Chlamydia trachomatis protein GrgA activatestranscription by contacting the nonconservedregion of σ66Xiaofeng Baoa, Bryce E. Nickelsb, and Huizhou Fana,1

aDepartment of Pharmacology, Robert Wood Johnson Medical School, University of Medicine and Dentistry of New Jersey, Piscataway, NJ 08854;and bDepartment of Genetics and Waksman Institute, Rutgers University, Piscataway, NJ 08854

Edited by Seth A. Darst, The Rockefeller University, New York, NY, and approved August 27, 2012 (received for review May 17, 2012)

The bacterial RNA polymerase holoenzyme consists of a catalyticcore enzyme in complex with a σ factor that is required for pro-moter-specific transcription initiation. Primary, or housekeeping,σ factors are responsible for most of the gene expression that occursduring the exponential phase of growth. Primary σ factors sharefour regions of conserved sequence, regions 1–4, which have beenfurther subdivided. Many primary σ factors also contain a noncon-served region (NCR) located between subregions 1.2 and 2.1, whichcan vary widely in length. Interactions between the NCR of theprimary σ factor of Escherichia coli, σ70, and the β′ subunit of theE. coli core enzyme have been shown to influence gene expression,suggesting that the NCR of primary σ factors represents a potentialtarget for transcription regulation. Here, we report the identifica-tion and characterization of a previously undocumented Chlamydiatrachomatis transcription factor, designated GrgA (general regulatorof genes A). We demonstrate in vitro that GrgA is a DNA-bindingprotein that can stimulate transcription from a range of σ66-depen-dent promoters. We further show that GrgA activates transcriptionby contacting the NCR of the primary σ factor of C. trachomatis, σ66.Our findings suggest GrgA serves as an important regulator ofσ66-dependent transcription in C. trachomatis. Furthermore, becauseGrgA is present only in chlamydiae, our findings highlight how non-conserved regions of the bacterial RNA polymerase can be targets ofregulatory factors that are unique to particular organisms.

Chlamydiae are obligate intracellular bacterial parasites, andtheir hosts range from single cellular eukaryotes to humans

(1). In humans, Chlamydia trachomatis is the most commonsexually transmitted bacterial pathogen (2). C. trachomatis isknown for its ability to efficiently ascend from the lower genitaltract to the upper genital tract, where chlamydial replicationcan have many devastating consequences, including infertilityand pelvic inflammatory disease in women. In addition, someC. trachomatis serovars cause eye infection, which is still a ma-jor cause of blindness in underdeveloped countries (2).Chlamydiae have a unique developmental cycle with two alter-

nating cellular forms. The metabolically inactive, infectious ele-mentary body (EB) enters a vacuole through host cell endocytosis(3). Inside the vacuole (termed inclusion), the EB develops into theproliferating but noninfectious reticulate body (RB). As RBs ac-cumulate, they reorganize, in an asynchronous manner, back toEBs, which exit the host cell at the end of the developmental cycle(4). Whereas a typical developmental cycle ranges from 2 to 4 d,the infection may enter a latent state, which is characterizedby accumulation of aberrant reticulate bodies and a lack of EBproduction (5).The C. trachomatis genome is ∼1 Mb in size and encodes

∼1,000 genes (6, 7). Genome-wide microarray analyses haverevealed that ∼80% of all genes are expressed a few hours afterinfection through the remaining developmental cycle. For theremaining genes, some are expressed immediately after chla-mydial entry into the cell, and others are not transcribed untila middle or late stage. (8, 9). Furthermore, certain gene tran-scripts exhibit increases or decreases in abundance as chlamydiae

enter latency (10). These observations suggest that alterations ingene expression manifest at the level of transcription contributeto the development of latency. Nevertheless, due to the limitednumber of genetic tools available in chlamydiae, only a fewtranscription regulators have been identified.Here we describe the identification and characterization of a

previously undocumented transcription factor that we call GrgA(general regulator of genes A). GrgA was identified based upon itsability to activate transcription in vitro from the promoter thatcontrols the expression of defA, which encodes the peptide defor-mylase (PDF). We demonstrate that efficient transcription activa-tion of the defA promoter by GrgA in vitro requires contact withboth DNA and a portion of a nonconserved region (NCR) of theprimary σ factor of C. trachomatis, σ66. We further show that GrgAcan stimulate σ66-dependent transcription in vitro of three othergenes that are expressed in vivo during different stages of thechlamydial developmental cycle. Our findings suggest that GrgA isnot only a regulator of defA expression in vivo, but also a generaltranscription activator of many σ66-dependent genes. Further-more, because GrgA is present only in chlamydiae, our findingshighlight how regulatory factors that are unique to particularorganisms can target nonconserved regions of the bacterialRNA polymerase (RNAP).

ResultsIdentification of GrgA as a defA Promoter-Binding Protein withTranscription Activation Activity. PDF catalyzes the removal of theN-formyl group from the leading methionine of newly synthesizedproteins. Most bacterial proteins require the deformylation andsubsequent removal of the N-terminal methionine to functionproperly. Furthermore, even for the small proportion of proteinsthat can function while carrying the N-formyl methionine, defor-mylation is necessary for the initiation of regulated degradation(11). Thus, PDF is a potential therapeutic and/or prophylactictarget for infectious diseases. Accordingly, small inhibitors of PDFhave shown effectiveness against a variety of pathogens, includingchlamydiae both in vitro and in vivo (12, 13).Our goal was to identify additional chlamydial proteins that

might serve as potential therapeutic targets. Given that smallmolecules targeting PDF inhibit chlamydial growth, we sought toidentify factors involved in the expression of defA, which mightthemselves serve as potential targets. To accomplish this goal, weused a DNA pull-down assay to identify chlamydial proteins thatbound to a DNA fragment containing the defA promoter. To dothis, we mixed cell extracts isolated from C. trachomatis-infected

Author contributions: X.B., B.E.N., and H.F. designed research; X.B. performed research;X.B., B.E.N., and H.F. analyzed data; and B.E.N. and H.F. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: fanhu@umdnj.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1207300109/-/DCSupplemental.

16870–16875 | PNAS | October 16, 2012 | vol. 109 | no. 42 www.pnas.org/cgi/doi/10.1073/pnas.1207300109

mouse L cells with streptavidin beads attached to either a bio-tinylated DNA fragment that carried sequences extending fromposition –144 to +52 of the defA promoter (pdflong) or a bio-tinylated DNA fragment that carried sequences from position –54to +52 (pdfshort). We then analyzed proteins associated with eitherthe beads bound to the pdflong fragment or the beads bound to thepdfshort fragment by liquid chromatography (LC)–MS/MS. Fromthis analysis we identified C. trachomatis proteins with no pre-viously assigned function that were enriched when the pull-downassay was performed with the pdflong fragment compared with thepdfshort fragment.We focused our attention on the Chlamydia-specific hypo-

thetical protein that showed the highest enrichment in bindingthe pdflong fragment compared with the pdfshort. This protein wasencoded by an ORF designated CTL0766 in the genome ofC. trachomatis L2. BLAST analysis did not detect significanthomology between the CTL0766 protein and any nonchlamydialproteins, and no bacterial protein motif was identified in a motifsearch. We expressed and purified a His-tagged derivative ofCTL0766 in Escherichia coli (Table S1 and Fig. S1 A and B) andtested its potential role in the regulation of defA gene tran-scription using an in vitro transcription assay (14) (Fig. 1 and Fig.S2). Addition of CTL0766 stimulated transcription of the defApromoter by chlamydial RNA polymerase (cRNAP) in a dose-dependent manner from the wild-type promoter (Fig. 1A) as wellas two mutant derivatives (Z100 and GR10) (14), which containsingle base-pair substitutions that increase basal transcription(Fig. 1 B and C and Fig. S2). These results demonstrate that thegene product of CTL0766 can activate transcription of the defAgene in vitro. As shown below, CTL0766 also acts as an activatorof three other chlamydial genes tested. Accordingly, we renamethe C. trachomatis gene CTL0766 grgA (general regulator ofgenes A).

