Chlamydia trachomatis enters a viable but non-cultivable (persistent) state within herpes simplex virus type 2 (HSV-2) co-infected host cells


*E-mail; Tel. (+1) 423 439 6295; Fax (+1) 423 439 8044.


Epidemiological and clinical studies have shown that double infection with herpes simplex virus type 2 (HSV-2) and Chlamydia trachomatis occurs in vivo. We hypothesized that co-infection would alter replication of these agents. To test this hypothesis, HeLa cells were infected with C. trachomatis serovar E, followed 24 h later by HSV-2 strain 333. Transmission electron microscopic (TEM) analyses indicated that, by 10 h after HSV addition, reticulate bodies (RBs) in co-infected cells were swollen, aberrantly shaped and electron-lucent. In infectious titre assays, HSV-2 co-infection abrogated production of infectious chlamydial progeny. Western blot analyses indicated that accumulation of chlamydial major outer membrane protein (MOMP) was decreased by HSV co-infection while accumulation of chlamydial heat-shock protein 60-1 (HSP60-1) was increased. Polymerase chain reaction (PCR) experiments indicated that chlamydial genome copy number was unaltered by HSV-2 superinfection. Semi-quantitative, reverse transcription PCR (RT-PCR) experiments demonstrated that levels of chlamydial groEL, ftsK, ftsW, dnaA and unprocessed 16S rRNA transcripts were not changed by HSV-2 super-infection. These data indicate that HSV-2 superinfection drives chlamydia into a viable but non-cultivable state, which is the hallmark of persistence. Because chlamydial HSP60-1 has been associated with immunopathology in vivo, these results also suggest that disease severity might be increased in co-infected individuals.


Sexually transmitted diseases (STDs) have reached epidemic proportions worldwide. This year, in the USA, there will be approximately 19 million new cases of STDs (Weinstock et al., 2004). The most commonly reported STD agents in the USA include Chlamydia trachomatis serovars D–K, with 4 000 000 new cases per year and herpes simplex virus (primarily HSV-2), with 200 000–500 000 new cases per year (Butler, 1997). Because both organisms establish asymptomatic or latent infections, these numbers underestimate the infectious burden.

The Chlamydia are a genus of Gram-negative, obligate intracellular bacteria. Multiple and persistent chlamydial infections are strongly associated with increased pathology in vivo (Beatty et al., 1994a). Genital chlamydial infections are associated with serious complications; these include pelvic inflammatory disease, tubal factor infertility and ectopic pregnancy in females and epididymitis and reactive arthritis in men (Darville, 2000). Extracellularly, chlamydiae exist as elementary bodies (EBs). Infection begins by contact of infectious EBs with the epithelial cell apical surface, followed by receptor-mediated endocytosis. Upon uptake, each EB is incorporated into an endocytic vesicle. Multiple EB-containing endocytic vesicles fuse to form an expanding, membrane-bound vacuole termed an inclusion. The small, spherical, metabolically inactive EBs (0.2 µm diameter) then transform into larger (0.8 µm), metabolically active, non-infectious, reticulate bodies (RBs). RBs use host metabolites and energy to synthesize macromolecules, grow and divide by binary fission. The RBs then mature into infectious EBs and exit the host cell (Wyrick, 2000).

Chlamydiae can establish persistent infection of host cells in vitro under specific environmental conditions. These include interferon-γ (IFN-γ), tumour necrosis factor-α (TNF-α) and penicillin G exposure as well as amino acid or iron deprivation and monocyte infection (Beatty et al., 1994a; Raulston, 1997; Darville, 2000; Gerard et al., 2001), all of which can occur in vivo as well. The persistent state is characterized by inclusion enlargement, the presence of swollen, diffuse RBs and decreased EB production (Matsumoto and Manire, 1970; Johnson and Hobson, 1977; Byrne et al., 1986; Beatty et al., 1993; 1994b). Persistent chlamydiae can remain in a viable but less metabolically active, non-replicative state for weeks or months. Chlamydial major outer membrane protein (MOMP) expression is often decreased during persistence while expression of the chlamydial proinflammatory antigen, heat-shock protein 60 (HSP60-1), is stable or increased (Cevenini et al., 1988; Beatty et al., 1993; 1994c). Persistent chlamydial forms shed MOMP and lipopolysaccharide (LPS) into membrane blebs, which find their way to the host cell surface and are released (Karimi et al., 1989; Wyrick et al., 1994; 1997; 1999). Although C. trachomatis persistence in vivo has been difficult to demonstrate, recent studies suggest that C. trachomatis establishes persistent infections in humans (Patton et al., 1994; Fortenberry et al., 1999; Dean et al., 2000; Gerard et al., 2001; 2002). Increased release of inflammatory mediators from persistently infected cells in vivo may enhance disease pathology (Darville, 2000).

The Herpesviridae are a family of enveloped DNA viruses. Primary human HSV-2 genital infections usually occur on genital skin or mucous membranes, where an intense inflammatory response is observed. The virus may then establish latency in the sacral ganglia. HSV-2-infected individuals experience an average of five reactivations per year throughout their lives, during which lesions and virions are present (Corey et al., 1983). HSV-2 may also cause several serious diseases, including keratitis, meningitis and disseminated herpes infection. At a single cell level, the HSV-2 replication cycle takes from 12 to 24 h and is initiated by viral attachment to one of several host cell receptors. Once viral DNA enters the nucleus, the viral genes are expressed in a specific temporal order. New virions are then assembled, enveloped and released (Roizman and Knipe, 2001).

