POLYPEPTIDE INVOLVED IN CELL ADHESION AND PHAGOCYTOSIS
AND ITS USE IN TREATMENT
The present invention relates to a new polyto- pic membrane protein involved in adhesion and phagocytosis in Dictyostelium and to its use for producing molecules for treating medical indications that involve cell adhesion and/or phagocytosis.
Phagocytosis is the process by which cells internalize large particles (typically >lμm diameter) , such as bacteria or cell debris. In higher eukaryotes, phagocytic cells are an essential element of the host defense mechanism against invading pathogens and tissue remodeling. Phagocytosis involves adhesion of the phago- cytic cell to the particle, and reorganization of the actin cytoskeleton to allow engulfment .
A number of receptors required for the recognition of particles to be phagocytosed have been identified in mammalian cells, for example, Fc receptors involved in the phagocytosis of opsonized particles. These receptors presumably transduce a local activation signal upon recognition of their ligand, leading to reorganization of the actin cytoskeleton. Protein kinases such as Syk as well as GTP-binding proteins of the Rho family have been implicated in the transduction of the activation signal. The cellular slime mold Dictyostelium discoi- deum has been used previously as a model organism to study phagocytosis. The vegetative Dictyostelium amoebae multiply as free-living, single cells that feed phagocy- tically on bacteria. The mechanisms involved in phagocytosis by Dictyostelium cells are very similar to those used by mammalian phagocytes, and involve notably the actin cytoskeleton and RacFl, a member of the Rho family of GTP-binding proteins. In several studies, mutants deficient for phagocytosis have been generated by random mutagenesis, however the mutated genes could not be identified.

H- PJ ct H- PJ rt 3 J rt CD ■d J μ- £ CD tr id μ- •d 0 tr μ- Ω^ 3 0 Ω 3 rt rt J Hi X p. 3 Ω tr o 0 rt tr Φ 0 tl 3 Φ Φ φ PJ Φ ts tr ts J ts Φ Φ ts Φ Φ C 0 μ-
<l rt Φ X H- Φ o\° ii φ Λ PJ id PJ id a < id J CD Ω <J CD 3 Λ rt 03 0
Φ H- Φ &. Φ H •=; 3 CQ μ- μ- rt Φ sQ - rt 0 μ- tr rt C PJ μ- ts ø tT i CD SD μ- 0 Φ 3 Φ 0 μ- rt Ω φ ts μ- Ω 0 φ IQ M tr μ- ts Q rt 0 0) H- Ω CD ts •d J φ ts CQ Ω rt ts cQ α. rt Ω CQ M < ts PJ μ- rt ii rt rt
H- P. rt 0 rt Φ < ii CD rt Ω ■d C PJ Φ Ω φ μ- "< φ μ- φ a 03 0 φ CD S s;
0 H- ts μQ Φ Φ Φ rt tr Φ rt Φ rt o C φ J 0 rt id PJ J φ Ω. Φ 0 ts Φ H rt C ts < o μ- CD Φ o\° ts μ- ts 0 rt Ω μ- •s ii
03 PJ PJ X φ rt φ Φ LΠ a- μ- P. Q. μ- μ- 3 CQ CQ μ- Φ rt •d μ- rt J Φ 3
Ω ti 1-3 Φ P μ- H ts M o |χ) m D Φ rt CD ts & PJ •d 1-3 • μ- Ω μ- M M ts 03 0 03 c H sn - I-1 M Ω 0 ts PJ o\° ii tr μ- φ Φ 3 tr CD ts Φ 0 & ii μ- rt ts
Hi H- • : "\ Φ i-α Φ ts *-o tr Φ P- PJ rt J ts Λ |fl ts μ- μ- φ CQ PJ rt μ- CD 03 PJ