GrgA Binds to DNA in a Sequence-Nonspecific Manner. To charac-terize the nature of the interaction between GrgA and the defApromoter, we first performed an EMSA using purified GrgA anda 32P-labeled promoter fragment extending from position –144 to+52 of the defA promoter. Surprisingly, initial experimentsdetected no clear shifted band (Fig. S3A), even though the samesequence of the defA promoter was used for initial pull-down

experiment that led to the identification of GrgA (see above). Wereasoned that the inclusion of poly(deoxyinosinic-deoxycytidylic)acid [poly(dI-dC)] in the reaction might have prevented GrgAfrom binding the defA promoter fragment. Consistent with thishypothesis, dose-dependent retardation of the defA promoterfragment by GrgA was readily detectable in the absence of poly(dI-dC) (Fig. 2A). However, GrgA-bound DNA did not migrate asa clear band in the resolving gel. At lower concentrations (0.2 and0.3 μM) of GrgA the DNA appeared as smear in the gel, whereasat higher concentrations of GrgA (≥0.6 μM) the protein–DNAcomplex largely or completely remained in the loading well (Fig.2A). We further performed GrgA pull-down assays to determinewhat effect shortening the DNA fragment carrying the defA pro-moter had on the binding of GrgA. Accordingly, we generatedbiotinylated defA promoter fragments of different lengths (Fig.2B) that were immobilized to streptavidin-conjugated agarosebeads and determined the amount of GrgA that precipitated withequimolar amounts of each DNA fragment. Progressive removalfrom the 5′-end of the promoter resulted in a steady decrease inthe amount of GrgA that was precipitated (Fig. 2 C and D), re-vealing a clear correlation between the length of the promoterDNA fragment and the amount of precipitated GrgA. Thesefindings, which are consistent with the enrichment of GrgA asso-ciation with the pdflong fragment compared with the pdfshortfragment as detected by MS (see above), suggest that GrgA doesnot bind preferentially to defA promoter sequences, but ratherbinds DNA in a sequence-nonspecific manner. Consistent withthis hypothesis, replacement of the sequences extending from –144to +5 of the defA promoter fragment with unrelated DNA se-quence had no effect on the amount of GrgA precipitated (Fig. S3B and C). Furthermore, we found that GrgA could bind to DNAlocated in the 5′-untranslated region of the defA gene and to twofragments containing portions of the GrgA ORF (Fig. S3 D andE). Taken together, data presented in Fig. 2 and Fig. S3 indicatethat GrgA binds DNA in a sequence-nonspecific manner.

GrgA-Dependent Transcription Activation Requires DNA Binding. Wenext pursued the identification of amino acid residues in GrgAthat were required for sequence-nonspecific DNA binding. Weconstructed a series of GrgA mutants with deletions of 50–82amino acids (Table S1 and Fig. S4 A and B) and assessed thesemutants’ abilities to associate with defA promoter DNA usingboth a pull-down assay (Fig. S4C) and an EMSA assay (Fig. 3A).GrgA mutants lacking amino acids 1–64 (Δ1–64), 65–113 (Δ65–113), 166–206 (Δ166–206), or 207–288 (Δ207–288) all exhibitedDNA binding activities that were identical to that of the wild-type GrgA protein. In contrast, the GrgA mutant lacking aminoacid residues 114–165 (Δ114–165) was severely compromised inits ability to bind DNA (Fig. 3A and Fig. S4C), suggesting thatresidues 114–165 comprise a portion of GrgA that is essential forsequence-nonspecific DNA binding. Noticeably, this region isrich in positively charged lysine and arginine residues (Fig. S4D),which may mediate the binding of GrgA to the negativelycharged DNA. Furthermore, the lysine/arginine-rich sequenceand a following region appeared to form a helix-turn-helix (Fig.S4D), a structural motif characteristic of DNA binding proteins.We next assessed the effects of the various deletions in GrgA

on the ability of GrgA to activate transcription from the defApromoter. Transcription activation was essentially that of thewild-type GrgA for the Δ65–113, Δ166–206, and Δ207–288mutants. In contrast, the Δ114–165 mutant, which could not bindDNA, suffered a significant (∼75%) loss of transcription acti-vation activity compared with wild-type GrgA (Fig. 3B), sug-gesting that the ability of GrgA to bind DNA is important forGrgA-dependent transcription activation. In addition, the Δ1–64mutant, which retained a strong DNA binding activity (Fig. 3Aand Fig. S4C), also displayed a significant (∼65%) loss of tran-scription activation activity (Fig. 3B). These results indicate that

Tran

scrip

t am

ount

CBA

0

2

4

6

PDF (WT)

**

**

**

0

2

4

6

PDF (Z100)

*

*

**

0

1

2

4

3

PDF (GR10)

*

***

0.3 0.7 1.80

μM NH·GrgA

0.3 0.7 1.80

μM NH·GrgA

0.3 0.7 1.80

μM NH·GrgA

Fig. 1. GrgA stimulates transcription of the defA promoter in vitro. Shownare in vitro transcription assays performed using cRNAP, a DNA templatecarrying the indicated defA promoter variant, and the indicated concentrationof NH·GrgA. The Z100 promoter derivative (used in B) carries a base pairsubstitution in the promoter –35 element, whereas the GR10 derivative (usedin C) carries a substitution in DNA upstream of the –35 element (14). Graphsshow the averages and SDs for three independent measurements. We notethat C-terminally His-tagged GrgA also demonstrated a dose-dependentstimulatory effect on defA promoter activity (Fig. S2). Single and doubleasterisks denote that the difference between control and GrgA-containingreactions were statistically significant (P < 0.05 and P < 0.01, respectively).

Bao et al. PNAS | October 16, 2012 | vol. 109 | no. 42 | 16871

BIOCH

EMISTR

Y

DNA binding is necessary but not sufficient for GrgA to effi-ciently activate transcription.

GrgA-Dependent Transcription Activation Requires Interaction withthe Nonconserved Region of σ66. We have previously demonstratedthat a hybrid RNAP holoenzyme consisting of the E. coli RNAPcore enzyme and chlamydial σ66 can transcribe from the defApromoter (14). We therefore tested whether or not GrgA couldactivate transcription of the defA promoter in reactions performedusing this hybrid holoenzyme. We found that addition of GrgA totranscription reactions performed using the hybrid holoenzymestimulated transcription (Fig. S5A) in a manner similar to thestimulatory effect observed in reactions performed using cRNAP(Fig. 1 and Fig. S2). In contrast, GrgA was unable to activatetranscription in reactions performed with E. coli RNAP core en-zyme reconstituted with E. coli σ70 (Fig. S5B) or C. trachomatis σ28

(Fig. S5C). These findings raised the possibility that GrgA maystimulate transcription through direct interaction with σ66. Con-sistent with this hypothesis, Strep-Tactin–immobilized, C-termi-nally Strep-tagged σ66 (CS·σ66) was able to precipitate GrgA (Fig.4A); furthermore, N-terminally Strep-tagged GrgA (NS·GrgA)also precipitated C-terminally His-tagged σ66 (CH·σ66; Fig. 4B).

We next used the GrgA mutants constructed for mapping theDNA binding region (Fig. S4A) to determine what residues ofGrgA were required for the interaction with σ66. The Δ65–113,Δ114–165, Δ166–206, and Δ207–288 GrgA mutants all retainedthe ability to interact with σ66 (Fig. 4C). Importantly, the findingthat the Δ114–165 mutant retained the ability to bind σ66 indi-cates that this mutant is stable, providing further support that thereduction in transcription activation observed with this mutant(Fig. 3B) is a consequence of its inability to bind DNA. In con-trast to the other GrgA mutants, the Δ1–64 mutant, which couldefficiently bind DNA (Fig. 3A and Fig. S4C), did not detectablyinteract with σ66 (Fig. 4C). This finding suggests that the inabilityof the Δ1–64 mutant to efficiently stimulate transcription fromthe defA promoter (Fig. 3B) is due to the inability of the mutantto contact σ66.To further explore the hypothesis that GrgA-dependent tran-

scription activation requires contact with σ66, we sought to identifyregions of σ66 that were required for the interaction with GrgA.Primary σ factors share four regions of conserved sequence,regions 1–4 (15–18). Furthermore, many primary σ factors carrya NCR between regions 1 and 2 that can widely vary in sequenceand length (15–17). We therefore constructed individual C. tra-chomatis σ66 fragments (Table S1 and Fig. S6A) that comprised σ

A

μM NH·GrgA

0.1 0.2 0.30

B

-144

-144

GrgA precipitated (% control)NH·GrgA

-108

-57

-46

-37

-16

B

B

B

B

B

B

-35 -10+52 0 20 40 60 80 100

0.6 1.0 2.0 3.0

C D

Fig. 2. GrgA binds DNA. (A) EMSA assays performed in the absence of poly(dI-dC) using a DNA fragment carrying sequences extending from position –144 to+52 of the defA promoter. (B) Schematic of the DNA fragments used to precipitate NH·GrgA (star indicates the presence of a 3′ biotin moiety). (C) Westernblot analysis of the amount of NH·GrgA precipitated by the corresponding DNA fragments in B. (D) Graph shows the amount of GrgA precipitated asa percentage of that precipitated by the biotinylated –144 to +52 fragment. Plotted are the averages and SDs for three independent measurements.

A B

Con

trol

1-64

65-1

13

114-

165

166-

206

207-

288

Grg

A

0

2

4

6

** *

Tran

scrip

t am

ount

0.2

μM

0.4

μM

1.0

μM

No

Grg

A

0.2

μM

0.4

μM

1.0

μM

0.2

μM

0.4

μM

1.0

μM

0.2

μM

0.4

μM

1.0

μM

0.2

μM

0.4

μM

1.0

μM

0.2

μM

0.4

μM

1.0

μM

1-64 65-113GrgA 166-206 207-288114-165

3.0

μM

Fig. 3. DNA binding by GrgA is essential but not sufficient for full transcription activation. (A) EMSA assays performed using a radiolabeled DNA fragmentcarrying sequences extending from position –144 to +52 of the defA promoter in the presence of the indicated concentrations of wild-type GrgA or theindicated GrgA mutant. (B) In vitro transcription assays performed in the presence or absence of 1.8 μM wild-type GrgA or the indicated GrgA mutant. Assayswere performed using cRNAP and a DNA template carrying the Z100 defA promoter variant. Graph shows the averages and SDs for three independentmeasurements.

