Molecular Analysis of the Multiple GroEL Proteins of Chlamydiae

Abstract

Genome sequencing revealed that all six chlamydiae genomes contain three groEL-like genes (groEL1, groEL2, and groEL3). Phylogenetic analysis of groEL1, groEL2, and groEL3 indicates that these genes are likely to have been present in chlamydiae since the beginning of the lineage. Comparison of deduced amino acid sequences of the three groEL genes with those of other organisms showed high homology only for groEL1, although comparison of critical amino acid residues that are required for polypeptide binding of the Escherichia coli chaperonin GroEL revealed substantial conservation in all three chlamydial GroELs. This was further supported by three-dimensional structural predictions. All three genes are expressed constitutively throughout the developmental cycle of Chlamydia trachomatis, although groEL1 is expressed at much higher levels than are groEL2 and groEL3. Transcription of groEL1, but not groEL2 and groEL3, was elevated when HeLa cells infected with C. trachomatis were subjected to heat shock. Western blot analysis with polyclonal antibodies raised against recombinant GroEL1, GroEL2, and GroEL3 demonstrated the presence of the three proteins in C. trachomatis elementary bodies, with GroEL1 being present in the largest amount. Only C. trachomatis groEL1 and groES together complemented a temperature-sensitive E. coli groEL mutant. Complementation did not occur with groEL2 or groEL3 alone or together with groES. The role for each of the three GroELs in the chlamydial developmental cycle and in disease pathogenesis requires further study.

The group I chaperonins, such as the GroEL protein of Escherichia coli and Hsp60 of eukaryotic mitochondria and plant chloroplasts, are a ubiquitous family of abundant proteins. These proteins play an essential role in ensuring that genome encoded proteins are expressed as fully functional molecules. In E. coli, homopolymers of GroEL interact with a ring-shaped cofactor composed of septamers of GroES or Hsp10 that forms the lid on an Anfinsen protein folding cage in which polypeptide substrates are assisted in folding into the correct three-dimensional (3-D) structure (24, 39). GroEL and GroES are present in the cytoplasm of unstressed E. coli cells, and both proteins are essential for bacterial growth at all temperatures (11). GroEL expression increases during a variety of conditions such as heat shock, nutrient deprivation, infection, and inflammatory reaction and functions to stabilize cellular proteins (40). GroEL proteins are highly conserved in sequence among bacteria and are recognized in hosts by Toll-like receptors as part of an innate defense system (38). GroELs are implicated in bacterial disease pathogenesis (41), and antibodies to chlamydial GroEL have been strongly associated with chlamydial disease sequelae (4). The mechanistic explanation underlying this epidemiological correlation remains undefined. We therefore undertook detailed bioinformatic, genetic, and immunologic studies of the three chlamydial groEL genes and proteins revealed during whole genome sequencing to determine whether this knowledge could shed light on their role(s) in chlamydial disease pathogenesis.

MATERIALS AND METHODS

Growth of C. trachomatis and E. coli.

The D serovar of C. trachomatis was propagated in and purified from HeLa cells. E. coli strain NL441 (25) was grown in Luria broth for liquid culture or on plates solidified with 1.5% agar.

Phylogenetic comparison of chlamydia GroELs.

The protein sets of 75 completed bacterial genomes (http://www.tigr.org/CMR2/) were searched against chlamydial GroEL1, GroEL2, and GroEL3 by using the BLASTP (1) program. All proteins that matched the Chlamydia GroELs with a minimum threshold P value of 10⁻⁵ were aligned by using CLUSTALW (37). Protein distances were calculated by using a maximum-likelihood method (PROTDIST in the PHYLIP 3.6 package, University of Washington, Seattle) and plotted as an unrooted tree by using a neighbor-joining algorithm.

Sequence comparison and analysis.

A database containing both nucleotide and amino acid sequences of all three groEL homologs of serovar D and biovar mouse pneumonitis (MoPn) of C. trachomatis, C. pneumoniae strain AR39, and the groEL gene of E. coli was collected from The Institute of Genomic Research database. Sequences were aligned by using the CLUSTALW program (37).

3-D structural predictions of C. trachomatis GroELs. (i) Threading.

