaginating bud is based on a different mechanism. Both

tentacle and bud outgrowth are associated with rapid epithelialmovements, which are independent of cell proliferation(Clarkson and Wolpert, 1967; Diehl and Burnett, 1965; Ottoand Campbell, 1977b). These movements appear to be driven by

lateral intercalation of individual epithelial cells (Philipp et al.,2009) (supplementary material Fig. S3), a cell behaviour thathas been shown to generate physical forces necessary for

morphogenesis in other systems (Bertet et al., 2004; Rauziet al., 2008). The movements are also accompanied by radicalchanges in the relationship of epithelial sheets to their underlying

substrate. In the head region, epithelial cells migrate over thestationary mesoglea until they reach the border between head andtentacle tissue (Fig. 10C, 1). Distal to this border, they step onto

an expanding mesoglea (Fig. 10C, 2). Even further distally alongthe tentacle, cells stably attach to unchanging mesoglea and bothare displaced together towards the tentacle tip (Fig. 10C, 3). Cell-matrix interactions during bud evagination are very similar

(Fig. 10D, 1–3). We hypothesize that physical forces generatedduring epithelial cell movements induce the observed localmesoglea shape change (Fig. 10C,D, top). It remains unclear,

however, how forces transmitted from migrating epithelial sheetsresult only in a very localized shape change of the mesoglea atthe tentacle base (Fig. 10C, 2) and in the evaginating bud after

passing the parent–bud border (Fig. 10D, 2). Again, localproteolytic remodeling might make the mesoglea in these areasmore susceptible to shape change.

Whereas mesoglea expansion in the tentacle base and the

evaginating bud are both associated with epithelial cellmovements, the rate at which cells enter tentacles and budsdiffers significantly. Tentacle tissue grows out by 10–12% of its

length per day, which corresponds to roughly 100 epithelial cells(Otto and Campbell, 1977a). By contrast, bud area multiplies fourto five times in size during the first 24 hours and more than 4000

cells are recruited from the parental body column in this timeinterval (Graf and Gierer, 1980; Otto and Campbell, 1977a). Inboth young buds and tentacle bases, high levels of laminin and

collagen expression reflect the need for new ECM material(Zhang et al., 2007; Zhang et al., 2002) (supplementary materialFig. S2). In the evaginating bud, however, the rate of expansion is

Journal of Cell Science 124 (23)4036

Journ

alof

Cell

Scie

nce

so great that ECM synthesis cannot keep up and for 24 hours,young buds show reduced levels of collagen staining (Fig. 5,

Fig. 6A,B; supplementary material Fig. S1). By contrast, no suchthinning of ECM was observed at the tentacle base, probablybecause of the slower rate of mesoglea expansion. Neverthelessde novo synthesis is not essential for rapid evagination, because

inhibition of collagen biosynthesis with DP did not inhibit theinitial phase of bud evagination. This suggests that remodeling ofthe existing ECM network, but not new ECM synthesis, is critical

for early bud evagination.

Materials and MethodsAnimal culture

Hydra vulgaris (strain Basel) was used in all experiments. The animals werecultured at 18 C in mass culture dishes of 1000–2000 individuals per dishcontaining 2 liters of M-solution (Lenhoff, 1983) and were fed with freshlyhatched Artemia salina larvae. During tracking experiments, individual animalswere kept in single wells containing 2 ml culture medium and were fed either dailyor every other day with 10–15 Artemia salina larvae.

In vivo mesoglea labeling

Monoclonal antibodies against Hydra collagen-1 (mAb39) and laminin (mAb52)(Sarras et al., 1991a; Sarras et al., 1993) were labeled with Alexa Fluor 488 andAlexa Fluor 546 Monoclonal Antibody Labeling Kits (Molecular Probes)following the manufacturer’s instructions. Fluorescently tagged antibodies wereloaded into pulled glass needles at a concentration of 1 mg/ml and injected directlyinto the mesoglea with a PLI-90 pressure injector (Harvard Apparatus). Singleinjections of 10–50 nl volumes resulted in labeling of 10–20% of the entiremesoglea of an animal. Multiple injections allowed us to label the entire animal.

