Epithelial membrane protein 2 modulates infectivity of Chlamydia muridarum (MoPn)

Microbes and Infection 9 (2007) 1003e1010 www.elsevier.com/locate/micinf

Original article

Epithelial membrane protein 2 modulates infectivity of Chlamydia muridarum (MoPn) Kaori Shimazaki a,1, Madhuri Wadehra b,1, Ashley Forbes a, Ann M. Chan c, Lee Goodglick b, Kathleen A. Kelly b, Jonathan Braun a,b, Lynn K. Gordon d,e,* a

Molecular Biology Institute, Los Angeles, CA 90095, USA Department of Pathology and Laboratory, Los Angeles, CA 90095, USA c Department of Physiological Science, Los Angeles, CA 90095, USA d Ophthalmology, University of California, Los Angeles, CA 90095, USA e Department of Surgery, Greater Los Angeles Veterans Affairs Healthcare System, Los Angeles, CA 90099, USA b

Received 8 February 2007; accepted 5 April 2007 Available online 18 April 2007

Abstract Chlamydiae are bacterial pathogens which have evolved efficient strategies to enter, replicate, and survive inside host epithelial cells, resulting in acute and chronic diseases in humans and other animals. Several candidate molecules in the host receptor complex have been identified, but the precise mechanisms of infection have not been elucidated. Epithelial membrane protein-2 (EMP2), a 4-transmembrane protein, is highly expressed in epithelial cells in sites of chlamydial infections. Here we show that infectivity of the Chlamydia muridarum (MoPn) is associated with host cellular expression of EMP2 in multiple cell lines. Recombinant knockdown of EMP2 impairs infectivity, whereas infectivity is augmented in cells recombinantly modified to over-express EMP2. An epithelial cell line without native expression of EMP2 is relatively resistant to MoPn infection, whereas infectivity is markedly increased by recombinant expression of EMP2 in that cell line. Blockade of surface EMP2 using a specific anti-EMP2 antibody significantly reduces chlamydial infection efficiency. In addition, MoPn infectivity as measured in the EMP2 overexpressing cell line is not heparin-dependent, suggesting a possible role for EMP2 in the non-reversible phase of early infection. These findings identify EMP2 as a candidate host protein involved in infection of C. muridarum (MoPn). Ó 2007 Elsevier Masson SAS. All rights reserved. Keywords: Chlamydia; Epithelial membrane protein-2; Infection; Hostepathogen interaction

1. Introduction Chlamydiae are obligate gram-negative intracellular prokaryotic pathogens that cause significant human morbidity through infections of multiple organ systems (reviewed in [1,5]). Some serovars of C. trachomatis produce human * Corresponding author. Jules Stein Eye Institute, David Geffen School of Medicine at UCLA, 100 Stein Plaza, Los Angeles, CA 90095, USA. Tel.: þ1 310 206 4803; fax: þ1 310 825 5674. E-mail address: [email protected] (L.K. Gordon). 1 These authors contributed equally to this work. 1286-4579/$ - see front matter Ó 2007 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.micinf.2007.04.004

genitourinary tract infections which are a significant cause of infertility, ectopic pregnancy, and chronic pelvic pain syndromes [13]. Trachoma, the most common cause of preventable infectious blindness, is caused by other specific serovars of C. trachomatis [24]. Defining specific hostepathogen interactions and identification of host cellular proteins involved in pathogenesis of these infections may reveal new potential strategies for disease control. Although all Chlamydia species seem to share a common life cycle, identification of interacting host receptors has been challenging due to the diversity of host entry mechanisms between differing Chlamydia serovars and multiple host cell

1004

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

types [2]. Successful host entry of Chlamydia requires an initial attachment phase that consists of reversible electrostatic interactions between chlamydial ligands and host cell surface molecules followed by specific interactions with putative cell surface receptors [3,4]. Glycosaminoglycans variably contribute to binding between Chlamydiae and host cells and may mask participation of other ligands [3,15,25,26], further complicating the study of chlamydial entry. Host heparin sulfate is implicated in the entry of the C. muridarum (MoPn) in HeLa cells [14] and CHO-K1 cells [25], but not the related C. trachomatis serovars B or E in McCoy, HeLa 229, HEC1B, or CHO-K1 cells [15]. Consistent with these findings, the host heparin sulfate dependent C. trachomatis serovars showed a significantly reduced infectivity in host surface heparin sulfate deficient mutant CHO-K1 cells [3,15,25]. Several chlamydial or host cell ligands are also shown to be involved in entry of Chlamydiae and include the major outer membrane protein (MOMP) [14], chlamydial glycans [2], OmcB [16] and heat shock protein [11], mannose-6-phosphate receptor [10], estrogen receptor beta [7], low density lipoprotein (LDL) receptor and LDL receptor-related protein (LRP) [6,8]. Overall, these findings indicate that the initial hoste pathogen interaction requires a well-coordinated interplay of multiple molecules on both the microbial and host cells. Epithelial membrane protein-2 (EMP2), a member of GAS3/PMP22 tetraspan protein family, is expressed at high levels in epithelial cells of the lung, eye, and genitourinary tracts [22,23]. EMP2 controls and coordinates cell surface trafficking and function of certain integrins, GPI-linked proteins, and class I MHC molecules, and reciprocally regulates caveolin expression [18e21]. One physiological role of EMP2 is in binding and penetration of endometrium by the blastocyst during blastocyst-endometrial implantation. Specifically, endometrial cell expression of EMP2 is required for endometrial display and function of avb3 integrin involved in blastocyst interaction. [17,18]. Thus, expression of EMP2 in epithelial cells can play an important role in the trafficking and organization of multipartite protein complexes that are critical for cellecell interaction. Accordingly, we wondered whether EMP2 might similarly play a role for entry of Chlamydia into host cells. Here we show that infectivity of MoPn correlates positively with the levels of EMP2 expression in multiple epithelial cell lines. Although MoPn is not a natural human pathogen, it is known to infect human epithelial cell lines from diverse tissues of origin [12]. Concordantly, blockade of the large extracellular loop of EMP2 using a specific anti-EMP2 antibody significantly decreases Chlamydia infectivity. Importantly, an EMP2 deficient cell line was relatively resistant to infectivity, and recombinant expression of EMP2 in the same cell line restored infectivity to comparable levels observed in other epithelial cell lines. Finally, temperature and heparin manipulations identified the interactions between MoPn and EMP2 as heparin-independent and irreversible early in the Chlamydia infection process. These findings suggest that EMP2 is a newly identified host molecule involved in hosteChlamydia interaction.