A number of studies have established that co- and super-infections with HSV-2 and C. trachomatis occur in the human population. HSV-2 and C. trachomatis have been simultaneously isolated from the genital tract of women suffering from endometritis and salpingitis (Paavonen et al., 1985) as well as cystitis (Tait et al., 1985). Several large serological studies indicate that HSV-2-positive individuals are likely to be C. trachomatis-positive as well (Paroli et al., 1990; Vetter et al., 1990; Silins et al., 2002). Also, seropositivity rates of >50% for both HSV-2 and C. trachomatis have been reported, suggesting that some study participants had been exposed to both (Duncan et al., 1992; Wagner et al., 1994). Although IgG seropositivity cannot establish concurrent infection, it is likely that both organisms were present simultaneously in some individuals.

Cell culture models of HSV/chlamydial co-infection have been established. Pontefract et al. (1989) observed that inclusions in Vero cells co-infected with C. trachomatis serovar L2 and HSV-2 were swollen and contained few RBs or EBs. Chiarini et al. (1996) reported that in HSV-2/C. trachomatis serovar D co-infected HeLa cervical epithelial cells, the number of cells positive for chlamydiae by immunofluorescence (IFA) was reduced. HSV-2 pre-infection of HT-1376 human bladder cells also reduced both inclusion number and production of infectious C. trachomatis EBs (Superti et al., 2001). These data suggested that HSV-2 co-infection might alter the chlamydial developmental cycle in a manner similar to that observed during persistent infections. Because no previous study directly addressed this possibility, we established a cell culture HSV-2/C. trachomatis serovar E co-infection system and evaluated this hypothesis utilizing multiple experimental indicators of chlamydial persistence. These data demonstrate that HSV-2 co-infection alters the chlamydial developmental cycle similarly to other inducers of chlamydial persistence and illuminate potential mechanisms by which co-infection could enhance pathology.


HeLa cells can be co-infected with C. trachomatis serovar E and HSV-2

HeLa monolayers were either mock, singly- or co-infected (Fig. 1A). At 20 h after HSV infection, HeLa cells were immunostained (Fig. 1B). Both C. trachomatis and HSV-2 singly-infected cells stained only with antibodies specific for either agent; control, mock-infected cells did not stain with either antibody. IFA using C. trachomatis MOMP-specific antibodies demonstrated that chlamydial inclusions were present in more than 60% of C. trachomatis-infected or co-infected cells. HSV-2 ICP5 and ICP8 antibody staining showed that at least 90% of the HSV singly infected and co-infected cells were HSV-2 infected (Fig. 1B). Similar results have been obtained in multiple experiments and demonstrate that co-infection did occur and that most chlamydiae-infected cells were super-infected with HSV-2.

Figure 1.

HSV-2 co-infection increases C. trachomatis inclusion size.
A. HeLa cells were plated and divided into four groups: mock (uninfected negative control), Ct-25 (infected only with C. trachomatis), HSV (infected only with HSV-2) and Ct-25/HSV (infected first with C. trachomatis and 24 h later with HSV-2). Each group was mock infected with 2SPG alone or C. trachomatis serovar E at 80% infectivity at T-25 and with HSV-2 at 10 PFU cell−1 or culture medium at T-1 as indicated. All cells were incubated at 35°C from T-25 until the end of the experiment. Following a 1 h attachment/entry period, all cells were refed with fresh culture medium following both C. trachomatis (T-25) and HSV-2 (T0) and mock (T-25 and/or T0) infection. Samples for IFA (immunofluorescence assay), TEM (transmission electron microscopy), plaque assays, EB (elementary body) titre and Western blot experiments were collected at various times after T0.
B. Mock, HSV-2, C. trachomatis and co-infected cells were fixed and stained using FITC-conjugated anti-C. trachomatis major outer membrane protein (Anti-C. trach.) or using antibodies against HSV-2 (Anti-HSV-2) as described in Experimental procedures. Photomicrographs are 320× magnification.
C. Average relative inclusion size was determined from fluorescence photomicrographs. Images were coded and analysed in double-blind fashion as described. Means were compared using a two-sample t-test for independent samples. Data are presented as average relative inclusion size ± the SEM. The Ct-25 and Ct-25/HSV inclusion size values are significantly different (P < 0.05)*.

HSV-2 super-infection increases chlamydial inclusion size

Because inclusion enlargement has been observed in several in vitro models of chlamydial persistence (Matsumoto and Manire, 1970; Johnson and Hobson, 1977; Byrne et al., 1986; Beatty et al., 1993; Raulston, 1997), image analysis was used to determine the average inclusion size in singly and co-infected cells at 20 h after HSV infection (Fig. 1C). Comparison of inclusion size indicated that inclusions were 28% larger in co-infected compared with singly infected cells (448.6 ± 27 versus 624.9 ± 49). Statistical analysis indicated that average inclusion areas in these two samples were significantly different (P < 0.05).