16872 | www.pnas.org/cgi/doi/10.1073/pnas.1207300109 Bao et al.

region 1 (residues 1–146), region 2 (residues 303–408), region 3(residues 386–490), region 4 (residues 473–571), or the NCR(residues 121–322), and attempted to express them in E. coli. All ofthe fragments except the one comprising region 3 were successfullyexpressed and purified (Fig. S6B). We next determined whether ornot we could precipitate these fragments with GrgA, and foundthat only the fragment encompassing the NCR could precipitateGrgA (Fig. 5A), suggesting that GrgA contacts the σ66 NCR.To further define the residues of the NCR that were important

for GrgA binding, we constructed a series of σ66 mutants that eachcontained 40–52 amino acid deletions in the NCR (ΔNCR1, whichlacked residues 132–183; ΔNCR2, which lacked residues 184–224; ΔNCR3, which lacked residues 224–268; and ΔNCR4,which lacked residues 269–316; Fig. S6 C and D). Among thesemutants, only ΔNCR4 lost the ability to bind to GrgA (Fig. 5B),indicating that residues 269–316 of the NCR of σ66 are requiredfor interaction with GrgA.We next tested the effect of deleting residues 269–316 of σ66 on

the ability of GrgA to activate transcription of the defA promoter.To do this, we performed in vitro transcription assays using a hy-brid RNAP holoenzyme consisting of the E. coli RNAP coreenzyme and either wild-type or mutant σ66. Removal of residues269–316 in σ66 had no effect on basal transcription but severelyreduced the ability of GrgA to stimulate transcription (Fig. 5C). Incontrast, removal of other NCR residues that did not affect theability of GrgA to bind σ66 did not significantly affect GrgA-dependent activation (Fig. 5C). Thus, disrupting the interactionbetween GrgA and the NCR of σ66 by either removing residues 1–64 of GrgA or residues 269–316 of σ66 severely impairs GrgA-de-pendent transcription activation (Figs. 3, 4, and 5C). Taken to-gether, these results establish that the interaction between GrgA

and the NCR of σ66 is required for efficient transcription activationby GrgA.

GrgA Is a General Activator of σ66-Dependent Genes in Vitro. Giventhat GrgA bound DNA in a sequence-nonspecific manner andcould activate transcription from the defA promoter, we nextdetermined if the transcription activation activity of GrgA waslimited to the defA gene. Accordingly, we assessed the effects ofGrgA in vitro on the activities of three additional σ66-dependentpromoters that are active at different stages of growth: ribosomalRNA promotor P1 (rRNA P1; early), major outer membraneprotein A (ompA; middle), and histone-like protein A (hctA;late) (8, 9, 19, 20). GrgA demonstrated significant stimulatoryeffects on all of the three promoters (Fig. 6). Furthermore, boththe Δ1–64 GrgA mutant (that does not interact with σ66) and theΔ114–165 GrgA mutant (that does not bind DNA) exhibitedsignificantly reduced levels of activation at each promoter com-pared with that observed with wild-type GrgA (Fig. 6). Thus, the

A

-CS

·66

Inpu

t

+ C

S·66

55

43NH·GrgA

Precip: StrepTactin WB: α His

C

Inpu

t

-NS

·Grg

A

+ N

S·G

rgA

55

72 CH·σ66

Precip: StrepTactin WB: α His

B

55

34

43

26

-CS

.66

+ C

S.66

-CS

.66

+ C

S.66

-CS

.66

+ C

S.66

-CS

.66

+ C

S.66

NH

·1-

64

NH

·65

-113

NH

·11

4-16

5

NH

·16

6-20

6

NH

·20

7-28

8

NH

·Grg

A

-CS

.66

+ C

S.66

-CS

.66

+ C

S.66

Precip: StrepTactin; WB: GrgA

Fig. 4. GrgA binds σ66. (A) Precipitation of NH·GrgA by Strep-Tactin–immobilized CS·σ66. Shown is a Western blot detecting His-tagged GrgA. (B)Precipitation of CH·σ66 by NS·GrgA. Shown is a Western blot detecting σ66.(C) Precipitation of wild-type GrgA or the indicated mutant GrgA derivativesby Strep-Tactin–immobilized CS·σ66. Shown is a Western blot detecting GrgAvariants. Anti-GrgA instead of anti-His was used for detection because theanti-His recognized GrgA variants with greatly varying efficiency, as dem-onstrated in Fig. S8.

CH

·N

CR

1

CH

·N

CR

2

CH

·N

CR

3

CH

·N

CR

4-N

S·G

rgA

+ N

S·G

rgA

-NS

·Grg

A+

NS

·Grg

A

-NS

·Grg

A+

NS

·Grg

A

-NS

·Grg

A+

NS

·Grg

A

CH

·66

-NS

·Grg

A+

NS

·Grg

A

B72

55

-NS

·Grg

A+

NS

·Grg

A

-NS

·Grg

A+

NS

·Grg

A

-NS

·Grg

A+

NS

·Grg

A

-NS

·Grg

A+

NS

·Grg

A

NH

·NC

RPrecip: StrepTactin; WB: His

A

NH

·1

NH

·2

NH

·4

26

17

34

10

-NH

·Grg

A

+ N

H·G

rgA

-NH

·Grg

A

+ N

H·G

rgA

-NH

·Grg

A

+ N

H·G

rgA

-NH

·Grg

A

+ N

H·G

rgA

-NH

·Grg

A

+ N

H·G

rgA

CH

·N

CR

1

CH

·N

CR

2

CH

·N

CR

3

CH

·N

CR

4

CH

·66

2

4

6

0

Tran

scrip

t am

ount

**

* *C

Precip: StrepTactin; WB: His

Fig. 5. Interaction between GrgA and the nonconserved region of σ66 is re-quired for efficient transcription activation. (A) Precipitation of His-taggedfragments of σ66 by Strep-Tactin–immobilized GrgA (NS·GrgA). Shown isa Western blot detecting the His-tagged σ66 fragments recovered after pre-cipitation. (B) Precipitation of His-tagged derivatives of σ66 by Strep-Tactin–immobilized GrgA (NS·GrgA). Shown is a Western blot detecting the His-tagged σ66 derivatives recovered after precipitation. ΔNCR1 lacks residues132–183; ΔNCR2 lacks residues 184–223; ΔNCR3, lacks residues 224–268; andΔNCR4 lacks residues 270–316 (Fig. S6). (C) In vitro transcription assays per-formed in the presence or absence of 1.8 μM wild-type GrgA using a hybridholoenzyme consisting of E. coli core and the indicated σ66 derivative. TheDNA template used for these assays was the Z100 defA promoter variant.Graph shows the averages and SDs for three independent measurements.

Bao et al. PNAS | October 16, 2012 | vol. 109 | no. 42 | 16873

BIOCH

EMISTR

Y

GrgA-dependent activation observed in the context of the rRNAP1, ompA, and hctA promoters, like that observed in the contextof the defA promoter, requires GrgA to contact the DNA andthe NCR of σ66.

DiscussionHere we identify a transcription factor from C. trachomatis, GrgA,which activates transcription of several σ66-dependent promotersin vitro by binding DNA and contacting the σ66 NCR. Thesefindings strongly suggest GrgA serves as an important regulator ofσ66-dependent transcription in C. trachomatis.

GrgA Is a General Regulator of σ66-Dependent Transcription. Chla-mydia is an important pathogen that has a unique developmentalcycle with two alternating cellular forms. Knowledge of the ex-tent to which the regulation of transcription contributes to thechlamydial life cycle has been limited, in part, by the lack ofsuitable experimental tools. In this regard, only a handful oftranscription inhibitors (21–25) and three transcription activators(26–28) had been identified in chlamydiae before our study.Our work identifies GrgA as a previously undocumented

transcription factor in C. trachomatis. Although we originallyidentified GrgA on the basis of its ability to stimulate tran-scription from the σ66-dependent defA promoter (Fig. 1), wefound that GrgA stimulates transcription from several σ66-de-pendent promoters that are active at different stages of thechlamydial developmental cycle (Fig. 6). Based upon these invitro findings, we propose that GrgA functions as a generaltranscription activator in C. trachomatis that up-regulates theexpression of a broad spectrum of genes during all developmentalphases. Consistent with this proposal, Western blot analysis indi-cates that GrgA is present in both chlamydial cellular forms, theEB and RB (Fig. S7). Nevertheless, although our in vitro dataprovide strong evidence that GrgA functions in vivo as an im-portant regulator of σ66-dependent gene expression, a direct test ofthe functional role that GrgA plays in vivo awaits the developmentof methods to perform targeted mutagenesis in Chlamydia.