In vivo labeling of epithelial cells

Ectodermal epithelial cells were labeled with Fluoresbrite polychromatic 1.0micron microspheres (Polysciences, Warrington, PA) either by incubation for12 hours in a 0.026% solution (Technau and Holstein, 1992) or by direct injectionof a 0.13% solution together with the fluorescently tagged antibodies. Endodermalepithelial cells were labeled with Fluoresbrite polychromatic microspheres byinjection of a 2.6% solution into the gastric cavity or by injection of 20 mg/mlRhodamine Dextran (Sigma). These two labeling methods resulted in uniformlabeling of the endodermal cell layer. Alternatively, labeling of small clusters ofendodermal cells with CM-DiI (Molecular Probes) was carried out according to themethod described previously (Kishimoto et al., 1996).

Tissue grafting

Grafting experiments were performed 24 hours after mesoglea and cell labelingand were carried out according to the procedure described earlier (Teragawa andBode, 1990).

Imaging of living animals

Living animals were relaxed in 2% urethane in culture medium for 2–5 minutesbefore observation under a Nikon SMZ 1500 stereomicroscope equipped a filter-set for 450–490 nm excitation and 515 nm long pass emission. The long-passemission filter allowed us to observe double- and triple-labeled specimens with theAlexa Fluor 488 in green emission and Fluoresbrite polychromatic microspheresand DiI in yellow–orange emission by using only blue excitation. Thestereomicroscope was equipped with an EXFO X-cite series 120 fluorescentlight source. Images were taken with an Olympus DP70 digital camera.

Confocal microscopy

Immunofluorescent staining with monoclonal antibodies mAb39 against collagen-1 and JK-2 against collagen-4 was performed as described earlier (Shimizu et al.,2008). Animals labeled in vivo by injection of fluorescently tagged antibodieswere relaxed in 2% Urethane for 2 minutes and fixed in 4% paraformaldehyde for30 minutes at room temperature. After three washes in PBS, they were mounted invectashield (Vector Laboratories) and observed under a Zeiss 510 LSM or a Nikoneclipse 50i microscope.

Mesoglea isolation

Isolation of mesoglea of single buds was performed as described for whole animals(Day and Lenhoff, 1981).

Morphometry and measurements of fluorescence intensity

Morphometric measurements and measurements of the fluorescence intensity wereperformed on digital images using ImageJ software. For measurements of

fluorescence intensities of collagens in buds, average grey values along a centralline were compared with average grey values along a central line in the upper bodycolumn (Fig. 5D).

ECM degradation essay

ECM turnover was monitored on protein levels by assaying the proteolytic stabilityof laminin in isolated mesoglea samples incubated with different tissue lysates. Foreach set of experiments, the mesogleas of 15 animals were extracted andsolubilized in 30 ml PBS. Cell lysates from different body parts of Hydra (15–25animals) were generated by passing the tissue through a syringe needle severaltimes on ice. The lysate was cleared by centrifugation and the proteinconcentration of each supernatant was adjusted to 1 mg/ml by adding PBS. Thecell lysates were immediately mixed with the mesoglea samples in a 1:1 ratio andincubated at room temperature. Samples were taken at different time points, mixedwith sample buffer and boiled. Laminin stability was assayed by western blottingusing monoclonal Hydra laminin antibody (mAB 52) at 1:1000 and secondaryanti-mouse HRP at 1:10,000. Mouse anti-a-tubulin was applied at 1:2000,secondary anti-mouse HRP at 1:10,000. The mesoglea samples had to bepreincubated in the respective tissue lysates for 1 hour to release the laminin fromthe ECM network. After preincubation, the levels of laminin were approximatelyequal in all samples (time point50).

In situ hybridization

Experiments were carried out using probes to Hydra collagen-1 and collagen-4(Deutzmann et al., 2000; Fowler et al., 2000). The procedure for in situhybridization is the same as described previously (Zhang et al., 2007).

Drug treatment with 2,29-dipyridyl

Treatment was performed by soaking the animals in 0.05 mM or 0.1 mM DP(Sigma) and 0.1% DMSO in culture medium at 1 ml solution per individual. Allcontrols were incubated in 0.1% DMSO in culture medium. Inhibitor treatmentwas performed in the dark and solutions were changed routinely every 24 hours.

AcknowledgementsThe authors would like to thank Dale R. Abrahamson for providingthe Hydra mesoglea specific antibodies and Stefanie Pontasch forcritical reading and helpful comments on the manuscript. This workwas initiated and carried out mainly in the lab of X.Z. by R.A. whenR.A. was a visiting PhD candidate.

FundingFunding for this work was provided by a program project grant fromthe National Institutes of Health NIDDK [grant number 5 P01DK065123] to D.R.A. and M.P.S. and by an award from theG. Harold and Lila Y. Mathers Charitable Foundation to C.D.L.Deposited in PMC for release after 12 months.

