Rh

EI

a

ARRAA

KCP3A

1

eeahmtCuaeoetfh2

r

0h

Molecular Immunology 52 (2012) 88– 95

Contents lists available at SciVerse ScienceDirect

Molecular Immunology

j ourna l ho me pag e: www.elsev ier .com/ locate /mol imm

oles of promoter and 3′ untranslated motifs in expression of theuman C5a receptor

lizabeth Palmer, Lisa C. Gray, Matthew Stott, Derrick J. Bowen, Carmen W. van den Berg ∗

nstitute of Molecular and Experimental Medicine, School of Medicine, Cardiff University, Cardiff CF14 4XN, UK

r t i c l e i n f o

rticle history:eceived 6 March 2012eceived in revised form 27 April 2012ccepted 29 April 2012vailable online 17 May 2012

eywords:5a receptor/CD88romoter′UTR

a b s t r a c t

The C5a receptor (C5aR) is a 7 transmembrane G-protein coupled receptor (GPCR) that mediates thepowerful pro-inflammatory effect of the complement activation product C5a. Excess C5a generatedunder pathological conditions has been implicated in a variety of conditions including sepsis, asthmaand rheumatoid arthritis, but very little is known about the regulation of expression of the C5aR. The 5′

promoter region and 3′ untranslated region (UTR) of the C5aR mRNA were cloned, generating enhancedgreen fluorescent protein (EGFP)-reporter plasmids, which were transfected into the monocytic cell lineU937. Most of the cloned 2 kb 5′ region was dispensable for the expression of the reporter constructsand the majority of regulatory sequences are in the first 200 bp. Three motifs, a NF�B, a CCAAT and aNFAT site, were identified to be of importance by site directed mutagenesis for basal expression. Analysis

′

U-rich element of the 3 UTR of the C5aR mRNA showed that it contained two AU-rich elements (AREs), however sitedirected mutagenesis showed that these had no effect on basal expression. While the phorbol ester PMAand dibutyryl cAMP increased C5aR protein expression, these agents had no effect on the regulation ofexpression via the promoter or the 3′UTR. This is the first study to investigate the role of both the pro-moter and 3′UTR in regulating C5aR expression and our results show that regulation of the human C5aRis similar but not identical to that of the mouse C5aR.

. Introduction

Complement activation results in the generation of C5a, a pow-rful chemoattractant, anaphylatoxin and general cell activator. Itxerts its effect via interaction with the C5a Receptor (C5aR/CD88),

seven transmembrane G-protein coupled receptor, which isighly expressed on cells of myeloid origin such as monocytes andacrophages but has also been described on a variety of other cell

ypes (Klos et al., 2009; Monk et al., 2007; Ward, 2009). Excess5a generation and its interaction with the C5aR has been doc-mented in a variety of pathological conditions including sepsis,sthma and ischaemia/reperfusion injury (Klos et al., 2009; Monkt al., 2007; Ward, 2009) and therapeutic drugs are being devel-ped to inhibit the generation of C5a and to inhibit the signallingvents induced via interaction with the C5aR. However, very lit-le is known about the regulation of expression of the C5aR. Onlyew studies have investigated the transcriptional regulation of theuman C5aR (Gerard et al., 1993) and the mouse C5aR (Hunt et al.,

005; Martin, 2007).

The first study investigating the role of the human promoteregion of the C5aR, showed that a CAT-reporter construct consist-

∗ Corresponding author. Tel.: +44 2920744824; fax: +44 2920748316.E-mail address: vandenbergcw@cardiff.ac.uk (C.W. van den Berg).

161-5890/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.molimm.2012.04.012

© 2012 Elsevier Ltd. All rights reserved.