Superinfection with HSV-2 induces morphological changes in C. trachomatis

Persistent forms of C. trachomatis have a characteristic electron microscopic appearance (Matsumoto and Manire, 1970; Johnson and Hobson, 1977; Byrne et al., 1986; Beatty et al., 1993; Raulston, 1997). Thus, chlamydial morphology in singly and co-infected cells was compared by transmission electron microscopy (TEM) (Fig. 2). At 30 h after chlamydial infection, chlamydial inclusions in C. trachomatis singly infected cells contained morphologically normal RBs (black arrow, Fig. 2A); by 45 h after infection, these inclusions contained chlamydiae progressing through the standard developmental cycle with normally sized RBs (Fig. 2B, black arrow) as well as intermediate and elementary bodies (IBs and EBs, black arrow with star). The C. trachomatis singly infected control cells in Fig. 2A and B are labelled 5 h and 20 h (after HSV infection) for ease of comparison with the appropriate co-infected samples. In co-infected cells, the RBs appeared slightly swollen as early as 5 h after HSV-2 infection (Fig. 2D, black arrow). At 10 h (Fig. 2E), 15 h (Fig. 2F) and 20 h (Fig. 2G and H) after HSV-2 infection, the inclusions contained enlarged, irregular and diffuse RBs (Fig. 2E–H, black arrows). At 20 h after viral infection, very few IBs and no EBs were observed in co-infected cells (Fig. 2G and H). Increased accumulation of membrane blebs was also apparent in co-infected cells at both 15 h (Fig. 2F, black boxes) and 20 h (Fig. 2G and H, black boxes) after virus infection. HSV-2 virions were visible within nearly all co-infected cells containing aberrant chlamydial forms, beginning at 10 h after HSV-2 infection (Fig. 2E–G and I, black circles). Neither chlamydial inclusions nor HSV-2 virions were visible in mock-infected cells (Fig. 2C). Similar data were obtained in multiple experiments and indicated that HSV-2 superinfection induced readily apparent alterations in C. trachomatis serovar E ultrastructure.

Figure 2.

HSV-2 co-infection induces a ‘persistent-like’ chlamydial morphology. HeLa cells were mock infected (Mock, C), C. trachomatis infected (Ct-25, A and B) and C. trachomatis/HSV-2 co-infected (Ct-25/HSV, D–I) as described in Fig. 1A. At 5 (A and D), 10 (E), 15 (F) or 20 h (B, C, G–I) after HSV infection, cells were pre-fixed with glutaraldehyde-paraformaldehyde, processed and then subjected to high contrast transmission electron microscopy. (A) through (H) are at 12 101× magnification. (I) is at 29 388× magnification.

HSV-2 superinfection abolishes production of infectious chlamydial EBs

Previous studies have demonstrated that chlamydial persistent forms are non-infectious (Johnson and Hobson, 1977; Beatty et al., 1993). The dearth of EBs in co-infected cells observed by TEM suggested that HSV-2 co-infection interfered with RB maturation to EB. Therefore, subpassage was used to determine whether infectious chlamydiae were produced by co-infected cells (Table 1). Neither mock nor HSV-2 singly infected cells produced detectable infectious EBs, whereas C. trachomatis singly infected cells produced abundant EBs. Co-infected cells produced few, or no, infectious EBs, demonstrating that production of infectious chlamydial progeny was essentially abolished by HSV-2 co-infection.

Table 1.  Co-infected cells do not produce infectious chlamydial EBs.
Experimental groupbIFU per 106 cellsa
Experiment 1cExperiment 2c
  • a

    . Inclusion counts from triplicate wells were averaged and used to calculate the number of inclusion-forming units (IFU) per 106 cells.

  • b

    . HeLa cells were mock infected (Mock), C. trachomatis infected (Ct-25), HSV-2 infected (HSV) and C. trachomatis/HSV-2 co-infected (Ct-25/HSV) as described in Fig. 1A.

  • c

    . Results from two independent co-infection experiments are shown.

  • d

    . Because the lowest dilution of cell lysate tested was 1/5, inclusion counts of ‘0’ cannot be used to calculate the exact number of IFU present in undiluted cell lysates and only indicate that the number of IFU present must be less than 5.

Mock< 5d< 5d
HSV< 5d< 5d
Ct-25586918 000
Ct-25/HSV< 5d   120

Co-infection alters neither chlamydial nor HSV-2 genome copy number

One characteristic of persistent chlamydiae is that they do not divide but continue to replicate their DNA (Gerard et al., 2001). Total DNA from co-infected and singly-infected cells (Fig. 3) was subjected to polymerase chain reaction (PCR) using primers specific for the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene, the chlamydial 16S rRNA gene and the HSV-2 G2 glycoprotein gene. All amplimers were the expected size (Fig. 3A) and the identity of each was confirmed by DNA sequencing (data not shown). The quantity of chlamydia DNA-specific product is similar in both co-infected and C. trachomatis singly infected samples (Fig. 3A, lanes 4 and 5). Likewise, the amount of HSV-2-specific amplimer appears unchanged by C. trachomatis co-infection (Fig. 3A, lanes 3 and 5). Amplification was never observed in negative (no template) controls (Fig. 3A, lane 1). GAPDH amplification products were observed in all samples at similar amounts (data not shown). HSV-2-specific and chlamydia-specific products were observed only in the appropriate samples, confirming amplification specificity. The experiment shown in Fig. 3A was repeated three times; six log dilution series of known DNA template controls were also amplified and used to generate amplification standard curves. All experimental samples quantified fell within the linear range of the PCR amplification as determined from the appropriate, co-amplified, standard curve. Chlamydial 16S rRNA and HSV G2 amplimer quantities were then normalized to host genome copy number as ascertained by PCR with human GAPDH-specific primers and plotted in Fig. 3B. Statistical analyses indicated that there was no difference in either chlamydial or HSV genome copy number in co-infected versus singly-infected cells (Fig. 3B).

Figure 3.