2. Materials and methods 2.1. Cell lines and Chlamydiae The human endometrial adenocarcinoma cell line HEC1A (HTB112, ATCC, Manassas, VA), human conjunctival cell line HCJE (gift of Dr Ilene K. Gipson at Schepens Eye Research Institute and Department of Ophthalmology, Harvard Medical School, Boston, MA), and human breast cancer epithelial cell line Hs578T (gift of Dr John Colicelli at Department of Biological Chemistry, University of California, Los Angeles, CA) were cultured in appropriate media supplemented with 10% fetal calf serum at 37  C in a humidified 5% CO2 and passaged every 7 days. EMP2 expression levels in each cell line were determined by Western blot. In addition to HEC1A wild type cells (HEC1A-WT), EMP2-modulated HEC1A sublines, both over-expressing and knockdown, were stably transfected cell lines with expression plasmids for GFP, a human EMP2-GFP fusion protein, or a human EMP2-specific ribozyme (HEC1A-GFP, HEC1A-hEMP2, and HEC1A-hRZ2) [18]. Similarly, Hs578T wild type cells (Hs578T-WT) were stably infected with retroviral constructs for GFP or a human EMP2-GFP fusion protein expression (Hs578T-GFP and Hs578T-hEMP2). C. muridarum (MoPn) and a mixture of human serovars D-K (gift of Dr Harlan Caldwell; Laboratory of Intracellular Parasites, Rocky Mountain Laboratories, National Institutes of Health, Hamilton, MT) was purified, aliquoted, and stored at 80  C as previously described [9]. 2.2. Antibodies and peptides Antibodies against human EMP2 were produced by immunization of rabbits with EDIHDKNAKFYPVTREGSYG, a peptide in the second extracellular loop of human EMP2 [23]. In blocking experiments, the immunogenic peptide was used to specifically bind the antibody. The negative control peptide was an unrelated 20-mer peptide from the first extracellular loop of human EMP2. Antibody from the pre-immune rabbits was used as a negative control. For immunohistochemical detection of Chlamydia inclusions, an anti-Chlamydia LPS mouse antibody (clone EV1-H1) was kindly provided by Dr Harlan Caldwell [27]. FITC anti-mouse IgG and horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG antibody were from Southern Biotechnology Associates (Birmingham, AL). Rabbit anti-human b-actin was from Sigma (St. Louis, MO). 2.3. Chlamydial infection HEC1A, HCJE, and Hs578T cells were plated at a concentration of 2.5  105 cells/ml and incubated overnight to establish monolayers. Infection with MoPn was performed as previously reported and quantified after an incubation of 24 h [9]. Cells were fixed in cold methanol and inclusion bodies were identified immunohistochemically using mouse antiChlamydia LPS (EV1-H1) and FITC anti-mouse IgG

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

1005

recombinant expression of EMP2 or EMP2 expression in the HEC1A-hRZ2 knock down cell line or the control HEC1AGFP cell line is expressed as a percentage of native EMP2 expression in control, untransfected HEC1A cells. Experiments were repeated at least three times.

secondary antibody. Cells were counter stained with Evans Blue, mounted in glycerol, and scored using fluorescence microscopy. For the antibody blocking study, cells were incubated with antibody for 1 h at 37  C before the infection step. For peptide blocking, antibody was mixed with peptide at indicated concentrations for 1 h at room temperature prior to addition to cell cultures. Heparin competition experiments were performed using a modified protocol described by Carabeo et al. [3]. Briefly, cells were incubated at 4  C or 37  C for 1 h immediately after the infection and rinsed three times with ice-cold Hanks balanced salt solution (HBSS) (Invitrogen, Carlsbad, CA) or heparin-HBSS with various heparin concentrations (0, 10, 150, and 250 mg/ml) (Sigma). Heparin-HBSS was removed and replaced by media containing cyclohexamide according to previously published protocol [9]. After an additional 24 h incubation at 37  C, cells were fixed, stained, and inclusions were counted as described in the previous section.

