XPOX2-peroxidase expression and the XLURP-1 promoter reveal the siteof embryonic myeloid cell development in Xenopus

Stuart J. Smith, Surendra Kotecha, Norma Towers, Branko V. Latinkic, Timothy J. Mohun*

Division of Developmental Biology, National Institute for Medical Research, The Ridgeway, Mill Hill, London NW7 1AA, UK

Received 3 May 2001; received in revised form 4 June 2002; accepted 5 June 2002

Abstract

Phagocytic myeloid cells provide the principle line of immune defence during early embryogenesis in lower vertebrates. They may also

have important functions during normal embryo morphogenesis, not least through the phagocytic clearance of cell corpses arising from

apoptosis. We have identified two cDNAs that provide sensitive molecular markers of embryonic leukocytes in the early Xenopus embryo.

These encode a peroxidase (XPOX2) and a Ly-6/uPAR-related protein (XLURP-1). We show that myeloid progenitors can first be detected

at an antero-ventral site in early tailbud stage embryos (a region previously termed the anterior ventral blood island) and transiently express

the haematopoetic transcription factors SCL and AML. Phagocytes migrate from this site along consistent routes and proliferate, becoming

widely distributed throughout the tadpole long before the circulatory system is established. This migration can be followed in living embryos

using a 5 kb portion of the XLURP-1 promoter to drive expression of EGFP specifically in the myeloid cells. Interestingly, whilst much of

this migration occurs by movement of individual cells between embryonic germ layers, the rostral-most myeloid cells apparently migrate in

an anterior direction along the ventral midline within the mesodermal layer itself. The transient presence of such cells as a strip bisecting the

cardiac mesoderm immediately prior to heart tube formation suggests that embryonic myeloid cells may play a role in early cardiac

morphogenesis. q 2002 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Xenopus; Peroxidase; Ly-6/uPAR-related protein; Myeloid; Phagocyte; Macrophage; Ventral blood island; Heart development; Embryo

1. Introduction

The adaptive immune system appears relatively late

during embryonic development of the anuran amphibian

Xenopus laevis (Nagata, 1977; Tochinai, 1980). By the

time differentiated lymphocytes emerge to the periphery

from the newly formed thymus (Kau and Turpen, 1983;

Maeno et al., 1985), the feeding tadpole is already 12

days old. Prior to this, the embryo must rely on the innate

immune system of phagocytic myeloid blood cells (these are

sometimes referred to as phagocytes, non-lymphoid leuko-

cytes, or abbreviated to leukocytes). It is commonly held

that myeloid cells can offer protection from at least 98% of

all pathogens encountered (Jones, 2000), and so their forma-

tion during early development may be fundamental to survi-

val of lower vertebrate embryos growing in a hostile aquatic

environment.

In addition to this role, myeloid cells are likely to have

important functions during normal embryogenesis. In

Drosophila embryos, phagocytic clearance of cell corpses

resulting from apoptosis is primarily mediated by macro-

phages that originate from the embryonic haemocyte popu-

lation (Franc et al., 1999; Tepass et al., 1994). In a similar

manner, macrophages actively induce apoptosis of endothe-

lial cells in the pupillary membrane of the developing

mammalian eye (Diez-Roux et al., 1999). Such findings

suggest that myeloid cells could play an important role

facilitating embryo morphogenesis, through their ability to

engulf apoptotic corpses (Savill and Fadok, 2000). Unfortu-

nately, such a possibility has proved difficult to explore

since few molecular markers are available either to identify

embryonic myeloid cells, or to trace their ontogeny.

The embryonic origin of myeloid blood cells in verte-

brates is poorly characterised, but is probably distinct

from that of other blood lineages. In zebrafish embryos,

early macrophages are first visualised in the most rostrally

located region of ventrolateral mesoderm, just anterior to

the cardiac field (Herbomel et al., 1999, 2001). In amphi-

bians, a leukocyte-specific antibody has identified a widely

dispersed population of non-lymphoid leukocytes in X.

laevis tadpoles, prior to the onset of cardiovascular circula-

tion (Ohinata et al., 1989). Grafting experiments indicate

Mechanisms of Development 117 (2002) 173–186

0925-4773/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved.

PII: S0925-4773(02)00200-9

www.elsevier.com/locate/modo

* Corresponding author. Tel.: 144-20-8913-8621; fax: 144-20-8906-

4477.

E-mail address: tmohun@nimr.mrc.ac.uk (T.J. Mohun).

that such cells arise from the head region of early tailbud

embryos (Ohinata et al., 1990), unlike erythroid and

lymphoid blood cell lineages which originate from ventral

mesoderm below the gut (Ciau-Uitz et al., 2000; Kau and

Turpen, 1983; Maeno et al., 1985; Turpen and Knudson,

1982). Their precise origins could not be traced directly

since the antigenic determinants recognised by the antibody

only appear at the tadpole stage.

Here we present studies that establish the embryonic

origin of myeloid cells in the amphibian X. laevis. We

have identified two cDNAs, a peroxidase family member

(XPOX2) and Ly-6/uPAR-Related Protein (XLURP-1),

that provide molecular markers for myeloid cells from the

early tailbud stage in Xenopus embryos. Expression of these

cDNAs is restricted to a small population of cells, first

detected in the antero-ventral region of the early tailbud

embryo. These cells appear to proliferate and distribute

rapidly throughout the embryo along consistent routes, at

least 20 h before the cardiovascular circulation of the

tadpole is established. Their subsequent location and beha-

viour matches that of non-lymphoid leukocytes, previously

identified in the tadpole by antibody staining. Interestingly,

we find that the early myeloid cells migrate through the

ventral-most portion of the prospective heart myocardial

tissue during the onset of heart tube formation, and transi-

ently co-express the haematopoietic transcription factors

SCL and AML whilst they reside in the anterior ventral

blood island (VBI). Finally, as a first step towards investi-

gating the potential functions of these cells in living

embryos, we describe a transgenic line of X. laevis whose

myeloid cells can be observed directly, using a myeloid-

specific gene promoter (XLURP-1) to drive expression of

the green fluorescent protein (EGFP).

2. Results

2.1. Isolation of XPOX2 and XLURP-1 cDNAs

In a search for cDNAs expressed within the developing

heart field of the tailbud embryo, we used a subtraction

strategy to construct a heart field-enriched cDNA library.

A combination of DNA sequencing and in situ hybridisation

was used to identify candidate cDNAs. This screen identi-

fied a 1448 bp cDNA encoding the partial sequence of a

peroxidase, Xenopus POX2 (GenBank AF364820). A

search of the Xenopus EST database identified a closely

related 2534 bp cDNA corresponding to a second XPOX2 0

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186174

Fig. 1. Comparison of Xenopus POX2 (A) and LURP-1 (B) deduced protein sequences with other representative family members. Hyphens (-) represent amino

acid identity with XPOX2 0 or XLURP-1 while dots (·) indicate gaps inserted in the sequences to facilitate alignment. (A) The XPOX2 0 allele sequence aligned

with other peroxidases using the MegAlign program (DNASTAR Inc). (B) XLURP-1 sequence compared with other Ly-6/uPAR and snake venom toxin ‘three-

finger’ proteins. To make the most appropriate comparison of these structurally related proteins, the positions of eight of the ten disulphide-forming cysteines

(one, and four to ten) were manually aligned first. These four disulphides maintain the structure of the hydrophobic core from which protrude the three loops or

fingers (Tsetlin, 1999). The position of the disulphide formed between the second and third cysteines can vary within loop 1 of Ly-6/uPAR proteins and is

absent from the snake venom toxins. The arginine located between the fifth and sixth cysteines interacts with the mature C-terminal asparagine and so was also

aligned manually (Fleming et al., 1993). Having made these constraints (indicated by asterisks above the sequences), the remaining regions of the sequences

were aligned using the MegAlign program (DNASTAR Inc). GenBank accession numbers: Xenopus POX2, AF364820; Xenopus POX2 0, AY069942; Xenopus

polysomal ribonuclease PMR-1, AAC94959; mouse myeloperoxidase, P11247; mouse eosinophil peroxidase, P49290; Danio rerio myeloperoxidase,

AAK83239; Xenopus LURP-1, AF364819; mouse ThB, A46528; human SLURP-1, A59031; western green mamba short toxin 2, T5EP2V.

allele (GenBank AY069942, IMAGE clone 4203899) that

encoded the complete 725 amino acid polypeptide (Fig.