Co-infection alters neither chlamydial nor HSV-2 genome copy number.
A. HeLa cells were mock infected (Mock), C. trachomatis infected (Ct-25), HSV-2 infected (HSV) and C. trachomatis/HSV-2 co-infected (Ct-25/HSV) as described in Fig. 1A. Total DNA was purified as described and subjected to PCR analyses using host genome (GAPDH), chlamydial genome (16S rRNA gene) and HSV-2 genome (G2 glycoprotein gene) specific DNA oligonucleotides. Chlamydial 16S rRNA (107 bp) and HSV-2 G2 (72 bp) amplimers were electrophoresed as described; representative photographs are shown. The position of DNA size markers are shown to the left of each gel image in base pairs (bp).
B. Specific DNA amplimers were quantified as described. Chlamydial (black bars) and HSV (grey bars) amplimer quantities were normalized to host genome copy number as ascertained by PCR with human GAPDH-specific primers (data not shown). The values plotted are averages of three independent experiments; the value obtained for each independent experiment is the average obtained from three biological replicates. The average integrated intensity for each amplimer was plotted ± SEM versus sample identity (x-axis).

Co-infection does not alter accumulation of chlamydial dnaA, ftsW, ftsK, groEL-1 or unprocessed 16S rRNA transcripts

Several investigators have demonstrated that expression of one or both of the chlamydial cell division genes, ftsK and ftsW, changes after persistence induction. In contrast, expression of the DNA replication gene dnaA remains unchanged (Byrne et al., 2001; Gerard et al., 2001; Nicholson and Stephens, 2002). Also, by definition, persistent chlamydiae are viable but non-infectious. As a result, transcription of unprocessed 16S rRNA continues during persistence and, thus, would be expected to remain unchanged compared with actively dividing, viable chlamydiae. Co-infected and singly infected cell cDNAs (Fig. 4) were subjected to reverse transcription PCR (RT-PCR) using primers specific for the chlamydial dnaA, ftsW, ftsK, groEL-1 and unprocessed 16S rRNA transcripts. All amplimers were the expected size (Fig. 4A) and the identity of each was confirmed by DNA sequencing (data not shown). The quantity of each specific product was similar in both co-infected and C. trachomatis singly infected samples (Fig. 4A, lanes 4 and 5). Amplification was not observed in negative (no template) controls (Fig. 4A, lane 1) or in RT(–) controls (data not shown). Specific amplimers were observed only in C. trachomatis singly and co-infected cells, confirming amplification specificity. Amplimers were quantified, compared with amplification standard curves and then normalized to relative chlamydial genome copy number from that same sample as ascertained by PCR in Fig. 3. Statistical analyses indicated that there was no significant difference in accumulation of any of the chlamydial transcripts in co-infected versus C. trachomatis singly-infected cells (Fig. 4B).

Figure 4.

Co-infection does not alter accumulation of chlamydial dnaA, ftsW, ftsK, groEL-1 or unprocessed 16S rRNA transcripts.
A. HeLa cells were mock infected (Mock), C. trachomatis infected (Ct-25), HSV-2 infected (HSV) and C. trachomatis/HSV-2 co-infected (Ct-25/HSV) as described in Fig. 1A. Total cellular RNA was collected from each group of cells at 20 h after HSV-2 infection and subjected to RT-PCR analyses using transcript-specific oligonucleotide probes. Chlamydial dnaA (501 bp), ftsW (274 bp), ftsK (351 bp), groEL-1 (223 bp) and 16 rRNA primary transcript (466 bp) amplimers were electrophoresed and photographed; representative photographs are shown. The position of DNA size markers are shown to the left of each gel image in base pairs (bp).
B. Specific DNA amplimers were quantified as described. Amplimer quantities were normalized to chlamydial genome copy number (Fig. 3B). The values plotted are averages of three independent experiments; the value obtained for each independent experiment is the average obtained from three biological replicates. The average integrated intensities for dnaA (white bars), ftsW (grey bars), ftsK (black bars), groEL-1 (vertically striped bars) and 16S rRNA primary transcript (diagonally striped bars) were plotted versus sample identity (x-axis). Mock and HSV singly-infected reactions were negative for these amplimers, therefore, only data for Chlamydia singly-infected and co-infected cells are shown.

HSV-2 superinfection reduces accumulation of chlamydial MOMP and increases accumulation of HSP60-1

Accumulation of chlamydial MOMP is often decreased when RBs enter a persistent state; in contrast, HSP60-1 levels are usually stable or increased (Cevenini et al., 1988; Beatty et al., 1993; 1994c). To determine whether HSV-2 superinfection altered MOMP or HSP60-1 accumulation, Western blot analyses were performed at 0, 5, 10, 15 and 20 h after HSV infection. Single bands, of the correct size for MOMP (40 kDa) and HSP60-1 (60 kDa), were observed in C. trachomatis singly infected and co-infected cell lysates but not in mock or in HSV singly infected cell lysates (Fig. 5A). The HSV-2 ICP8 and ICP5 proteins were observed in only HSV singly and co-infected samples (Fig. 5A). Bands were quantified, normalized to focal adhesion kinase (FAK – a host cellular control protein) and plotted in Fig. 5B. Statistical analyses have demonstrated that FAK accumulation does not change over the course of the experiment (data not shown). MOMP accumulation was clearly decreased at 5, 10, 15 and 20 h after HSV infection in co-infected cells. In contrast, HSP60-1 accumulation was increased in co-infected cells, particularly at 10 h after HSV infection. Accumulation of the HSV-2 proteins ICP8 and ICP5 was either unchanged or diminished by chlamydial co-infection (Fig. 5B). Similar results were obtained in two other independent time-course experiments.

Figure 5.

