• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. May 2006; 188(9): 3415–3419.
PMCID: PMC1447467

Functional Characterization of Escherichia coli GlpG and Additional Rhomboid Proteins Using an aarA Mutant of Providencia stuartii


The Providencia stuartii AarA protein is a member of the rhomboid family of intramembrane serine proteases and required for the production of an extracellular signaling molecule that regulates cellular functions including peptidoglycan acetylation, methionine transport, and cysteine biosynthesis. Additional aarA-dependent phenotypes include (i) loss of an extracellular yellow pigment, (ii) inability to grow on MacConkey agar, and (iii) abnormal cell division. Since these phenotypes are easily assayed, the P. stuartii aarA mutant serves as a useful host system to investigate rhomboid function. The Escherichia coli GlpG protein was shown to be functionally similar to AarA and rescued the above aarA-dependent phenotypes in P. stuartii. GlpG proteins containing single alanine substitutions at the highly conserved catalytic triad of asparagine (N154A), serine (S201A), or histidine (H254A) residues were nonfunctional. The P. stuartii aarA mutant was also used as a biosensor to demonstrate that proteins from a variety of diverse sources exhibited rhomboid activity. In an effort to further investigate the role of a rhomboid protein in cell physiology, a glpG mutant of E. coli was constructed. In phenotype microarray experiments, the glpG mutant exhibited a slight increase in resistance to the β-lactam antibiotic cefotaxime.

In the gram-negative pathogen Providencia stuartii, the AarA protein is a member of the rhomboid family of intramembrane serine proteases that are widely distributed in prokaryotes and eukaryotes (6, 7, 13, 15, 19, 20). AarA is required for the production of an extracellular signaling molecule that activates gene expression in a manner that may involve quorum sensing (4, 5, 12, 13-15). The first rhomboid protein, rhomboid-1, was identified in Drosophila melanogaster as a locus required for pattern formation in the ventral ectoderm. The active site serine of rhomboid-1 is within the membrane bilayer and allows rhomboid-1 to carry out regulated intramembrane proteolysis (1, 19). The rhomboid-1-mediated cleavage of the membrane protein Spitz, Gurken, or Keren releases a cleavage product that serves as a ligand for the epidermal growth factor receptor (19, 20). Rhomboid proteins can recognize substrates via helix-breaking residues in the transmembrane domain or by recognition of sequences in the cytoplasmic domain (9, 21).

Studies have demonstrated that the rhomboid-1 protein of Drosophila and the AarA protein of P. stuartii can functionally substitute for each other (7). Proteins related to AarA/rhomboid are widespread in bacteria, suggesting that important functions may be dependent on these serine proteases (7, 20). Homology comparisons with AarA suggest that the Escherichia coli GlpG protein is a member of the rhomboid family. Despite this similarity, the chromosomal context of these genes is quite different. The glpG gene is part of the glpEGR operon (17, 23). In contrast, aarA appears to be monocistronic, with the surrounding genes having no similarity to glpEGR (13). The glpE gene encodes a sulfur transferase that transfers sulfane from thiosulfate to cyanide and dithiols (16). The glpR gene encodes a repressor that regulates various members of the glp regulon in the absence of glycerol (17). With respect to GlpG, the native substrates and its role in E. coli are unknown (10, 23). Expression of GlpG in eukaryotic CHO cells resulted in cleavage of the rhomboid substrates Spitz, Gurken, and Keren (20), and purified GlpG exhibited proteolytic activity in vitro (8, 10, 22). Recent studies in E. coli have shown that GlpG can cleave an artificial substrate composed of a periplasmic β-lactamase fused to a transmembrane region from LacY (10).

The GlpG protein of E. coli can functionally replace AarA in P. stuartii.