The threading approach uses pair potentials to score for the propensity of two amino acid residues to occur at a specified distance. The methodology relies on the assumption that different proteins fold into a limited number of shapes (estimated to be ca. 4,000). The strategy is to identify the folding pattern of the test protein sequence by fitting it onto a library of known structures by using pseudoenergy as a measure of fit (5, 14, 17, 18, 28, 29, 33, 36). The library of folds is ranked in ascending order of total energy, with the lowest energy fold being taken as the most probable match. The approach is sufficient to identify proteins that share common folds in the “twilight zone” (<25%) of sequence identity in which sequence-based approaches normally fail (29, 33). The amino acid sequences of C. trachomatis serovar D GroEL1 to GroEL3 were analyzed by using the THREADER2 package (18), and a score representing a statistical Z score of threading pseudoenergy was obtained for each of the fold matches from the library of known protein structures. The Z score reflects the “goodness” of threading of a given sequence into the corresponding model fold: a Z-score value of <2.0 means that the analyzed sequence does not correspond to known protein folds in the library; scores between 2.0 and 3.5 are considered possibly significant and may correctly predict fold identification; scores of >3.5 are very significant and are likely to represent correct fold prediction.

(ii) Homology modeling.

The best fit of the threading results was used as a template for homology modeling. Homology modeling was carried out with MOE (for Molecular Operation Environment package, version 2001.01; Chemical Computing Group, Inc., Montreal, Quebec, Canada). Three dimensional structures for each of the C. trachomatis serovar D GroEL1 to GroEL3 proteins were derived as a Cartesian average of the 10 best intermediate models build by MOE.

Molecular cloning, expression, and purification of recombinant GroEL1, GroEL2, and GroEL3.

groEL1, groEL2, and groEL3 DNA fragments were generated by PCR by using genomic DNA isolated from D serovar of C. trachomatis. In order to subclone the PCR product as KpnI fragment into pET30b(+) vector (Novagen), forward and reverse primers used for amplification were designed as shown in Table 1. PCRs were carried out with Pfu DNA polymerase (Stratagene). The reaction mixture contained 2 mM MgSO₄, 200 μM concentrations of deoxynucleoside triphosphates, 2.5 U of Pfu DNA polymerase enzyme, and 25 pmol of each oligonucleotide primer in a total volume of 50 μl. The PCR cycling conditions were as follows: one cycle of 95°C for 2 min and 35 cycles of 94°C for 15 s, 58°C for 30 s, and 72°C for 2 min. This step was followed by strand elongation for 10 min at 72°C. The PCR product was purified with the QIAquick PCR purification kit (Qiagen), and the purified DNA fragment was cloned into pET30b(+) after restriction enzyme digestion by standard molecular biology techniques. The sequence of the subcloned groEL1, groEL2, and groEL3 genes were confirmed by sequencing with dye-labeled terminators by using the ABI PRISM kit (PE Biosystems). Plasmids containing the groEL1, groEL2, and groEL3 genes were transformed into the E. coli BL21(DE3) or BL21 Codonplus(DE3)RIL (Strategene), where GroEL1, GroEL2, and GroEL3 expression was carried out by inducing the lac promoter for expression of T7 RNA polymerase with IPTG (isopropyl-β-d-thiogalactopyranoside). The expressed GroEL1, GroEL2, and GroEL3 proteins with N-terminal His tags were purified by nickel column chromatography by using the His Bind purification system (Qiagen).

TABLE 1.

Primer sequences^a

Analysis and gene	Sequence
	Forward primer (5′-3′)	Reverse primer (5′-3′)
pET30b(+) cloning
groEL1	CGGGGTACCTTAAGGAGCGCATCAATGG	CGGGGTACCGGCTCGAAGAATCTATTTGTTCC
groEL2	CGGGGTACCTTGTCTAATAGTTTTCGAGACCAAG	CGGGGTACCAAATGATCAAGAAATTCTCCTCAG
groEL3	CGGGGTACCTACCCCGCTATGCCTCAC	CGGGGTACCTCGGTATGATAGGCGGATAG
pBAD24 cloning
groEL1	CGGGGTACCTTTAAGGAGCGCATCAATGG	CGGGGTACCGGCTCGAAGAATCTATTTGTTCC
groEL2	CGGGGTACCTTTGTCTAATAGTTTTCGAGACCAAG	CGGGGTACCAAATGATCAAGAAATTCTCCTCAG
groEL3	CGGGGTACCATACCCCGCTATGCCTCAC	CGGGGTACCTCGGTATGATAGGCGGATAG
RT-PCR
16S rRNA	CGGTAATACGGAGGGTGCTAGC	GAATTAAACCACATGCTCCACTGC
pKPK cloning
groES	CGGAATTCCATGTCAGATCAAGCAACGACC	CCGTCGAGTTATTGCAGAACTGCGATAACTT

^aThe specific primers used for RT-PCR of groEL1 to groEL3 were the same as those used for pET30b(+) cloning. 16S rRNA primers were designed in such a way to amplify the 16S rRNA genes of serovar D and biovar MoPn of C. trachomatis, as well as that of C. pneumoniae.