Chlamydial major outer membrane protein (MOMP) accumulation is reduced in co-infected cells.
A. HeLa cells were mock infected (Mock), C. trachomatis infected (Ct-25), HSV-2 infected (HSV) and C. trachomatis/HSV-2 co-infected (Ct-25/HSV) as described in Fig. 1A. Cell lysates were collected from each group of cells at 0, 5, 10, 15 and 20 h after HSV-2 infection and Western blotted using anti-C. trachomatis MOMP, anti-C. trachomatis HSP60-1, anti-HSV-2 ICP5, anti-HSV-2 ICP8 or anti-human FAK sera. Three independent experiments were performed; a representative autoradiogram is shown.
B. MOMP, HSP60-1, ICP5 and ICP8 proteins present in cell lysates from the experiment in (A) were quantified by phosphorimage analysis. The integrated intensities were normalized to the internal control protein (FAK) and plotted in (B) as integrated intensity versus time after HSV-2 infection.

Chlamydia trachomatis co-infection does not significantly alter HSV-2 progeny virion release or viral ultrastructure

To determine whether chlamydial pre-infection altered subsequent viral replication, HSV-2 production/release from singly and co-infected cells was quantified using plaque assays (Fig. 6A). Mock-infected and C. trachomatis singly infected cells produced no virions (Fig. 6A). While co-infected cells appeared to release a greater quantity of HSV-2 progeny virions at the later time intervals (22–30 h after viral infection), at no time was there greater than 10% difference between the two groups. These data suggested that co-infection of HSV-infected HeLa cells with chlamydiae did not significantly affect HSV-2 replication kinetics or release of infectious HSV-2 progeny virions.

Figure 6.

Chlamydial pre-infection does not significantly effect HSV-2 replication or release.
A. HeLa cells were mock infected (Mock), HSV-2 infected (HSV), C. trachomatis infected and C. trachomatis/HSV-2 co-infected (Ct-25/HSV) as described (Fig. 1A). Supernatants (sup.) were collected from each group of cells 2, 12, 14, 16, 18, 20, 22, 24, 26, 28 and 30 h after HSV-2 infection. Plaque assays were performed as described. Plaques were counted and the number of PFU per ml supernatant at the end of each time interval was calculated. These values were plotted in (A). C. trachomatis singly infected supernatants were identical to the mock values and are not shown.
B. HeLa cells were infected as described (Fig. 1A). Culture supernatants and cell pellets were collected from each group of cells at 20 h after HSV-2 infection. Plaque assays were performed and the total number of PFU per culture was calculated by adding the values obtained for cell associated and supernatant virus. Average PFU per culture from four independent experiments ± the SEM were plotted on the histogram in (B).

It was possible that co-infection might have decreased HSV-2 virion production and increased release, such that the supernatant virus levels would have appeared unaffected. Therefore, cell pellets and culture supernatants were collected at 20 h after virus infection and used for virus titre assays. Total plaque-forming units (PFU) produced in each culture were determined by adding the cell-associated and supernatant virus titres; the data obtained from four independent experiments are plotted in Fig. 6B. Again, the total amount of virus obtained from co-infected cells tended to be slightly higher than that obtained from HSV singly-infected cells; however, the difference was not statistically significant. These data indicated that chlamydial co-infection had little, if any, effect on HSV-2 replication.

Viral morphology was also studied by TEM. In HSV-2 singly and co-infected cells, viral capsids could be observed as early as 10 h after infection (Fig. 2E). By 20 h after infection, arrays of capsids were readily apparent in the nucleus of HSV-2 singly infected and co-infected cells (Fig. 2I). Comparison of viral particles in singly and co-infected cells indicated that virion morphology was unaltered by chlamydial co-infection (data not shown). Considered together, these data indicated that chlamydial co-infection did not significantly alter HSV-2 replication.


While previous studies (Pontefract et al., 1989; Chiarini et al., 1996; Superti et al., 2001) suggested that HSV-2 co-infection alters the chlamydial developmental cycle, none of the previous studies determined whether a persistent chlamydial phenotype was induced. A number of experimental parameters have been used as ‘markers’ of chlamydial persistence. These include: altered inclusion size, abnormal RB morphology, decreased progeny EB infectivity, decreased MOMP and stable/increased HSP60 accumulation, unchanged DNA copy number and decreased ftsK and/or ftsW expression (Beatty et al., 1994a; Darville, 2000; Hogan et al., 2004). However, downregulation of ftsK and ftsWTends to be variable in the persistent state. For example, accumulation of ftsK and ftsW RNA is reduced in C. trachomatis serovar K-infected monocytes as well as in IFN-γ-exposed Chlamydia pneumoniae (Byrne et al., 2001; Gerard et al., 2001). In contrast, Nicholson and Stephens (2002) demonstrated that in penicillin-exposed, C. trachomatis serovar D-infected cells, ftsW remains unchanged while ftsK is elevated. Finally, in C. pneumoniae continuous infection, ftsK expression is unchanged (Hogan et al., 2003). It is likely that many ‘markers’ of persistence vary according to the inducer, host cell, chlamydial species and timing used (Hogan et al., 2004), which increases the importance of examining as many of the different markers as is feasible.

Because most of the previously mentioned markers were observed during HSV co-infection, it seems likely that HSV superinfection of C. trachomatis-infected HeLa cells induces development of a persistent chlamydial state. By definition, persistent chlamydiae are viable but non-infectious. As a result, transcription of unprocessed 16S rRNA continues during persistence and, thus, would be expected to remain unchanged compared with that in actively dividing, viable chlamydiae. The observation that accumulation of unprocessed 16S rRNA is unaltered but chlamydial infectivity is abrogated, thus, indicated that HSV co-infection renders the chlamydae viable but non-infectious, the hallmark of the persistent state. As co-infected cells quickly die in culture (approximately 30 h after HSV addition), this is not long-term persistence in the classical sense. However, because it is similar in many respects, we shall continue to refer to HSV-induced, chlamydial developmental changes as persistence.