To determine whether GlpG or another E. coli protein could function as a rhomboid protein and substitute for AarA, the P. stuartii strain XD37.A (cma37::lacZ ΔaarA) was electroporated with an E. coli (PB103) genomic library of 2- to 5-kb partial Sau3A fragments in pET21a that was obtained from P. deBoer (Case Western Reserve University). Plasmids that complemented the aarA mutation were identified as colonies with restored production of an extracellular yellow pigment. One class of inserts contained overlapping fragments of the glpG region of the chromosome. One plasmid, pLibG, was used for further studies and contained a 3.8-kb insert with the entire glpG gene. A library of random Tn7Cm insertions in pLibG was constructed in vitro using a Tn7Cm transposon contained in the GPS-LS Linker Scanning System (New England Biolabs). A pool of random insertions was prepared in E. coli, and the resulting plasmid DNA was used to transform P. stuartii XD37.A (ΔaarA) by electroporation. Plasmids containing Tn7Cm insertions that inhibited the ability of pLibG to complement the aarA allele in P. stuartii XD37.A (ΔaarA) were identified as transformants that failed to produce pigment. Three noncomplementing plasmids were isolated, and DNA sequence analysis using primers that read outward from the end of Tn7Cm indicated that the transposon had inserted into the glpG coding sequence in all three plasmids.

The ability of glpG to complement several additional aarA-dependent phenotypes was investigated. First, the production of extracellular activating signal was tested using an aarA-dependent cma37::lacZ reporter gene fusion (Fig. (Fig.1).1). Conditioned medium was prepared from P. stuartii strain XD37.A containing pLibG or from XD37.A containing the vector control pET21a. These plasmids are unstable in the absence of selection, and LB broth was supplemented with ampicillin at 100 μg/ml to maintain pET21a and pLibG. The conditioned medium prepared from P. stuartii XD37.A/pLibG exhibited significant production of extracellular signal with a sixfold activation of the cma37::lacZ reporter gene fusion, compared to the activation observed with P. stuartii XD37.A/pET21a containing only the vector. For comparison purposes, conditioned medium prepared from wild-type P. stuartii XD37/pET21a under the same conditions activated the cma37::lacZ reporter gene fusion 31-fold and XD37.A ΔaarA containing the aarA gene in a high copy number activated the cma37::lacZ fusion 47-fold (Fig. (Fig.1).1). This difference in conditioned medium activity may result from less efficient rhomboid activity of GlpG in P. stuartii relative to the native AarA protein. The ability of the E. coli glpG gene to rescue additional phenotypes that resulted from the loss of aarA in P. stuartii was also examined. The chain-forming phenotype of P. stuartii XD37.A was rescued by the glpG gene, as was pigment production and the ability of P. stuartii XD37.A to grow on MacConkey medium (Table (Table11).

FIG. 1.
Ability of GlpG and various mutants to restore signal production to a P. stuartii aarA mutant. Conditioned medium was prepared at an optical density of A600 = 1.0 from P. stuartii strain XD37 wild type and XD37.A (ΔaarA) containing pET21a ...
E. coli glpG can complement an aarA mutation in P. stuartii

Mutagenesis of glpG.

The rhomboid family of proteins contains three highly conserved residues that are proposed to form a catalytic triad for protease activity (19). In GlpG, these residues are Asn154, Ser201, and His254. Single alanine substitutions were made at these residues in the GlpG protein using plasmid pLibG as a template and the QuikChange mutagenesis system (Stratagene, La Jolla, CA). DNA sequence analysis of the entire glpG gene confirmed that only the desired change was present in each mutant.

Each of the glpG mutant constructs containing N154A, S201A, and H254A was unable to restore signal production to the aarA mutant based on activation of the cma37::lacZ reporter gene fusion (Fig. (Fig.1).1). Each of the mutant glpG genes was also unable to rescue the cell division defect, growth on MacConkey plates, and the lack of pigment production in aarA mutant P. stuartii (Table (Table11).

Use of a P. stuartii aarA mutant to test rhomboid activity from diverse organisms.

