* Department of Biochemistry, School of Medicine, Fukuoka University, Fukuoka, Japan; and
Suntory Biomedical Research Limited, Osaka, Japan
Correspondence: Dr. Motomu Kuroki, Department of Biochemistry, School of Medicine, Fukuoka University, 7-45-1 Nanakuma, Jonan-ku, Fukuoka 814-0180, Japan. E-mail: mokuroki{at}fukuoka-u.ac.jp
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: CEA family • CD66b • CD66c • homologue-scanning mutagenesis
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Each CEACAM protein consists of several domains similar to the immunoglobulin (Ig) V or C domains: one NH2-terminal IgV-like domain (N-domain) of 108 amino acids and zero to six IgC2-like domains (A- and B-domains) followed by a glycolipid membrane anchor or a transmembrane-spanning region with a cytoplasmic tail [1 , 3 ]. The N-domains of CEACAM family proteins reveal sequence similarity with 70–90% identity to each other. CEA [5 ], CEACAM1 [7 , 17 ], and CEACAM6 [6 ] exhibit homophilic adhesion and heterophilic adhesion amongst them. Conversely, CEACAM8 binds only CEACAM6 relatively strongly compared with the binding activities of the other CEACAM members without showing homophilic-adhesion activity [8 ]. In contrast to these CEACAM antigens, CEACAM3 exhibits neither homophilic adhesion nor heterophilic adhesion to the other CEACAM members [18 ]. However, CEACAM3 binds pathogenic bacteria belonging to Neisseria through recognition of the cell-surface Opa proteins as do CEA, CEACAM1, and CEACAM6 [14 , 15 ]. No adhesion activity has been demonstrated for CEACAM4 or CEACAM7. It has been demonstrated that the homophilic adhesion of human [19 ] and mouse CEACAM1 [20 ] is mediated by direct interaction between their N-domains. Conversely, double-reciprocal interaction between the N-domain of one molecule and the A3-domain of the other was proposed for the homophilic adhesion of CEA [21 ]. In our previous studies [8 , 22 ], we have demonstrated that the N-domains of CEACAM6 and CEACAM8 are important for their heterophilic adhesion by using blocking monoclonal antibodies (mAbs) and chimeric proteins between CEA and CEACAM6. Recently, several amino acid residues in the N-domains of CEA [23 ] and CEACAM1 [24 ] have been identified as critical residues for binding to the neisserial Opa proteins. However, the structural basis is still unclear for the homophilic and heterophilic interactions between the CEACAM family molecules as well as their binding to the bacterial proteins.
In the present study, to elucidate the structure/function relationships in the homophilic adhesion of CEACAM6 and in the heterophilic adhesion between CEACAM6 and CEACAM8, we first analyzed the importance of their N-domains and then located and compared the critical amino acid residues for their adhesion activities by introducing single and multiple homologue substitutions into their sequences with particular emphasis on the residues that are unique to CEACAM6 or CEACAM8.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Preparation of Chinese hamster ovary (CHO) cells expressing
recombinant CEACAM proteins
N-domains of the CEACAM6 and CEACAM8 were exchanged to each
other as follows: cDNA fragments encoding CEACAM6 [25
]
and CEACAM8 [26
] were inserted into the EcoRI
restriction enzyme site of the plasmid pUC118, yielding pUC118-CEACAM6
and pUC118-CEACAM8, respectively. The 313-bp fragment that covers the
leader sequence and the residues 1–71 in the N-domain of each CEACAM
antigen was obtained by digestion of the plasmids with NcoI
and NsiI restriction enzymes. Each fragment of CEACAM6 and
CEACAM8 was inserted into the NcoI and
NsiI-digested pUC118-CEACAM8 and pUC118-CEACAM6,
respectively. The resultant plasmid pUC118-CEACAM6/8 encodes a chimeric
protein comprised of the residues 1–71 of the CEACAM6 N-domain
followed by the residues 72–108 of the CEACAM8 N-domain and the
AB-domains of CEACAM8 (see Fig. 1 ). Similarly, the plasmid pUC118-CEACAM8/6 encodes a chimeric protein
consisting of the residues 1–71 of the CEACAM8 N-domain followed by
the residues 72–108 of the CEACAM6 N-domain and the AB-domains of
CEACAM6. For preparation of CEACAM3/6, a chimeric protein consisting of
the residues 1–93 of the CEACAM3 N-domain, followed by the residues
94–108 of the CEACAM6 N-domain and the AB-domains of CEACAM6, the
NcoI and BgIII-digested fragment of
pUC118-CEACAM3 [25
], was replaced with the
NcoI and BgIII-digested fragment of
pUC118-CEACAM6. The EcoRI-digested insert of each plasmid
was ligated into the unique EcoRI site of the mammalian
expression vector pdKCR-neo [27
]. CHO cells were
transfected with the expression vectors by the calcium-phosphate method
using a kit (Cell-Phect Kit; Amersham Pharmacia Biotech, Uppsala,
Sweden) and selected by culture in alpha-MEM containing 10% FBS and
500 µg/ml geneticin. Expression of the recombinant proteins was
analyzed by flow cytometry on a FACSCaliver (Becton Dickinson, San
Jose, CA) after successive treatments with CD66 mAbs and FITC-labeled
anti-mouse IgG and by immunoblotting after sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described
[16
].