An alternate interpretation of the observation that co-infection reduces chlamydial infectivity and MOMP expression is that HSV super-infection simply delays the chlamydial developmental cycle. If HSV causes a chlamydial developmental cycle delay, one would expect chlamydial DNA replication to be slowed or halted. In contrast, persistent chlamydiae continue to replicate their chromosome (Gerard et al., 2001). Thus, the observation that chlamydial genome copy number was similar in co-infected and singly-infected cells strongly supports the conclusion that HSV super-infection of C. trachomatis-infected HeLa cells induces development of chlamydial persistence rather than simply delaying the chlamydial developmental cycle. Additionally, delayed chlamydial development would not be expected to induce the alterations in EB morphology observed in co-infected cells.

There are a number of mechanisms by which HSV-2 superinfection could induce chlamydial persistence. One simple explanation is that HSV-2-induced cell death aborts the chlamydial developmental cycle. Chlamydial morphologic alterations are observed as early as 5 h after HSV infection, much earlier than HSV-induced cell death usually occurs (Roizman and Knipe, 2001). Trypan blue staining experiments indicate that >99% of co-infected HeLa cells are viable at 20 h after HSV infection (R.V. Schoborg, unpubl. data). Also, TEM analysis of co-infected cells at 20 h after HSV infection does not reveal any ultrastructural alterations associated with cell death. These observations argue that cell death is not inducing this effect.

Amino acid starvation is a well-studied inducer of C. trachomatis persistence (Beatty et al., 1994a; Darville, 2000), suggesting that HSV-2 might ‘out-compete’ the chlamydiae for cellular amino acid pools. However, persistent-like chlamydial morphological changes are observed at 5 and 10 h after HSV addition, well before the bulk of HSV-2 protein synthesis occurs. Additionally, the HSV-2 VHS (virion host shut-off) and ICP27 proteins act early in the replication cycle to diminish host gene expression (Roizman and Knipe, 2001). Thus, at early times after HSV infection, it seems likely that amino acids would be more readily available to support both HSV-2 and chlamydial replication. These observations reduce the likelihood that HSV-induced chlamydial persistence is simply an effect of amino acid depletion.

A more interesting possibility is that a specific viral protein(s) could induce chlamydial persistence. If so, this protein is most likely a virion component or an early gene product as the effect is evident by 5 h after HSV infection. The HSV-2 VHS and ICP27 proteins are known to profoundly influence host cell gene expression. VHS is a virion component that suppresses host transcription beginning immediately after entry (Kwong et al., 1988). ICP27 is a viral inhibitor of host cell mRNA splicing/transport that is expressed early in the replication cycle (Smith et al., 1992; Sandri-Goldin and Mendoza, 1992; Hardy and Sandri-Goldin, 1994). Although either protein could induce chlamydial persistence, it is not likely to be due to their effects on host gene expression as cyclohexamide exposure is known to enhance chlamydial development (Ripa and Mardh, 1977).

A previously published study suggested that pre-infection of host cells with C. trachomatis increases HSV-2 replication by up to 40%, as quantified by IFA (Chiarini et al., 1996). In contrast, we did not observe any difference in staining intensity or frequency in co-infected compared with HSV singly infected cells. Also, plaque assays show no significant difference in total virus production from singly and co-infected cells. Because any major effect of co-infection on viral replication would be readily visible as an increase or decrease in viral progeny production, these data indicate that chlamydial pre-infection does not significantly effect HSV-2 replication.

Induction of chlamydial persistence by HSV-2 co-infection has significant implications for pathogenesis. The presence of membrane blebs similar to those observed in chlamydia/HSV co-infected cells has been associated with release of chlamydial LPS from persistently infected cells (Karimi et al., 1989; Wyrick et al., 1994; 1999). Because complications of chlamydial infection are thought to be due, at least in part, to release of inflammatory mediators from infected cells, any stimulus that increases release of these compounds might enhance disease pathology. Thus, it is possible that co-infection will result in increased release of inflammatory molecules and more severe immunopathology compared with that in an individual infected with either agent alone.

Co-infection in vivo could also influence transmission or immunity to either organism. Both C. trachomatis and HSV-2 have been shown to modulate immune cell function. For example, HSV infection suppresses dendritic cell function (Pollara et al., 2003). C. trachomatis infection increases IL-11 production, which may suppress local immune responses and aid in establishment of infection (Dessus-Babus et al., 2000). It is easy to envision how an organism that manipulates the local immune response in the genital tract might also suppress or alter the host response to other pathogens that are present at the same time. Also, super- or co-infection could increase genital tract susceptibility or increase shedding of either organism by induction of inflammatory lesions, in a manner analogous to that proposed in co-infections with HIV (Cameron et al., 1989; Plummer et al., 1991; Laga et al., 1993; McClelland et al., 2001; 2002). Determination of the effect of co-infection on disease severity and transmission is dependent on determination of the exact mechanism by which HSV interferes with the chlamydial developmental cycle and development of an animal co-infection model, studies that are currently ongoing in our laboratory.

Experimental procedures

Chlamydia, HSV-2 and host cells

A human urogenital isolate of C. trachomatis E/UW-5/CX was originally obtained from S.P. Wang and C.C. Kuo (University of Washington, Seattle, WA). The same standardized inoculum of C. trachomatis serovar E EBs, propagated in McCoy cells, was used for all experiments (Wyrick et al., 1996). Herpes simplex type 2 strain 333 stock was prepared in monolayers of CV-1 simian kidney cells using standard techniques (Duff and Rapp, 1971).