Previous studies have demonstrated that expression of various prokaryotic rhomboids in eukaryotic CHO cells resulted in cleavage of the Drosophila rhomboid substrates Spitz, Gurken, and Keren (20). However, the ability of these rhomboid proteins to function in a prokaryote has not been tested. The use of the P. stuartii aarA mutant provides a powerful screening approach to identify proteins from other bacteria that have rhomboid-like activity. Rhomboid genes from Pseudomonas aeruginosa (NP251776), Bacillus subtilis (NP390367), Aquifex aeolicus (NP213910), Methanococcus jannaschii (NP247593), Pyrococcus horikoshii (NP143361), Streptococcus pyogenes (NP268586), and RHBDL2, a human rhomboid (NM017821), were introduced into P. stuartii XD37.A (ΔaarA) and tested for rhomboid activity by restored production of the AarA-dependent extracellular activating signal. Table Table22 shows the ability of conditioned medium from P. stuartii XD37.A containing various plasmid-encoded rhomboids to activate the cma37::lacZ reporter gene fusion. All the proteins examined exhibited various degrees of rhomboid activity based on restoration of aarA mutant phenotypes, with the P. aeruginosa rhomboid exhibiting the strongest activity based on the 14-fold activation of the cma37::lacZ reporter gene fusion (Table (Table2).2). Next, the rhomboid protein YqgP (GluP) from Bacillus subtilis in XD37.A restored signal production, as evidenced by the sixfold activation of cma37::lacZ. Interestingly, the human rhomboid RHBDL2 exhibited significant activity, with a fivefold activation of cma37::lacZ by conditioned medium from this strain. The ability of these rhomboid proteins to restore the other AarA-dependent phenotypes, such as loss of pigment and growth on MacConkey agar, was also examined. Pigment production was restored to various degrees and correlated well with rhomboid activity based on extracellular signal production (Table (Table2).2). However, the rescue of growth on MacConkey agar was observed only with rhomboid proteins from P. aeruginosa and B. subtilis, both of which appeared to have the strongest activity based on restoration of signal activity and pigment production (Table (Table22).

Rhomboid activity from diverse organisms

Phenotype microarray analysis of a glpG mutant.

To address the role of glpG in E. coli, the PCR-mediated allelic replacement procedure of Datsenko and Wanner was used to construct a glpG::cat allele that disrupted glpG at position 67 of the 276-amino-acid protein (3). This insertion point was chosen because it was known to inactivate function based on Tn7 insertions that disrupted glpG function on a plasmid, and it was upstream of the two promoters within glpG that transcribe the downstream glpR gene and would not be polar (23). The resulting PCR fragment was used to electroporate E. coli MG1655/pKD46 carrying genes for an arabinose-inducible, lambda red recombination system (3). Transformants were selected on LB medium containing chloramphenicol (25 μg/ml). The glpG::cat allele was verified by PCR using the glpG ORFmers (Sigma-Genosys) and also by Southern blot analysis.

The role of GlpG in E. coli was investigated by the use of phenotype MicroArrays (Biolog, Hayward, CA). These experiments were conducted by Michael Zimon at Biolog. A comprehensive set of 20 plates was used to test metabolic differences with respect to growth under a variety of conditions and sensitivities to a variety of compounds. The only parameter that was verified by our lab as consistently different in the glpG mutant was an increased resistance to cefotaxime, a β-lactam antibiotic. These results were followed up by testing the levels of resistance by Kirby-Bauer disk diffusion assays according to CLSI (formerly NCCLS) guidelines (2). These experiments were repeated in triplicate with all experiments exhibiting the pattern of resistance. The cefotaxime zone diameter for wild-type E. coli MG1655 was 32.5 ± 0.7 mm, and for the glpG::cat mutant it was 30.5 ± 0.7 mm. The zone diameters for ampicillin and ceftriaxone (additional β-lactams) and for the structurally unrelated antibiotics ciprofloxacin (a fluoroquinolone) and amikacin (an aminoglycoside) were the same for both wild-type E. coli MG1655 and the glpG mutant.