RT-PCR analysis.

Total RNA from HeLa cells infected with C. trachomatis serovar D was isolated by using Trizol (Life Technologies) and treated with RNase-free DNase (Roche). These RNA samples were used as templates for reverse transcription (RT) in a 20-μl reaction mixture containing 2-μg portions of random hexamers (Roche), 1 μl of 10 mM deoxyribonucleoside triphosphates, 2 μl of 0.1 M dithiothreitol, 1 μl of RNasin (Promega), and 200 U of Superscript II (Life Technologies). RT was carried out at 42°C for 50 min. PCR was performed with PTC-200 Peltier Thermal Cycler (MJ Research). The reaction mixture contained 1.5 mM MgCl₂, 200 μM concentrations of the deoxynucleoside triphosphates, 5 U of Taq DNA polymerase enzyme, 25 pmol of each oligonucleotide primer specific for each gene (Table 1), and 2 μl of RT product in a total volume of 50 μl. The PCR cycling conditions were as follows: one cycle of 95°C for 3 min and 35 cycles of 94°C for 15 s, 55°C for 30 s, and 72°C for 2 min. This step was followed by strand elongation for 10 min at 72°C. No PCR product could be detected in controls that lack reverse transcriptase.

Quantification of gene expression after heat shock by using a DNA microarray. (i) Primer design and PCR.

PCR primers were designed to amplify the full-length open reading frame of 96 selected genes of C. trachomatis serovar D by using a computer algorithm written in our laboratory. The program adjusts the length of the primers to achieve a specified melting temperature. All N-terminal primers contained an “adaptamer” sequence with a KpnI restriction site and the C-terminal primers with a NotI site to facilitate directional cloning in later stages for functional analysis. The PCR amplification was the same as described above except that the final PCR products were ethanol precipitated and air dried.

(ii) Array procedure.

DNA was dissolved in 3× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and divided into aliquots in a 384-well plate. Microarrays were fabricated on CMT-GAPS-coated microscope slides (Corning) with MicroGrid Microarrayer (BioRobotics, Cambridge, United Kingdom).

(ii) Probe preparation.

Total RNA was extracted by using Trizol (Life Technologies) and treated with RNase-free DNase (Roche). Heat shock samples were subjected to a 10-min heat pulse to 45°C prior to RNA isolation. A total of 20 μg of RNA was labeled with cyanine 3 (Cy3)-dUTP or Cy5-dUTP by using Superscript II (Life Technologies) with randon hexamers (Roche).

(iii) Hybridization.

Cy3- and Cy5-labeled cDNA were combined with DIG Easy Hyb solution (Roche), yeast tRNA (Life Technologies), and calf thymus DNA in a final volume of 30 μl and then hybridized to a microarray at 37°C for 18 h. Slides were then washed in 1× SSC with 0.1% sodium dodecyl sulfate (SDS) at 50°C and dried.

(iv) Analysis and quantification.

Arrays were scanned on a Virtek Chipreader version 2.0 (Virtek Vision Corp., Waterloo, Ontario, Canada) and the Cy3 and Cy5 images were combined for quantification by using Imagine software (version 4.0; BioDiscovery, Inc., Los Angeles, Calif.). The quantified data for each spot were analyzed by using GeneSpring software (version 4.1; SiliconGenetics, San Carlos, Calif.). After background subtraction, the data were normalized by using a global normalization technique, and the signal intensities were calculated as the fold change of expression values. This experiment was performed twice with similar results.

Preparation of antisera against recombinant GroELs.

Antisera to GroEL1, GroEL2, and GroEL3 were raised in BALB/c mice (Charles River Canada) by intraperitoneal injection of the recombinant proteins (100 μg of protein in incomplete Freund adjuvant), followed by two booster injections at 2-week intervals. Sera were collected and pooled 4 weeks after the final boost.