Supplementary material available online at

http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.087239/-/DC1

ReferencesBenazeraf, B., Francois, P., Baker, R. E., Denans, N., Little, C. D. and Pourquie, O.

(2010). A random cell motility gradient downstream of FGF controls elongation of an

amniote embryo. Nature 466, 248-252.

Bertet, C., Sulak, L. and Lecuit, T. (2004). Myosin-dependent junction remodelling

controls planar cell intercalation and axis elongation. Nature 429, 667-671.

Bosch, T. C. and David, C. N. (1984). Growth regulation in Hydra: relationship

between epithelial cell cycle length and growth rate. Dev. Biol. 104, 161-171.

Burnett, A. L. and Hausman, R. E. (1969). Mesoglea of Hydra. 2. Possible Role in

Morphogenesis. J. Exp. Zool. 171, 15-24.

Campbell, R. D. (1967a). Tissue dynamics of steady state growth in Hydra littoralis. I.

Patterns of cell division. Dev. Biol. 15, 487-502.

Campbell, R. D. (1967b). Tissue dynamics of steady state growth in Hydra littoralis. II.

Patterns of tissue movement. J. Morphol. 121, 19-28.

Campbell, R. D. (1973). Vital marking of single cells in developing tissues: India ink

injection to trace tissue movements in hydra. J. Cell Sci. 13, 651-661.

Campbell, R. D. (1974). Cell movements in Hydra. Amer. Zool. 14, 523-535.

Chapman, J. A., Kirkness, E. F., Simakov, O., Hampson, S. E., Mitros, T.,

Weinmaier, T., Rattei, T., Balasubramanian, P. G., Borman, J., Busam, D. et al.

(2010). The dynamic genome of Hydra. Nature 464, 592-596.

Clarkson, S. G. and Wolpert, L. (1967). Bud morphogenesis in hydra. Nature 214,

780-783.

Czirok, A., Rongish, B. J. and Little, C. D. (2004). Extracellular matrix dynamics

during vertebrate axis formation. Dev. Biol. 268, 111-122.

Hydra ECM dynamics in vivo 4037

Journ

alof

Cell

Scie

nce

Davidson, L. A., Dzamba, B. D., Keller, R. and Desimone, D. W. (2008). Liveimaging of cell protrusive activity, and extracellular matrix assembly and remodelingduring morphogenesis in the frog, Xenopus laevis. Dev. Dyn. 237, 2684-2692.

Day, R. M. and Lenhoff, H. M. (1981). Hydra mesoglea: a model for investigatingepithelial cell-basement membrane interactions. Science 211, 291-294.

Deutzmann, R., Fowler, S., Zhang, X., Boone, K., Dexter, S., Boot-Handford, R. P.,Rachel, R. and Sarras, M. P., Jr (2000). Molecular, biochemical and functionalanalysis of a novel and developmentally important fibrillar collagen (Hcol-I) in hydra.Development 127, 4669-4680.

Diehl, F. A. and Burnett, A. L. (1965). The role of interstitial cells in the maintenanceof Hydra. II. Budding. J. Exp. Zool. 158, 283-295.

Dubel, S. (1989). Cell differentiation in the head of Hydra. Differentiation 41, 99-109.Epp, L., Smid, I. and Tardent, P. (1986). Synthesis of the Mesoglea by Ectoderm and

Endoderm in Reassembled Hydra. J. Morphol. 189, 271-279.Fowler, S. J., Jose, S., Zhang, X., Deutzmann, R., Sarras, M. P., Jr and Boot-

Handford, R. P. (2000). Characterization of hydra type IV collagen. Type IVcollagen is essential for head regeneration and its expression is up-regulated uponexposure to glucose. J. Biol. Chem. 275, 39589-39599.

Graf, L. and Gierer, A. (1980). Size, shape and orientation of cells in budding hydraand regulation of regeneration in cell aggregates. Dev. Genes Evol. 188, 141-151.

Hausman, R. E. and Burnett, A. L. (1969). The mesoglea of Hydra. I. Physical andhistochemical properties. J. Exp. Zool. 171, 7-14.

Hausman, R. E. and Burnett, A. L. (1971). Mesoglea of Hydra. IV. Qualitativeradioautographic study of protein component. J. Exp. Zool. 177, 435-446.

Hobmayer, E., Holstein, T. W. and David, C. N. (1990). Tentacle morphogenesis inhydra. II. Formation of a complex between a sensory nerve cell and a battery cell.Development 109, 897-904.