M 0 PJ X J o CD φ t rt C μ- rt ts Ω tr Φ PJ rt Φ ts c rt 0 o tl rt
X Φ 0 M rt CQ J "i *•<: Φ Φ CD CD PJ Ω Φ LQ μ- 0 PJ > CD 0 Ω α 3 μ- cQ tr tr c rt tr
Φ Ω Ω 3 H CD φ >d • Λ ii Ω 0 rt PJ φ P rt rt i Ω rt tr Φ tι 03 φ ti PJ 03 Φ
ΓT Ω ■d Φ I-1 J J C ii ii ti Ω ω ^ J φ Φ Ω X μ- Φ ti rt rt rt
£ Φ £ t a Id Ω- PJ i rt Φ Φ μ- J μ- a> > Ω & ω tr O CD rt J tr φ Φ O M to M PJ rt i Φ rt rt ts Ox tr ts CD •s μ- Φ ts μ- ts φ J rt Ω- φ φ i-S iQ 0 »d Φ μ- Ω Φ J PJ J μ- C μ- |Ω- μ- ts J PJ 0 rt 0 03 - rt H- 0 0 rt H- CD Ω Φ Φ Φ a id iCb Ω- id rt ts rt CD ts ^ . ts 0 μ- Φ μ- ϋ Φ
13- 0 Ω Φ Ω C PJ 3 0 J CD tr Φ id tr CD Ω^ ii J ts Ω Φ ι-3 μ- PJ
PJ Φ CQ •<: H- rt rt h-> CD μ- tr μ- iQ M 0 Φ ts φ φ Ω μ- tr ■d ti 0 μ- ts tr Ω M ω fi rt H- 3 Φ 0 J rt Si 0 ts 0 rt CD rt H rt (Ω φ ts PJ J CD < PJ φ φ Ω φ rt Ω
Φ 3 0 3 • M Φ 0. c μ- id μ- μ- tr < Ω 0 M c PJ ts ts 0 < tr rt rt Φ CD 0* J I-1 m μ- rt CD PJ PJ 0 rt M Φ φ Φ rt M PJ Ω 0 rt rt rt CQ 0 tr 0 3 H- H- H- •<: o rt 3 tr -x M ts • : 0 ts ts Φ CQ X μ- CQ 0 0 Φ Φ 03 rt
Φ • X 03 Ω 3 ■d μ- φ Φ μ- I-1 rt •C J Ω PJ rt ii μ- μ- 0 c i PJ ts rt X ill 0 ■d rt tl ts > PJ PJ 3 Φ & μ- μ- 3 rt J φ 03 id PJ φ Φ
JD J Ω rt *fl J rt μ- tr 0 PJ
^ a 0 tl ts 3 φ CQ CD CD tr tr rt
3 X 3 PJ H- H- ■ : Φ ts CD X X a <X Ό ts φ ti £ tr PJ φ 3 μ- rt* (ϋ Φ P 0 IQ •d P- PJ c 0 s 0 0 X CQ ~. ts μ- rt 3 rt PJ (Q CQ Φ μ- h-x
H- 3 Φ Φ id 0 3 ts 3 rr 0 PJ μ- μ- rt 0 ts C φ tr < 0 μ- 3 03 3 Φ
IQ Ti Ό tr 03 ω ■d 3 ^ id Ω 0 tr M ii 1
3 PJ Ω 0 X d 3 φ Φ Ω 0 tr . M
Φ 0 M φ rt ϋ 0 φ rt μ- Φ Ω t ts C PJ 3 Φ O μ- M |- ts M C
3 O rt μ- Φ CD rt Φ μ- o tn Φ μ- CD 0 Ω M Ω φ Ω rt PJ id φ rt s rt φ X p. P. rt φ tr • 0 I3 tr J i - CD PJ J Ω rt CD PJ 0 0 0 o ts rt 0 c-r Φ PJ •ϋ φ M φ ts 0 μ- PJ rt X ts ii CD C μ- - ti 03 Φ Φ φ
0 φ H- rt Ό Φ •d PJ ts CQ W μ- m 0 & Φ μ- 0 Ω φ Ω rt id P Φ Φ tr 3 M tr Ω CD <J 0 0 & Φ s; rt PJ ts μ- μ- Ω &. rr id rt tr
Ό rt Ω H- ii PJ tr φ 0 PJ Φ Ω Φ & CD 0 PJ Φ rt 01 ts ts PJ C tr M μ- μ- Φ
M H- tr rt ts <! 0 > ts μ- ts ■ : CD c ii rt μ- ^ &. ts & φ 0 ts <
0 P. ^ Φ < ii H- p. φ 3 ts rt rt 0 φ PJ Φ J 0 μ- Φ J rt φ id p. Φ P. 0 0 ts μ- ii PJ tr μ- 0 rt ts rt |p CD ts ts <! J Ω tr & ii id Φ rt ii
£ 3 S IQ 3 J rt Φ Φ PJ 0 CQ ts tr Ω tr & μ- 3 PJ Φ rt tr μ- tr Φ
Ω 0 φ X φ tr •d 3 ts μ- a. PJ Φ PJ rt 3 rt rt μ- J ts Φ 0 03
Φ PJ *• . φ rt ^ H H rt rt μ- CD rt Φ Q 0 <! 0 PJ PJ μ- £ tr Ω iQ 03 ii Φ
3 tr rt • ! Φ μ- φ 3 rt 0 J 0 μ- ii rt 0 CD PJ 0 ■ Φ ts
CD rt CD tr Ω X Φ PJ α, CD 0 21 PJ 0 tn CD Ω J M μ- ts Φ rt φ Ω tr id rt c X l-1 Φ H- rt PJ CD φ rt X ii tr rt CD Q CD o Ω PJ tr rt •< J μ- μ- 0 • μ- rt tr
Ω Φ o rt Φ PJ Φ φ tr o Tl rt P P ts rt . rt Ω ≤ J μ- tT Ω X 3 φ tr Ω rt Φ •s X 0 0 <! ra rt rt tr μ- X O IQ ts t μ- H cn PJ μ- X CQ o CQ 0 φ X 0 μ- ω Ω 0 <!