Western blotting.

Samples for Western analysis were prepared by boiling purified serovar D C. trachomatis elementary bodies for 5 min in the protein sample buffer. The samples were subjected to SDS-7.5% polyacrylamide gel electrophoresis according to the method of Laemmli (22). The gels were blotted onto nitrocellulose membranes (Bio-Rad) at 70 V for 1 h in blotting buffer acccording to the method of Maniatis et al. (23). The filters were blocked overnight with Tris-buffered saline containing 3% bovine serum albumin at 4°C prior to incubation with polyclonal anti-GroEL antibodies and peroxidase-conjugated sheep anti mouse immunoglobulin G secondary antibodies. The blots were processed for color detection by using the substrate BCIP (5-bromo-4-chloro-3-indolylphosphate)-nitroblue tetrazolium.

E. coli complementation.

groEL1, groEL2, and groEL3 DNA fragments were generated by PCR with genomic DNA isolated from C. trachomatis serovar D and cloned into pBAD24 (16) as a KpnI fragment downstream of an arabinose-inducible promoter to yield the constructs pBADgroEL1, pBADgroEL2, and pBADgroEL3. The groES gene was cloned into pKPK as EcoRI/XhoI fragment downstream of IPTG-inducible lacUV5 promoter to yield pKPKgroES. pKPK was derived from pBT (Stratagene) by deleting the phage λcI gene (as a NotI/PshAI fragment), followed by filling in and ligation. The primers used for these amplifications are listed in Table 1. These plasmid constructs were transformed into E. coli NL441 (25), and transformants were grown at 37°C overnight in Luria broth with ampicillin (100 μg ml⁻¹) and/or chloramphenicol (50 μg ml⁻¹). Cultures were streaked onto two sets of plates containing an appropriate combination of antibiotics and inducers (ampicillin at 100 μg ml⁻¹, chloramphenicol at 50 μg ml⁻¹, and IPTG at 50 μM or as indicated) and various amounts of arabinose as indicated. One set of plates was incubated at 37°C, and the other was incubated at 42°C. Scoring for growth was done after an overnight incubation.

RESULTS

Phylogeny of chlamydia GroELs.

Phylogenetic analysis of the GroEL proteins revealed that the groEL genes probably duplicated before divergence of the members of the Chlamydiaceae family. For each of the three GroELs, the closest protein match is the orthologous chlamydial protein in another species rather than the GroELs within the same genome. This gives rise to the clustering of each of the loci within the Chlamydia branch in the phylogenetic tree (Fig. 1). Additionally, the tree for each GroEL protein within the Chlamydiaceae matches the tree of 16S rRNA evolution (9). Comparison of the chlamydial GroEL proteins with those of other bacteria suggested that GroEL1 may have arisen from a different lineage than GroEL2 and GroEL3 (Fig. 1). Based on analysis of microbes with complete genome sequences, GroEL1 appears to be the most related, with a homolog from the deeply branched eubacterium Aquifex aeliocus, whereas GroEL2 and GroEL3 cluster in an offshoot from GroELs of the Mycoplasma species.

FIG. 1.

Phylogenetic relationships of chlamydial GroELs. The figure shows an unrooted tree of GroEL homologs in 75 completed genomes (see Materials and Methods for details regarding how the tree was constructed). Branches consisting of GroEL1, GroEL2, and GroEL3 cluster from six chlamydial genomes (C. trachomatis serovar D and biovar MoPn; C. pneumoniae strains AR39, CWL029, and J138; and C. psittaci) are indicated by shaded boxes. Bacterial genera at key branch points are also included.

Sequence analysis of chlamydial GroELs.

At the nucleotide level the identity observed for groEL1 among all three chlamydiae was 29 to 33%. However, comparison of groEL2 and groEL3 among chlamydiae revealed only 5 to 16% and 11 to 25% identities, respectively. The identity for E. coli groEL with chlamydial groEL1 varied between 15 and 18% and was even less (5 to 10%) for groEL2 and groEL3. A pairwise comparison of amino acid sequence identity among the GroELs of three different chlamydial species and E. coli GroEL is shown in Table 2. At the amino acid level, GroEL1 was the most conserved GroEL among the chlamydiae (91 to 98%). Although GroEL2 and GroEL3 of serovar D and biovar MoPn are strongly related (71 and 77%), a comparison of serovar D or biovar MoPn with C. pneumoniae reveals lower values (26 to 37% identity). E. coli GroEL, on the other hand, shows 60% identity with GroEL1 of all three chlamydiae but ranges between 19 and 32% in identity for GroEL2 and GroEL3.