ing of the 350 bp upstream of the start codon, the start codon and∼450 bp of intron 1 possessed promoter activity when transientlytransfected into the rat basophilic leukaemia cell line RBL-1, butnot in the nonmyeloid cells (human neuroblastoma cell line SK-N-SH) (Gerard et al., 1993). Deletion analysis showed that betweenpositions −346 and −225 (relative to the A of the ATG start codon)a repressor element prevented the expression in the nonmyeloidcells, while deletion of −82 to −49 resulted in abrogation of thereporter activity, suggesting the presence of strong promoter activ-ity within this region. Furthermore, treatment with the phorbolester PMA of RBL cells, doubled the reporter activity, but not inthe SK-N-SH cell line, however the mechanism was not furtherinvestigated (Gerard et al., 1993). The only other published studyinvestigating the human C5aR promoter region was from Barneset al. (2004) who studied the role of the single nucleotide poly-morphism (SNP) T/C (rs10404456) at −253 bp. The authors foundno difference in promoter activities using a luciferase reporterconstruct containing the 2711 bp sequence upstream of the startcodon, containing either the T or C isoform, when transiently trans-fected into U937 cells. Despite C5aR expression in humans beingmost commonly observed on monocytes and neutrophils, no stud-

ies have investigated the role of the regulatory sequences in theexpression of the C5aR in these cell types.

Two other publications, by the group of Martin, have investi-gated the transcriptional regulation of the mouse C5aR (Hunt et al.,

r Imm

2itsemrbr−sfwscn

tlagsaKttDbdstGac(eAGIrsaAe

ohsctoeuat

2

2

(SDH

E. Palmer et al. / Molecula

005; Martin, 2007). Using luciferase reporter constructs contain-ng 5′ deletions of the mouse C5aR promoter region (from −2240 bpo +38 bp, relative to the initiation of transcription start site), theyhowed that the majority of this sequence was dispensable forxpression in the mouse RAW.274 macrophage cell line and twoouse endothelial cell lines (Hunt et al., 2005). A −232 promoter

eporter construct was fully transcriptionally active in these cells,ut the activity of a −91 bp reporter construct was very mucheduced. Mutation of the consensus sequence for NF-Y(CCAAT) at96 bp abrogated the transcriptional activity and supershift assays

howed the binding of NF-YA to this sequence. Their other studyound higher levels of the mouse C5aR in microglia than astrocytes,hich was attributed to strong binding of NF-YA to the NF-Y/CCAAT

ite in microglia cells, while this site had little activity in astro-ytes suggesting different mechanism of regulation in myeloid andonmyeloid cells (Martin, 2007).

Protein expression is not just regulated by sequences inhe promoter region but also by sequences in the 3′ untrans-ated region (3′UTR). In particular certain sequences containingdenylate–uridylate rich elements (AREs), typically displaying sin-le representations or repeats of the ATTTA sequence, have beenhown to influence the stability of the mRNA transcript, thusffecting protein synthesis (Bakheet et al., 2001, 2003, 2006;uersten and Goodwin, 2003). We have recently shown that

he 3′UTR of the complement regulator decay accelerating fac-or (DAF/CD55) contains such a motif, which destabilised theAF mRNA, and that expression of DAF can be modulated byinding of HuR to its ARE (Gray et al., 2010). Using the AREDatabase a search was carried out for the presence of AU-richequences in the C5aR 3′UTR and one such sequence was iden-ified (ATGTATTTATTTTAT: 372–384 bp after the stop codon inenBank sequence: AC 099491; core sequence ATTTA underlined)nd C5aR was grouped in Cluster 5 of the AU-rich element-ontaining mRNA (ARED) database assembled by Bakheet et al.2001, 2003, 2006) (URL http://rc.kfshrc.edu.sa/ARED). Furtherxamination of the C5aR 3′UTR sequence also identified a secondRE-like motif (AGTTAATTTAAA: 474–483 bp after stop codon ofenBank sequence: AC 099491, core sequence ATTTA underlined).

t has not been investigated whether either of these AREs play aole in the stabilisation or destabilisation of the C5aR mRNA buteveral other G-protein coupled receptor (GPCR), including the �2-drenergic receptor and the angiotensin II type 1 receptor, containREs which regulate their expression (Danner et al., 1998; Pendet al., 1999).

The aim of this study was to investigate the presence and rolef regulatory sequences in the promoter and 3′UTR region of theuman C5aR in myeloid cells. The promoter region and 3′UTRequences of the C5aR were cloned and enhanced green fluores-ent protein (EGFP) reporter constructs were generated and stablyransfected into the human promyeloid cell line U937. The rolef various sequences of putative importance in the regulation ofxpression was investigated by deletion or mutation. A betternderstanding of the regulation of expression of the C5aR couldid the development of therapies where excessive C5a contributeso pathology.