1A). The existence of two POX2 transcripts in X. laevis

most probably results from the pseudotetraploid character

of this species (Bisbee et al., 1977). XPOX2 0 resembles the

one other Xenopus peroxidase family member characterised

to date, the polysomal ribonuclease PMR-1 (Chernokals-

kaya et al., 1998). XPOX2 0 protein is 66% identical to

XPMR-1, shows 55% identity to the mammalian haemato-

poietic peroxidases, myeloperoxidase and eosinophil perox-

idase, and 47% identity to a zebrafish myeloperoxidase

polypeptide (Bennett et al., 2001).

The 456 bp cDNA encoding Xenopus LURP-1 (GenBank

accession AF364819) encodes an 88 amino acid polypeptide

(Fig. 1B) that belongs to the Ly-6/uPAR immune cell

surface antigen and snake venom a-neurotoxin groups of

‘three-finger’ proteins (Fleming et al., 1993; Tsetlin,

1999). While XLURP-1 is most closely related to the

mouse Ly-6 protein ThB (36% amino acid identity), it

lacks the cleavable GPI-anchor signal sequence (Wang et

al., 1999) common to the extreme C-terminal region of Ly-6

proteins. We presume that the Xenopus protein is therefore

secreted rather than tethered to the cytoplasmic membrane.

A human secreted Ly-6/uPAR-related protein, SLURP-1,

has previously been identified as a secreted rather than teth-

ered polypeptide, but the weaker sequence similarity

between XLURP-1 and human SLURP-1 (31% identity,

and differing ‘finger-lengths’) suggests they are not ortho-

logous genes (Andermann et al., 1999). More likely, there

are further secreted Ly-6 family members that await char-

acterisation.

In addition to XPOX2 and XLURP-1, we have identified

four novel Xenopus cDNAs that exhibit similarity to

mammalian ADAM23, the CD34 1 stem cell protein

HSPC280, a kallikrein 15-like elastase, and a pentraxin-

like protein (data not shown, though GenBank accession

numbers of mammalian homologs are AAD25099,

AAF28958, NP_059979, P49263, respectively). Together,

the cDNAs represent a family of markers expressed in the

same cell population as XPOX2 and XLURP-1 (see Section

2.2).

2.2. Embryonic expression of XPOX2 and XLURP-1

XPOX2 mRNA expression is first detected by whole-

mount in situ hybridisation in the antero-ventral region of

stage 19 embryos (Fig. 2A, B). Transverse sections of these

embryos reveal that when first detected, the ventral XPOX2

expression domain is mesodermal (Fig. 3A, B). By mid-

tailbud stage, however, profound changes in the morphol-

ogy and location of the XPOX2-expressing cells have

occurred. Instead of residing together in the mesodermal

layer, individual XPOX2-expressing cells are observed

(Fig. 2C, D). Moreover, XPOX2-expressing cells are now

detected in the anterior-two-thirds and the ventral-half of

stage 24 embryos (Fig. 2C, D).

We analysed progressive transverse sections through

mid-tailbud embryos after whole-mount in situ hybridisa-

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186 175

Fig. 2. Whole-mount in situ hybridisation analysis of Xenopus POX2 (A–H,L), LURP-1 (I,J) and L-plastin (K) expression. Right-lateral views of stage 19 (A),

24 (C) and 27 (E) embryos showing XPOX2-expressing cells. (B,D,F) Ventral views of the same embryos depicted in (A,C,E). (G) Higher magnification image

of the stage 27 embryo depicted in (F) to illustrate the streams of XPOX2-expressing cells that emanate from a focal point (red arrow) located at the ventral

midline within the heart-forming region. (H) Right-lateral view of a stage 30 embryo. (I,J) Left-lateral views of embryos at stages 24 and 36, respectively,

showing cells that express XLURP-1. (K,L) Right-lateral view of the posterior trunk of a stage 35 tadpole after sequential, double whole-mount in situ

hybridisation staining to reveal first (K), Xenopus L-plastin (pale blue) and second (L), POX2 (magenta) expression. The co-localisation of the two

chromogenic reagents results in a dark blue colour (L). A, anterior; P, posterior.

tion to characterise the precise distribution of cells expres-

sing XPOX2. From the level of the cement gland to the

trunk, punctate XPOX2-expressing cells reside between

the ectodermal and mesodermal layers, and also between

mesodermal and endodermal layers in the ventro-lateral

portion of the embryo (Fig. 3C–G). However, the ventral-

most cells, particularly those near the heart-forming region,

form part of the mesodermal layer itself (Fig. 3F). Inspec-

tion of more-posterior sections also reveals XPOX2-expres-

sing cells within the endoderm, close to the remnant of the

blastocoel cavity (Fig. 3H).

As the tailbud embryo develops, XPOX2-positive cells

are detected in progressively larger numbers, distributed

ever more widely throughout the embryo. By stage 27, a

few expressing cells can be detected dorsally near the roof

plate (Fig. 3J) and many are located in proximity to the

posterior region of the VBI (Fig. 3I). Interestingly, such

cells reside on both ectodermal and endodermal sides of

the erythropoietic mesoderm that constitutes the posterior

VBI. The anterior pattern of XPOX2-cell distribution is

especially noteworthy around stage 27 with a particular

concentration on the ventral midline, located between the

newly forming progenitors of heart myocardium (Fig. 2E–

G). From this point, the punctate XPOX2-expressing cells

are distributed along a characteristic pattern of radial lines

or streams that extend towards the head (Fig. 2G). From

stage 30 onwards, the cells visualised by XPOX2 expression

are distributed throughout the embryo (Fig. 2H), with a high

concentration observed surrounding the proctodeum at

swimming tadpole stages (data not shown).

The expression of XLURP-1 (Fig. 2I, J) appears similar

to XPOX2 with the important exception that XLURP-1

mRNA is detected by whole-mount in situ hybridisation

only in the punctate, cell population from stage 24 onwards

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186176

Fig. 3. Transverse sections (10 mm) through Xenopus tailbud stage embryos after whole-mount in situ hybridisation for POX2 (A–J) or LURP-1 (K,L). (A)

Section through the centre of the XPOX2 expression domain of a stage 20 embryo, showing the location of the high magnification detail (B). A stage 25 embryo

(C) shows the relative location of five ventral-half view images (D–H) that are representative sections marking progressively posterior slices. Sections are

numbered (top right of each panel) commencing from the posterior limit of the cement gland. Red arrows mark the ventral-most XPOX2-expressing

mesodermal cells within the heart-forming region. (I,J) Sections of stage 27 embryos showing a ventral-half view of the posterior trunk (I) and a dorsal-

half view at the level of the eye (J). Ventral-half view of a stage 28 embryo (K) gives the location of a detail (L), which depicts a cell that expresses XLURP-1

located between mesodermal and endodermal layers that was fixed while it apparently phagocytosed cell material from the endoderm (L, black arrow).

Additionally, this phagocyte formed such tight contacts with two mesodermal cells that they were prised away from the bulk of the mesoderm as the germ

layers became separated during the sectioning procedure. All sections have been counterstained to reveal cell nuclei. Red boxes mark the location of detail

images. Ect/Mes, expressing cells located between ectodermal and mesodermal germ layers; Mes/End, expressing cells located between mesodermal and

endodermal layers; CG, cement gland; Bla, remnant of the blastocoel cavity; RP, roof plate; NT, neural tube; E, eye.

and not in the small anterior population of the early tailbud

embryo.