 View larger version (54K): [in a new window] |
Figure 1. Involvement of N-domains in the cell-adhesion activities of CEACAM6 and
CEACAM8. (A) Portion (residues 1–71) of each N-domain of CEACAM6
(indicated by 6) and CEACAM8 (8) was replaced by recombination of their
cDNAs to yield chimeric proteins CEACAM8/6 (8/6) and CEACAM6/8 (6/8),
respectively. The domain structures of these recombinant proteins are
schematically shown on the right-hand side (N, N-terminal IgV-like
domain; A and B, IgC2-like domains; GPI, glycosylphosphatidyl
inositol). CHO cells expressing the native and chimeric proteins were
mixed for 30 min, and cell aggregation was determined by calculating
single cells, which were expressed as percentage of total cells. Data
are means ± SD of one representative experiment
performed in triplicate out of three similar ones. (B) Expression
levels of the recombinant proteins on CHO cells shown by flow cytometry
with F34-187, reactive with N-domain of CEACAM6; 80H3, reactive with
N-domain of CEACAM8; and MOPC21, negative control.
|
Cell-adhesion assays
We used two different methods, a cell-aggregation assay and a
cell-attachment assay, to evaluate cell-adhesion activities. For the
cell-aggregation assay, CHO transfectants expressing CEACAM proteins
were dissociated by incubation in phosphate-buffered saline (PBS)
containing 0.1% trypsin and 5 mM ethylenediaminetetraacetate (EDTA)
and were resuspended in alpha-MEM at 1 x 106
cells/ml. Cell aggregation was performed in aliquots of 200 µl in
wells of round-bottom 96-well plates (Nunc, Roskilde, Denmark), which
were rotated at 240 rpm on a platform shaker at room temperature. After
30 min, the number of nonaggregated cells was microscopically counted
with a hemacytometer, and the percentage of cell aggregation was
calculated. Alternatively, cell-adhesion activity was assessed by
attachment of CHO transfectants to CEACAM protein-coated plates.
Recombinant CEACAM6 and CEACAM8 proteins or bovine serum albumin (BSA;
1 µg/well) were dried up onto the wells of 96-well flat-bottom plates
(Greiner, Frickenhausen, Germany) as described previously
[22
]. CHO transfectants were labeled with H33342 (50
µg/ml in serum-free medium) for 30 min at 37°C, plated into the
CEACAM-coated wells (5x104 cells/100 µl/well), and
allowed to attach to the bottom for 20 min at room temperature. Wells
were gently washed twice with 100 µl PBS, and cells adhered to the
bottom were solubilized in 100 µl of 0.2% (w/v) Nonidet P-40.
Fluorescence of the cell lysates was determined by using a Fluoroskan
II reader (Labsystems Oy, Helsinki, Finland). Percent-adherence was
calculated by the following formula: (fluorescence of adherent
cells - fluorescence of solubilizing buffer)/fluorescence of
total cells added - fluorescence of solubilizing buffer) x
100.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
N-domain residues critical for the homophilic adhesion of CEACAM6
and its heterophilic adhesion to CEACAM8
To locate the amino acid residues critical for the homophilic
adhesion of CEACAM6, we carried out homologue mutagenesis screening of
the N-domain. The CEACAM family antigens show similar amino acid
sequences of their N-domains to each other with about 70–90%
identities (Fig. 2
). Each residue unique to the N-domain (the residues 1–71) of
CEACAM6 was replaced with the respective residue of CEA by
site-directed mutagenesis (Fig. 3A
). As shown in Figure 4A
, CHO cells expressing CEACAM6-H27, a CEACAM6 mutant whose 27th
residue of the N-domain is histidine, were reacted with F34-187 and
F106-88, mAbs for epitopes on the N- and the AB-domains of CEACAM6,
respectively [2
]. Further, 4/3/17, which recognizes CEA
and CEACAM1 but hardly CEACAM6 [2
], bound to
CEACAM6-H27, indicating that the epitope for this mAb was newly
generated on CEACAM6 by substitution of N27 with histidine. Similar
reactivities with F34-187 and F106-88 were observed for CEACAM6-N32
(Fig. 4B) and the other CEACAM6 mutants (unpublished results).