We could find no additional phenotypes resulting from the glpG::cat mutation with respect to colony morphology, growth at 30 to 42°C, and growth on minimal medium with either glucose or glycerol.

glpG is not required for extracellular signal production in E. coli.

In a previous study, the ability of E. coli to produce a factor biologically similar to that of P. stuartii was demonstrated by the ability of conditioned medium from E. coli to activate the aarA-dependent cma37::lacZ fusion in P. stuartii (18). Signal production was examined in conditioned medium prepared from six cultures each from an independent glpG::cat mutant and six colonies of wild-type MG1655. The degree of cma37::lacZ activation in the P. stuartii biosensor varied from 18- to 29-fold with wild-type MG1655, with an average of 24-fold (±4-fold). For the glpG::cat mutants, the degree of activation ranged from 15- to 39-fold (average, 27-fold ± 10-fold). The basis for the high variability is unknown. However, the glpG::cat allele does not appear to significantly alter signal production. Moreover, the frequencies of transduction of the glpG::cat allele into E. coli MG1655/pET21a or MG1655/pLibG were similar, and these transductants exhibited the same variability in signal production (data not shown).

To rule out the possibility that the inability to detect a role for glpG in signal production was due to its lack of expression under laboratory growth conditions, we performed Northern blot analysis of glpG mRNA in cells at mid-log phase in LB-only medium and LB supplemented with 0.4% glycerol (Fig. (Fig.2).2). The glpG mRNA was clearly detectable, and the levels were similar in LB with and without glycerol. The accumulation of glpG mRNA was also examined in M9 salts containing glucose (0.2%) or glycerol (0.4%). The levels of glpG mRNA were similar under each condition, although the absolute levels were lower than in cells grown in LB (Fig. (Fig.2).2). As a control, the levels of mRNA for the housekeeping gene secD did not significantly vary under the growth conditions tested (Fig. (Fig.22).

FIG. 2.
Accumulation of glpG mRNA under various growth conditions. Cells of MG1655 were grown to mid-log phase (OD600 = 0.6) under the following growth conditions: LB, LB plus 0.4% glycerol, M9 plus 0.2% glucose, and M9 plus 0.4% glycerol. Total RNA was ...

The effect of glycerol on signal activity in conditioned medium was examined from cells grown in LB containing either glucose (0.2%) or glycerol (0.4%). For these experiments, conditioned medium was harvested at an optical density at 600 nm (OD600) of 1.0. In addition, cells grown in conditioned medium were harvested at an OD600 of 0.4, a point where the optimal response to the extracellular signal has been observed previously (18). The activation of cma37::lacZ by conditioned medium prepared from LB plus 0.4% glycerol was 1.3-fold (±0.03-fold) above the levels seen in LB-only-grown cells (data not shown). The addition of glucose had no effect. To rule out the possibility that residual glycerol from conditioned medium stimulated expression, 0.4% glycerol was added to LB-only medium and β-galactosidase expression from cma37::lacZ was examined in cells grown to an OD600 of 0.4. There was no effect of glycerol on cma37::lacZ expression (data not shown).

Concluding remarks.

The results from this study provide further evidence that GlpG in E. coli functions as a rhomboid-like protein. First, GlpG was functionally similar to AarA, a protein previously shown to function as a rhomboid-like serine protease (7, 20). GlpG was capable of complementing the phenotypes in aarA mutants of P. stuartii including (i) loss of production of an extracellular activating signal, (ii) defective cell division, (iii) loss of an extracellular yellow pigment, and (iv) growth on MacConkey agar. Rhomboid proteins contain conserved asparagine, serine, and histidine residues that comprise a catalytic triad for serine proteases. Single alanine substitutions at these residues in GlpG completely abolished function based on complementation of the above phenotypes (Table (Table11 and Fig. Fig.1).1). This result differs from that reported by Meagawa at al., where GlpG containing an N254A substitution still exhibited activity (10). This difference may reflect the use of different substrates and organisms. For example, in our study, GlpG function was addressed in P. stuartii, not in E. coli. Interestingly, in eukaryotic CHO cells, GlpG N254A was also nonfunctional (20); however, the purified GlpG N254A was active in vitro (8, 10).