H- μ- ts Φ Φ o <*. s; Φ PJ H H- ■> Ω PJ CQ J μ- Ω Φ
P CO 0 0 PJ o\° Φ μ- rt CQ 1 < rt rt CQ 03 * ; ts tQ CD rt PJ Φ μ- Φ 1 1 rt tr J CQ 1
clearly retain their capacity to be implicated in cell adhesion and/or phagocytosis.
A further means of treatment of tumors, inflammations, bacterial infections, parasite infections, viral infections, septic shock, and for immuno-modulation are agonists or antagonists of the polypeptides of the invention.
The invention further relates to methods of treating tumors, inflammations, bacterial infections, parasite infections, viral infections, septic shock, and for immuno-modulation comprising the administration to a subject in need of treatment of a suitable amount of one or more antibodies directed against the membrane protein, and/or one or more soluble polypeptides of the invention and/or one or more agonists or antagonists of the membrane protein.
The present invention will be further illustrated in the following Example. In the Example reference is made to the following figures : Fig. 1. Phagocytosis mutants in Dictyostelium discoideum. Wild-type (thin line) or phgl-1 mutant (thick line) cells were incubated for lhr with lμm diameter FITC-labelled latex beads (A) , or FITC-dextran (B) . The amount of internalized fluorescence was analyzed using a fluorescence-activated cell sorter (FACS) . The fluorescence corresponding to 0 , 1, 2 or approximately 30 internalized beads is indicated in (A) .
Fig. 2. Structure of PHQ1 gene. (A) Schematic drawing of the PHG1 gene . Sequencing of the genomic DNA and cDNA revealed the presence of one intron, the position of which is indicated (amino acid 98) . The position of the vector insertion site in mutants phg*l-l (amino acid 329) and phgl-2 (amino acid 479) is also indicated. (B) Southern blot analysis of mutants phgl-1 and phgl-2. Genomic DNA was purified and digested with Clal . The plasmid recovered by rescue from phgl-1 was used as a probe. Molecular weight standards are indicated (kb) .
Fig. 3. Sequence analysis of the Dictyostelium Phgl protein. (A) The amino acid sequences of the Phgl protein and its closest human homologue (KIAA0255=D87444) were aligned using the ALIGN program. (B) Hydrophobicity plots (Kyte-Doolittle) of the two proteins.
Fig. 4. Internalization by phgl mutant cells. Cells were incubated with the indicated substrates (lμm diameter FITC-labelled latex beads, FITC-dextran, rhoda- mine-labelled K. aerogenes or DH5c- bacteria) for lhr with or without shaking. The internalized fluorescence was then analyzed using a fluorescence-activated cell sorter (FACS) . The results are expressed as a percentage of the internalization by wild-type cells. Three independent clones of each strain were analyzed. Strains used : phgl- 1 (black) and phgl-2 (hatched) .
Fig. 5. Mutant phgl cells are unable to feed phagocytically on bacteria. Growth of phgl-1, phgl-2 or wild-type cells was measured in HL5 medium (A) 2.104 cells were seeded in 2ml of medium. At the indicated times lOμl of the culture were recovered and cells were counted.
Each point represents the mean of 3 independent clones.