Holstein, T. W., Hobmayer, E. and David, C. N. (1991). Pattern of epithelial cellcycling in hydra. Dev. Biol. 148, 602-611.

Ingber, D. E. (2006). Mechanical control of tissue morphogenesis during embryologicaldevelopment. Int. J. Dev. Biol. 50, 255-266.

Kawano, H., Li, H. P., Sango, K., Kawamura, K. and Raisman, G. (2005). Inhibitionof collagen synthesis overrides the age-related failure of regeneration of nigrostriataldopaminergic axons. J. Neurosci. Res. 80, 191-202.

Kishimoto, Y., Murate, M. and Sugiyama, T. (1996). Hydra regeneration fromrecombined ectodermal and endodermal tissue. I. Epibolic ectodermal spreading isdriven by cell intercalation. J. Cell Sci. 109, 763-772.

Kivirikko, K. I., Myllyla, R. and Pihlajaniemi, T. (1989). Protein hydroxylation:prolyl 4-hydroxylase, an enzyme with four cosubstrates and a multifunctional subunit.FASEB J. 3, 1609-1617.

Larsen, M., Wei, C. and Yamada, K. M. (2006). Cell and fibronectin dynamics duringbranching morphogenesis. J. Cell Sci. 119, 3376-3384.

Lenhoff, H. M. (1983). Culturing large numbers of Hydra. In Hydra: Research Methods

(ed. H.M. Lenhoff), pp. 53-62. New York: Plenum Press.Leontovich, A. A., Zhang, J., Shimokawa, K., Nagase, H. and Sarras, M. P., Jr

(2000). A novel hydra matrix metalloproteinase (HMMP) functions in extracellularmatrix degradation, morphogenesis and the maintenance of differentiated cells in thefoot process. Development 127, 907-920.

Mizoguchi, H. and Yasumasu, I. (1983). Inhibition of archenteron formation by theinhibitors of prolyl hydroxylase in sea urchin embryos. Cell Differ. 12, 225-231.

Munder, S., Kasbauer, T., Prexl, A., Aufschnaiter, R., Zhang, X., Towb, P. and

Bottger, A. (2010). Notch signalling defines critical boundary during budding inHydra. Dev. Biol. 344, 331-345.

Ohashi, T., Kiehart, D. P. and Erickson, H. P. (1999). Dynamics and elasticity of thefibronectin matrix in living cell culture visualized by fibronectin-green fluorescentprotein. Proc. Natl. Acad. Sci. USA 96, 2153-2158.

Otto, J. J. and Campbell, R. D. (1977a). Budding in Hydra attenuata: bud stages andfate map. J. Exp. Zool. 200, 417-428.

Otto, J. J. and Campbell, R. D. (1977b). Tissue economics of hydra: regulation of cellcycle, animal size and development by controlled feeding rates. J. Cell Sci. 28, 117-132.

Page-McCaw, A., Ewald, A. J. and Werb, Z. (2007). Matrix metalloproteinases andthe regulation of tissue remodelling. Nat. Rev. Mol. Cell Biol. 8, 221-233.

Philipp, I., Aufschnaiter, R., Ozbek, S., Pontasch, S., Jenewein, M., Watanabe, H.,Rentzsch, F., Holstein, T. W. and Hobmayer, B. (2009). Wnt/beta-catenin andnoncanonical Wnt signaling interact in tissue evagination in the simple eumetazoanHydra. Proc. Natl. Acad. Sci. USA 106, 4290-4295.

Rauzi, M., Verant, P., Lecuit, T. and Lenne, P. F. (2008). Nature and anisotropy of

cortical forces orienting Drosophila tissue morphogenesis. Nat. Cell Biol. 10, 1401-

1410.

Rongish, B. J., Drake, C. J., Argraves, W. S. and Little, C. D. (1998). Identification of

the developmental marker, JB3-antigen, as fibrillin-2 and its de novo organization

into embryonic microfibrous arrays. Dev. Dyn. 212, 461-471.

Sarras, M. P., Jr and Deutzmann, R. (2001). Hydra and Niccolo Paganini (1782-

1840)--two peas in a pod? The molecular basis of extracellular matrix structure in the

invertebrate, Hydra. Bioessays 23, 716-724.

Sarras, M. P., Jr, Madden, M. E., Zhang, X. M., Gunwar, S., Huff, J. K. and

Hudson, B. G. (1991a). Extracellular matrix (mesoglea) of Hydra vulgaris. I.