Co-infection experimental design

HeLa cells, a cervical adenocarcinoma epithelial cell line (ATCC ♯CCL2), were used for all infection experiments. They were propagated in Minimal Essential Medium (MEM) with Earle's salts containing l-glutamine, 10% fetal calf serum (Atlanta Biologicals) and 1 µg ml−1 gentamicin at 37°C in an atmosphere of 5% CO2. HeLa cells were divided into four groups, 1 × 106 cells per 60 mm culture dish, for mock infection, chlamydial infection, HSV-2 infection, and both C. trachomatis and HSV-2 double infection (Fig. 1A). Host cells were infected with a titre of crude EB stock (200 µl) calculated to infect 80% of the HeLa cells and with HSV-2 at a multiplicity of infection (moi) of 10 PFU per cell; mock-infected cells were treated similarly except they were exposed to 200 µl of either 2SPG (0.2 M sucrose, 6 mM NaH2PO4, 15 mM Na2HPO4, 5 mM l-glutamine, pH 7.2; mock C. trachomatis infected) or growth medium (mock viral infection).

Microscopy and image analysis

For fluorescence analyses, infected cells were fixed and permeabilized as described (Saltarelli et al., 1994). A pool of FITC-conjugated monoclonal antibodies generated against C. trachomatis MOMP (SYVA MicroTrak, Wampole Laboratories) was used to visualize chlamydial inclusions. Monoclonal antibodies against HSV-2 ICP5 (the major capsid protein; Virusys) and ICP8 (the viral DNA-binding protein; Virusys) were combined, diluted 1/200 in PBT buffer (1× PBS, 1% BSA, 0.05% Tween 20) and used to immunostain for HSV-2 for 1 h at 37°C. After washing, HSV-2 antigens were visualized by staining with a Texas red-conjugated secondary anti-murine IgG (Jackson Immuno Research) diluted 1/100 in PBT buffer for 1 h at 37°C. Stained monolayers were washed, mounted and photographed at 320× using a Zeiss Axiovert S100 inverted microscope and Axiocam camera. FITC-labelled cells were photographed at an exposure of 8 s using a FITC-band pass filter (excitation 470 nm and emission 540 nm); Texas red-labelled cells were photographed at an exposure of 2 s using a Rhodamine-long pass filter (excitation 546 nm and emission 590 nm).

Selected photomicrographs were subjected to image analysis using the Quantity One software package (Version 4.2.1, Bio-Rad). The relative area of each fluorescent focus was determined using the software volume analysis function. Ten random fields (40–50 inclusions per field) from each experimental group (mock, HSV-2-infected, C. trachomatis-infected and co-infected cells) were analysed blind and the average inclusion size for each set was determined.

Duplicate samples of infected HeLa cells were processed at 5, 10, 15 and 20 h for high-contrast TEM as described (Wyrick et al., 1994). Counterstained gold thin sections were examined using a Tecnai 10 (FEI) transmission electron microscope operating at 60–80 kV.

Chlamydial titrations by subpassage

At 20 h after HSV-2 infection, monolayers were scraped into 1 ml of cold growth medium, pelleted and resuspended in 1 ml of fresh medium. Infected host cells were lysed by freeze/thaw and sonication and centrifuged (500 g for 5 min) to pellet cell debris. Supernatants were centrifuged twice (8000 g for 30 min). Control experiments demonstrated that >95% of the ‘contaminating’ HSV-2 was removed by discarding supernatants from the two high-speed centrifugation steps, while >75% of the input chlamydial EBs were recovered in the pellet (data not shown). Final EB pellets were resuspended in 200 µl of 2SPG, diluted (1/5–1/200) and used to infect HeLa cells plated on glass coverslips in triplicate. After infection, the HeLa monolayers were refed with medium containing 1 µg ml−1 cyclohexamide and 400 µM phosphonoformate (Cheng et al., 1981). Neither drug interferes with chlamydial inclusion development while completely abrogating productive replication of HSV-2 (data not shown). HeLa monolayers were incubated at 35°C for 48 h, fixed and stained with SYVA stain as described above. The number of inclusions in 10 random microscopic fields per sample was determined using a Zeiss Axiovert S100 Microscope at 320×. Triplicate coverslips were counted and the counts averaged. The number of inclusion-forming units (IFU) in the undiluted inoculum was then calculated and expressed as IFU per 106 cells.

RNA and DNA isolation

Total DNA and RNA were simultaneously isolated from the same plate of cells without manipulation of the cells before the lysis step using modifications to the RNeasy Mini (Qiagen) and QIAmp DNA Blood Mini (Qiagen) kits. After aspiration of the culture medium, 750 µl of RLT (lysis) buffer (RNeasy Mini Kit) was immediately added to each monolayer. The plates were scraped and the lysate passed 10 times through a 20-gauge needle using a 1 cc syringe. Six hundred microlitres of cell lysate were used for total RNA isolation and DNase treatment using the Qiagen RNase-free DNase Kit following the manufacturer's protocol. One-hundred and fifty microlitres of each cell lysate were diluted with 50 µl of 1× PBS and used for total DNA isolation as per the kit instructions. Control experiments with known quantities of RNA and DNA demonstrated that this modification did not alter recovery efficiency or quality of isolated host cellular, chlamydial or HSV nucleic acids (data not shown). Total RNA and DNA preparations were quantified using optical density at 260 and 280 nm (OD260 and OD280); all samples had OD260/OD280 ratios > 1.9. The concentration and integrity of each RNA sample was further confirmed by formaldehyde-agarose gel electrophoresis and ethidium bromide (EtBr) staining (data not shown).