Moreover, these mutants showed similar molecular masses on SDS-PAGE to
that of the CEACAM6 (about 80 kDa; unpublished results). These findings
imply that the extracellular domains of the CEACAM6 mutants were
expressed properly on the cell surfaces and that their expression
levels are similar enough to compare their cell-adhesion activities.
 View larger version (33K): [in a new window] |
Figure 2. Amino acid sequences of the N-domains of CEACAM antigens deduced from
their cDNA sequences. The residues identical to those of CEA are
indicated by dashes.
|
 View larger version (24K): [in a new window] |
Figure 3. Amino acid sequences of CEACAM mutants and their adhesion activities.
Residues 1–64 of N-domains are shown. (A) CEACAM8 mutants; (B) CEACAM6
mutants; (C) CEACAM3/6 mutants. Dashes indicate the residues identical
to those of each parental protein. ++, Strong binding; +, weak binding;
-, no binding.
|
 View larger version (37K): [in a new window] |
Figure 4. Flow cytometric analysis of CHO cells expressing CEACAM mutants. CHO
transfectants expressing CEACAM mutants—(A) CEACAM6-H27; (B)
CEACAM6-N32; (C) CEACAM8-S32; (D) CEACAM8-Q44—were reacted with mAbs
for the N-domain (F34-187, 4/3/17) or AB-domains (F106-88) of CEACAM6
(A and B) and for the N-domain (80H3 and B13.9) or AB-domains (TET-2)
of CEACAM8 (C and D). See Figure 3A
and 3B
, for the locations of
mutation. MOPC21 IgG was used for negative control. Binding of mAbs was
analyzed by flow cytometry after incubation with FITC-conjugated
anti-mouse IgG.
|
 View larger version (70K): [in a new window] |
Figure 5. Effects of mutations in the N-domain amino acid sequence on the
homophilic-adhesion activity of CEACAM6. CHO cells expressing CEACAM6
(indicated by 6) or CEACAM6 mutants, CEACAM6-V21 (6-V21), CEACAM6-H27
(6-H27), CEACAM6-L28 (6-L28), CEACAM6-F29 (6-F29), CEACAM6-N32 (6-N32),
or CEACAM6-Q44 (6-Q44), were mixed for 30 min. Cell aggregation was
determined by calculating single cells, which were expressed as
percentage of total cells. See Figure 3A
for the locations of mutation.
Data are means ± SD of one representative experiment
performed in triplicate out of three similar ones.
|
 View larger version (65K): [in a new window] |
Figure 6. Effects of mutations in the N-domain amino acid sequence on the
heterophilic-adhesion activity of CEACAM6 to CEACAM8. CHO cells
expressing CEACAM6 mutants, CEACAM6-V21 (indicated by 6-V21)
CEACAM6-H27 (6-H27), CEACAM6-L28 (6-L28), CEACAM6-F29 (6-F29),
CEACAM6-N32 (6-N32), or CEACAM6-Q44 (6-Q44), were mixed with CHO cells
expressing CEACAM8 (8) for 30 min. Cell aggregation was determined by
calculating single cells, which were expressed as percentage of total
cells. See Figure 3A
for the locations of mutation. Data are means ± SD of one representative experiment performed in
triplicate out of three similar ones.
|
 View larger version (48K): [in a new window] |
Figure 7. Effects of mutations in the N-domain amino acid sequence on the
heterophilic-adhesion activity of CEACAM8 to CEACAM6. CHO cells
expressing CEACAM8 mutants, CEACAM8-N27, R28, I29 (indicated by 8-NRI),
CEACAM8-S32 (8-S32), or CEACAM8-Q44 (8-Q44), were mixed with
CHO cells expressing CEACAM6 (6) for 30 min. Cell aggregation was
determined by calculating single cells, which were expressed as
percentage of total cells. Homophilic adhesion of CEACAM8 (8; negative)
and CEACAM6 (positive) is also shown. See Figure 3B
for the locations
of mutation. Data are means ± SD of one
representative experiment performed in triplicate out of three similar
ones.