TABLE 2.

Percent identity of amino acid sequences of GroEL1, GroEL2, and GroEL3^a

Strain and protein	% Sequence identity
	CtD			AR39			MoPn			E. coli
	1	2	3	1	2	3	1	2	3
CtD
1		23	18	91	33	24	98	24	17	60
2			17	23	27	17	23	71	13	22
3				19	20	37	19	18	77	20
AR39
1					33	24	92	22	18	60
2						22	33	26	19	32
3							24	18	34	20
MoPn
1								24	17	60
2									18	22
3										19

^aAbbreviations: CtD, C. trachomatis serovar D; AR39, C. pneumoniae strain AR39; MoPn, C. trachomatis biovar MoPn. The numbers 1, 2, and 3 represent GroEL1, GroEL2, and GroEL3.

A comparison of the multiple GroELs within each species of chlamydia revealed 17 to 23, 17 to 24, and 22 to 33% amino acid identities for serovar D and for the MoPn biovars of C. trachomatis and C. pneumoniae strain AR39, respectively. These values were even much lower when comparisons were made at the nucleotide level (5 to 11%). Like chlamydiae, other bacteria also contain multiple copies of GroELs (Table 3). Notably, Bradyrhizobium japonicum has five copies of GroELs. Strikingly, the percent identity between paralogous copies of GroEL among other bacteria is much higher (>60%) than that observed for chlamydiae.

TABLE 3.

Identity among multiple copies of GroEL in different bacteria

Organism	No. copies of GroEL	Identity (%)	Source or reference
Bradyrhizobium japonicum	5	80-90a	13
Rhizobium meliloti	3	78-85	32
Sinorhizobium meliloti	2	79	6
Rhodobacter sphaeroides	2	75	19
Synechococcus vulcanus	2	65	35
Synechocystis sp.	2	62	15
Streptomyces lividans	2	64	7
Mycobacterium tuberculosis	2	61	21
Mycobacterium leprae	2	59	31
Chlamydia trachomatis serovarD	3	17-23	This study
Chlamydia trachomatis biovar MoPn	3	17-24	This study
Chlamydia pneumoniae strain AR39	3	22-33	This study

^aThe percent identity was determined using only three of the five copies of GroELs available in GenBank.

Functional regions of E. coli GroEL amino acid sequences that are related to polypeptide binding, GroES contact, and ATP/ADP binding (12) were aligned to the chlamydial GroEL amino acid residues (Table 4). Selected GroEL1 residues of all three chlamydiae matched perfectly, whereas for GroEL2 and GroEL3 only the polypeptide binding site residues are conserved. Residues important for GroES binding and ATPase activity in E. coli GroEL are not well conserved in chlamydial GroEL2 and GroEL3.

TABLE 4.

Alignment of functional regions of E. coli GroEL sequence with those of chlamydial GroEL homologs^a

^aAmino acid residues that are substituted with completely different amino acids are in red, and similar substitutions are in purple. Residues important for polypeptide binding (red asterisks), GroES contact (blue asterisks), and ATPase activity (green asterisks) that were identified in previous studies are marked at the bottom of the alignment (12). The percentage identity (%I) and the percent similarity (%S) were calculated based on the 22 amino acids aligned above. The numbering of the amino acid residues was the same as that one used for E. coli GroEL (12).

^bAbbreviations: Ct1, -2, and -3, C. trachomatis serovar D GroEL1, -2, and -3; Mo1, -2, and -3, C. trachomatis biovar MoPn GroEL1, -2, and -3; Cp1, -2, and -3, C. pneumoniae GroEL1, -2, and -3; Ec, E. coli.

3-D structural conservation of chlamydial GroELs.

A threading approach was used to search for proteins structurally related to the chlamydial GroELs. For the three C. trachomatis serovar D GroEL sequences analyzed by using the THREADER2 package, we obtained highly significant threading Z scores for at least two model folds (1derA0 and 1derB0). Values of Z that were greater than 15 were obtained for all three chlamydial GroEL proteins and identified the protein folds (1derA0 and 1derB0) as belonging to E. coli chaperonin GroEL (2).