Reverse transcription, PCR and RT-PCR

One microgram of total RNA was subjected to reverse transcription using SuperScript II RNase H minus reverse transcriptase (RT; Invitrogen) and random hexamer primers (Pharmacia) as per the manufacturer's suggested protocol. Each reaction also contained 20 U of Super RNase Inhibitor (Ambion). Duplicate reactions containing ddH2O substituted for RT (RT– reactions) were performed in parallel. PCR was performed using total cellular DNA or cDNA for RT+ and RT– reactions as appropriate. Experimental template DNAs were used at dilutions ranging from 1/10 to 1/1000 (in ddH2O) such that each reaction was in the linear amplification range. Amplification was performed with MasterTaq (Eppendorf) under the manufacturer's recommended reaction conditions. Human GAPDH, chlamydial dnaA and HSVG2 reactions were run without MasterTaq enhancer (Eppendorf); all other reactions contained enhancer at 1× concentration. Published primer sets included chlamydial dnaA, ftsK, ftsW (Gerard et al., 2001), chlamydial 16S rRNA primary transcript (Gerard et al., 1997) and HSV-2 glycoprotein gene 2 (HSV G2; Filen et al., 2004). We designed specific primers for chlamydial CT-110/groEL-1 (forward: 5′-GAG TTA AGA CTT TAG CTG AAG CTG T-3′, reverse: 5′-GTA GCT GTT GTA GTT CCG TCT CCA-3′), chlamydial 16S rRNA (forward: 5′-GGA CGG AAG TCT GAC GAA-3′, reverse: 5′-TCA AATCCA GCG GGT ATT-3′) and human GAPDH (forward: 5′-GTC CAC CAC TGA CAC GTT G-3′, reverse: 5′-GGG AAA CTG TGG CGT GAT-3′). Most PCR reactions were heated to 94°C for 1 min and then cycled under the following conditions 30 times: 94°C, 1 min; 58°C, 1 min; 72°C, 1 min. Human GAPDH reactions were performed under identical conditions but were cycled 35 times; chlamydial dnaA PCR reactions were performed for 30 cycles but annealed at 54°C. After PCR, all reactions were electrophoresed on 1.5% agarose/TBE gels. Gels were stained with EtBr and visualized using a Bio-Rad Chemi Doc XRS Image Capture System. Amplimers were quantified on the image analysis system using Quantity One V4.5.0 software (Bio-Rad).

SDS-PAGE and Western blotting

At various times after HSV-2 infection, monolayers were lysed in 1 ml of 1× PBS + 0.1% SDS + 0.1% Nonidet NP-40. Equal cell equivalents of each lysate (2 × 104 cells) were electrophoresed; gels were then Coomassie stained, fixed and dried. Densitometric analysis indicated that the variation in protein concentrations between samples was < 5%. Total protein concentrations for each set of lysates were also determined using the Micro-BCA method (Pierce), which confirmed the results obtained by Coomassie stains (data not shown). For determination of relative protein accumulation, 1 × 104 cell equivalents of each lysate were electrophoresed on 4–12% NuPage pre-cast gels using the MES-SDS buffer system (Invitrogen). Gels were electrophoresed at 150 V for 25 min and blotted to PVDF membranes (Pall) for 1 h at 200 mA. Blots were blocked in 1× PBS + 0.1% Tween 20 + 5% non-fat dried milk (BLOTTO) for 1 h at 25°C. Replicate blocked blots were incubated with at 1/1000–1/5000 dilutions the following primary antibodies in BLOTTO: polyclonal goat anti-MOMP (OEM Concepts); polyclonal, monospecific rabbit anti-HSP60-1 (a kind gift from J. Raulston); polyclonal rabbit anti-human FAKC-20 (SantaCruz Biologicals) or monoclonal antibodies against HSV-2 ICP5 or ICP8 (Virusys). Blots were incubated for 1 h, washed five times in 1× PBS + 0.1% Tween 20 (PBS-T) and reprobed with peroxidase-conjugated rabbit anti-goat, rabbit anti-mouse or goat anti-rabbit secondary antibodies (Cappel) at 1/20 000 dilution. After five additional washes, blots were subjected to chemiluminescent detection using SuperSignalWest Pico reagent (Pierce). Specific bands were quantified using an FX phosphorimager and Quantity One V2.5.0 software (Bio-Rad).

HSV-2 plaque assay

At various times after HSV-2 infection, culture supernatants or frozen/thawed cell pellets were centrifuged at 4000 g for 5 min and at 4°C to remove cell debris. Plaque assays were carried out on the resulting supernatants as described (Duff and Rapp, 1971). Duplicate, infected cultures were incubated for 72 h at 37°C, fixed, stained and counted in a blind fashion. Average plaque counts from each set of duplicate plates were used to calculate the PFU ml−1 present in the original supernatant.

Statistical analyses

Statistical analyses were performed using MiniTab (Version 9). Comparison of means was performed by using a two-sample t-test for independent samples. P-values of ≤0.05 were considered significant.


The authors would like to thank Dr Jane Raulston, Dr John Laffan and Dr Michelle Duffourc for helpful discussion of these experiments. We would also like to thank Elizabeth Neely for her technical help and advice. Finally, we would like to acknowledge the excellent work of the Electron Microscopy Core Facility, Department of Pathology, James H. Quillen College of Medicine. This work was supported by NIH Grant ♯5R21AI59563 to R.V.S., ♯P01-AI-37829-08 to M.K.H. and P.B.W. and ETSU RDC Grant ♯04-024M to R.V.S.