GrgA-Dependent Transcription Activation Requires Contact with bothDNA and the σ66 NCR. We found that efficient GrgA-dependentactivation requires GrgA to retain the ability to contact both DNAand the σ66 NCR. In particular, we found that a GrgA mutantlacking amino acid residues required for DNA binding (Δ114–165)

and a GrgA mutant lacking amino acid residues required forcontact with the σ66 NCR (Δ1–64) were significantly impaired fortranscription activation (Figs. 3, 4, and 6). The importance of theGrgA–σ66 NCR interaction is further indicated by the inability ofwild-type GrgA to efficiently activate transcription when reactionswere performed with RNAP reconstituted with a σ66 mutantlacking amino acids required for the interaction with GrgA (Fig. 5).Many bacterial proteins have been identified that bind DNA

and activate transcription through direct contact with conservedregion 4 of the primary σ factor (for review, see ref. 29). To ourknowledge, GrgA represents the first DNA binding protein thatactivates transcription through direct contact with the NCR re-gion of a primary σ factor. Prior studies have identified a role forthe NCR of the primary σ factor of E. coli, σ70, in modulatingboth promoter escape and early elongation pausing (30). Inparticular, interaction between the σ70 NCR and the β′ subunit ofthe E. coli core enzyme has been shown to facilitate escape frompromoter DNA during initial transcription as well as escape fromσ70-dependent pausing during early elongation. We found thatGrgA can stimulate transcription with both chlamydial RNAPand with a hybrid RNAP consisting of chlamydial σ66 and E. colicore (Figs. 1 and 6 and Fig. S5). Furthermore, the NCR of σ66and the NCR of σ70 lack significant sequence similarity, sug-gesting that the NCR of σ66 is unlikely to interact with the β′subunit of the E. coli core enzyme. Thus, we consider it unlikelythat GrgA activates transcription by influencing interactionsbetween the σ66 NCR and the RNAP core enzyme. We proposeinstead that GrgA activates transcription by stabilizing thebinding of RNAP to promoter DNA. Nevertheless, defining theprecise mechanism by which GrgA activates transcription awaitsfurther investigation.In conclusion, we have identified GrgA as a Chlamydia-specific

transcription activator that exerts its stimulatory effect throughinteractions with the NCR of σ66 and sequence-nonspecificinteractions with DNA. Furthermore, because GrgA can stimu-late transcription in vitro from several promoters that control theexpression of genes that are critical for chlamydial growth, GrgAlikely represents a promising antichlamydial target.

Materials and MethodsPurification and Identification of PDF Promoter-Binding Proteins. RBs werepartially purified from 12 L of suspension culture and disrupted by sonication.The lysate was clarified by centrifugation. The clarified RB lysate was mixedwith biotinylated PDF promoter DNA immobilized to streptavidin-conjugatedagarose beads, which were then packed into a column and washed withbuffer I (SI Materials and Methods) supplemented with 200 mM NaCl. Boundproteins were eluted using buffer I containing 600 mM NaCl. Proteins weremixed with SDS/PAGE sample buffer, resolved by electrophoresis, anddigested by trypsin. Tryptic peptides were identified by nanoLC-MS/MS asdetailed in SI Materials and Methods.

Purification of Recombinant GrgA. For transcription assays and protein–pro-tein interaction assays, N- or C-terminally His-tagged GrgA (Table S1) waspurified from E. coli extracts prepared in guanidine hydrochloride. Dena-tured proteins were purified with metal TALON affinity resin and renaturedas detailed in SI Materials and Methods. Strep-tagged GrgA or σ66 werepurified using Strep-Tactin beads following manufacturer’s instruction.

In Vitro Transcription Assay. The ability of GrgA to regulate transcription fromchlamydial promoters was determined using a previously reported in vitrotranscription assay (14, 31) with modifications, as detailed in SI Materialsand Methods.

GrgA–DNA Interaction. EMSA was performed with or without poly(dI-dC).Streptavidin-immobilized biotinylated DNA fragments were used to pre-cipitate GrgA. Alternatively, antibody immobilized GrgA was used to pre-cipitate DNA. Experimental conditions for each of these assays are provided inSI Materials and Methods.

Tran

scrip

t am

ount

rRNA-P1

0

2

4

ompA hctA

1-64

Grg

A

114-

165

No

Grg

A

1-64

Grg

A

114-

165

No

Grg

A

1-64

Grg

A

114-

165

No

Grg

A

6**

* *

Fig. 6. GrgA is a general activator of σ66-dependent transcription in vitro. Invitro transcription assays performed in the presence or absence of 1.8 μMwild-type GrgA or the indicated GrgA mutant. Assays were done usingcRNAP and a DNA template carrying the indicated promoter. Graph showsthe averages and SDs for three independent measurements.

16874 | www.pnas.org/cgi/doi/10.1073/pnas.1207300109 Bao et al.

Protein–Protein Interaction. A lysate of E. coli expressing either NS·GrgA orCS·σ66 (Table S1), which contained 50 mM Hepes and 300 mM NaCl, wasdiluted with equal volume of H2O. A total of 200 μL of the diluted lysate wasmixed with 10 μL of Strep-Tactin beads (20 μL suspension) on a nutatorfor 1 h at 4 °C. The beads were washed 4× with 25 mM Hepes (pH 7.0)containing 150 mM NaCl and 1% Nonidet P-40 (HeNN buffer), and mixedwith 5 μg of CH·σ66, a CH·σ66 mutant, NH·GrgA, or a NH·GrgA mutant for1 h at 4 °C. After four additional washes with HeNN, His-tagged proteinswere resolved by SDS/PAGE and visualized by Western blotting usinganti-His or anti-GrgA.

ACKNOWLEDGMENTS. We thank Dr. Peter Lobel and his colleagues, Drs.Haiyan Zheng and Meiqian Qian, at the Robert Wood Johnson MedicalSchool (RWJMS) Proteomics Core for advice and help with mass spectrom-etry; Dr. Marc Gartenberg (RWJMS) for access to his phosphorimager; Dr.Masayori Inouye for access to his French press used for the preparation ofnon-denatured E. coli extracts; Dr. Guangming Zhong (University of TexasHealth Sciences Center at San Antonio) for producing mouse anti-GrgA; Dr.Ming Tan (University of California, Irvine) for transcriptional reporter plas-mids; Christopher Oey for assistance in the purification of recombinant pro-teins; and Dr. John Kerrigan for in silico sequence analyses of GrgA. Thiswork was supported in part by National Institutes of Health GrantsAI071954 (to H.F.) and GM088343 (to B.E.N.).

1. Stephens RS, Myers G, Eppinger M, Bavoil PM (2009) Divergence without difference:Phylogenetics and taxonomy of Chlamydia resolved. FEMS Immunol Med Microbiol55:115–119.

2. Schachter J (1999) Infection and disease epidemiology. Chlamydia IntracellularBiology, Pathogenesis, ed Stephens RS (ASM Press, Washington, DC), pp 139–169.

3. Hybiske K, Stephens RS (2007) Mechanisms of Chlamydia trachomatis entry intononphagocytic cells. Infect Immun 75:3925–3934.

4. Hybiske K, Stephens RS (2007) Mechanisms of host cell exit by the intracellular bac-terium Chlamydia. Proc Natl Acad Sci USA 104:11430–11435.

5. Todd WJ, Caldwell HD (1985) The interaction of Chlamydia trachomatis with hostcells: Ultrastructural studies of the mechanism of release of a biovar II strain fromHeLa 229 cells. J Infect Dis 151:1037–1044.

6. Stephens RS, et al. (1998) Genome sequence of an obligate intracellular pathogen ofhumans: Chlamydia trachomatis. Science 282:754–759.

7. Thomson NR, et al. (2008) Chlamydia trachomatis: Genome sequence analysis oflymphogranuloma venereum isolates. Genome Res 18:161–171.

8. Nicholson TL, Olinger L, Chong K, Schoolnik G, Stephens RS (2003) Global stage-specific gene regulation during the developmental cycle of Chlamydia trachomatis.J Bacteriol 185:3179–3189.

9. Belland RJ, et al. (2003) Genomic transcriptional profiling of the developmental cycleof Chlamydia trachomatis. Proc Natl Acad Sci USA 100:8478–8483.

10. Haranaga S, Ikejima H, Yamaguchi H, Friedman H, Yamamoto Y (2002) Analysis ofChlamydia pneumoniae growth in cells by reverse transcription-PCR targeted tobacterial gene transcripts. Clin Diagn Lab Immunol 9:313–319.

11. Capecchi MR (1966) Initiation of E. coli proteins. Proc Natl Acad Sci USA 55:1517–1524.12. Chen D, Yuan Z (2005) Therapeutic potential of peptide deformylase inhibitors.

Expert Opin Investig Drugs 14:1107–1116.13. Balakrishnan A, et al. (2009) Inhibition of chlamydial infection in the genital tract of

female mice by topical application of a peptide deformylase inhibitor. Microbiol Res164:338–346.