Homology modeling of C. trachomatis serovar D GroEL1, GroEL2, and GroEL3 gave the predicted structures shown in Fig. 2A, B, and C, respectively. The estimated 3-D structures of GroEL1 to GroEL3 are very similar, except for the irregular N and C termini. Structural superposition of the predicted structures of C. trachomatis serovar D GroEL1 to GroEL3 proteins yielded the corresponding RMSD (i.e., the root mean square deviation of the atomic coordinates) value of only 0.9 Å (Fig. 2D).

FIG. 2.

Predicted 3-D structures of C. trachomatis serovar D GroEL1 (A), GroEL2 (B), and GroEL3 (C) (see Materials and Methods for details of how the 3-D prediction was done through homology modeling) and the structural superposition of the three predicted structures for GroEL1, GroEL2, and GroEL3 (D) (represented by green, blue, and yellow, respectively).

Expression of groEL genes in vivo. (i) Expression during developmental cycle.

cDNA templates were made by RT with total RNA isolated from HeLa cells 2, 6, 12, 24, and 36 h after infection with C. trachomatis serovar D. These cDNAs were used as templates for subsequent PCR with specific primers to determine the expression levels of the groEL genes at the five time points. All three genes were constitutively expressed at all tested time points (data not shown). However, the expression level of groEL1 was much greater compared to groEL2 or groEL3 (Fig. 3A).

FIG. 3.

(A) The top panel shows control PCR amplification of groEL1, groEL2, and groEL3 with serovar D DNA as the template as a measure to compare primer efficiency. The bottom panel shows an RT-PCR analysis of total RNA from HeLa cells infected with C. trachomatis serovar D isolated at 12 h after infection. (B) The top panel shows a Western blot analysis of C. trachomatis serovar D GroEL1, GroEL2, and GroEL3. Boiled samples of purified serovar D C. trachomatis elementary bodies were subjected to SDS-7.5% polyacrylamide gel electrophoresis, and the blotted nitrocellulose membranes were incubated with polyclonal anti-GroEL1 (first lane), anti-GroEL2 (middle lane), or anti-GroEL3 (third lane) antibodies, followed by incubation with alkaline phosphatase-conjugated secondary antibodies and color detection. In the bottom panel is shown a quantification of GroEL1, GroEL2, and GroEL3 by using the procedure described above except that 1-, 10-, and 100-ng portions of recombinant proteins were loaded per lane.

(ii) Induction of groEL expression.

HeLa cells infected with C. trachomatis serovar D for 18 h were given a 10-min heat pulse at 45°C. Quantification of the mRNA expression levels of groEL1, groEL2, and groEL3 under heat shock and non-heat shock conditions was done by using a microarray procedure. The results (Fig. 4) show a >5-fold increase for groEL1 expression for the heat shock condition compared to the non-heat shock condition. Strikingly, the expression levels for groEL2 and groEL3 did not change after heat shock. There was significant upregulation of two other genes after heat shock: (i) groES, the gene located upstream of groEL1 and probably expressed together as a single transcript, and (ii) dnaK, encoding the 70-kDa heat shock protein homolog of chlamydia. This finding is in agreement with an immunoblot analysis showing elevated DnaK and GroEL1 expression after heat shock (8). None of the other 91 studied genes exhibited upregulated mRNA expression after heat treatment (data not shown).

FIG. 4.

Quantification of C. trachomatis serovar D groEL1, groEL2, and groEL3 expression after heat shock by using microarray. (groES and dnaK genes were included for comparison purposes). Heat-shocked and non-heat-shocked RNA samples were reverse transcribed, labeled with Cy3 or Cy5, and hybridized to the microarray. The fluorescence intensity was measured for each spot and then normalized to the average fluorescence intensity for the entire microarray. Data analysis was performed with the GeneSpring software (see Materials and Methods for more details).

(iii) Expression of recombinant GroEL1, GroEL2, and GroEL3 in E. coli.