|
 View larger version (42K): [in a new window] |
Figure 8. Acquisition of adhesion activity by CEACAM3/6 after introduction of the
residues critical for the binding between CEACAM6 and CEACAM8. N27 and
I29 of CEACAM6 were introduced into CEACAM3/6, which consists of the
N-domain (residues 1–93) of CEACAM3 followed by residues 94–108 of
the CEACAM6 N-domain and the AB-domains of CEACAM6, yielding
CEACAM3(6)/6. N32 and Q44 of CEACAM8 were also introduced into
CEACAM3/6, yielding CEACAM3(8)/6. See Figure 3C
for the locations of
mutation. (A) Homophilic-adhesion activity of CHO cells expressing
CEACAM3(6)/6 and CEACAM3(8)/6 was examined by cell aggregation. Single
cells were expressed as percentage of total cells. (B) CHO cells
expressing CEACAM3(6)/6 and CEACAM3(8)/6 were labeled with H33342,
their attachment to the plates was coated with BSA, and CEACAM6 and
CEACAM8 proteins were assessed as described in Materials and Methods.
Data are means ± SD of one representative experiment
performed in triplicate out of two similar ones. *P <
0.05.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We next located the amino acid residues in the N-domains crucial for the homophilic adhesion of CEACAM6 and the heterophilic adhesion between CEACAM6 and CEACAM8 by homologue-scanning mutagenesis. It was shown that N27, S32, and L44 are key residues for the homophilic adhesion of CEACAM6, whereas N27 and I29, and to a lesser degree, L28 and S32, also are important for the heterophilic adhesion of CEACAM6 to CEACAM8. Conversely, in CEACAM8, N32 and R44 were essential for its heterophilic adhesion to CEACAM6. However, in contrast to the 27th–29th residues of CEACAM6, those of CEACAM8 did not appear important for its heterophilic-adhesion activity to CEACAM6. The region containing these 27th–29th residues in CEACAM8 is of a strong hydrophilic nature in contrast to the hydrophobic or weak hydrophilic nature of this region in the other CEACAM antigens. This might be partly associated with the lack of homophilic-adhesion activity for CEACAM8.
As demonstrated in this paper, substitution of one residue in CEACAM6 induced strong inhibition of its homophilic-adhesion activity. Introduction of each critical residue alone into CEACAM8 did not result in acquisition of homophilic-adhesion activity. However, as shown in Figure 8 , CEACAM3(6)/6, which had the four key residues N27, I29, S32, and L44 of CEACAM6, acquired homophilic- and heterophilic-adhesion activities to CEACAM6. The key residues located in different positions in the N-domains of CEACAM6 and CEACAM8 are thus equally prerequisites for their full adhesion activities.
The residues critical for the homophilic and heterophilic adhesion of these CEACAM antigens may be located on the potential GFCC'C'' face of their N-domains predicted from the three-dimensional model of CEA based on the structures of CD2, CD4, and REI [23 , 24 , 32 , 33 ]. The possible secondary structure predicted from the sequence, however, suggested that the N-domain of CEACAM6 lacks the C strand that is predicted to exist in CEA and CEACAM1, both of which display no adhesion activity to CEACAM8 [8 ]. The absence of the C strand might be related to the specific affinity of CEACAM6 to CEACAM8. As we have suggested in a previous paper, that the carbohydrate portion is not required for the adhesion between CEACAM6 and CEACAM8 [22 ], the N-linked carbohydrate chains possibly existing in the N-domains of CEACAM6 and CEACAM8 [22 , 26 ] are not located in the regions analyzed in the present study.