In E. coli, a chromosomal glpG::cat null allele resulted in only one detectable phenotype, a slightly increased resistance to cefotaxime. Presently, there are only two examples where a function for the rhomboid family of proteins has been identified in prokaryotes. In addition to the AarA-dependent phenotypes in P. stuartii (13, 15), the GluP (formerly YqgP) protein in B. subtilis is required for normal cell division and glucose export (11). Additional studies to identify functions for GlpG in E. coli under conditions that differ from those of typical laboratory growth are under way. Interestingly, in addition to GlpG, clinical E. coli isolates contain a second protein (examples include accession numbers ZP00704902 in E. coli HS and NP752671 in E. coli CFT073) with features that strongly suggest rhomboid activity. These features include a conserved GASG active site embedded within a membrane bilayer and conserved triad residues asparagine and histidine at positions similar to those of other rhomboids. This protein is missing in K-12 isolates, and its presence in these clinical isolates may indicate a role in virulence.

In summary, this study demonstrates the utility of the P. sturtii aarA mutant as a biosensor strain to assess rhomboid activity and conduct structure-function studies. This strain will be useful in identifying or verifying new rhomboids from both prokaryotic and eukaryotic genomes. For example, genomic libraries can be introduced into a P. stuartii XD37.A ΔaarA cma37::lacZ strain and plasmids encoding strong rhomboid activity can be directly selected by growth on MacConkey plates. Plasmids encoding rhomboids with weaker activity can be identified by restored pigment production. In addition, inhibitors of rhomboid activity can be identified using the wild-type P. stuartii strain XD37 cma37::lacZ by sensitivity to growth on MacConkey agar, inhibition of pigment production, or decreased extracellular signal production. Specific inhibitors of rhomboid proteins from other sources can be addressed by placing the respective genes in P. stuartii XD37.A ΔaarA cma37::lacZ and then screening for inhibition based on the above phenotypes.


We thank Matthew Freeman for the gift of rhomboid genes from various organisms. The technical assistance of Julian Wong is acknowledged.

This work was supported by a National Science Foundation award, MCB 0406047, to P.N.R.