For the anti-EMP2 antibody and EMP2 peptide studies, groups were analyzed by two-tailed Student’s paired t-test, with a confidence level of p  0.05. The statistical significance of infection rate on stably transfected cells was tested using two-tailed two-sample equal variance t-test with a confidence level of p  0.05. Linear correlation between EMP2 expression levels and Chlamydia infectivity was calculated using GraphPad Prism Program (GraphPad, San Diego, CA). Linear regression analysis was used to obtain the correlation coefficient (R2).

2.4. Fluorescence microscopy

3. Results

Chlamydia inclusions were identified with an epiillumination fluorescent microscope (Olympus, Melville, NY) using a FITC filter. Chlamydia inclusions were defined by round, regular shape, with a diameter of approximately 1/3 of cell size. Inclusions were scored in a masked fashion by at least two independent observers and the percent of infected cells was calculated using the number of cells with inclusions (C1) and the number of cells without inclusions (C0) (C1/ (C1 þ C0)  100). Nuclei were visualized using DAPI (Invitrogen). Experiments were performed with 2e3 replicate samples, and repeated at least three times.

The tissue (lung, eye, and genitourinary tract) localization of EMP2 and its role in protein trafficking and coordination of protein complexes [17e22] raised the possibility that EMP2 might directly or indirectly affect chlamydial infectivity. To begin testing this hypothesis, the ability of MoPn to infect endometrial cell lines, in which EMP2 expression was recombinantly modified, was evaluated. HEC1A cell lines were stably transfected with expression plasmids to overexpress EMP2 (HEC1A-hEMP2), reduce expression of native EMP2 via an EMP2-specific ribozyme (HEC1A-hRZ2), or serve as a control GFP transfectant (HEC1A-GFP) (Fig. 1a). Quantitation of Western blots showed that EMP2 levels in these over-expressing and ribozyme HEC1A sublines were respectively 2.5-fold and 0.5-fold compared to GFP control cells (Fig. 1a). In order to determine if EMP2 levels were associated with infectivity, experiments were initially performed to determine the multiplicity of infection (MOI) of MoPn required to reproducibly infect about 60% of the wild type HEC1A cells (data not shown). This level of infection would then permit experimental observations of either an increase, decrease, or no change in infectivity as correlated with EMP2 expression. Based on this criteria, MoPn at an MOI of 2 was selected for use in subsequent experiments with the three HEC1A sublines (Fig. 1b), and chlamydial infection was quantitated as chlamydial inclusions, expressed as ‘‘infection efficiency’’, percent inclusions relative to the HEC1A-GFP control cells. Nearly 100% of the HEC1A-hEMP2 (EMP2 over-expressing cells) were infected by MoPn at an MOI of 2, indicating a significant increase in infectivity as compared to control HEC1AGFP cells (>60% increase, p < 0.005). Reciprocally, only about 30% of the HEC1A-hRZ2 were infected by MoPn at an MOI of 2, indicating that these cells were more resistant to infection than the control HEC1A cells (>40% decrease, p < 0.0001). A similar relationship was observed between levels of EMP2 expression and chlamydial infectivity in

2.5. Western immunoblots Cellular lysates in Laemmli buffer were treated with peptide-N-glycosidase F (PNGase; New England Biolabs, Beverly, MA) to remove N-linked glycans to convert the heterogeneously glycosylated EMP2 protein into a single w20 kDa species [23]. Protein concentrations were determined using the BCA protein assay kit according to the manufacturer’s directions (Pierce, Rockford, IL). A total of 25 mg of protein was loaded per lane and was separated by SDSPAGE (Invitrogen) as previously described [22,23]. Blots were probed with anti-EMP2 (1:1000 dilution) or, as an internal control, with anti-b-actin (1:5000 dilution, Sigma), followed by incubation with a horseradish peroxidaseconjugated anti-rabbit or anti-mouse IgG antibody (1:2000 dilution for each antibody, Southern Biotechnology Associates). Proteins were visualized by chemiluminescence (ECL; Amersham Biosciences, Piscataway, NJ). Negative controls (secondary antibodies alone) produced no signal. ECL films (Amersham Biosciences) were scanned, and band intensities of the native EMP2, recombinant EMP2, and b-actin were calculated for each lane using the Image J program (National Institute of Health, Bethesda, MD). EMP2 bands, either native or recombinant, were normalized to the b-actin band and the

2.6. Statistical analysis

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

1006

(a) 300

Relative Infection Efficiency (%)

***

250 200 150

***

100

***

***

80

Control

60 Anti-EMP2

40 20

0

1

2

3

4

5

6

Antibody Concentration (%) 50

(b) 0 HEC1A-GFP

HEC1A-hEMP2

HEC1A-hRZ2

***

(b) Relative Infection Efficiency (%)

120

0

100

***

200

Relative Infection Efficiency (%)

Relative Expression of EMP2 (%)

(a)