In addition to the adhesion among the family members, most of the CEACAM antigens bind Opa proteins of N. meningitides [14 ] and N. gonorrhoeae [15 , 16 ]. This is likely to be related to infectious activity of the bacteria or recognition of the bacteria by phagocytes that express the CEACAM antigens. The recent studies using site-directed mutagenesis demonstrated several amino acid residues critical for the binding activity of the CEACAM family antigens to the Opa proteins. Virji et al. [24 ] showed by alanine-scanning mutagenesis that the substitution of Y34 of CEACAM1 resulted in a complete loss of the binding of CEACAM1 to OpaA, OpaB, and OpaC and that differential effects against these three Opa variants were seen by alanine substitution of S32, V39, or Q44. The positions of two of these residues, S32 and Q44, are the same as those of the residues prerequisite for the homophilic adhesion of CEACAM6 (S32 and L44). In a similar study, Bos et al. [23 ] showed that crucial residues in the N-domain of CEA for the adhesion of OpaB, OpaC, and OpaI include F29 and S32, by using homologue-scanning mutagenesis between CEA and CEACAM8, which shows no affinity for any Opa proteins. It is interesting that the 29th and 32nd residues in CEACAM6 were essential for its heterophilic adhesion to CEACAM8, which lacks any adhesion activity except to CEACAM6. It is intriguing to speculate that the three-dimensional structures of the Opa proteins required for the binding to CEACAM6 may resemble that of CEACAM8.
The recent study on the mechanism of homophilic binding of CEA [33 ] demonstrated that three regions in the N-domain, the residues 30–35, 42–46, and 80–84, are important for the binding, similar to our present findings on the adhesion activity of CEACAM6 and CEACAM8. It remains to be elucidated, however, whether the residues critical for the adhesion activity of CEACAM6 to self or to CEACAM8 are also important for its heterophilic adhesion to CEA and CEACAM1. Our present study thus suggests that subtle variations on several restricted regions in the N-domains alter the specificity of the adhesion activity of the CEACAM antigens against self, the same family antigens, and the bacterial proteins.
Received March 7, 2001; revised June 4, 2001; accepted June 4, 2001.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  V. Kolla, L. W. Gonzales, N. A. Bailey, P. Wang, S. Angampalli, M. H. Godinez, M. Madesh, and P. L. Ballard Carcinoembryonic cell adhesion molecule 6 in human lung: regulated expression of a multifunctional type II cell protein Am J Physiol Lung Cell Mol Physiol, June 1, 2009; 296(6): L1019 - L1030. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Riley, K. P. Engelhardt, J. W. Saldanha, W. Qi, L. S. Cooke, Y. Zhu, S. T. Narayan, K. Shakalya, K. D. Croce, I. G. Georgiev, et al. Design and Activity of a Murine and Humanized Anti-CEACAM6 Single-Chain Variable Fragment in the Treatment of Pancreatic Cancer Cancer Res., March 1, 2009; 69(5): 1933 - 1940. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ozeki and J. E. Shively Differential cell fates induced by all-trans retinoic acid-treated HL-60 human leukemia cells J. Leukoc. Biol., September 1, 2008; 84(3): 769 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Tanaka, J. Huang, S. Hirai, M. Kuroki, M. Kuroki, N. Watanabe, K. Tomihara, K. Kato, and H. Hamada Carcinoembryonic Antigen-Targeted Selective Gene Therapy for Gastric Cancer through FZ33 Fiber-Modified Adenovirus Vectors. Clin. Cancer Res., June 15, 2006; 12(12): 3803 - 3813. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Stern, G. Markel, T. I. Arnon, R. Gruda, H. Wong, S. D. Gray-Owen, and O. Mandelboim Carcinoembryonic Antigen (CEA) Inhibits NK Killing via Interaction with CEA-Related Cell Adhesion Molecule 1 J. Immunol., June 1, 2005; 174(11): 6692 - 6701. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Markel, R. Gruda, H. Achdout, G. Katz, M. Nechama, R. S. Blumberg, R. Kammerer, W. Zimmermann, and O. Mandelboim The Critical Role of Residues 43R and 44Q of Carcinoembryonic Antigen Cell Adhesion Molecules-1 in the Protection from Killing by Human NK Cells J. Immunol., September 15, 2004; 173(6): 3732 - 3739. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Comegys, S.-H. Lin, D. Rand, D. Britt, D. Flanagan, H. Callanan, K. Brilliant, and D. C. Hixson Two Variable Regions in Carcinoembryonic Antigen-related Cell Adhesion Molecule1 N-terminal Domains Located in or Next to Monoclonal Antibody and Adhesion Epitopes Show Evidence of Recombination in Rat but Not in Human J. Biol. Chem., August 13, 2004; 279(33): 35063 - 35078. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Finkenzeller, B. Fischer, S. Lutz, H. Schrewe, T. Shimizu, and W. Zimmermann Carcinoembryonic Antigen-Related Cell Adhesion Molecule 10 Expressed Specifically Early in Pregnancy in the Decidua Is Dispensable for Normal Murine Development Mol. Cell. Biol., January 1, 2003; 23(1): 272 - 279. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