Fig. 6. The phenotype of phgl mutants is reversed by the expression of the Phgl protein (A) Approximately 100 cells were mixed with K. aerogenes bacteria and spread on SM plates. Growing clones of Dictyostelium were visible after seven days as clear areas in the bacterial lawn. (B) Expression of the Phgl protein was assessed by western blot in the same cell lines. Cells were lysed in sample buffer and the equivalent of 10s cells loaded on each lane and analyzed by immunoblotting with an antise- rum directed to a peptide of the Phgl protein. Molecular weight standards are indicated (kDa) . Truncated forms of Phgl are visible in mutant cells as a band of about 57kDa for phgl-1 cells and 50kDa for phgl-2 cells. (C) Cells were incubated for lhr in the presence of lμm diameter FITC-labelled latex beads with 200rpm shaking. Internalized fluorescence was then analysed as described in Fig.
1. Strains used : 1: wild-type; 2: phgl-1 ,- 3: phgl-1 + PHG1 ; 4: phgl-2 ; 5: Phgl-2 + PHG1.
Fig. 7. Adhesion of wild-type and phgl mutant cells to their substrate. Cells were grown on sterile glass plates for three days, fixed, dehydrated and coated with gold. They were visualised in a scanning electron microscope. Scale bar = lμm.
Fig. 8. Phgl protein is present in phagosomes . Cells were allowed to phagocytose latex beads, and phago- somes were purified by flotation on two successive sucrose gradients . Equivalent amounts of proteins from either the total cellular lysate (1) or the purified phagosomal preparation (2) were analyzed by Western blot for their content in PDI, a marker of the endoplasmic reticulum or in Phgl protein. Molecular weight standards are indicated (kDa) .
EXAMPLE
Identification of a Dictyostelium membrane protein invol- ved in cell adhesion and phagocytosis
1. Introduction
To dissect the molecular mechanisms involved in phagocytosis, random insertion mutants of Dictyostelium discoideum were generated and two mutants defective for phagocytosis selected. Both represented insertions in the same gene, named PHG1. This gene encodes a polytopic membrane protein with an N-terminal lumenal domain and nine potential transmembrane segments. Homologous genes can be identified in many species. Disruption of PHG1 caused a marked defect in phagocytosis of latex beads and bacteria, but did not noticeably effect fluid phase endocytosis. This defect in phagocytosis was caused by a decrease in the adhesion of mutant cells to phagocytosed particles. These results indicate that the Phgl protein is involved in the adhesion of Dictyostelium to various substrates, a crucial event of phagocytosis.
2. Materials and methods
2.1 Cell culture and internalization assays
Wild-type cells used in this study are DH1-10 cells, a subclone of DH1 cells isolated by the present inventors. They were grown at 21°C in HL5 modified medium (Cornillon et al , J. Cell Sci . 107, 2691-2704, 1994) in 90mm diameter petri dishes (Bibby Sterilin Ltd, Stone, England) and subcultured twice a week. Cells were typically not allowed to reach a density of more than 2xl06 cells/ml.
To assess the growth of various strains on a bacterial lawn, approximately 100 cells of each cell type were mixed with 400μl of an overnight culture of K. aerogenes and plated on SM plates (Kay, Meth. Cell Biol . 28, 433-448, 1987) . Cells were allowed to grow for 7 days before plates were photographed. For growth in suspension, 2xl04 cells were seeded either in HL5 medium or in liquid medium A supplemented with K. aerogenes as described (Raper, Cultivation. In The Dictyostelids . Princeton University Press, Lawrenceville, NJ, 48-76, 1984) . The cells were grown at 21°C with shaking (200rpm) , and an aliquot counted every day.
To obtain rhodamine-labeled bacteria, an overnight culture was centrifugated and resuspended in PBS. Bacteria were boiled 30min in a water bath under mild stirring, washed 4 times with PBS and once with SB (2mM Na2HP04, 14.7mM KH2P04, pH 6.0) . Cells were then resuspended in a filtered solution of 50mM Na2HP04 pH 9.2/5mg of rhodamine-isothiocyanate (ICN Biomedicals Inc.) at 2xl09 cells/ml and incubated for 30min under mild agitation. Cells were then washed twice in SB/40mM NH.C1 , twice in SB and frozen in aliquots .