14. Bao X, et al. (2011) Non-coding nucleotides and amino acids near the active siteregulate peptide deformylase expression and inhibitor susceptibility in Chlamydiatrachomatis. Microbiology 157:2569–2581.

15. Gross CA, et al. (1998) The functional and regulatory roles of sigma factors in tran-scription. Cold Spring Harb Symp Quant Biol 63:141–155.

16. Paget MS, Helmann JD (2003) The sigma70 family of sigma factors. Genome Biol 4:203.

17. Lonetto M, Gribskov M, Gross CA (1992) The sigma 70 family: Sequence conservation

and evolutionary relationships. J Bacteriol 174:3843–3849.18. Murakami KS, Darst SA (2003) Bacterial RNA polymerases: The whole story. Curr Opin

Struct Biol 13:31–39.19. Hackstadt T, Baehr W, Ying Y (1991) Chlamydia trachomatis developmentally regu-

lated protein is homologous to eukaryotic histone H1. Proc Natl Acad Sci USA 88:

3937–3941.20. Perara E, Ganem D, Engel JN (1992) A developmentally regulated chlamydial gene

with apparent homology to eukaryotic histone H1. Proc Natl Acad Sci USA 89:

2125–2129.21. Wyllie S, Raulston JE (2001) Identifying regulators of transcription in an obligate in-

tracellular pathogen: A metal-dependent repressor in Chlamydia trachomatis. Mol

Microbiol 40:1027–1036.22. Akers JC, Tan M (2006) Molecular mechanism of tryptophan-dependent transcrip-

tional regulation in Chlamydia trachomatis. J Bacteriol 188:4236–4243.23. Rao X, et al. (2009) A regulator from Chlamydia trachomatismodulates the activity of

RNA polymerase through direct interaction with the beta subunit and the primary

sigma subunit. Genes Dev 23:1818–1829.24. Akers JC, HoDac H, Lathrop RH, Tan M (2011) Identification and functional analysis of

CT069 as a novel transcriptional regulator in Chlamydia. J Bacteriol 193:6123–6131.25. Wilson AC, Tan M (2002) Functional analysis of the heat shock regulator HrcA of

Chlamydia trachomatis. J Bacteriol 184:6566–6571.26. Koo IC, Walthers D, Hefty PS, Kenney LJ, Stephens RS (2006) ChxR is a transcriptional

activator in Chlamydia. Proc Natl Acad Sci USA 103:750–755.27. Koo IC, Stephens RS (2003) A developmentally regulated two-component signal

transduction system in Chlamydia. J Biol Chem 278:17314–17319.28. Zhong J, Douglas AL, Hatch TP (2001) Characterization of integration host factor (IHF)

binding upstream of the cysteine-rich protein operon (omcAB) promoter of Chla-

mydia trachomatis LGV serovar L2. Mol Microbiol 41:451–462.29. Dove SL, Darst SA, Hochschild A (2003) Region 4 of sigma as a target for transcription

regulation. Mol Microbiol 48:863–874.30. Leibman M, Hochschild A (2007) A sigma-core interaction of the RNA polymerase

holoenzyme that enhances promoter escape. EMBO J 26:1579–1590.31. Tan M, Engel JN (1996) Identification of sequences necessary for transcription in vitro

from the Chlamydia trachomatis rRNA P1 promoter. J Bacteriol 178:6975–6982.

Bao et al. PNAS | October 16, 2012 | vol. 109 | no. 42 | 16875

BIOCH

EMISTR

Y

Supporting InformationBao et al. 10.1073/pnas.1207300109SI Materials and MethodsReagents. DNA primers, including a biotinylated DNA primer,were custom-synthesized at Sigma. Streptavidin-conjugated agarosebeads and the Colloidal Blue Staining Kit were purchased fromInvitrogen. The TALON Metal Affinity Resin was purchasedfrom Clontech. PfuUltra High-Fidelity DNA Polymerase and theQuikChange Site-DirectedMutagenesis Kit were purchased fromAgilent Technologies. Taq DNA polymerase, T4 polynucleotidekinase, and DNA endorestriction enzymes were purchased fromNew England Biolabs. The 4-(2-aminoethyl) benzenesulfonylfluoride (AEBSF) was purchased from Merck. Strep-TactinSepharose and D-desthiobiotin were purchased from IBA Bio-TAGnology. [α-32P]CTP and [γ-32P]ATP were purchased fromMP Biochemicals. σ70-Saturated Escherichia coli RNA poly-merase (RNAP) holoenzyme (σ70-RNAP) and E. coli RNAPcore enzyme (eCore) was purchased from EPICENTRE. Poly-clonal mouse anti-GrgA antibody was a generous gift fromGuangming Zhong (University of Texas Health Sciences Center,San Antonio, TX). Mouse anti-His antibody and horseradish-conjugated mouse anti-His antibody were purchased from Sigma-Aldrich and Clontech, respectively. The TEN buffer contained40 mM Tris·HCl (pH 7.5), 1 mM EDTA, and 150 mM NaCl.Buffer I contained 10 mM Tris·HCl (pH 8.0), 10 mM MgCl2,1 mM EDTA, 10 mM 2-mercaptoethanol, and 7.5% (vol/vol)glycerol. The RB lysis buffer contained 10 mM Tris·HCl (pH8.0), 0.6 M NaCl, 10 mM MgCl2, 1 mM EDTA, 7.5% (vol/vol)glycerol, 0.1% (vol/vol) Nonidet P-40, 0.3 mM DTT, 1 mMAEBSF, 10 mg/mL pepstatin, and 150 mg/mL lysozyme; DTT,AESBF, pepstatin A, and lysosome were added just before use.The chlamydial RNAP (cRNAP) storage buffer contained 10mM Tris·HCl (pH 8.0), 10 mM MgCl2, 0.1 mM EDTA, 0.1 mMDTT, 100 mM NaCl, and 30% (vol/vol) glycerol. The TNGbuffer contained 25 mM Tris acetate (pH 8.0), 150 mM NaCl,and 10% (vol/vol) glycerol.

Strains and Culture Conditions. Chlamydia trachomatis serovar L2(L2; strain 434/Bu) was purchased from ATCC. HeLa cells, grownas adherent cultures, were used for the preparation of elementarybody (EB) stocks (1, 2). For the preparation of cRNAP, and pu-rification of PDF promoter-binding proteins, mouse L929 cellsuspension cultures were infected with EBs at an inoculating dos-age of 3 inclusion-forming units per cell (2, 3). E. coliArcticExpresswas purchased from Agilent and cultured with the LB media.

Vectors. Plasmids for expressing His- or Strep-tagged GrgA and σ66and their mutants are listed in Table S1. For the construction ofwild-type expression vectors, DNA fragments were amplified usingPfuUltra, digested with one or two endorestriction enzymes, andligated to properly prepared plasmids. Most deletion mutants withtruncation at either one or both termini were constructed througha similar strategy, but some were made using the QuikChange site-directed mutagenesis kit. Point mutations were constructed usingthe QuikChange kit.

Preparation of Biotinylated DNA.Biotinylated DNA fragments wereproduced by Taq polymerase using a nonbiotinylated 5′ primerand a biotinylated 3′ primer designated PDF-TF-R-(5′-biotin),which carried a biotin residue at the 5′ position. The PCRproducts were purified using Qiagen PCR purification columns.

Preparation of PDF Promoter-Bound Agarose Beads. The above-pu-rified biotinylated DNA was diluted 1:5 in TEN buffer. Strep-

tavidin-conjugated agarose beads were washed twice with TENbuffer and thenmixed with biotinylated DNA on a nutator at 4 °C.At 30 min later, beads were washed three times with TEN bufferto remove unbound DNA.

Preparation of Partially Purified Chlamydiae from Bulk Culture. At22 h after inoculation, L2-infected L929 cells were collected bycentrifugation at 2,500 × g for 10 min at 4 °C. All subsequentsteps were performed on ice or at 4 °C. The pellet was re-suspended in 8 mL of cold PBS (pH 7.4) per liter culture. Cellswere disrupted using a Daigger GEX130 Ultrasonic processor.The cell suspension was subjected to three 10-s cycles of soni-cation with the energy intensity level set at 30%. The lysate wascentrifuged at 1,000 × g for 10 min, and the supernatant wascollected. The pellet was resuspended in another 8 mL PBS, andsubjected to an additional round of sonication and centrifugationas described above. The two supernatants were combined andcentrifuged at 18,000 × g for 15 min. The resulting pellet con-tained partially purified chlamydiae.