We subcloned C. trachomatis serovar D groEL1, groEL2, and groEL3 genes into an expression vector [pET30b(+)] in E. coli. GroEL1 and GroEL3 were expressed well, but the expression of GroEL2 was poor. Codon usage analysis showed that C. trachomatis serovar D GroEL2 had four arginine (AGG) and seven isoleucine (AUA) codons that are rarely used (0.1 in 1,000) in E. coli. Use of a genetically engineered E. coli expression host [BL21 Codonplus(DE3)RIL] containing extra copies of tRNA genes encoding these rare codons allowed the production of sufficient GroEL2 for purification and preparation of antisera.

(iv) Expression of GroEL1, GroEL2, and GroEL3 in C. trachomatis serovar D.

Western blot analysis with antichlamydial GroEL polyclonal antibodies revealed positive bands for all three proteins in an elementary body lysate of C. trachomatis serovar D. All three GroELs were similar in size (ca. 60 kDa), and GroEL1 was more abundant than the other two GroELs, a finding consistent with the mRNA expression levels (Fig. 3B).

Complementation of an E. coli groEL mutant.

To examine whether chlamydial groEL genes are functional, complementation tests were performed with each of the chlamydial groEL genes alone or with each of the chlamydial groEL genes and groES expressed from two different plasmids in the E. coli groEL mutant strain NL441 (25). NL441 is a temperature-sensitive groEL mutant strain that can grow up to 37°C but not at 42°C. The results are summarized in Table 5 and Fig. 5. The E. coli groEL mutant NL441 regained viability at 42°C only with the expression of C. trachomatis groEL1 and groES. The expression of groEL1 alone did not complement the defective E. coli groEL. NL441 transformed with plasmids carrying C. trachomatis groEL2 or groEL3 alone or together with C. trachomatis groES did not regain viability.

TABLE 5.

Summary of results from the complementation experiments with E. coli groEL mutant strain NL441

Strain	Growthc at:
	37°C (no arabinose)	42°C with:
		No arabinose	Arabinose at:
			0.02%	0.1%
NL441a	+	−	−	−
NL441/pBADgroEL1	+	−	−	−
NL441/pBADgroEL2	+	−	−	−
NL441/pBADgroEL3	+	−	−	−
NL441/pBADgroEL1/pKPKgroESb	+	−	(+)	+
NL441/pBADgroEL2/pKPKgroESb	+	−	−	−
NL441/pBADgroEL3/pKPKgroESb	+	−	−	−

^aAgar plates used for the growth of strain NL441 did not contain ampicillin and/or chloramphenicol.

^bPlates used for the growth of strain NL441 containing pKPKgroES were also supplemented with 50 μM IPTG.

^c+, Heavy growth; −, no growth; (+), light growth.

FIG. 5.

Complementation of a temperature-sensitive mutant of E. coli groEL by C. trachomatis groEL1 and groES. Two independent transformants of NL441 strains containing the plasmid pairs indicated were streaked onto Luria-Bertani agar plates supplemented with ampicillin (100 μg/ml), chloramphenicol (50 μg/ml), and arabinose (as indicated) and then incubated at the temperatures indicated for 18 h.

DISCUSSION

We show here that all three GroELs of chlamydiae likely belong, based on sequence and structural analyses, to the same family of proteins, the group I chaperonins. Phylogenetic analysis indicates that GroEL1 may have arisen from a different lineage than did GroEL2 and GroEL3. mRNA expression analysis at different time points of the chlamydial developmental cycle and after heat shock suggest that GroEL1 is regulated differently than GroEL2 and GroEL3. As expected from the close similarity of chlamydial GroEL1 to the E. coli GroEL, C. trachomatis serovar D GroEL1 was expressed in an E. coli mutant and complemented a temperature-sensitive defect in E. coli groEL. To our knowledge, this is the first time a chlamydial GroEL has been shown to be functional in a heterologous system.

As a general guideline, proteins with more than 30% sequence identity are usually similar in structure and function (3). On the other hand, a threshold of 25% identical residues has been used to define a “twilight zone” of sequence similarity, wherein the structural and functional homology between proteins is no longer certain (30). The chlamydial GroEL1, with 60% amino acid identity to E. coli chaperonin GroEL, is clearly a member of the GroEL family. However, chlamydial GroEL2 and GroEL3 fall near the twilight zone threshold value, and therefore speculation on their structural and functional identity based on sequence information alone is uncertain. However, the results of sequence analysis (Table 4) of conserved active-site amino acid residues (with reference to the E. coli GroEL), as well as protein modeling (Fig. 2), reliably group GroEL2 and GroEL3 proteins into the GroEL family.