To assess internalization, 105 cells were transferred in 1ml of fresh HL5 medium containing lμl of lμm-diameter fluoresbrite YG carboxylate microspheres (Polysciences Inc., Warrington, PA), or 0.5mg/ml FITC- dextran (Molecular Probes, Eugene, Oregon), or 0.25mg/ml Lucifer Yellow CH (Sigma Chemical Co., St Louis, MO), or
5.107 rhodamine-labeled bacteria. The cells were incubated with or without shaking (200rpm) for lhr, then washed twice with ice-cold HL5 and analyzed using a fluorescence spectro luorometer (FACSCalibur, Beckton Dickinson, San Jose, CA) .
2.2 Isolation of phagocytosis mutants
Cells were transformed with the pUCBsrΔBamHI vector by the REMI procedure essentially as described elsewhere (Cornillon et al . , Cell Death Diff . 5, 416-425, 1998; Adachi et al., Biochem. Biophys . Res. Commun. 205, 1808-1814, 1994; Kuspa and Loomis, Proc . Natl . Acad. Sci . USA, 89, 8803-8807, 1992) . Briefly, cells were washed once in sterile ice-cold electroporation buffer (lOmM NaP04 pH 6.1, 50mM sucrose), mixed with lOμg of BamHI linearized vector and 10 units of DpnII restriction enzyme and electroporated using a BioRad Gene Pulser (0.4cm cuvettes, lkV, 3μF) , then resupsended in 30ml of HL5 medium. Blasticidin S hydrochloride (lOFg/ml; ICN Biomedicals Inc., Aurora, Ohio) was added 24hrs later.
Ten days later the cells were grouped into five pools, incubated with lμm diameter FITC-labelled latex beads in HL5 medium for lhr and the cells having phagocytosed no beads were sorted in a Fluorescence activated cell sorter (FACSstar plus) . Cells were subjected to a second round of selection 7 days later and cloned in 96 wells plates at the exit of the FACS . Finally, individual clones were retested for their ability to phagocytose lμm diameter FITC-labelled latex beads. Forty seven individual clones were identified as deficient for phagocytosis. Three of them corresponded to phgl-1 mutants, and one to a phgl-2 mutant.
Genomic DNA from selected clones was extracted as described (Chang et al . , Plasmid 34, 175-183, 1995) digested with various enzymes (Clal, Ndel, EcoRV) and analyzed by Southern blot as previously described (Cornillon et al . , supra , 1998) , using radiolabelled pUCBsrΔBamHI as a probe. Genomic DNA from phgl mutants
(4μg) was digested with Clal, purified by repeated phenol-chloroform extractions, precipitated in EtOH and resuspended in lOOμl dH20. Fifteen microliters of DNA were then ligated overnight at 16°C in a final volume of lOOμl, in the presence of 400 units of ligase (New England Biolabs, Beverly, MA) . Ligation products were precipitated in the presence of lOOng/μl glycogen, resuspended in lOμl dH20, and 5μl was used to transform DH10B bacteria by electroporation. Plasmids recovered by plasmid rescue of phgl mutants (lOμg) were linearized with Clal and electroporated into DH1-10 cells (see above) . After blasticidin S hydrochloride selection of cells insertion of plasmid at the homologous site in the genome was confirmed by southern blot, using the vector obtained from plasmid rescue as a probe. Disruption mutants thus obtained were then used for all the experiments described here .
A cDNA encoding Phgl protein (SSL630) was obtained from Morio et_al . , DNA Res. 5, 335-340, 1998) and subcloned into pDAX-3C expression vectors (Manstein et al. , Gene 162, 129-134, 1995) to obtain the pSC3A vector .
For complementation experiments, lOμg of pSC3A were digested with Pvul and electroporated into phgl mutant cells. After G418 selection performed with lOμg/ml of G418, cells were cloned and tested as described.
2.3 Adhesion of cells to substrate Fifteen thousand mid-log phase cells (2 to 3xl06 cells/ml) of each cell type in triplicate were allowed to adhere for 20min in 96 well plates. Plates were then shocked three times before discarding the supernatant . Cells were resuspended in lOOμl of HL5 and counted. To visualize adherent cells by scanning electron microscopy, cells were seeded in HL5 medium at low density on sterile glass coverslips and allowed to grow for three days. The coverslips were transfered
delicately to HL5 containing 1% glutaraldehyde, and fixed for 30min at room temperature. They were then rinsed with PBS and dehydrated by successive lOmin incubations with increasing concentrations of ethanol (30%, 50%, 70%, 90%, 100%) . Dehydrated cells were vacuum coated with gold and photographed with a Siemens Autoscan scanning microscope .