2.3. Expression of XPOX2 is restricted to antero-ventral

embryo explants

The dynamic distribution of XPOX2 and XLURP-1-

expressing cells revealed by in situ hybridisation suggests

active migration and proliferation of a small group of foun-

der cells. However, it could also be explained by rapid de

novo differentiation of XPOX2 and XLURP-1-expressing

cells at dispersed sites throughout the embryo. To investi-

gate this further, we bisected early tailbud (stage 21)

embryos diagonally, yielding anterior-ventral fragments

containing the first detectable XPOX2-expressing cells,

and also the corresponding posterior-dorsal fragments

(Fig. 4A–C). These embryo explants were cultured until

stage 33, by which stage XPOX2/XLURP-1-cells are

normally distributed throughout the entire embryo (Fig.

4F). The results are consistent with migration/proliferation

of these cells during tailbud stages of development. Of 47

embryos dissected and cultured, all anterior-ventral explants

contained numerous XPOX2-expressing cells at stage 33

(Fig. 4E) whilst 37 of the 47 posterior-dorsal explants

lacked any XPOX2-positive cells (Fig. 4D). A further

eight posterior-dorsal explants contained fewer than ten

positive cells, all of which were located close to the poster-

ior region of the forming VBI, suggesting that these were

the product of inaccurate dissections.

These results support the view that the appearance of

XPOX2-expressing cells in the dorsal half of Xenopus

embryos from late tailbud stages onwards is due to cell

migration from their anterior-ventral site of origin. This

experiment does not exclude the possibility that posterior-

dorsal tissue of stage 21 embryos makes some contribution

to the XPOX2-cell population observed at stage 33, but is

incapable of doing so when cultured in isolation as a tissue

explant. We therefore sought a more direct way to follow

the location of XPOX2 and XLURP-1-expressing cells as

development proceeds.

2.4. Cells expressing an XLURP-1-EGFP transgene are

migratory

In order to visualise XPOX2 and XLURP-1-expressing

cells, we screened Xenopus genomic DNA fragments

derived from the two genes for their ability to drive expres-

sion of an EGFP transgene in a pattern similar to that of the

endogenous mRNA transcripts. From this, we identified a

5 kb fragment of genomic DNA immediately upstream of

the XLURP-1 gene that, when integrated into the frog

genome, appears to recapitulate the entire expression

pattern of its endogenous counterpart (Figs. 5 and 6).

Using the EGFP reporter, fluorescent cells are first observed

at stage 24 in an antero-ventral location (Fig. 5A), and their

dispersal throughout the embryo over the next 9 h can be

viewed directly (Fig. 5B–F). Their routes of migration and

the timing of their dispersal are entirely consistent with the

distribution of XPOX2 and XLURP-1-expressing cells

evident from whole-mount in situ hybridisation analyses.

For example, the fluorescent cell observed at the dorsal

midline at stage 28, immediately caudal to the head (Fig.

5D, E), appeared to follow an anterior migration route iden-

tical to that suggested by the distribution of expressing cells

detected in Fig. 2F, G.

In stage 41 tadpoles, different cell types are apparent

among those that express EGFP. Approximately half of

the widely dispersed cells are distinguished by their irregu-

lar cell shape and steady migration (Fig. 5G–J), suggesting

that they may be tissue-resident macrophages. Other fluor-

escent cells are found in the peripheral blood vessels,

moving erratically within the circulation as they constantly

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186 177

Fig. 4. XPOX2 expression within embryo explants. Left-lateral views of representative posterior-dorsal (A) and anterior-ventral (B) embryo dissections fixed at

stage 21 and assayed for XPOX2 expression by in situ hybridisation. (C) Ventral view of the same fragment shown in (B). Arrows indicate the posterior limit of

the XPOX2-expressing cells. Left-lateral views of posterior-dorsal (D) and anterior-ventral (E) dissections that were cultured till stage 33 and assayed as

before. (F) Control stage 33 embryo with XPOX2-positive cells present in posterior and dorsal locations (arrows). The position of the myocardium (Mc) of the

forming heart was also assayed by MLC2 expression (Chambers et al., 1994) in this experiment (E,F). Pigmented-wildtype embryos were used for this

experiment. CG, cement gland; A, anterior; P, posterior.

attach to, then release themselves from, the vascular

endothelial walls (Fig. 5G–J).

2.5. XPOX2 and XLURP-1 are co-expressed

The punctate appearance of XPOX2 and XLURP-1-

expressing cells, along with their detection in progressively

more of the embryo in the early tadpole is similar to that

previously reported for myeloid blood cells, identified with

antibodies recognising Xenopus tadpole leukocyte lineages

(Miyanaga et al., 1998; Ohinata et al., 1989). Since both

transcripts can be detected much earlier than the leukocyte

antigens, one or both of them could provide a marker to

examine the ontogeny of embryonic myeloid cells. As

shown earlier, expression of XLURP-1 appears similar to

that of XPOX2 (Fig. 2I, J), at least from stage 24 onwards.

We therefore sought to establish if XPOX2 and XLURP-1

were indeed expressed in the same cells.

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186178

Fig. 6. XPOX2 expression in XLURP-1-EGFP transgenic embryos. (A–D) Ventral views of the anterior-half of a stage 29 transgenic Xenopus embryo showing

fluorescent cells (A), and after it was fixed, the whole-mount in situ hybridisation for XPOX2 (D). To facilitate a comparison of EGFP reporter, and XPOX2

expression, duplicate images are presented (A ¼ B, C ¼ D), with 20 representative XLURP-1-EGFP-positive/XPOX2-positive (double-positive) cells high-

lighted by red circles (B,C). (E,F) Ventral views of a second stage 29 XLURP-1-EGFP transgenic embryo showing fluorescent cells (E) and XPOX2 expression

(F). The location of the details (G–J) is marked by red boxes (E,F). Again, duplicate images are presented (G ¼ H, I ¼ J), with 18 double-positive cells marked

by red circles (H,I). In this embryo, a subset of cells is apparently XLURP-1-EGFP-negative/XPOX2-positive (cells not marked with red circles in I).

Fig. 5. EGFP expression directed by the XLURP-1 promoter in living, transgenic embryos. (A–E,G–J) Left-lateral views showing time-points in the

development of a single XLURP-1-EGFP transgenic Xenopus embryo at stage 24 (A), 26 (B, and detail C), 28 (D, and detail E) and 41 (G, and details H–

J). (F) Right-lateral view showing fluorescent cells distributed throughout the tail region of a second transgenic animal at stage 33. White arrows (A–E) indicate

the dorsal-most limit of the fluorescent cell dispersal, though it is not the same cell labelled at each stage. Oedemas are commonly found in animals raised by

the Xenopus transgenesis procedure, and one develops as a swelling on the ventral surface of the first embryo (A–E). Nevertheless, the fluid drains from the

oedema once the blood circulation is initiated, and this animal appears normal at stage 41 (G). The red box (G) gives the location of the high magnification

images (H–J), which represent a time-series taken at 10 s intervals. The red arrows (H–J) follow the path of a spherical blood cell that resides within the

peripheral circulation. This cell resists the blood flow, pausing three times as it forms transient contacts with the vascular endothelium. The output levels of the

different colour channels have been adjusted individually while rendering these dark-field images (A–G) in Adobe Photoshop. This approach enables the

silhouette of the embryo to be observed, while giving the EGFP fluorescence a false-blue colour. A, anterior; P, posterior.