Purification of PDF Promoter-Binding Proteins. All procedures wereperformed on ice or at 4 °C. Chlamydiae partially purified from12 L of suspension culture were resuspended in 24 mL reticulatebody (RB) lysis buffer, and disrupted by ten 10-s sonication at45% energy intensity. The lysate was centrifuged at 14,000 × g for30 min. The supernatant was further clarified by four additionalrounds of 15-min centrifugation. The supernatant resulted fromthe final round of centrifugation was diluted with 3 vol of buffer I.One-half of the diluted lysate was mixed with PDF promoterDNA-coated beads, prepared as described above, on a nutator(Clay Adams). At 1 h later, the beads were packed into a column,which were then washed with 20 resin vol of buffer I supple-mented with 200 mMNaCl. The bound proteins were eluted using4 vol of buffer I containing 600 mM NaCl. The elution was con-centrated to 20 μL using Amicon centrifugal filters with a cutoffsize of 3 kDa (Millipore). As control, nonbiotinylated PDF pro-moter DNA generated using the 5′ primer, PDF P1-F, and 3′,PDF-TF-R were added to the other half of the lysate, which werethen mixed with plain streptavidin beads prewashed with TENbuffer. The remaining procedures for handling mock pull-downwere the same as real pull-down.

LC-MS/MS. The concentrated pull-down samples were mixed withSDS/PAGE sample buffer and run ∼1 cm into a Bis-Tris 10%polyacrylamide gel (Novex Biotech). The entire band was ex-cised, and proteins in the gel were reduced, carboxymethylated,and digested with trypsin using standard protocols. Peptideswere extracted, solubilized in 0.1% trifluoroacetic acid, and an-alyzed by nanoLC-MS/MS using rapid-separation LC (Dionex)interfaced with a LTQ Orbitrap Velos (ThermoFisher). Sampleswere loaded onto a self-packed 100-μm × 2-cm trap packed withMagic C18AQ, 5 μm, 200 Å (Michrom Bioresources Inc.) andwashed with buffer A (0.2% formic acid) for 5 min with a flowrate of 10 μL/min. The trap was brought in-line with the ana-lytical column (Magic C18AQ, 3 μm, 200 Å, 75 µm × 50 cm) andpeptides fractionated at 300 nL/min with a 30-min linear gradientof 2–45% buffer B [0.08% formic acid, 80% (vol/vol) acetonitrile].MS data were acquired using a data-dependent acquisition pro-cedure with a cyclic series of a full scan acquired in Orbitrap withresolution of 60,000 followed by MS/MS scans (acquired in linearion trap) of 20 most-intense ions with a repeat count of two andthe dynamic exclusion duration of 60 s.

Bao et al. www.pnas.org/cgi/content/short/1207300109 1 of 10

The LC-MS/MS data were searched against the TrEMBL L2using a local version of the Global Proteome Machine (GPMUSB; Beavis Informatics Ltd.) with carbamidoethyl on cysteine asfixed modification and oxidation of methionine and tryptophan asvariable modifications using a 10-ppm precursor ion toleranceand a 0.4-Da fragment ion tolerance.

MALDI-TOF/TOF Mass Spectrometry. Protein sample was dilutedwith 10 vol of matrix solution [10 mg/mL sinapinic acid in 50%(vol/vol) acetylnitrole/0.1% trifluoric acid]. A total of 1 μL of themix was loaded onto an Opti-TOF 384-well plate and air-dried.Spectra were acquired with a 4800 MALDI TOF/TOF analyzer(AB Sciex) using positive linear high-mass mode from 15 K to100 K. Each spectrum reported was the average of spectra gener-ated from 2,000 lacer shots. BSA was used as external calibration.

Preparation of cRNAP. cRNAP was prepared using a publishedprotocol (4) with modifications. A typical cRNAP purificationexperiment starts with 4 L of suspension culture. Chlamydial or-ganisms were partially purified as described above, resuspended in8 mL of freshly prepared RB lysis buffer with the omission ofNaCl, and disrupted by sonication. Cell debris was removed bycentrifugation as described for the preparation of lysate for thepurification of promoter-binding proteins. To the final superna-tant was added one-ninth vol of buffer I containing 1.5 M NaCl,and then 4 mL of the heparin-conjugated agarose beads. Fol-lowing 2 h of mixing, each milliliter of the beads was packed intoa column and washed with 20 mL of buffer I containing 200 mMNaCl. cRNAP was eluted with buffer I containing 1.5 M NaCl andcollected as 250-μL fractions. Corresponding fractions collectedfrom different columns were combined and dialyzed overnightagainst the cRNAP storage buffer. Small aliquots were made andstored at −80 °C. Typically, fraction 5 was used for experiments.

Purification and Refolding of GrgA, σ70, σ28 from Denatured CellExtracts. ArcticExpress E. coli cells expressing His-tagged pro-teins were resuspended and lysed in 6 M guanidine hydrochloridesolution containing 50 mM Hepes (pH 7.4) and 300 mM NaCl.After the removal of cell debris by centrifugation, the supernatantwas incubated with TALON metal affinity resin on a nutator for1 h at room temperature. Resin was washed five times with theguanidine hydrochloride solution, packed into a column, andeluted with 45 mM Hepes (pH 7.4) containing 270 mM NaCl and150 mM imidazole. To refold GrgA, the elution was dialyzed firstovernight against a solution containing 2 M urea, 25 mM Hepes(pH 7.4), 300 mM NaCl, 0.5 mM AEBSF, 1 mM reduced gluta-thione, and 10% (vol/vol) glycerol, and then for additional 4 hagainst 25 mM Hepes (pH 7.4) containing 300mMNaCl and 10%(vol/vol) glycerol. The dialyzed GrgA was concentrated and ex-changed to theTNGbuffer before itwas aliquotedand storedat−80 °C. Refold of σ70 and σ28 was accomplished by following proceduresdescribed by Panaghie et al. (5) and Yu and Tan (6), respectively.

In Vitro Transcription Assay. The assay in a total volume of 30 μLcontained 400 ng supercoiled plasmid DNA, 1 mM potassiumacetate, 8.1 mMmagnesium acetate, 50 mM Tris acetate (pH 8.0),27 mM ammonium acetate, 1 mM DTT, 3.5% (wt/vol) poly-ethylene glycol (average molecular weight, 8,000), 330 μM ATP,330 μM UTP, 1 μM CTP, 0.2 μM [α-32P]CTP (3,000 Ci/mmol),100 μM 3′-O-methyl-GTP, 36 units of RNasin, RNAP, and in-dicated amount of GrgA or GrgA mutant, purified by proceduresinvolving denaturing and refolding as described above. For re-actions using cRNAP, the amount of cRNAP was 3.0 μL/reaction.For reactions using eCore and σ66, their concentrations were20 nM and 100 nM, respectively. When different amounts ofGrgA were used, the amount of GrgA storage buffer (i.e., theTNG buffer) remained constant. All control reactions involving

GrgA received equal volume of TNG buffer. The reaction wasallowed to pursue at 37 °C for 30 min and terminated by theaddition of 70 μL of 2.86 M ammonium acetate containing 4 mg ofglycogen. After ethanol precipitation, 32P-labeled RNA was re-solved by urea–polyacrylamide gel electrophoresis, and visualizedon a phosphorimager, and the intensities of the 157 base transcriptbands were determined by ImageQuant software. Relative amountsof transcript were presented with that of the control reaction set as1 unit. Data shown in bar graphs represent averages ± SDs fromthree or more independent experiments. Pairwise, two-tailed Stu-dent t tests were used to compare data from groups. Single anddouble asterisks indicate P ≤ 0.05 and P ≤ 0.01, respectively.

EMSA. EMSA was performed following a standard protocol (7).A 5′ primer designated PDF P1-F of the PDF promoter, wasreacted with γ-[32P]ATP in the presence of the T4 polynucleotidekinase. The resulting 32P-labeled primer was used in conjunctionwith an unlabeled 3′ primer designated PDF-TF-R to amplifya DNA fragment containing the PDF promoter. The PCR-am-plified fragment was purified with a Qiagen column. The GrgA-DNA binding reaction contained, in a total volume of 10 μL,10 nM promoter fragment, an indicated amount of NH·GrgA,1 mM potassium acetate, 8.1 mM magnesium acetate, 50 mMTris acetate (pH 8.0), 27 mM ammonium acetate, 1 mM DTT,and 3.5% (wt/vol) polyethylene glycol (average molecular weight,8,000), with or without 0.5 μg poly(dI-dC). After mixing for 1 h at4 °C, thebindingmixturewas loadedonto 6%(wt/vol) nondenaturingpolyacrylamide gel. Free and GrgA-bound DNA fragments werevisualized on a Storm Phosphorimager (Molecular Dynamics).

DNA Pull-Down of GrgA. The 50-μL streptavidin-conjugated aga-rose beads were washed twice with buffer TEN and mixed with40 pmol of a biotinylated DNA fragment on a nutator for 30 minat 4 °C. Beads were washed three times with buffer TEN to re-move unbound DNA, and twice with transcription buffer. A 5-μgnative purified NH·GrgA (full-length or deletion mutant) wasadded to the beads. The mixtures were incubated on a nutatorfor 1 h at 4 °C. After two washes with transcription buffer andthree times with PBS containing 1% (vol/vol) Triton X-100 (PBST),GrgA was resolved by SDS/PAGE and detected by colloidal bluestain or Western blotting using HRP-conjugated anti-His.