Phylogenetic analyses of the chlamydial GroEL proteins indicate that GroEL1 may have arisen from a different lineage than did GroEL2 and GroEL3 (Fig. 1). The assignment of GroELs to different branches may be an artifact of phylogenetic reconstruction caused by the large evolutionary distance between chlamydiae and other bacteria and also to the small number of closely related genomes that have been sequenced to date. There are several possible explanations for the divergence among GroELs that may have a significant bearing on the understanding of the nature and function of these proteins. There may have been a period of rapid sequence evolution of GroEL2 and GroEL3 after duplication of the GroELs in a chlamydia-like ancestor; perhaps this change was related to adaptation for distinct cellular functions in a particular niche. Possibly when chlamydia established a niche as an obligate intracellular pathogen, with little opportunity for horizontal gene transfer, additional chaperonin activity was required to compensate for the acquisition of mildly deleterious mutations (10, 27). Since the origin of the Chlamydiaceae, the rates of change appear to have stabilized since each protein is equidistant from the other two. Alternative suggestions are that GroEL2 and GroEL3 may represent an ancestral lineage deleted from most other bacteria (but possibly conserved in Mycoplasma spp.) or that they were acquired by the chlamydiae through horizontal gene transfer in the remote past. We cannot discriminate among these possibilities.

In support of the phylogenetic analysis, groEL1 also differed from groEL2 and groEL3 in response to heat shock. Only the expression of groEL1—and not groEL2 or groEL3—increased when C. trachomatis serovar D was exposed to elevated temperature (Fig. 4). Although the expression of groEL1, groEL2, and groEL3 is constitutive throughout the developmental cycle of C. trachomatis, the level of groEL1 expression was highest (Fig. 3A). This finding was supported by a Western blot analysis (Fig. 3B) in which GroEL1 was found to be substantially more abundant than GroEL2 and GroEL3. Gene expression and protein abundances in bacteria can be regulated by different factors such as promoter strength, transcription factors, mRNA stability, and codon usage. We analyzed the codon usage for GroEL1, GroEL2, and GroEL3 in chlamydia, and our analysis revealed many rare codons in GroEL2, suggesting that the GroELs may be in part translationally regulated. This difference is consistent with the studies of Karlin and Mrazek (20), who carried out a statistical analysis of the expression of many bacterial proteins, including chlamydial GroEL1, GroEL2, and GroEL3. On the basis of codon usage, GroEL1, but not GroEL2 and GroEL3 was predicted to be a highly expressed gene. We also analyzed the potential promoter regions of groEL1, groEL2, and groEL3 in C. trachomatis serovar D, and only the groE operon (groES and groEL1) was found to contain the CIRCE (controlling inverted repeat of chaperone expression), as well as the putative σ⁶⁶ promoter element, as described for C. trachomatis MoPn strain (34). These data suggest that GroEL2 and GroEL3 are likely to have different regulatory mechanisms than does GroEL1.

Complementation of an E. coli groEL mutant by C. trachomatis groEL1 and groES genes together, but not with groEL1 alone, indicates that E. coli groES and chlamydial groEL are not functionally compatible. A similar phenomenon was also reported for Vibrio cholerae (26). In contrast, C. trachomatis groEL2 and groEL3, together with groES, were not able to complement the E. coli groEL mutant. Thus, even though GroEL2 and GroEL3 are expressed as proteins in chlamydiae and therefore are not pseudogenes, their physiologic role remains unclear.

In conclusion, we determined the grouping of the three chlamydial groEL genes within the groEL family based on sequence and structural properties. Transcriptional and translational expression analyses showed that all three genes are expressed in chlamydia and that they are not pseudogenes. Phylogenetic analysis, heat shock, and complementation assays indicated that groEL2 and groEL3 may have functions distinct from groEL1 and are differentially regulated. The low sequence similarity among groEL1, groEL2, and groEL3 compared to the level of identity reported for bacterial groELs in general also supports this speculation (Table 3). If chlamydial GroEL proteins play distinct roles in the accurate expression of information encoded in the genome, it will be of interest to determine whether each chaperonin plays a different role in the development cycle and in disease pathogenesis. Furthermore, it will be of interest to determine whether the chlamydial GroELs form homo- or heteropolymeric structures with variable stoichiometries in these processes.