Isolation and characterization. Dev. Biol. 148, 481-494.

Sarras, M. P., Jr, Meador, D. and Zhang, X. M. (1991b). Extracellular matrix

(mesoglea) of Hydra vulgaris. II. Influence of collagen and proteoglycan components

on head regeneration. Dev. Biol. 148, 495-500.

Sarras, M. P., Jr, Zhang, X., Huff, J. K., Accavitti, M. A., St John, P. L. and

Abrahamson, D. R. (1993). Extracellular matrix (mesoglea) of Hydra vulgaris III.

Formation and function during morphogenesis of hydra cell aggregates. Dev. Biol.

157, 383-398.

Shimizu, H., Aufschnaiter, R., Li, L., Sarras, M. P., Jr, Borza, D. B., Abrahamson,

D. R., Sado, Y. and Zhang, X. (2008). The extracellular matrix of hydra is a porous

sheet and contains type IV collagen. Zoology (Jena) 111, 410-418.

Shostak, S. and Globus, M. (1966). Migration of epithelio-muscular cells in Hydra.

Nature 210, 218-219.

Shostak, S. and Kankel, D. R. (1967). Morphogenetic movements during budding in

Hydra. Dev. Biol. 15, 451-463.

Shostak, S., Patel, N. G. and Burnett, A. L. (1965). The role of mesoglea in mass cell

movement in Hydra. Dev. Biol. 12, 434-450.

Sivakumar, P., Czirok, A., Rongish, B. J., Divakara, V. P., Wang, Y. P. and Dallas,

S. L. (2006). New insights into extracellular matrix assembly and reorganization from

dynamic imaging of extracellular matrix proteins in living osteoblasts. J. Cell Sci.

119, 1350-1360.

Sudhop, S., Coulier, F., Bieller, A., Vogt, A., Hotz, T. and Hassel, M. (2004).

Signalling by the FGFR-like tyrosine kinase, Kringelchen, is essential for bud

detachment in Hydra vulgaris. Development 131, 4001-4011.

Technau, U. and Holstein, T. W. (1992). Cell sorting during the regeneration of Hydra

from reaggregated cells. Dev. Biol. 151, 117-127.

Teragawa, C. K. and Bode, H. R. (1990). Spatial and temporal patterns of interstitial

cell migration in Hydra vulgaris. Dev. Biol. 138, 63-81.

Tyler, S. (2003). Epithelium–The primary building block for metazoan complexity.

Integr. Comp. Biol. 43, 55-63.

Wainwright, S. A., Biggs, W. D., Currey, J. D. and Gosline, J. M. (1976). Support in

Organisms. In Mechanical Design in Organisms, pp. 287-344. New York: Halsted

Press, John Wiley and Sons.

Wittlieb, J., Khalturin, K., Lohmann, J. U., Anton-Erxleben, F. and Bosch, T. C.

(2006). Transgenic Hydra allow in vivo tracking of individual stem cells during

morphogenesis. Proc. Natl. Acad. Sci. USA 103, 6208-6211.

Zamir, E. A., Czirok, A., Rongish, B. J. and Little, C. D. (2005). A digital image-

based method for computational tissue fate mapping during early avian morphogen-

esis. Ann. Biomed. Eng. 33, 854-865.

Zamir, E. A., Czirok, A., Cui, C., Little, C. D. and Rongish, B. J. (2006). Mesodermal

cell displacements during avian gastrulation are due to both individual cell-

autonomous and convective tissue movements. Proc. Natl. Acad. Sci. USA 103,

19806-19811.

Zamir, E. A., Rongish, B. J. and Little, C. D. (2008). The ECM moves during

primitive streak formation–computation of ECM versus cellular motion. PLoS Biol. 6,

e247.

Zhang, X., Fei, K., Agbas, A., Yan, L., Zhang, J., O’Reilly, B., Deutzmann, R. and

Sarras, M. P., Jr (2002). Structure and function of an early divergent form of laminin

in hydra: a structurally conserved ECM component that is essential for epithelial

morphogenesis. Dev. Genes. Evol. 212, 159-172.

Zhang, X., Boot-Handford, R. P., Huxley-Jones, J., Forse, L. N., Mould, A. P.,

Robertson, D. L., Lili, Athiyal, M. and Sarras, M. P., Jr (2007). The collagens of

hydra provide insight into the evolution of metazoan extracellular matrices. J. Biol.

Chem. 282, 6792-6802.

Journal of Cell Science 124 (23)4038

Journ

alof

Cell

Scie

nce