Late tailbud stage embryos expressing the XLURP-1-

EGFP transgene were photographed immediately before

fixation and then assayed for XPOX2 expression by in situ

hybridisation. Two such embryos are shown in Fig. 6. In the

first (Fig. 6A–D), the number and location of the fluorescent

cells matches the data for XPOX2-expressing cells

precisely. In the second transgenic embryo, all the fluores-

cent cells also express XPOX2 (Fig. 6E–J), but there are

also a number of punctate, XPOX2-positive cells, especially

within the antero-ventral region of the embryo, that do not

express detectable EGFP (Fig. 6E–J). These results demon-

strate that XLURP-1-expressing cells also express XPOX2,

and support the results of double whole-mount in situ hybri-

disation to detect XPOX2 and XLURP-1 expression in non-

transgenic embryos (data not shown).

The observation of XPOX2-positive/EGFP-negative cells

in the transgenic embryos may indicate that XLURP-1 is

expressed only in a subset of the punctate, XPOX2-positive

cells or it may reflect a limitation in the fidelity of expres-

sion of the transgene (see Section 3).

2.6. XPOX2/XLURP-1-expressing cells are myeloid

In other vertebrate species, the expression of L-plastin

has been used to identify leukocyte populations in the

embryo. Consequently, we compared the expression of a

Xenopus homolog of L-plastin (IMAGE clone 4175686)

with POX2, using a sequential, double whole-mount in

situ hybridisation procedure. L-plastin mRNA can be

detected only weakly in Xenopus tailbud embryos, with

stronger expression observed by stage 35. Subsequent

development of the XPOX2 probe demonstrated unequivo-

cally that in stage 35 tadpoles, all cells expressing L-plastin

also express POX2 (Fig. 2K, L).

Further evidence that XPOX2/XLURP-1-expressing cells

are indeed embryonic myeloid cells is provided by histolo-

gical studies. The lower intensity of staining for XLURP-1

mRNA has made it possible to identify expressing cells that

had been fixed as they phagocytosed adjacent tissue. For

example, Fig. 3K (detail Fig. 3L) shows a cell with

XLURP-1 mRNA staining its cytoplasm blue. The cytoplas-

mic membrane of this cell, on the left (Fig. 3L, black arrow),

is apparently engulfing material from the endoderm. Further-

more, within the cytoplasm, a number of other phagosome-

like vesicles are visible since they do not stain for XLURP-1

mRNA. Taken together, the distribution of XPOX2/XLURP-

1-expressing cells in the tadpole, co-expression of L-plastin,

their behaviour when observed in XLURP-1-EGFP trans-

genic embryos, and their apparent phagocytic activity, all

indicate that these cells are indeed embryonic leukocytes.

2.7. Expression of XPOX2 within the heart field

The observation of myeloid cells within a region of the

tailbud stage embryo previously thought to contain progeni-

tors of the heart myocardium (Fig. 3F) is intriguing, espe-

cially since their appearance coincides with the onset of

heart tube formation. We therefore studied the relationship

between myeloid cells and cardiac precursors in greater

detail, by characterising the expression of XPOX2 using

double whole-mount in situ hybridisation in combination

with two markers of cardiogenesis, the tinman homolog,

XNkx2-5, and a myosin light chain, XMLC1av. In early

tailbud embryos, XPOX2 expression is first detected imme-

diately posterior to the Nkx2-5 domain that encompasses

cardiogenic mesoderm and the underlying pharyngeal endo-

derm (Fig. 7A) (Tonissen et al., 1994). By stage 24, when

individual myeloid cells can be observed, anterior XPOX2

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186 179

Fig. 7. XPOX2 expression adjacent to the heart-forming region. Left-lateral views of Xenopus stage 20 (A), 24 (B) and 26 (D) embryos after double whole-

mount in situ hybridisation for Nkx2-5 (pale blue) and POX2 (magenta). Higher magnification ventral views (C,E) of the heart field of the same embryos

depicted in (B) and (D), respectively. Red arrow indicates expressing cells at the focal point located posterior of the cement gland (CG) on the ventral midline.

(F) High magnification ventral view of the myocardial plate (Mc.P) of a stage 30 embryo after double whole-mount in situ hybridisation for MLC1av (pale

blue) and POX2 (magenta). The stage 20 embryo (A) is a pigmented-wildtype example while the rest are albino embryos. A, anterior; P, posterior.

expression is detected between bilateral Nkx2-5 domains

(Fig. 7B, C). We analysed progressive transverse sections

through these mid-tailbud embryos at the level of the heart

field. Significantly, at stage 24, XPOX2 expression in this

region is observed in the ventral-most portion of the meso-

dermal layer and not in isolated migratory cells (Fig. 8A–C).

This observation clarifies earlier reports on the expression

of Nkx2-5 and other tinman homologs in Xenopus, whose

mesodermal domains fuse in a narrow stripe across the

ventral midline before stage 20 yet are detected as separate

bilateral domains from stage 23 through to the onset of

myocardial differentiation (Evans et al., 1995; Sparrow et

al., 2000a; Tonissen et al., 1994). Of course, we cannot

exclude a model in which the ventral-most mesoderm that

at first expresses Nkx2-5 loses its cardiogenic potential and

differentiates into myeloid progenitors. Indeed the resolution

of our data cannot discount weak Nkx2-5 expression in those

cells that express XPOX2 within the heart field. Neverthe-

less, we believe it more likely that Nkx2-5-expressing cells

are displaced from the ventral midline by mesoderm that is

fated to produce leukocytes. That is, the rostral portion of

myelopoietic mesoderm comes to occupy a ventral strip

within the heart field as a result of the movement of these

cells relative to the cardiogenic mesodermal cells. The timing

of this remodelling coincides with the onset of developmen-

tal growth of the tailbud embryo along the antero-posterior

axis (Larkin and Danilchik, 1999).

At stage 26, the pattern of XPOX2 staining within the

heart field has changed once more. Individual myeloid

cells are now observed having migrated around the anterior

limit of the lateral plate mesoderm between the heart field

and the cement gland (Figs. 7D, E and 8D). This results in

streams of myeloid cells at the ventral edge of each Nkx2-5

domain that emanate from a focal point at the ventral

midline, posterior to the cement gland (Fig. 7E). Where

the streams meet, transverse sections show that the cardio-

genic mesoderm is split into two by the XPOX2 expression,

which still resembles a cluster of cells rather than individual

leukocytes (Fig. 8E). The XPOX2-expressing cells caudal

of the focal point more typically resemble migratory

myeloid cells at this stage (Fig. 8F).

Finally, we used XPOX2 expression to study the location

of myeloid cells within developing hearts at swimming

tadpole stages in combination with XMLC1av (GenBank

AF364821). The mRNA of this cardiac myosin light chain

is robustly expressed in all regions of the forming myocar-

dium, including prospective atrial and ventricular chambers,

with low level somitic expression also detected (S.J.S.,

unpublished data). After the onset of myocardial differentia-

tion, fewer XPOX2-expressing cells are detected in the

heart-forming region of the embryo (Fig. 7F). Prior to

heart tube formation, they are present primarily at the ante-

rior and dorsolateral edges of the myocardial plate (Fig. 7F).

Additional XPOX2-expressing myeloid cells can be

detected between the myocardial region and the endoderm

(Fig. 8G). At linear and looping heart tube stages, transverse

sections reveal individual leukocytes associated with the

ventral side of the outer surface of the endocardium, resid-

ing between the endocardium and myocardium (Fig. 8H–K).