***

120

***

100

*** EMP2 Peptide

80 60

Control Peptide 40 20

160

0

120

0

20

40

60

80

100

120

Peptide (nM) 80 40 0 HEC1A-GFP

HEC1A-hEMP2

HEC1A-hRZ2

Fig. 1. EMP2 expression levels positively correlate with MoPn infection efficiency. (a) EMP2 levels were compared by anti-EMP2 Western immunoblot in HEC1A cells stably transfected with plasmids for expression of a human EMP2-GFP fusion protein (HEC1A-hEMP2), GFP (HEC1A-GFP), or a EMP2-specific ribozyme (HEC1A-hRZ2). Relative expression level of EMP2 was calculated relative to the b-actin control and compared between sublines. Data is presented from 3 independent analyses. (b) Cells were infected with C. muridarum (MoPn) serovar at an MOI of 2, and Chlamydia inclusions were identified and the percentage of infected cells calculated. About 60% of the HEC1A-GFP cells had inclusions and, as the control standard, the mean value of 3 replicates in 5 independent experiments was identified as an infection efficiency of 100%. The percent of infected cells was calculated for the HEC1A-hEMP2 and HEC1A-hRZ2 cell lines and the relative infection efficiency (as a percent of HEC1A-GFP) calculated. Statistical analysis of the data was performed using a Student’s t-test (***p < 0.005). Data in are presented from 5 independent experiments, and each experiment had at least three replicate groups.

genetically modified HEC1A cell lines using a serovar mixture (human serovars D-K, gift of Dr Harlan Caldwell; data not shown). These results suggest a potential role, either direct or indirect, for EMP2 in chlamydial infectivity. To examine whether EMP2 was directly related to MoPn infectivity we tested whether specific anti-EMP2 antibodies could block MoPn infectivity. Either the specific anti-EMP2 antibody or a control antibody was added to wild type HEC1A cell cultures prior to incubation with MoPn (MOI of 2). The resultant infection was then quantified as percent inclusions relative to HEC1A cells that had not been exposed to any antibody (Fig. 2a). In contrast to the control antibody,

Fig. 2. Anti-EMP2 antibody inhibits chlamydial infection. (a) Effect of antiEMP2 antibody. HEC1A were infected with MoPn at an MOI of 2 in the presence of indicated concentrations of anti-EMP2 or control pre-immune antibody. Chlamydial infection efficiency (% Chlamydia inclusions compared to untreated cells) was determined by immunostaining (mean þ SEM), and compared at each antibody concentration by Student’s t-test. (b) Effect of EMP2 peptide on anti-EMP2 inhibition. Anti-EMP2 (5%) was pretreated with indicated concentrations of specific EMP2 (second extracellular loop) peptide or control peptide (non-immunogenic EMP2 peptide, 1st extracellular loop), and then co-incubated with cells during MoPn infection. Infection efficiency was normalized to chlamydial inclusions in cells without peptide treatment, and compared at each concentration of EMP2 or control peptide. Results are representative of 2 or more independent experiments. ***p < 0.05.

the anti-EMP2 antibody produced a significant dose-dependent inhibition of infectivity and the resultant infection efficiency was less than 50% of wild type HEC1A cells. To confirm specificity for EMP2, the anti-EMP2 antibody was pre-incubated with a peptide prior to exposure to the HEC1A cells. The anti-EMP2 antibody was pre-incubated with either the immunogenic EMP2 peptide used to create the antibody (EMP2, 2nd extracellular loop) or a control EMP2 peptide that was not used to generate the antibody (EMP 2, 1st extracellular loop) (Fig. 2b). Pre-incubation of anti-EMP2 antibody with the specific EMP2 peptide neutralized the blocking effect of the anti-EMP2 antibody, resulting in significant restoration of MoPn infectivity. In contrast, exposure of the anti-EMP2 antibody to the control peptide did not significantly restore MoPn infectivity. Thus, the antiEMP2 antibody effect on MoPn infectivity reflected its specificity for EMP2, providing supporting evidence for a possible role for EMP2 in the cell surface host receptor complex for MoPn infection. We wanted to test whether the relationship between EMP2 and MoPn infectivity was unique to the HEC1A cell line or

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

***

(a)

Relative Expression of EMP2 (%)

800

600

400

200

0 HEC1A-WT

Relative Infection Efficiency (%)

(b)

120

HCJE

HEC1A-WT

***

100 80 60 40 20 0 No Ab

(c)

120

anti-EMP2 Ab

Control Ab

HCJE

***

100

Relative Infection Efficiency (%)