. Materials and methods

.1. Reagents

Antibodies were from the following sources: mouse anti-C5aR

S5/1; IgG2a) and; polyclonal rabbit anti-C5aR (H-100; SC-25774):antaCruz (Wembley, UK); goat anti-mouse IgG FITC (GAM-FITC):ako (Ely, UK); goat anti-rabbit IgG Horseradish Peroxidase (GAR-RPO): Stratech (Soham, UK); other reagents: human IgG, Phorbol

unology 52 (2012) 88– 95 89

12-myristate 13-acetate (PMA), and N6,2′-O-dibutyryladenosine3′,5′-cyclic monophosphate sodium salt (Bt2cAMP) were fromSigma (Poole, UK).

2.2. Plasmid constructs

All reporter constructs were made using the promoterlesspEGFP plasmid (Clontech) as described previously for the DAFpromoter and 3′UTR studies (Gray et al., 2010). For the pro-moter constructs of the C5aR, based on the GenBank sequenceAC099491.2, the region −2035 bp upstream of the translation startcodon down to, but not including, the ATG was cloned and insertedin the promoterless pEGFP plasmid (2 kb construct). Numbering ofthe basepairs is relative to the initation of translation start site. Fordeletion constructs, primers were designed to decrease the pro-moter region to positions −1525, −1006, −504, −200, −100 (forease of labelling they were designated −1500, −1000, −500, −200and −100 bp, respectively). The −355 bp promoter construct wasgenerated by restriction enzyme digestion using the 2 kb constructand PstI.

Site directed mutagenesis within the 355 bp reporter constructwas carried out to mutate the following putative sites: NF�B(−238 bp to −232 bp; GCTTCCT (wt) replaced by CCATGGT (mut)),CCAAT (−123 bp to −119 bp; CCAATG (wt) replaced by GAGATC(mut)), and NFAT (−93 bp to −87 bp; and ATTTCCTCC (wt) replacedby CTTAAGGCC (mut)).

For the 3′UTR constructs the SV40 polyA region of the EF1� pro-moter plasmid was removed by excision using NotI and AflII (Grayet al., 2010), and replaced by the C5aR 3′UTR starting at the firstnucleotide after the stop codon until the end of the polyadenyla-tion site (bp 1233) based on GenBank sequence: M62505.1. Thecore sequence (ATTTA) of the ARE sequences was mutated bysite directed mutagenesis which changed each ARE from ARE1:ATTTATTTT (wt) to ACCGGTTTT (mut) and ARE2: TAATTTAA (wt)to TGGTACCA (mut). All plasmids were sequenced at least twicein both directions using ABI Big Dye 3.1 (Applied Biosystems) andsequences were compared using Blast align.

2.3. Cell culture and transfection

Two U937 cell lines were used in this study. One was obtainedfrom the European Collection of Animal Cell Cultures (ECACC) (Por-ton Down, UK; here designated U937-EC), the other was obtainedfrom Dr Peter Monk (Sheffield, UK; here designated U937-PM). Cellswere cultured in RPMI1640, 10% (v/v) foetal calf serum, penicillin G(100 U/mL), streptomycin (100 �g/mL), glutamine (2 mM), sodiumpyruvate (2 mM). Media and supplements were all from Invitro-gen (Paisley, UK). Reporter constructs were stably transfected intothe U937 cell lines by electroporation as described (Powell et al.,1997) with the modification that 1.25% (v/v) DMSO was added tothe medium for the overnight incubation after transfection, whichaccelerated the recovery of cells and the generation of reporterconstruct-expressing cells by at least 1 week. A promoterless plas-mid (pEGFP) served as a negative control (Gray et al., 2010). G418antibiotic (500 �g/mL) was added to the medium for selection;medium was refreshed 2–3 times per week and after approximately2–3 weeks, when cells were at least 95% viable, EGFP fluores-cence was assessed by flow cytometry using a FACSCalibur (BectonDickinson, San Jose, CA). Viable cells were gated and fluorescencerecorded and expressed as median fluorescent intensity (MFI). Alltransfections were carried out at least four times.