GrgA Pull-Down of DNA. DNA fragments corresponding to dif-ferent portions of the PDF gene were produced by PCR. Primerswere removed by PCR purification columns (Qiagen). DNA wereeluted with water and diluted in transcription buffer. A total of2 μg NH·GrgA were incubated on a nutator with 0.5 μL mouseanti-His (Sigma) or 1.0 μL control normal mouse serum for 1 hat 4 °C, and then with 20 μL of protein A/G agarose (Sigma) for2 h at 4 °C. The beads were washed four times with PBST, re-suspended in the transcription buffer, and mixed with a DNAfragment (40 pmol) on a nutator for 2 h at 4 °C. After four washeswith PBST, the reaction was heated at 95 °C for 5 min to dissociateDNA from GrgA. DNA was visualized after electrophoresis ina 1.5% (wt/vol) agarose gel containing ethidium bromide.

Preparation of Highly Purified EBs and RBs.HeLa cells were infectedwith L2. EBs and RBs were released from infected cells 36 h laterby sonication. Host DNA and RNA were removed by incubatingthe lysates at 37 °C for 30 min after the addition of DNase (finalconcentration: 10 μg/mL) and RNase (100 μg/mL). The lysateswere layered over 8 mL of 35% (vol/vol) RenoCal and centri-fuged at 43,000 × g for 60 min in an SW28 rotor (BeckmanCoulter). The EBs and RBs in the pellet were further purified bycentrifugation through a RenoCal gradient (13 mL of 40%, 8 mLof 44%, and 5 mL of 52%) at 43,000 × g for 90 min in an SW28rotor (8).

Bao et al. www.pnas.org/cgi/content/short/1207300109 2 of 10

1. Balakrishnan A, et al. (2006) Metalloprotease inhibitors GM6001 and TAPI-0 inhibit theobligate intracellular human pathogen Chlamydia trachomatis by targeting peptidedeformylase of the bacterium. J Biol Chem 281:16691–16699.

2. Bao X, et al. (2011) Non-coding nucleotides and amino acids near the active siteregulate peptide deformylase expression and inhibitor susceptibility in Chlamydiatrachomatis. Microbiology 157:2569–2581.

3. Fan H, Brunham RC, McClarty G (1992) Acquisition and synthesis of folates by obligateintracellular bacteria of the genus Chlamydia. J Clin Invest 90:1803–1811.

4. Tan M, Engel JN (1996) Identification of sequences necessary for transcriptionin vitro from the Chlamydia trachomatis rRNA P1 promoter. J Bacteriol 178:6975–6982.

5. Panaghie G, Aiyar SE, Bobb KL, Hayward RS, de Haseth PL (2000) Aromatic amino acidsin region 2.3 of Escherichia coli sigma 70 participate collectively in the formation of anRNA polymerase-promoter open complex. J Mol Biol 299:1217–1230.

6. Yu HH, Tan M (2003) Sigma28 RNA polymerase regulates hctB, a late developmentalgene in Chlamydia. Mol Microbiol 50:577–584.

7. Galas DJ, Schmitz A (1978) DNAse footprinting: A simple method for the detection ofprotein-DNA binding specificity. Nucleic Acids Res 5:3157–3170.

8. Caldwell HD, Kromhout J, Schachter J (1981) Purification and partial characterization of themajor outer membrane protein of Chlamydia trachomatis. Infect Immun 31:1161–1176.

C

HeL

a

HeL

a/L2

NH

·Grg

A

72

4355

34

26

96130

A

NH

·Grg

A

CH

·Grg

A

1017

263443

72

170

55

13096

NH

·Grg

A

CH

·Grg

AB

Mass (m/z)

% In

tens

ity

34047

83468663304919232054149170

100

80

60

40

20

D

Fig. S1. Expression, purification, and molecular weight confirmation of GrgA. (A) Coomassie blue stain shows that the 34-kDa N-terminally His-tagged GrgA(NH·GrgA) purified from nondenatured cell extract migrated as a 47-kDa protein in SDS/PAGE, whereas no major protein band was observed for C-terminallyHis-tagged GrgA (CH·GrgA) using the same purification procedures. (B) NH·GrgA and CH·GrgA, purified from guanidine chloride-denatured cell extracts,migrated as a 47-kDa protein and a 46-kDa protein, respectively. Note that in NH·GrgA there is a 10-aa linker between the His-tag and GrgA, whereas no linkersequence existed between GrgA and the tag in CH·GrgA. (C) Western blotting showing that endogenous GrgA migrated as a 45-kDa protein. (D) MALDI-TOF/TOF mass spectrometry reveals that the major component in the NH·GrgA preparation had an expected 34,047-da molecular weight.

Bao et al. www.pnas.org/cgi/content/short/1207300109 3 of 10

0.3 0.7 1.80

µM CH·GrgA

0

1

2

4

3

Tran

scrip

t am

ount

Fig. S2. GrgA with a C-terminal His-tag stimulates transcription from the defA promoter. Shown are in vitro transcription assays performed using cRNAP,a DNA template carrying the GR10 defA promoter variant, and the indicated concentration of C-terminally His-tagged GrgA (CH·GrgA). Graphs show theaverages and SDs for three independent measurements.

Bao et al. www.pnas.org/cgi/content/short/1207300109 4 of 10

pMT1125

-35 -1025+441-

140

pMT1125

pMT1125 B

43 553426 7295

130

B

+5

C

B

DTranslation

120 336

NP

442 666

CSMouse Serum anti-His

P1

NP

CSP1

NP

CS

Transcription

+1 ORF

52

P1

-144

-144

E

µM NH·GrgA

0.1 0.2 0.30 0.6 1.0 2.0 3.0

A

Fig. S3. Promoter-independent binding of DNA by GrgA. (A) Electrophoresis mobility shift of PDF promoter fragment by NH·GrgA in the presence of poly(deoxyinosinic-deoxycytidylic). (B) Schematic of the DNA fragments generated from transcriptional reporter plasmid vector pMT1125 that were used toprecipitate NH·GrgA (star indicates the presence of a 3′ biotin moiety). (C) Western blot analysis of the amount of NH·GrgA precipitated by the correspondingDNA fragments in B. (D) Location of DNA fragments P1, NP, and CS with respect to the defA gene. (E) Precipitation of the P1, NP, and CS DNA fragments byanti-His protein A/G agarose-immobilized NH·GrgA. Substitution of anti-His with normal mouse serum resulted in loss of DNA precipitation. Left lane is the 100-bpmolecular ladder.

Bao et al. www.pnas.org/cgi/content/short/1207300109 5 of 10

10

17

263443

72

170

55

13096

1017

263443

72

170

55

13096

NS

-Btn

·PD

F-P

1

+ B

tn·P

DF-

P1

-Btn

·PD

F-P

1

+ B

tn·P

DF-

P1

-Btn

·PD

F-P

1+

Btn

·PD

F-P

1-B

tn·P

DF-

P1

+ B

tn·P

DF-

P1

-Btn

·PD

F-P

1

+ B

tn·P

DF-

P1

-Btn

·PD

F-P

1+

Btn

·PD

F-P

1

NH

·1-

64

NH

·65

-113

NH

·11

4-16

5

NH

·16

6-20

6

NH

·20

7-28

8

NH

·Grg

A

DF

NH

·1-

64

NH

·65

-113

NH

·11

4-16

5

NH

·16

6-20

6

NH

·20

7-28

8

NH

·Grg

A

C

B

A

8821NH·GrgA

NH· 1-64

NH· 114-165

NH· 166-206

NH· 207-288

NH· 65-113

D

helixstrandcoilconfidence of prediction

Fig. S4. Amino acids 114–165 in GrgA are required for DNA binding. (A) Schematic of the GrgA mutants lacking the indicated regions. (B) Coomassie blue stainof purified GrgA deletion constructs. (C) Deletion of amino acids 114–165 from NH·GrgA resulted in loss of precipitation by the PDF promoter fragment.Precipitated proteins were visualized by colloidal blue. DF, dye front; SN, nonspecific bands. (D) Prediction of GrgA as a potential helix-turn-helix protein byPSIPRED (1). Note that the predicted helix ranging from residues 139–158 is rich in positively charged lysine and arginine residues.

1. Jones DT (1999) Protein secondary structure prediction based on position-specific scoring matrices. J Mol Biol 292:195–202.

Bao et al. www.pnas.org/cgi/content/short/1207300109 6 of 10

1.80.3 0.70

0

0.5

1.5

1.0

BR

elat

ive

trans

crip

t am

ount

0.3 0.7 1.80

eCore + 66 + µM GrgA

0

2

6

4

*

**

*

CA

0.3 0.7 1.8 8.17.03.00 0

0

2

6

4

**

**

**

eCore + 70 + µM GrgA

Btch)001Z( FDP)001Z( FDP)001Z( FDP

eCore + 66 + µM GrgA eCore + 28 + µM GrgA

Fig. S5. Activation of PDF gene transcription by GrgA is σ66-dependent. In vitro transcription assays performed using a hybrid holoenzyme consisting of E. colicore RNAP reconstituted with C. trachomatis σ66 (A), E. coli core RNAP reconstituted with E. coli σ70 (B), or E. coli core RNAP reconstituted with C. trachomatisσ28 (C). Reactions were done using a DNA template carrying the Z100 defA promoter variant (A and B, and C Left) or the hctB promoter (1) (C Right) in thepresence of the indicated concentration of NH·GrgA. Graphs show the averages and SDs for three independent measurements.