whether EMP2 played a role in MoPn infectivity in other cell lines. In order to further test the relationship between EMP2 and MoPn infectivity, we chose two additional cell lines for investigation. One cell line exhibited native EMP2 expression and the effect of EMP2 blockade on MoPn infectivity was tested using the specific anti-EMP2 antibody. The second cell line lacks any detectable EMP2 expression and was tested for its ability to be infected by MoPn and the effect of recombinantly produced EMP2 expression on MoPn infectivity. There is a linear correlation between EMP2 expression levels and infectivity of MoPn in the HS578T and HEC1A cell lines (R2 ¼ 0.96; Fig. 4c). A human conjunctival cell line, HCJE, was tested for EMP2 protein expression in comparison to the wild type HEC1A (HEC1A-WT) cell line. Quantitative analysis of EMP2 expression levels in these two cell lines was performed using Western blot and revealed high EMP2 expression levels in the HCJE cell line (Fig. 3a). In comparison to HEC1A-WT, HCJE cells had about a 7.1-fold increase in EMP2 expression. Infection studies with MoPn (MOI of 2) showed higher baseline infectivity in the HCJE ocular cell lines (about 100% infection) as compared to the HEC1A-WT cell line (about 60% infection) (data not shown). Consistent with other results (Fig. 2a), pre-incubation of HEC1A-WT cells with the antiEMP2 antibody resulted in a dramatic decrease in MoPn infectivity (Fig.3b; p < 0.0005). Similarly, pre-incubation of the HCJE cells with anti-EMP2 antibody reduced infection efficiency (Fig. 3c; p < 0.0005), although the level of reduction was less than observed using the HEC1A cell line. These findings showed that inhibition of infection using an antibody against EMP2 was a feature of two independent cell lines from different tissues of origin. The relationship between EMP2 and MoPn infectivity was further tested using a wild type cell line Hs578T (Hs578TWT), a cell line that lacks detectable native EMP2 expression. Infectivity in the Hs578T-WT cell line was compared to infectivity in a stable subline that expressed EMP2 through a retroviral expression vector, HS587t-hEMP2; and, a negative vector control cell line, Hs578T-GFP. EMP2 expression in each cell line was determined by Western blot. As shown in Fig. 4a, in comparison to an internal b-actin control, EMP2 expression was undetectable in both Hs578T-WT and Hs578T-GFP cell lines, whereas EMP2 was expressed at about 40% of the level of b-actin in the HS578-hEMP2 cell line. Cell lines were infected with MoPn at multiple MOIs ranging from 0.5 to 5.0, and the percent of infected cells was quantified as described above. At an MOI of 3, 80% of HEC1A-WT cells were infected with MoPn, and 100% infection was observed at an MOI of 5 (data not shown). In contrast, MoPn infectivity was dramatically reduced in both Hs578T-WT (1.4%, 8.6%, and 18.9% at MOIs of 0.5, 3, and 5, respectively) and Hs578T-GFP (1.1%, 5.0%, and 19.3% at MOIs of 0.5, 3, and 5, respectively) cell lines (Fig. 4b). Notably, infectivity was significantly increased in the EMP2 expressing Hs578ThEMP2 cells (2.3%, 17.3%, and 52.5% at MOI of 0.5, 3, and 5, respectively), further implicating EMP2 as a candidate host protein in MoPn infectivity.

1007

80 60 40 20 0 No Ab

anti-EMP2 Ab

Control Ab

Fig. 3. Anti-EMP2 antibody inhibits chlamydial infection in an ocular cell line HCJE. (a) EMP2 protein expression levels in HEC1A and HCJE were identified using Western immunoblot. Relative expression level of EMP2 was calculated relative to the b-actin control and compared between HEC1A and HCJE cell lines. EMP2 expression of HEC1A was normalized to 100%. (b) HCJE cells were infected with MoPn at an MOI of 2, with or without the prior incubation with anti-EMP2 antibody or control antibody. Chlamydial infection efficiency was calculated (mean þ SEM). % infection efficiency of antiEMP2 antibody-treated cells was compared to that of control antibody-treated samples using Student’s t-test (***p < 0.0005). Data are compiled from 3 independent experiments, and each experiment had at least three replicate groups.

These findings with three independent cell lines indicate that the ability of MoPn to infect epithelial cells is modulated by EMP2 expression. While EMP2 might affect various stages of the infection process, one potential stage is at the time of initial encounter of the elementary body (EB) with the host cell. Successful Chlamydia internalization requires an initial attachment of Chlamydia EB through a reversible electrostatic interaction followed by a temperature sensitive irreversible binding step to candidate host receptors which initiate internalization of Chlamydia [3]. When EBs and host cells are

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

1008

Relative Expression of EMP2 (%)

(a)

50 40 30 20

▼

10

▼

0

Hs578T-WT 60

***

Hs578T-WT Hs578T-GFP Hs578T-EMP2

50

Hs578T-GFP

40

(a) ***

20 10

***

0 0

1

2

3

4

5

6

MOI

(c)

80

HEC1A-GFP 37 °C 4 °C

30

Relative Infection Efficiency (%)

Relative Infection Efficiency (%)

(b)

Hs578T-hEMP2

HEC1A cell lines which were modified for EMP2 expression levels. The different cell lines were incubated with MoPn EBs for 1 hour at either 4  C or at 37  C, competitive heparin washes were performed at various concentrations ranging from 0 to 250 mg/ml, and infection efficiency was quantified after 24 h of additional incubation at 35  C (modified protocol described by Carabeo et al. [3]). Consistent with our previous studies and in comparison to HEC1A-GFP controls, Chlamydia infectivity was greatly increased by about 2-fold in HEC1A-hEMP2 cells, and reduced by 2-fold in HEC1A-hRZ2 (Fig. 5). When incubated at 37  C, infectivity of each tested cell line was universally resistant to heparin; consistent with the previously reported irreversible