2.8. Transient expression of haematopoietic transcription

factors by myeloid cells

The transcription factor SCL plays a critical role in

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186180

Fig. 8. Transverse sections through the heart-forming region of Xenopus embryos after double whole-mount in situ hybridisation for the two markers of

cardiogenesis (in pale blue), Nkx2-5 (A–F) and MLC1av (G–K), in combination with POX2 (magenta). (A–C) Three representative sections marking progres-

sively posterior slices through the same stage 24 embryo depicted in Fig. 7(B,C). (D–F) Three progressively posterior sections through the stage 26 embryo

depicted in Fig. 7(D,E). Red arrows mark the ventral-most XPOX2-expressing mesodermal cells within the heart field. (G) Transverse section through the

myocardial plate of a stage 30 embryo. (H–J) Three progressively posterior sections through the myocardium of a stage 32 embryo. (K) Transverse section

through the looping heart tube of a stage 35 embryo. Sections (10 mm) are numbered (top right of each panel) commencing from the posterior limit of the cement

gland (A–C and D–F) or the anterior limit of the myocardium (G–K). Ect, ectoderm; Mes, mesoderm; End, endoderm; Mc, myocardium; Ec, endocardium.

haematopoiesis and vasculogenesis (Barton et al., 1999) and

in Xenopus, it is expressed in a broad antero-ventral, meso-

dermal domain of the early tailbud embryo (Mead et al.,

1998) that would encompass the XPOX2-expressing

myeloid progenitor cells. As development proceeds, expres-

sion of SCL becomes restricted to the VBI, posterior to the

developing heart (Mead et al., 1998). In order to establish

the precise relationship between the myeloid progenitors

and the other haematopoietic precursors in the tailbud

embryo, we compared the expression of Xenopus POX2

with SCL, using sequential, double whole-mount in situ

hybridisation.

In the early tailbud (stage 20), a major portion of the SCL

expression domain co-expresses XPOX2 (Fig. 9A, E).

However, the two expression domains soon resolve from

each other and by stage 22, XPOX2-positive myelopoietic

mesoderm still occupies the ventral midline but the same

cells express only low levels of SCL (Fig. 9B, F). Strong

SCL expression is now observed in cells immediately poster-

ior of the myeloid progenitors and also in two anterior-bilat-

eral domains that in part surround the myeloid progenitors.

By stage 24, XPOX2 expression is restricted rostrally to a

narrow midline strip (residing between bilateral Nkx2-5

domains) and caudally to a more dispersed group of cells.

Moreover, repression of SCL expression has occurred in

most of the region occupied by caudal myeloid cells (Fig.

9C, G). The few XPOX2-positive, punctate cells that appear

to retain SCL expression are most probably migratory leuko-

cytes that reside between the embryonic germ layers (Fig.

9C, G). By late tailbud stage, leukocytes that express XPOX2

are detected dispersed across the erythropoietic mesoderm of

the posterior VBI (Fig. 9D, H).

In summary, the myeloid progenitor cells transiently

express the haematopoietic transcription factor SCL, and

lie within a broader SCL expression domain that expands

caudally as the tailbud embryo develops. The caudal portion

of this comprises erythroid progenitor cells of the VBI that

form below the gut endoderm, whilst the anterior, bilateral

horns of SCL expression that flank the myeloid precursors

(Fig. 9C, G) probably comprise vascular endothelial

progenitors that will form the vitelline veins and perhaps

the endocardium (Mead et al., 1998). Qualitatively similar

results were obtained with a second haematopoietic tran-

scription factor, AML (CBFa2) (Tracey et al., 1998; Tracey

and Speck, 2000). AML expression detected by double

whole-mount in situ hybridisation persists in myeloid

progenitors until after stage 23, whereupon it is down-regu-

lated in a similar fashion to SCL (data not shown).

3. Discussion

3.1. XPOX2 and XLURP-1 are myeloid cell markers

Embryonic myeloid cells are likely to play an important

role in embryo morphogenesis, through their phagocytic

clearance of cell corpses. They are thus key participants in

the process of apoptotic cell death, regulated patterns of

which occur during development of the embryo. However,

to date the dearth of molecular markers has hindered studies

of the origins and functions of these cells in vertebrates.

Several lines of evidence indicate that XPOX2 and

XLURP-1 provide markers for embryonic myeloid cells in

Xenopus. Cells expressing these mRNAs are distributed

throughout the embryo in a pattern similar to that reported

for non-lymphoid leukocytes (Miyanaga et al., 1998;

Ohinata et al., 1989). XPOX2 is initially expressed in an

antero-ventral region of stage 19 embryos, consistent with

earlier grafting experiments which suggested that myeloid

cells originate rostral to the gill rudiments (Ohinata et al.,

1990). The morphology of the cells, after detection by

XLURP-1 expression, suggests that they are phagocytes.

Our tissue explant experiments suggest that XPOX2-expres-

sing cells are migratory, and imply that the entire population

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186 181

Fig. 9. XPOX2 expression in the anterior VBI. Double whole-mount in situ hybridisation for the haematopoietic transcription factor, SCL (pale blue), in

combination with POX2 (magenta). Embryos were photographed successively after the colour reactions for the SCL (A–D) and POX2 (E–H) probes were

developed. High magnification ventral views of the site of myeloid cell development at stages 20 (A,E), 22 (B,F) and 24 (C,G). (D,H) High magnification

ventral views of the trunk of a stage 28 embryo showing the posterior VBI. LS, lateral SCL expression; PS, posterior SCL expression; pVBI, posterior ventral

blood island; A, anterior; P, posterior; red arrow, narrow rostral POX2 expression within heart field.

of positive cells found in swimming tadpoles originates

from the domain of early XPOX2 expression. Moreover,

in transgenic animals, the promoter of the XLURP-1 gene

specifically directs the expression of fluorescent reporter

proteins to cells that are actually observed to migrate from

their antero-ventral origin. These cells co-express XPOX2

mRNA and after their dispersal, they resemble both tissue-

resident macrophages, and other leukocytes circulating in

the peripheral blood. Finally, the nature of XPOX2 and

XLURP-1 gene products, along with those of five other

cDNAs we have identified (including L-plastin) that exhibit

the same pattern of expression, is suggestive of their being

leukocyte or myeloid cell proteins.

Earlier studies of tadpole myeloid cells, identified by

leukocyte-specific antibodies, suggested the presence of

both mononuclear macrophage and polymorphonuclear

granulocyte lineages on the basis of cellular morphology

(Miyanaga et al., 1998; Ohinata et al., 1989). In our Xenopus

transgenesis experiments, we observed a subset of myeloid

cells that expressed XPOX2, yet did not exhibit detectable

XLURP-1 promoter activity (Fig. 6G–J). We are investigat-

ing whether variability in the intensity of EGFP fluores-

cence among the myeloid cells of a single embryo could

be an anomaly associated either with the transgenesis proce-

dure or with the choice of XLURP-1 promoter fragment.

Alternatively, it is possible that XPOX2-positive/XLURP-

1-EGFP-negative cells constitute a distinct cell type among

the myeloid population. As far as we know, the transgene

recapitulates the entire expression pattern of the endogenous

XLURP-1 gene with no evident ectopic activity. For exam-

ple, like the endogenous gene, the transgene does not appear

to be expressed in the rostral-most myeloid cells that remain

as part of the mesodermal layer within the heart field.

Among the fluorescent cells of XLURP-1-EGFP

embryos, we observed tissue macrophages but were not

able to identify distinct leukocyte lineages for the fluores-

cent cells in the peripheral blood. The limited histological

resolution available after whole-mount in situ hybridisation

also prevented us from assigning these cell lineages. For

these reasons, we have used the more general term,

‘myeloid cell’, until more is known of phagocyte popula-

tions in the embryo. Our study has only identified such cells

up to swimming tadpole stages and we have not yet studied

feeding tadpoles in which the thymic rudiment forms

(Nagata, 1977; Tochinai, 1980). We do not therefore

know if markers we have identified remain specific for

myeloid leukocytes in later stages of development or

whether they will identify definitive lymphoid cells of the

adult. Consequently, we are raising EGFP-expressing,

XLURP-1 promoter transgenic tadpoles to adulthood in

order to examine questions of leukocyte differentiation

and function at later stages of vertebrate development.