1. Brown, M. S., J. Ye, R. B. Rawson, and J. L. Goldstein. 2000. Regulated intramembrane proteolysis: a control mechanism conserved from bacteria to humans. Cell 100:391-398. [PubMed]
2. CLSI. 2005. Performance standards for antimicrobial susceptibility testing. Fifteenth informational supplement 25, p. 34. CLSI, Wayne, Pa.
3. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97:6640-6645. [PMC free article] [PubMed]
4. Dunny, G. M., and B. A. Leonard. 1997. Cell-cell communication in gram-positive bacteria. Annu. Rev. Microbiol. 51:527-564. [PubMed]
5. Fuqua, C., S. C. Winans, and E. P. Greenberg. 1996. Census and consensus in bacterial ecosystems: the LuxR-LuxI family of quorum sensing transcriptional regulators. Annu. Rev. Microbiol. 50:727-751. [PubMed]
6. Gallio, M., and P. Kylsten. 2001. Providencia may help find a function for a novel, widespread protein family. Curr. Biol. 10:693-694. [PubMed]
7. Gallio, M., G. Sturgill, P. Rather, and P. Kylsten. 2002. A conserved mechanism for extracellular signaling in eukaryotes and prokaryotes. Proc. Natl. Acad. Sci. USA 99:12208-12213. [PMC free article] [PubMed]
8. Lemberg, M. K., J. Menendez, A. Misik, M. Garcia, C. M. Koth, and M. Freeman. 2005. Mechanism of intramembrane proteolysis investigated with purified rhomboid proteases. EMBO J. 24:464-472. [PMC free article] [PubMed]
9. Lohi, O., S. Urban, and M. Freeman. 2004. Diverse substrate recognition mechanisms for rhomboids; thrombomodulin is cleaved by mammalian rhomboids. Curr. Biol. 14:236-241. [PubMed]
10. Maegawa, S., K. Ito, and Y. Akiyama. 2005. Proteolytic action of GlpG, a rhomboid protease in the Escherichia coli cytoplasmic membrane. Biochemistry 44:13543-13552. [PubMed]
11. Mesak, L. R., F. M. Mesak, and M. K. Dahl. 2004. Expression of a novel gene, gluP, is essential for normal Bacillus subtilis cell division and contributes to glucose export. BMC Microbiol. 4:1-16. [PMC free article] [PubMed]
12. Miller, M. B., and B. L. Bassler. 2001. Quorum sensing in bacteria. Annu. Rev. Microbiol. 55:165-199. [PubMed]
13. Rather, P. N., and E. Orosz. 1994. Characterization of aarA, a pleiotrophic negative regulator of the 2′-N-acetyltransferase in Providencia stuartii. J. Bacteriol. 176:5140-5144. [PMC free article] [PubMed]
14. Rather, P. N., M. M. Parojcic, and M. R. Paradise. 1997. An extracellular factor regulating expression of the chromosomal aminoglycoside 2′-N-acetyltransferase of Providencia stuartii. Antimicrob. Agents Chemother. 41:1749-1754. [PMC free article] [PubMed]
15. Rather, P. N., X. Ding, R. R. Baca-DeLancey, and S. Siddiqui. 1999. Providencia stuartii genes activated by cell-cell signaling and identification of a gene required for the production or activity of an extracellular factor. J. Bacteriol. 181:7185-7191. [PMC free article] [PubMed]
16. Ray, W. K., G. Zeng, M. B. Potters, A. M. Mansuri, and T. J. Larson. 2000. Characterization of a 12-kilodalton rhodanese encoded by glpE of Escherichia coli and its interaction with thioredoxin. J. Bacteriol. 182:2277-2284. [PMC free article] [PubMed]
17. Schweizer, H., W. Boos, and T. J. Larson. 1985. Repressor for the sn-glycerol-3-phosphate regulon of Escherichi coli K-12: cloning of the glpR gene and identification of its product. J. Bacteriol. 161:563-566. [PMC free article] [PubMed]
18. Sturgill, G., C. M. Toutain, J. Komperda, G. A. O'Toole, and P. N. Rather. 2004. Role of CysE in production of an extracellular signaling molecule in Providencia stuartii and Escherichia coli: loss of cysE enhances biofilm formation in Escherichia coli. J. Bacteriol. 186:7610-7617. [PMC free article] [PubMed]
19. Urban, S., J. R. Lee, and M. Freeman. 2001. Drosophila Rhomboid-1 defines a family of putative intramembrane serine proteases. Cell 107:173-182. [PubMed]
20. Urban, S., D. Schlieper, and M. Freeman. 2002. Conservation of intramembrane proteolytic activity and substrate specificity in prokaryotic and eukaryotic rhomboids. Curr. Biol. 12:1507-1512. [PubMed]
21. Urban, S., and M. Freeman. 2003. Substrate specificity of rhomboid intramembrane proteases is governed by helix-breaking residues in the substrate transmembrane domain. Mol. Cell 11:1425-1434. [PubMed]
22. Urban, S., and M. S. Wolfe. 2005. Reconstitution of intramembrane proteolysis in vitro reveals that pure rhomboid is sufficient for catalysis and specificity. Proc. Natl. Acad. Sci. USA 102:1883-1888. [PMC free article] [PubMed]
23. Yang, B., and T. J. Larson. 1998. Multiple promoters are responsible for transcription of the glpEGR operon of Escherichia coli K-12. Biochim. Biophys. Acta 1396:114-126. [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...