60

40

20

0

120

0

100

200

300

Heparin (µg/ml)

80

40 R 2 = 0.96

20 0 0

50

100

150

200

250

Relative EMP2 Expression (%) Fig. 4. EMP2 expression is required for chlamydial infection. (a) EMP2 expression levels were determined by anti-EMP2 Western immunoblot in Hs578T (Hs578T-WT), stably transfected cells which over-express EMP2 (Hs578T-hEMP2), and GFP only control plasmids (Hs578T-GFP). Relative expression level of EMP2 was calculated relative to the b-actin control and compared between sublines. In Hs578T-WT and Hs578T-GFP cells, EMP2 expression was undetectable (arrowheads). (b) Cells were infected with MoPn at an MOI of 0.5, 3.0, and 5.0, and Chlamydial infection efficiency was calculated (mean þ SEM) as described in Section 2. Data from Hs578T-hEMP2 were compared by Student’s t-test to Hs578T-GFP. Data is representative for three or more experiments. ***p < 0.0001. (c) Mean infectivity of MoPn in the Hs578T and HEC1A cell lines were plotted against relative expression levels of EMP2, wild type EMP2 expression in the HEC1A cell line was set at 100% and EMP2 expression in the genetically modified cell lines was adjusted relative to the wild type levels. Correlation coefficient (R2) was calculated using linear regression analysis as described in Section 2.

Relative Infection Efficiency (%)

(b)

60

80

HEC1A-hEMP2

37 °C 4 °C

60

40

20

0 0

100

200

300

Heparin (µg/ml)

(c) Relative Infection Efficiency (%)

Infectivity (%)

100

80

HEC1A-hRZ2 37 °C 4 °C

60

40

20

0 0

100

200

300

Heparin (µg/ml)

incubated at 4  C, the reversible, electrostatic interaction stage is prolonged and exogenous heparin is able to dislodge the EBs, decreasing infectivity. In contrast, the EBehost cell interaction is not sensitive to exogenous heparin following the rapid irreversible binding and internalization that occurs at 37  C [3]. Thus, we investigated the contribution of heparin in the initial Chlamydiaehost interaction by using the

Fig. 5. EMP2 mediates heparin-independent Chlamydiaehost interaction. (a) HEC1A-GFP, (b) HEC1A-hEMP2, and (c) HEC1A-hRZ2 cells were incubated at 4  C and 37  C for 1 hour after infection with MoPn. Cells were washed with HBSS or heparin-HBSS with a various heparin concentrations and incubated at 35  C for 24 h. Infection efficiency was calculated for each condition (mean þ SEM) as described in Section 2. Data from cells washed with heparin-HBSS were compared by Student’s t-test to those washed with HBSS only. Data is representative of three or more experiments. ***p < 0.05.

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

binding step at 37  C incubation [3]. Concordantly, incubation of the HEC1A-GFP cell line at 4  C exhibited reduced infectivity after exposure to exogenous heparin in a dose-dependent response. When incubated at 4  C the effect of heparin washes on infectivity correlated with the levels of expression of EMP2. Higher concentrations of heparin (up to 250 mg/ml), which almost completely suppressed infectivity in the HEC1A-GFP and HEC1A-hRZ2 cell lines (Fig. 5a,c; 4  C line graphs), only minimally reduced infectivity in HEC1AhEMP2 cells (Fig. 5b; 4  C line graph). These data indicate that EMP2 over-expression helps overcome sensitivity to heparin during the initial early hostepathogen interactions. Accordingly, the results support a separate role for EMP2 in the binding step of the MoPn EB to the host cell. 4. Discussion This study supports a role for EMP2 as a newly identified host protein in the pathogenesis of C. muridarum (MoPn) infectivity. Here, we present several lines of data to support the role of EMP2 in increased MoPn infectivity in host cells of epithelial origin. Recombinant over-expression of EMP2 was associated with increased infectivity, whereas recombinantly decreasing EMP2 expression reduced infectivity. Blockade of EMP2 using a specific antibody against the 2nd extracellular loop of EMP2 decreased infection efficiency of MoPn in two independent cell lines. As compared to HEC1A cells, MoPn infectivity was significantly decreased in an EMP2-deficient epithelial cell line, but infectivity was largely restored by recombinant over-expression of EMP2. Finally, at low temperatures exogenous heparin produced a dosedependent infectivity reduction in cell lines with low EMP2, but had a minimal effect in cells with a high level of EMP2 expression. These lines of evidence support a role for EMP2 as a candidate host molecule involved in a heparin-independent, early hosteChlamydia interaction. In this discussion, we address the unifying idea that EMP2 influences infection through its role in the attachment step, perhaps by affecting the formation of the host receptor complex for the Chlamydia EB. Our studies demonstrate that knockdown of EMP2 expression, or blocking with anti-EMP2 antibody, only partially inhibits Chlamydia infection. With regard to EMP2 expression, a 2e2.5-fold change in EMP2 levels in the HEC1A cell line resulted in 2e5-fold change in infection or attachment. Thus, the partial change in attachment and infection may be fully explained by the quantitatively concordant change in EMP2 levels. With regard to antibody treatment, blocking saturated at about 50% infection. This finding might indicate that antibody binding to the second extracellular domain only partially blocked EMP2 function. Alternatively, the EMP2 expression and antibody findings may also be explained, as others have suggested, that Chlamydia infection may be dependent on multiple receptoreligand pairs and not only expression of EMP2. In the anti-EMP2 antibody blockage experiments, a greater effect was seen in the HEC1A-WT cells as compared to the