3.2. Myeloid cells and the heart field

An intriguing feature of the expression pattern for

XPOX2 is its spatial relationship with cells destined to

form the heart. In early tailbud stage embryos, mesoderm

fated to produce myeloid cells is located posterior to the

cardiac field of Nkx2-5 expression, but as the tailbud

embryo grows, the rostral extent of the myelopoietic cells

comes to lie in a narrow, ventral, mesodermal domain

within the heart field, Nkx2-5-expressing cardiogenic meso-

derm apparently being displaced on either side. This

arrangement coincides with the onset of morphogenetic

changes that convert the myocardial mesoderm into a linear

heart tube and is transitory, since the myeloid cells subse-

quently appear to migrate anteriorly towards the head, leav-

ing few in the heart primordium.

It is noteworthy that the first evidence of the endocardium

is a cluster of cells with a central lumen, dorsal to the

myocardial plate (Mohun et al., 2000; Nieuwkoop and

Faber, 1956). Interestingly, this is precisely the same loca-

tion that the myeloid cells occupy just a few hours earlier.

Whilst it is possible that such rostral XPOX2-expressing

cells of the mid-tailbud stage embryo themselves subse-

quently form the endocardial tube, we favour a model in

which cell movements account for the sequential appear-

ance of first myeloid and then vascular endothelial/endocar-

dial progenitors within the cardiac field (Fig. 10). The

anterior-bilateral horns of SCL expression observed in

Fig. 9C would thus be the candidate cells of the endocar-

dium and vitelline veins (Mead et al., 1998). In this model,

there could be an active role for myeloid cells in assisting

heart tube formation by phagocytosing apoptotic cells or

inducing localised changes to the extracellular matrix

(Kolker et al., 2000).

The initial site of myeloid differentiation in Xenopus

apparently conflicts with evidence in zebrafish, where

early macrophages arise from ventro-lateral mesoderm

located rostral of the cardiac field (Herbomel et al., 1999).

This difference can perhaps be explained by considering the

atrio-ventricular patterning evident during zebrafish cardiac

development, which is also reversed at linear heart tube

stages with respect to amphibians (Yelon et al., 1999).

Thus, in both lower vertebrate models, myeloid cell devel-

opment occurs adjacent to the side of the cardiac field that

will ultimately form the inflow tract and atrial chamber(s).

3.3. Myeloid cells and the VBI

In Xenopus, the first peripheral blood cells arise exclu-

sively from the VBI of tailbud stage embryos, while

progeny of the dorsolateral plate (DLP) contribute most

cells to the circulation during late larval stages and in the

adult frog (Kau and Turpen, 1983; Maeno et al., 1985).

Recently, the embryonic origin of the VBI has been deter-

mined and appears more complex than was first thought. In

addition to a well-characterised domain that arises from

progeny of the ventral blastomeres at the four-cell stage,

studies of the haematopoietic transcription factors, AML,

SCL, and GATA-2 have revealed an anterior domain that

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186182

is derived from dorsal blastomeres of the early embryo

(Ciau-Uitz et al., 2000; Tracey et al., 1998).

Early studies demonstrated that in addition to supplying

circulating erythrocytes, cells from the posterior compart-

ment of the VBI invade the thymic rudiment between stages

43 and 47 (Nagata, 1977; Tochinai, 1980), with differen-

tiated lymphocytes emerging to the periphery some time

after stage 49 (Kau and Turpen, 1983; Maeno et al.,

1985). More recently, lineage-labelling experiments have

shown that progeny cells of dorsal blastomeres overlap

with the a-globin expression domain, indicating that the

anterior compartment of the VBI is also a source of

erythroid cells (Tracey et al., 1998). In the present study,

we have identified XPOX2-expressing cells at stage 19 at an

antero-ventral site that correlates with the location of the

anterior VBI. Specifically, we have found that when first

detected, XPOX2-positive myeloid progenitors also express

SCL and AML. This co-expression is transient, with first

SCL and then AML mRNA being repressed as myeloid

differentiation proceeds. Moreover, in cell-fate mapping

experiments, we have found that the myeloid progenitors

detected in the early tailbud (stages 22–23) are entirely

the progeny of the dorsal-vegetal blastomeres of the eight-

cell stage embryo (Fig. 11, see figure legend for experiment

description). Together, these results indicate that a signifi-

cant portion of the anterior VBI is in fact myelopoietic.

Furthermore, our finding that such anterior-derived myeloid

cells subsequently migrate over the surface of the ventrally

derived posterior VBI indicates a hitherto unsuspected

complexity in the interpretation of lineage-labelling experi-

ments. As a result, it will be important to re-examine the

extent to which the presence of dorsally derived cells in the

a-globin expression domain actually results from migratory

phagocytes rather than reflecting the contribution of the

anterior VBI to embryonic erythropoiesis.

4. Experimental procedures

4.1. Isolation and sequencing of XPOX2 and XLURP-1

cDNA

X. laevis POX2 and LURP-1 clones were isolated from a

stage 22 cDNA library in pSPORT1 made from an anterior-

ventral tissue dissection that included the heart field and

some adjacent cement gland tissue. cDNAs were sequenced

using an Applied Biosystems 377 sequencer and analysed

using the Lasergene suite of programs (DNASTAR Inc.).

4.2. Whole-mount in situ hybridisation

Albino, but also some pigmented-wildtype, embryos

were fixed in MEMFA prior to RNA whole-mount in situ

hybridisation using digoxigenin-labelled antisense probes

(Harland et al., 1991; Sive et al., 2000). Double-labelled

in situ hybridisations used fluorescein-, and digoxigenin-

incorporated probes, and BCIP (without NBT) (Roche)

and Magenta-Phos (Biosynth AG) were employed for the

chromogenic reactions. A sequential procedure for double

whole-mount in situ hybridisation was adopted, where

embryos were photographed successively after the chromo-

genic reactions for each probe were developed. Antisense

probes corresponded to the full-length cDNAs for XLURP-

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186 183

Fig. 10. Myeloid cells and early morphogenesis of the heart field. A model illustrating the movement of cells within the anterior-ventral mesoderm of tailbud

stage Xenopus embryos. At stage 24, there are cardiac and vascular endothelial mesodermal domains, with the myeloid cells occupying the ventral-most

portion of mesoderm. In the space of 9 h, three processes occur during the early events of cardiac morphogenesis. (1) Myeloid cells disperse from their site of

origin, with the rostral-most cells migrating out from the ventral midline of the heart field. (2) Formation of the endocardial endothelium. Vascular endothelial

progenitors invade the area vacated by anterior-migrating myeloid cells, in this model, forming the endocardium as an extension of the vitelline veins. (3)

Fusion of the bilateral cardiogenic mesoderm on the ventral midline. These progenitors ultimately form the muscular myocardial tube, and also the non-muscle

pericardial and mesocardial components of the heart. As a direct consequence of the myeloid cell dispersal, the region adjacent and posterior to the developing

tadpole heart is devoid of mesoderm. This region accommodates the liver rudiment as it grows out from the gut endoderm. Finally, it should be noted that the

origin of the endocardium is not certain. Alternative models are possible with the endocardium forming from the anterior, arterial pole of the heart (not shown

in this model), or even resulting from differentiation of cells within the cardiac or myeloid domains. A, anterior; P, posterior; Mc.P, myocardial plate; Ec,

endocardium.

1, XPOX2 0, XMLC1av, XMLC2a (Chambers et al., 1994)

and XNkx2-5 (Tonissen et al., 1994). Shorter XPOX2-allele

probes (nucleotides 1–1448, 390–1448, and also 1099–1448

of the XPOX2 allele sequence) were also used and gave

identical results. Full-length cDNAs were used for XSCL

(Mead et al., 1998) and XAML (Tracey et al., 1998), and

were kindly supplied by Roger Patient. For histological

analysis, 10 mm sections were cut from fixed embryos

embedded in Paraplast wax (BDH). The normal table of

X. laevis (Nieuwkoop and Faber, 1956) was used to assess

the developmental stage of embryos.