1. Yu HH, Tan M (2003) Sigma28 RNA polymerase regulates hctB, a late developmental gene in Chlamydia. Mol Microbiol 50:577–584.

Bao et al. www.pnas.org/cgi/content/short/1207300109 7 of 10

NH

·1

NH

·NC

R

26

17

34

10

B

NH

·2

NH

·3

NH

·4

CH

·N

CR

1

CH

·N

CR

2

CH

·N

CR

3

CH

·N

CR

4

CH

·N

CR

72

55

CH

·66

DC

1CH· 66

571

A

1NH· 66

571

1NH· 1

146

NH· 3386 490

NH·NCR121 322

NH· 4473 571

CH· NCR131 317

CH· NCR1131 184

CH· NCR2183 224

CH· NCR3223 269

CH· NCR4269 317

NH· 2303 408

Fig. S6. Expression analyses of σ66 mutants in E. coli. (A) Schematic of the N-terminally His-tagged σ66 fragments constructed on the basis of the functionaldomains of the E. coli homolog σ70. (B) Western blot analysis of the σ66 fragments indicate that all individual domains but σ3 were successfully expressed.Shown are Western blots of crude E. coli extracts detected with an anti-His antibody. (C) Schematic of the σ66 mutants with nonconserved region (NCR)deletions. (D) Western blot analysis of the σ66 variants with subregion deletions indicate that only the variant lacking the entire NCR could not be expressed.Shown are Western blots of crude E. coli extracts detected with an anti-His antibody.

EB

RB

72

4355

34

26

96130

72

4355

34

26

96130

GrgA

MOMP

Fig. S7. Detection of GrgA in EBs and RBs. The two chlamydial cellular forms were highly purified by two rounds of RenoCal gradient ultracentrifugation.Western blot analysis was done using an antibody against GrgA or an antibody against the chlamydial major outer membrane protein (MOMP).

Bao et al. www.pnas.org/cgi/content/short/1207300109 8 of 10

A

4355

34

26

72

4355

34

26

72B

NH

·1-

64

NH

·65

-113

NH

·11

4-16

5

NH

·16

6-20

6

NH

·20

7-28

8

NH

·Grg

A

GrgA

NH

·1-

64

NH

·65

-113

NH

·11

4-16

5

NH

·16

6-20

6

NH

·20

7-28

8

NH

·Grg

A

His

Fig. S8. His-tagged GrgA deletion constructs are recognized by the anti-His antibody with varying efficiencies. Western blot analysis of indicated His-taggedGrgA derivative using an anti-His antibody (A) or a polyclonal mouse anti-GrgA antibody (B). A total of 20 ng of each recombinant protein was loaded.

Bao et al. www.pnas.org/cgi/content/short/1207300109 9 of 10

Table S1. Vector information

Type/designation Description Source

Transcriptional reporterpMT1125 Plasmid harboring a promoterless guanine (G)-less cassette (1)pMT1125-WT pMT1125 with wild-type cPDF promoter leading the G-less cassette (2)pMT1125-GR10 Same as pMT1125-WT except it contains a C→A UP element mutation (2)pMT1125-Z100 Same as pMT1125-WT except it contains a C→A −35 element mutation (2)pMT1125-rRNA-P1 pMT1125 with chlamydial rRNA P1 promoter leading the G-less cassette This studypMT1187 pMT1125 with ompA promoter leading the G-less cassette (3)pMT1185 pMT1125 with chlamydial hctA promoter leading the G-less cassette (3)pMT1125-WT-Δ(−144

through −105)pMT1125-WT derivative with deletion of −144 through −105 from cPDF promoter This study

pMT1125-WT-Δ(−144through −65)

pMT1125-WT derivative with deletion of −144 through −65 from cPDF promoter This study

pMT1125-WT-Δ(−144through −29)

pMT1125-WT derivative with deletion of −144 through −29 from cPDF promoter This study

pMT1212 Transcriptional reporter plasmid for the chlamydial hctB promoter recognized by s28 (4)Wild-type GrgA and mutant

expressionpET28a For expression of N-terminally His-tagged proteins in E. coli Novogen, Ltd.pET28a-GrgA GrgA ORF inserted between NdeI and XhoI sites of pET28a for expression of N-(His)6-tagged

GrgA (NH·GrgA)This study

pET28a-GrgAΔ(1–64) For expression of NH·Δ1–64; derived by deleting residues 1–64 from NH·GrgA in pET28a-GrgA This studypET28a-GrgAΔ(65–113) For expression of NH·Δ65–113; derived by deleting residues 65–113 from NH·GrgA in

pET28a-GrgAThis study

pET28a-GrgAΔ(114–165) For expression of NH·Δ114–165; derived by deleting residues 114–165 from NH·GrgA inpET28a-GrgA

This study

pET28a-GrgAΔ(166–206) For expression of NH·Δ166–206; derived by deleting residues 106–206 from NH·GrgA inpET28a-GrgA

This study

pET28a-GrgAΔ(207–288) For expression of NH·Δ207–288; derived by deleting residues 207–288 from NH·GrgA inpET28a-GrgA

This study

pET21c For expression of C-terminally His-tagged proteins in E. coli Novogen, Ltd.pET21c-GrgA For expression of CH·GrgA; GrgA ORF inserted between NdeI and XhoI sites of pET21c This studypNS·GrgA For expression of NS·GrgA; derived by inserting a strep tag to the BamHI site of pET21c-GrgA

and deleting its C-terminal His-tagThis study

Expression of σ factorspCS·σ66 For expression of CS·σ66 using NdeI and BamHI site in pET21c with deleted His-tag This studypCOLADuet For expression His-tagged and/or untagged proteins in E. coli Novogen, Ltd.pCOLADuet-σ66R1 For expression of region 1 (residues 1–146) of σ66 (NH·R1) using BamHI and NotI sites of

pCOLADuetThis study

pCOLADuet-σ66R2 For expression of region 2 (residues 308–408) of σ66 (NH·R2) using BamHI and NotI sites ofpCOLADuet

This study

pCOLADuet-σ66R3 For expression of region 3 (residues 386–490) of σ66 (NH·R3) using BamHI and NotI sites ofpCOLADuet

This study

pCOLADuet-σ66R4 For expression of region 4 (residues 473–571) of σ66 (NH·R4) using BamHI and NotI sites ofpCOLADuet

Novogen, Ltd.

pCOLADuet-σ66NCR For expression of nonconserved region (residues 121–322) of σ66 (NH·NCR) using BamHI andNotI sites of pCOLADuet

This study

pET21c-σ66 For expression of C-terminal His-tagged σ66 (CH·σ66); σ66 ORF was inserted between NdeIand XhoI sites of pET21c

This study

pET21c-σ66-ΔNCR For expression of CH·ΔNCR, σ66 lacking NCR (residues 132–316); derived from pET21c-σ66 This studypET21c-σ66-ΔNCR1 For expression of CH·ΔNCR1, σ66 lacking residues 132–183 within the NCR; derived from

pET21c-σ66This study

pET21c-σ66-ΔNCR2 For expression of CH·ΔNCR1, σ66 lacking residues 184–223 within the NCR; derived frompET21c-σ66

This study

pET21c-σ66-ΔNCR3 For expression of CH·ΔNCR3, σ66 lacking residues 224–268 within the NCR; derived frompET21c-σ66

This study

pET21c-σ66-ΔNCR4 For expression of CH·ΔNCR4, σ66 lacking residues 269–316 within the NCR; derived frompET21c-σ66

This study

pLHN12-N-His-σ70 For expression of N-terminally His-tagged E. coli σ70 (5)pET28a-σ28 For expression of N-terminally His-tagged chlamydial σ28 This study

1. Yu HH, Di Russo EG, Rounds MA, Tan M (2006) Mutational analysis of the promoter recognized by Chlamydia and Escherichia coli sigma(28) RNA polymerase. J Bacteriol 188(15):5524–5531.2. Bao X, et al. (2011) Non-coding nucleotides and amino acids near the active site regulate peptide deformylase expression and inhibitor susceptibility in Chlamydia trachomatis.

Microbiology 157:2569–2581.3. Niehus E, Cheng E, Tan M (2008) DNA supercoiling-dependent gene regulation in Chlamydia. J Bacteriol 190(19):6419–6427.4. Yu HH, Tan M (2003) Sigma28 RNA polymerase regulates hctB, a late developmental gene in Chlamydia. Mol Microbiol 50:577–584.5. Panaghie G, Aiyar SE, Bobb KL, Hayward RS, de Haseth PL (2000) Aromatic amino acids in region 2.3 of Escherichia coli sigma 70 participate collectively in the formation of an RNA

polymerase-promoter open complex. J Mol Biol 299:1217–1230.

Bao et al. www.pnas.org/cgi/content/short/1207300109 10 of 10