1009

HCJE cells in terms of MoPn infection. It is possible that the effect of antibody blockade of in the ocular cell line is limited by its significantly high expression of EMP2 (about 7-fold increase compared to HEC1A-WT). It would be interesting to test whether a knock-down of EMP2 proteins can further decrease Chlamydia infectivity in this HCJE cell line. Infection studies with EMP2-deficient Hs578T cells showed up to a 10-fold decrease in infection efficiency compared to other cell lines such as HEC1A and HCJE at a similar MOI. Although the infectivity of different cell lines is not necessarily expected to be concordant, the significant increase in infectivity of this cell line when modified recombinantly to express EMP2 helps confirms its identity as a potentially important host receptor. However, the loss of infectivity in the EMP2-deficient Hs578T cells also suggests that this dramatic but incomplete reduction in infectivity suggests that EMP2 is a critical protein involved in MoPn host interaction, but EMP2 is not absolutely required for the infection process. The low level of infection seen in the wild type Hs578T cells was perhaps mediated by an alternate Chlamydia invasion pathway. Recently, EMP2 expression levels in HEC1A were shown to be limiting for surface expression of a necessary component of the endometrial receptor complex for blastocyst implantation [18]. There is a striking analogy between the role of EMP2 in blastocyst attachment and implantation, and its potential role in Chlamydia attachment during cellular invasion with resultant infection. The evidence that EMP2 controls surface expression of physiologically important proteins in the endometrial epithelium supports the idea that it may play a similar role for proteins involved in the Chlamydia entry process. Acknowledgments This work was supported by NIH HD48540 (JB), CA009120 (MW), GM007185 (AF), AI26328 (KK), 2-T32AI-07323 (KS) and CA016042 (UCLA Jonsson Comprehensive Cancer Center flow cytometry core); the Lalor Foundation (MW), and the Giannini Family Foundation (MW), and the Oppenheimer Family Foundation Grant Center for the Prevention of Eye Disease (LKG). We appreciate the helpful comments of Dr Carmen Williams. We are grateful to Dr Ilene K. Gipson, Dr John Colicelli, and Dr Harlan Caldwell for gifts of reagents (HCJE cell lines, Hs578T cell line, a mixture of Chlamydia trachomatis serovars D-K, and anti-Chlamydia LPS mouse antibody, respectively). We also thank Li Zhang for making Hs578T recombinants used in this study. References [1] R.C. Brunham, J. Rey-Ladino, Immunology of Chlamydia infection: implications for a Chlamydia trachomatis vaccine, Nat. Rev. Immunol. 5 (2005) 149e161. [2] L.A. Campbell, C. Kuo, Horizon bioscience, interactions of Chlamydia with the host cells that mediate attachment and uptake, in: P.M. Bavoil, P.B. Wyrick (Eds.), Chlamydia, Genomics and Pathogenesis, Norfolk, UK, 2006, pp. 505e522.

1010

K. Shimazaki et al. / Microbes and Infection 9 (2007) 1003e1010

[3] R.A. Carabeo, T. Hackstadt, Isolation and characterization of a mutant Chinese hamster ovary cell line that is resistant to Chlamydia trachomatis infection at a novel step in the attachment process, Infect. Immun. 69 (2001) 5899e5904. [4] A. Dautry-Varsat, M.E. Balana, B. Wyplosz, Chlamydia-host cell interactions: recent advances on bacterial entry and intracellular development, Traffic 5 (2004) 561e570. [5] K.A. Fields, T. Hackstadt, The chlamydial inclusion: escape from the endocytic pathway, Annu. Rev. Cell Dev. Biol. 18 (2002) 221e245. [6] Gerard, H.C., Whittum-Hudson, J.A., Hudson, A.P., C. pneumoniae utilizes APOe and the LDL receptor family for host cell attachment, in: Chernesky, M, Caldwell, H., Christiansen, G., Clarke, I.N., Kaltenboeck, B., Knirsch, C., Kuo, C.C., Mahony, J., Rank, R.G., Saikku, P., Schachter, J., Stamm, W.E., Stephens, R.S., Summersgill, J.T., Timms, P., Wyrick, P.B. (Eds.), Chlamydial Infections: Proceedings of the 11th International Symposium on Human Chlamydial Infections, San Francisco, 2006, pp. 153e156. [7] N.V. Guseva, S.T. Knight, J.D. Whittimore, P.B. Wyrick, Primary cultures of female swine genital epithelial cells in vitro: a new approach for the study of hormonal modulation of Chlamydia infection, Infect. Immun. 71 (2003) 4700e4710. [8] Mahony, J., Tu, X., Blonde, K., Brown, A., Uy, P., Do Chlamydiae use Apolipoprotein E and LDL receptor family to infect epithelial cells?, in: Chernesky, M, Caldwell, H., Christiansen, G., Clarke, I.N., Kaltenboeck, B., Knirsch, C., Kuo, C.C., Mahony, J., Rank, R.G., Saikku, P., Schachter, J., Stamm, W.E., Stephens, R.S., Summersgill, J.T., Timms, P., Wyrick, P.B. (Eds.), Chlamydial Infections: Proceedings of the 11th International Symposium on Human Chlamydial Infections, San Francisco, 2006, pp. 157e160. [9] H.K. Maxion, W. Liu, M.H. Chang, K.A. Kelly, The infecting dose of Chlamydia muridarum modulates the innate immune response and ascending infection, Infect. Immun. 72 (2004) 6330e6340. [10] M. Puolakkainen, C.C. Kuo, L.A. Campbell, Chlamydia pneumoniae uses the mannose 6-phosphate/insulin-like growth factor 2 receptor for infection of endothelial cells, Infect. Immun. 73 (2005) 4620e4625. [11] J.E. Raulston, C.H. Davis, D.H. Schmiel, M.W. Morgan, P.B. Wyrick, Molecular characterization and outer membrane association of a Chlamydia trachomatis protein related to the hsp70 family of proteins, J. Biol. Chem. 268 (1993) 23139e23147. [12] C. Roshick, H. Wood, H.D. Caldwell, G. McClarty, Comparison of gamma interferon-mediated antichlamydial defense mechanisms in human and mouse cells, Infect. Immun. 74 (2006) 225e238. [13] W.E. Stamm, M.E. Guinan, C. Johnson, T. Starcher, K.K. Holmes, W.M. McCormack, Effect of treatment regimens for Neisseria gonorrhoeae on simultaneous infection with Chlamydia trachomatis, N. Engl. J. Med 310 (1984) 545e549.