2.4 Western blot analysis
Cells were washed once in PBS, resuspended at 106 cells/20μl in sample buffer (0.103g/ml sucrose, 5xl0"2M Tris pH 6.8, 5xlO"3M EDTA, 0.5mg/ml bromophenol blue, 2% SDS) , and 20μl of each sample were run on a 9% acrylamide gel in non-reducing conditions. The gel was then transfered onto a Protran BA 85 membrane cellulosenitrate (Schleicher & Schuell, Dassel, Germany) in a Mini Trans- Blot II Module (Bio-Rad) according to manufacturer instructions. The membrane was incubated sequentially with an antipeptide antiserum (YC1, 1/300) directed to a sequence in the lumenal domain of the Phgl protein (YKKVENWKGDTGDDC) , and with an HRP-coupled donkey anti- rabbit Ig (Amersham) , washed and revealed by ECL. Monoclonal antibody to protein disulfide isomerase (221- 135-1) was according to Monnat et al . , FEBS Lett. 418, 357-362, 1997) .
2.5 Purification of phagosomes
Approximately 7xl08 exponentially growing cells were allowed to internalize lμm diameter latex beads for 90 min in HL5 medium, washed twice with HL5 and once with ice-cold homogenization buffer (Sucrose 250mM, Hepes 5mM, Imidazol 3mM pH 7.4, aprotinin 2μg/ml, leupeptin 2μg/ml, PMSF lOOμM) . Cells were resuspended in 2ml of homogenization buffer, and disrupted by 25 passages in a ball-bearing cell cracker. Intact cells were removed by centrifugation for 5min at 2000rpm, and the phagosomal fraction was isolated as already described by flotation on a sucrose step gradient (Desjardins et al . , J. Cell Biol. 124: 677-688, 1994). The phagosomal fraction was
collected at the interface of the 10 and 25% sucrose solutions, and submitted to a second round of flotation, a procedure which was found to reduce significantly the contamination by non-phagosomal markers. The phagosomal membranes were finally diluted with PBS and pelleted by centrifugation. Total protein content was tested by Micro BCA protein assay (Pierce, Rockford IL) in the unpurified cellular lysate and in the purified phagosomal fraction, and equivalent amounts of proteins (300ng) were loaded on SDS-polyacrylamide gels, and analyzed by Western blot as described above .
3. Results
3.1 Isolation of phgl mutants defective for phagocytosis To isolate mutants defective for phagocytosis, the classical method for the generation of mutants by restriction enzyme mediated integration (REMI; Kuspa and Loomis, supra , 1992) was followed. For this, a plasmid (pUCBsrΔBam) containing a selectable marker (resistance to blasticidin) was introduced into cells by electroporation together with the DpnII restriction enzyme. This results in the random integration of the plasmid at one of the many DpnII sites present in the genome of Dictyostelium. Using this technique approximately 10000 independent clones were obtained after blasticidin selection.
The cells were then incubated for lhr in the presence of Iμm diameter FITC-labelled latex beads, washed, and cells having phagocytosed no beads were sorted in a fluorescence activated cell sorter (FACS) . One week later the procedure was repeated, and the negative cells cloned into individual wells. The resulting clones were then individually tested and those found to be defective for phagocytosis of latex beads were identified (Fig. la) .
Two of these clones (phgl-1 and phgl-2 ) were selected for further analysis . They exhibited a marked decrease in their ability to phagocytose latex beads
(Fig. la) , but fluid-phase uptake of FITC-dextran appeared to be uneffected (Fig. lb) .
Digestion of the genomic DNA of these two clones with Clal yielded an identical plasmid-containing fragment of approximately 9kb (Fig. 2) . Genomic DNA digested with Clal was allowed to recircularize, then introduced into bacteria. This resulted in the selection of bacteria containing the original plasmid with the flanking sequences of the integration site. Upon sequencing, the two clones were found to represent distinct insertions in the same gene, named PHGl (Fig. 2) .
To ascertain that the observed phenotype was caused solely by the insertion in the PHGl gene, plasmids corresponding to the phgl-1 and phgl-2 insertions were retransfected into wild-type Dictyostelium cells. The insertion of the plasmid at the homologous site was verified by southern blot (not shown) . In both cases the insertion of the pUCBsrΔBamHI plasmid in PHGl gave rise to a mutated strain with a strong defect in phagocytosis.