4.3. Tissue explant culture

Microsurgical dissections were performed on a bed of 1%

agarose in 0.75 £ Normal Amphibian Medium (NAM)

(Sive et al., 2000), and resulting tissue explants were

cultured at 188C in 0.75 £ NAM with 40 mg/ml gentamicin

sulphate until they were fixed in MEMFA.

4.4. Cell-fate mapping

Approximately 3 nl of 1.25 mg/ml biotin-dextran

10,000 MW (Molecular Probes) was microinjected into

individual blastomeres of Xenopus eight-cell stage embryos,

in 4% ficoll 400,000, 0.75 £ NAM. Using the eight-cell

stage for injection, as opposed to four-cell stage, gives

more restricted targeting of the mesodermal germ layer.

Only regularly cleaving embryos were chosen for injection.

Embryos were allowed to develop in 0.1 £ NAM with

gentamicin sulphate until fixing. Alkaline phosphatase-

conjugated ExtrAvidin (Sigma) and BCIP (without NBT)

were employed for the chromogenic reaction.

4.5. Isolation of the XLURP-1 promoter

A X. laevis genomic DNA library in lFIX2 (Stratagene)

was screened using a 407 bp hybridisation probe corre-

sponding to a Sal1-Dra1, LURP-1 5 0-cDNA fragment.

The resulting Not1 digested genomic DNA fragments of

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186184

Fig. 11. XPOX2 expression in progeny cells of the dorsal-vegetal blastomeres. (A) Diagram illustrating the injection site of the cell-lineage marker, biotin-

dextran (pale blue), into an individual dorsal-vegetal blastomere of an eight-cell stage Xenopus embryo. The animal and vegetal poles, and the dorsal-ventral

axis, are indicated. The embryos were allowed to develop until stages 22–23, whereupon they were fixed and assayed for XPOX2 mRNA expression, and

stained to reveal the daughter cells of the injected blastomere. In 27 correctly targeted embryos, when viewed from the ventral side (Lane and Smith, 1999), we

observed two subtly different juxtapositions of myeloid gene expression and the posterior boundary of the dorsal-vegetal progeny cells, which are shown (B–

D,E–G). Nevertheless, with both observed configurations, the myeloid cells are wholly derived from the dorsal-vegetal blastomeres. The converse cell-fate

mapping experiment of ventral-vegetal blastomere cell-lineage injection never resulted in ventrally derived mesoderm present within the domain of myeloid

cells at stage 22 (data not shown). (B,E) Ventral views of the site of myeloid cell development for such representative tailbud embryos showing the progeny

cells of the dorsal-vegetal blastomere (pale blue) and XPOX2 expression (magenta). The stage 22 embryo (B) has left-sided lineage staining while the stage 23

embryo (E) has right-sided lineage staining. Black lines indicate the positions of the transverse sections illustrated (C,D,F,G). Red arrows (C,F,G) on the

sections mark mesodermal regions of coincident cell-lineage staining and XPOX2 expression, which appears a dark blue colour. Sections (10 mm) are

numbered (top right of each panel) commencing from the anterior limit of the XPOX2 expression domain. Pigmented-wildtype embryos were used for

this experiment. A, anterior; P, posterior; Ect, ectoderm; Mes, mesoderm; End, endoderm.

the 14 positively hybridising l phage clones were then

cloned intact into pBS2KS1. From these plasmid clones,

the DNA sequences located upstream of the XLURP-1

gene translation initiation site were PCR amplified with

proof-reading Advantage Polymerase Mix (Clontech)

using the antisense XLURP-1 oligonucleotide primer: 5 0-

CCCAggcgcgCCAAAACAACTGCAGTTTTtATAGTGG-

GTTTC-3 0 (in lower case are mismatches to create an Asc1

site and inactivate the endogenous XLURP-1 start codon),

and a primer specific to the flanking bluescript vector

sequence, either: 5 0-GTAACGCCAGGGTTTTCCCAGT-

CACGAC-3 0, or: 5 0-ACAATTTCACACAGGAAACAGC-

TATGAC-3 0. The clone that contained most sequence

upstream of XLURP-1 gave a 5 kb promoter fragment

with Not1 and Asc1 sites available for cloning purposes.

The XLURP-1 promoter fragment was cloned upstream

of the EGFP coding sequence using two derivatives of

pEGFP-1 (Clontech). pEGFP-1-8NPA contains a Not1-

Pac1-Asc1 linker between the Pst1 and BamH1 sites, and

allowed purification of an Asc1-Not1 EGFP fragment.

pEGFP-1-8PmN contains a Pme1-Not1 linker between the

same Pst1 and BamH1 sites, and allowed purification of an

‘empty’ Not1-vector fragment. The final expression

construct, pLURP1-g03-EGFP, was made using the Not1-

Asc1 XLURP-1 promoter fragment, the Asc1-Not1 EGFP

fragment, and the Not1 digested ‘empty’ vector fragment

of pEGFP-8PmN. The orientation of the cloned fragments

was such that the EGFP translation stop codon was adjacent

to the SV40 polyadenylation signal of the pEGFP-1-8PmN

vector.

4.6. Xenopus transgenesis procedure

Transgenic Xenopus embryos were generated as

described previously (Kroll and Amaya, 1996; Sparrow et

al., 2000b), except that pLURP1-g03-EGFP was simply

linearised with Pme1, rather than separating the promoter-

reporter sequence from the plasmid vector. The transgenesis

procedure has been repeated on five separate occasions,

with different batches of eggs giving identical results.

EGFP-detection of the full complement of myeloid cells,

as opposed to a small sub-population of fluorescent leuko-

cytes, occurred in approximately 40% of the transgenic

embryos. Successful transgenesis depends upon uniform

propagation of the integrated transgene from fertilised egg

to all daughter cells of the embryo. EGFP fluorescence was

detected using a Leica MZFLIII microscope equipped with

a GFP2 emission filter. Fluorescent images were captured

using a CoolSNAP camera and software from RS Photo-

metrics.

Acknowledgements

We thank Roger Patient for supplying XSCL and XAML

cDNAs, Duncan Sparrow for helpful discussions of

XLURP-1 and XPOX2 identities, Philippe Herbomel for

advice on macrophage identification, and Wendy Hatton

(NIMR Histology Service) for expert technical assistance.

References

Andermann, K., Wattler, F., Wattler, S., Heine, G., Meyer, M., Forssmann,

W.G., Nehls, M., 1999. Structural and phylogenetic characterization of

human SLURP-1, the first secreted mammalian member of the Ly-6/

uPAR protein superfamily. Protein Sci. 8, 810–819.

Barton, L.M., Gottgens, B., Green, A.R., 1999. The stem cell leukaemia

(SCL) gene: a critical regulator of haemopoietic and vascular develop-

ment. Int. J. Biochem. Cell. Biol. 31, 1193–1207.

Bennett, C.M., Kanki, J.P., Rhodes, J., Liu, T.X., Paw, B.H., Kieran, M.W.,

Langenau, D.M., Delahaye-Brown, A., Zon, L.I., Fleming, M.D., Look,

A.T., 2001. Myelopoiesis in the zebrafish, Danio rerio. Blood 98, 643–

651.

Bisbee, C.A., Baker, M.A., Wilson, A.C., Haji-Azimi, I., Fischberg, M.,

1977. Albumin phylogeny for clawed frogs (Xenopus). Science 195,

785–787.

Chambers, A.E., Logan, M., Kotecha, S., Towers, N., Sparrow, D., Mohun,

T.J., 1994. The RSRF/MEF2 protein SL1 regulates cardiac muscle-

specific transcription of a myosin light-chain gene in Xenopus embryos.