[14] H. Su, L. Raymond, D.D. Rockey, E. Fischer, T. Hackstadt, H.D. Caldwell, A recombinant Chlamydia trachomatis major outer membrane protein binds to heparan sulfate receptors on epithelial cells, Proc. Natl. Acad. Sci. USA 93 (1996) 11143e11148. [15] M. Taraktchoglou, A.A. Pacey, J.E. Turnbull, A. Eley, Infectivity of Chlamydia trachomatis serovar LGV but not E is dependent on host cell heparan sulfate, Infect. Immun. 69 (2001) 968e976. [16] L.M. Ting, R.C. Hsia, C.G. Haidaris, P.M. Bavoil, Interaction of outer envelope proteins of Chlamydia psittaci GPIC with the HeLa cell surface, Infect. Immun. 63 (1995) 3600e3608. [17] M. Wadehra, M. Dayal, M. Mainigi, T. Ord, R. Iyer, J. Braun, C.J. Williams, Knockdown of the tetraspan protein epithelial membrane protein-2 inhibits implantation in the mouse, Dev. Biol. 292 (2006) 430e441. [18] M. Wadehra, A. Forbes, N. Pushkarna, L. Goodglick, L.K. Gordon, C.J. Williams, J. Braun, Epithelial membrane protein-2 regulates surface expression of alphavbeta3 integrin in the endometrium, Dev. Biol. 287 (2005) 336e345. [19] M. Wadehra, L. Goodglick, J. Braun, The tetraspan protein EMP2 modulates the surface expression of caveolins and glycosylphosphatidyl inositol-linked proteins, Mol. Biol. Cell. 15 (2004) 2073e2083. [20] M. Wadehra, R. Iyer, L. Goodglick, J. Braun, The tetraspan protein epithelial membrane protein-2 interacts with beta1 integrins and regulates adhesion, J. Biol. Chem. 277 (2002) 41094e41100. [21] M. Wadehra, H. Su, L.K. Gordon, L. Goodglick, J. Braun, The tetraspan protein EMP2 increases surface expression of class I major histocompatibility complex proteins and susceptibility to CTL-mediated cell death, Clin. Immunol. 107 (2003) 129e136. [22] M. Wadehra, G.G. Sulur, J. Braun, L.K. Gordon, L. Goodglick, Epithelial membrane protein-2 is expressed in discrete anatomical regions of the eye, Exp. Mol. Pathol. 74 (2003) 106e112. [23] C.X. Wang, M. Wadehra, B.C. Fisk, L. Goodglick, J. Braun, Epithelial membrane protein 2, a 4-transmembrane protein that suppresses B-cell lymphoma tumorigenicity, Blood 97 (2001) 3890e3895. [24] S.K. West, Trachoma: new assault on an ancient disease, Prog. Retin. Eye Res. 23 (2004) 381e401. [25] F.N. Wuppermann, J.H. Hegemann, C.A. Jantos, Heparan sulfate-like glycosaminoglycan is a cellular receptor for Chlamydia pneumoniae, J. Infect. Dis. 184 (2001) 181e187. [26] J.P. Zhang, R.S. Stephens, Mechanism of C. trachomatis attachment to eukaryotic host cells, Cell 69 (1992) 861e869. [27] Y.X. Zhang, S. Stewart, T. Joseph, H.R. Taylor, H.D. Caldwell, Protective monoclonal antibodies recognize epitopes located on the major outer membrane protein of Chlamydia trachomatis, J. Immunol. 138 (1987) 575e581.