3.2 The Phgl protein belongs to a new family of polytopic membrane proteins
The genomic DNA of PHGl was sequenced, as well as a full-length cDNA clone recovered from the
Dictyostelium cDNA library (SSL630; Morio et al . , supra , 1998) . The Phgl protein (642 amino acid residues) exhibits a potential N-terminal signal sequence followed by a lumenal domain and 9 potential membrane-spanning segments (Fig. 3) . The alignment of the Phgl protein with its closest homologue in human (KIAA0255=D87444) is shown (Fig. 3a), as well as their hydrophobicity profiles (Fig. 3b) .
Numerous genes previously sequenced in other species show a high homology to PHGl gene. All of them exhibit a similar overall structure, with a rather variable potential lumenal domain followed by a more conserved membrane domain with nine putative
transmembrane domains . Three homologues of PHGl have been fully sequenced in human (D87444=KIAA0255 , U81006 and U94831; Chluba-de Tapia et al . , Gene 197, 195-204, 1997; Schimmδler et al . , Gene 216, 311-318, 1998), three in Saccharomyces cerevisiae (U53880, Z48758, U18916) , two in Caenorhabditis elegans (Z79759, AF026213) and three in Arabidopsis thaliana (AC005967, AC006532, U95973) . EST databases also reveal additional related genes in Dictyostelium and human, and homologues in other species (Drosophila melanogaster, Mus musculus) . The present invention now provides a novel function to these known homologues .
3.3 phgl mutants are defective for adhesion To determine more precisely the nature of the internalization defect in phgl mutants, the ability of the mutant cells to internalize various types of particles was tested. As seen during the isolation of the mutants, phgl mutant cells exhibited a very strong inability to phagocytose lμm diameter FITC-labelled latex beads (Fig. 4) . Phagocytosis of more physiological substrates by Dictyostelium mutants was also tested. A significant defect was also observed for the phagocytosis of DH5c- bacteria (Fig. 4) . The internalization of the fluid phase markers FITC-dextran (Fig. 4) or lucifer yellow (not shown) was comparable to that of wild-type. Thus phgl mutant cells are strongly defective for phagocytosis, but their ability to internalize fluid phase or membrane is not effected. The phenotype of the mutant cells was stable over a period of 3 months of culture, with only minor variation from experiment to experiment (compare Fig. 4, 6 and 7) .
A defect in phagocytosis would be expected to result in a loss of the ability of Dictyostelium cells to feed on bacteria. In agreement with this, phgl mutant cell lines did not grow in a bacterial suspension (Fig. 5b) , while their growth in liquid HL5 medium was normal (Fig. 5a) .
PJ Ω Ω rt
Φ Φ tr
X φ
Φ fi CD CD ts
Φ 0
Φ £ ti rt X Φ 1
0 X fi PJ μ- φ rt tr tr tr μ- Ω Φ φ rt 0 M
S3 φ
•s J tl tl φ rt rt - 3 Φ h-x P Ω
03 ii • Φ
. X
Φ >
03
P. 03 s;
Φ tr Φ
Ω o f ti ≤ Φ
Φ P
PJ $,
Ui μ- PJ
Φ ø CD
X μ- H φ ti PJ tr rt h-1 tr Φ PJ φ ts μ- H ii rt
PJ id tr tr tr Φ μ- Q
H ii μ- Φ rt 3 3 ; S3 PJ rt μ- rt PJ ts o ts μ- rt ts cQ
Table 1
Adhesion of phgl cells is defective
In addition, the cells were allowed to adhere to a glass plate, and the hydrodynamic stress necessary to detach 50% of the cells was determined as described above. The adhesion of phgl mutant cells in HL5 medium was markedly decreased compared to wild-type cells, as revealed by the smaller value of the hydrodynamic stress measured for phgl cells (<0.025 Pa) compared to wild-type cells (0.5 ± 0.1 Pa) .
Upon more prolonged culture in HL5 medium, some phgl mutant cells did adhere to their substrate. However examination of the cells by scanning electron microscopy revealed distinct differences between adherent wild-type and mutant cells. Whereas wild-type cells adhered tightly to the glass coverslip, phgl cells did not spread as extensively and could be seen to detach locally from the substrate (Fig. 7) . Together these results indicate that the primary defect of phgl mutant cells is a decrease in their adhesion capacity.
H I-1 o LΠ o ι-π