Genes Dev. 8, 1324–1334.

Chernokalskaya, E., Dubell, A.N., Cunningham, K.S., Hanson, M.N.,

Dompenciel, R.E., Schoenberg, D.R., 1998. A polysomal ribonuclease

involved in the destabilization of albumin mRNA is a novel member of

the peroxidase gene family. RNA 4, 1537–1548.

Ciau-Uitz, A., Walmsley, M., Patient, R., 2000. Distinct origins of adult and

embryonic blood in Xenopus. Cell 102, 787–796.

Diez-Roux, G., Argilla, M., Makarenkova, H., Ko, K., Lang, R.A., 1999.

Macrophages kill capillary cells in G1 phase of the cell cycle during

programmed vascular regression. Development 126, 2141–2147.

Evans, S.M., Yan, W., Murillo, M.P., Ponce, J., Papalopulu, N., 1995.

tinman, a Drosophila homeobox gene required for heart and visceral

mesoderm specification, may be represented by a family of genes in

vertebrates: XNkx-2.3, a second vertebrate homologue of tinman.

Development 121, 3889–3899.

Fleming, T.J., O’Huigin, C., Malek, T.R., 1993. Characterization of two

novel Ly-6 genes. Protein sequence and potential structural similarity to

alpha-bungarotoxin and other neurotoxins. J. Immunol. 150, 5379–

5390.

Franc, N.C., Heitzler, P., Ezekowitz, R.A., White, K., 1999. Requirement

for croquemort in phagocytosis of apoptotic cells in Drosophila.

Science 284, 1991–1994.

Harland, R.M., 1991. In Situ Hybridisation: An Improved Whole Mount

Method for Xenopus Embryos. In: Kay, B.K., Peng, H.B. (Eds.). Xeno-

pus laevis: Practical Uses in Cell and Molecular Biology, Academic

Press, London, pp. 658–695.

Herbomel, P., Thisse, B., Thisse, C., 1999. Ontogeny and behaviour of early

macrophages in the zebrafish embryo. Development 126, 3735–3745.

Herbomel, P., Thisse, B., Thisse, C., 2001. Zebrafish early macrophages

colonize cephalic mesenchyme and developing brain, retina, and

epidermis through a M-CSF receptor-dependent invasive process.

Dev. Biol. 238, 274–288.

Jones, G.E., 2000. Cellular signaling in macrophage migration and chemo-

taxis. J. Leukoc. Biol. 68, 593–602.

Kau, C.L., Turpen, J.B., 1983. Dual contribution of embryonic ventral

blood island and dorsal lateral plate mesoderm during ontogeny of

hemopoietic cells in Xenopus laevis. J. Immunol. 131, 2262–2266.

Kolker, S.J., Tajchman, U., Weeks, D.L., 2000. Confocal imaging of early

heart development in Xenopus laevis. Dev. Biol. 218, 64–73.

Kroll, K.L., Amaya, E., 1996. Transgenic Xenopus embryos from sperm

nuclear transplantations reveal FGF signaling requirements during

gastrulation. Development 122, 3173–3183.

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186 185

Lane, M.C., Smith, W.C., 1999. The origins of primitive blood in Xenopus:

implications for axial patterning. Development 126, 423–434.

Larkin, K., Danilchik, M.V., 1999. Ventral cell rearrangements contribute

to anterior–posterior axis lengthening between neurula and tailbud

stages in Xenopus laevis. Dev. Biol. 216, 550–560.

Maeno, M., Tochinai, S., Katagiri, C., 1985. Differential participation of

ventral and dorsolateral mesoderms in the hemopoiesis of Xenopus, as

revealed in diploid–triploid or interspecific chimeras. Dev. Biol. 110,

503–508.

Mead, P.E., Kelley, C.M., Hahn, P.S., Piedad, O., Zon, L.I., 1998. SCL

specifies hematopoietic mesoderm in Xenopus embryos. Development

125, 2611–2620.

Miyanaga, Y., Shiurba, R., Nagata, S., Pfeiffer, C.J., Asashima, M., 1998.

Induction of blood cells in Xenopus embryo explants. Dev. Genes. Evol.

207, 417–426.

Mohun, T.J., Leong, L.M., Weninger, W.J., Sparrow, D.B., 2000. The

morphology of heart development in Xenopus laevis. Dev. Biol. 218,

74–88.

Nagata, S., 1977. Electron microscopic study on the early histogenesis of

thymus in the toad, Xenopus laevis. Cell Tissue Res. 179, 87–96.

Nieuwkoop, P.D., Faber, J., 1956. Normal table of Xenopus laevis

(Daudin), North-Holland, Amsterdam.

Ohinata, H., Tochinai, S., Katagiri, C., 1989. Ontogeny and tissue distribu-

tion of leukocyte-common antigen bearing cells during early develop-

ment of Xenopus laevis. Development 107, 445–452.

Ohinata, H., Tochinai, S., Katagiri, C., 1990. Occurrence of nonlymphoid

leukocytes that are not derived from blood islands in Xenopus laevis

larvae. Dev. Biol. 141, 123–129.

Savill, J., Fadok, V., 2000. Corpse clearance defines the meaning of cell

death. Nature 407, 784–788.

Sive, H.L., Grainger, R.M., Harland, R.M., 2000. Early Development of

Xenopus laevis. A Laboratory Manual, Cold Spring Harbor Laboratory

Press, New York, NY.

Sparrow, D.B., Cai, C., Kotecha, S., Latinkic, B., Cooper, B., Towers, N.,

Evans, S.M., Mohun, T.J., 2000a. Regulation of the tinman homologues

in Xenopus embryos. Dev. Biol. 227, 65–79.

Sparrow, D.B., Latinkic, B., Mohun, T.J., 2000b. A simplified method of

generating transgenic Xenopus. Nucleic Acids Res. 28, E12.

Tepass, U., Fessler, L.I., Aziz, A., Hartenstein, V., 1994. Embryonic origin

of hemocytes and their relationship to cell death in Drosophila. Devel-

opment 120, 1829–1837.

Tochinai, S., 1980. Direct observation of cell migration into Xenopus

thymus rudiments through mesenchyme. Dev. Comp. Immunol. 4,

273–282.

Tonissen, K.F., Drysdale, T.A., Lints, T.J., Harvey, R.P., Krieg, P.A., 1994.

XNkx-2.5, a Xenopus gene related to Nkx-2.5 and tinman: evidence for

a conserved role in cardiac development. Dev. Biol. 162, 325–328.

Tracey, W.D., Speck, N.A., 2000. Potential roles for RUNX1 and its ortho-

logs in determining hematopoietic cell fate. Semin. Cell Dev. Biol. 11,

337–342.

Tracey, W.D., Pepling, M.E., Horb, M.E., Thomsen, G.H., Gergen, J.P.,

1998. A Xenopus homologue of aml-1 reveals unexpected patterning

mechanisms leading to the formation of embryonic blood. Development

125, 1371–1380.

Tsetlin, V., 1999. Snake venom alpha-neurotoxins and other ‘three-finger’

proteins. Eur. J. Biochem. 264, 281–286.

Turpen, J.B., Knudson, C.M., 1982. Ontogeny of hematopoietic cells in

Rana pipiens: precursor cell migration during embryogenesis. Dev.

Biol. 89, 138–151.

Wang, J., Maziarz, K., Ratnam, M., 1999. Recognition of the carboxyl-

terminal signal for GPI modification requires translocation of its hydro-

phobic domain across the ER membrane. J. Mol. Biol. 286, 1303–1310.

Yelon, D., Horne, S.A., Stainier, D.Y., 1999. Restricted expression of

cardiac myosin genes reveals regulated aspects of heart tube assembly

in zebrafish. Dev. Biol. 214, 23–37.

S.J. Smith et al. / Mechanisms of Development 117 (2002) 173–186186