2.4. Cell surface expression of C5aR

For incubation with differentiating agents, cells were seededat 1 × 105 cells/mL and stimulated for 48 h with PMA (10 nM)

9 r Immunology 52 (2012) 88– 95

o5a(a3(aRi

2

1StUobdW(

2

t

3

3C

cscno(fhitotS(twoslymhp

t(iccwb

50

60

70

80 U937-EC

U937-PM

I

EC PM pEGFP0

10

20

30

40

MF

EC PM pEGFP

Construct

Fig. 1. Comparison of EGFP fluorescence in U937 cell lines transfected with the 2 kb-EGFP reporter constructs. U937-EC and U937-PM cells were stably transfected witheither the 2 kb (EC) or 2 kb (PM) EGFP reporter construct or the promoterless plas-mid (pEGFP) and EGFP fluorescence was measured by flow-cytometry. Two-wayANOVA performed on the promoter EGFP reporter constructs showed no signifi-cant difference between the cell line used or the 2 kb promoter EGFP constructs,

0 E. Palmer et al. / Molecula

r Bt2cAMP (0.5 mM). Cells were washed and resuspended at × 106 cells/mL in FACS buffer (PBS 5% (w/v) BSA, 0.1% (v/v) sodiumzide). Cells were incubated with heat aggregated human IgG100 �g/mL; generated by heating human IgG (10 mg/mL) in PBS,t 63 ◦C for 20 min) followed by specific anti-C5aR (5 �g/mL) for0 min at 4 ◦C, washed and incubated for 30 min with GAM-FITC1/100). Cells were washed, fixed in 1% (v/v) paraformaldehyde/PBSnd analysed on a FACSCalibur. Results are expressed as MFI ±SD.esults are representative for at least three experiments carried out

n triplicate.

.5. SDS-PAGE and western blotting

For western blotting, cell pellets were resuspended at07 cells/mL in sample buffer and electrophoresed on 12% (w/v)DS-polyacrylamide gels under non-reducing conditions and blot-ed onto Hybond Nitrocellulose (GE Healthcare UK, Little Chalfont,K). Membranes were blocked with 5% (w/v) milk/PBS, incubatedvernight with rabbit anti-C5aR (1/200 dilution in milk/PBS). Mem-ranes were washed, incubated for 1 h with GAR-HRPO (1/1000ilution in milk/PBS), washed and developed using Supersignalest Pico substrate (Pierce, Cramlington, UK) and CL-Xposure film

Pierce) as recommended by the manufacturer.

.6. Statistics

Statistical analysis was carried out using Graphpad Prism usinghe tests indicated in the legends.

. Results

.1. Cloning and expression of the 2 kb promoter region of the5aR

To investigate the role of the sequence upstream from the startodon of the human C5aR in the expression of the C5aR, the 2 kbequence upstream of the start codon (C5aR promoter region) wasloned from two U937 cell lines. The two sequences obtained wereearly identical to GenBank sequence (AC099491.2), and to eachther, with the exception of a few base pair changes or insertionsTable 1). All base pair changes, except for two ‘A’ insertions, wereound to be known single nucleotide polymorphisms (SNPs), andad been previously documented in the SNP database. None of the A

nsertions changed putative transcription factor binding sites. Fur-her analysis of genomic DNA of the two cell lines by a combinationf RT-PCR, restriction enzyme digestion and sequencing, revealedhat the U937-EC and the U937-PM were heterozygous for the threeNPs closest to the start codon: −253 C/T (rs10404456), −396 G/Ars59259352) and −464 G/A (rs60925053) (Table 1). Analysis ofhe cloned promoter regions showed that one allele of each SNPas contained within the U937-EC promoter construct whilst the

ther allele of each SNP was contained within the U937-PM con-truct (Table 1). These data suggested that the three SNPs were ininkage disequilibrium. This was subsequently supported by anal-sis of several random human genomic DNAs isolated from humanonocytes, which showed that the three SNPs occurred in two

aplotypes corresponding to those found in the U937-EC and -PMromoter constructs, respectively (data not shown).

The 2 kb 5′ regions obtained from both the U937-EC andhe U937-PM cells were cloned into the EGFP-reporter vectorupstream of the EGFP coding region)(Gray et al., 2010). The result-ng plasmids were subsequently stably transfected into both U937

ell lines and changes in EGFP fluorescence were monitored by flowytometry. Promoter activity, as indicated by level of fluorescence,as independent of the cell line transfected, which suggested that

oth cell lines contain the necessary transcription factors to drive

while there was a significant increase in fluorescence of the cells transfected withthe 2 kb promoter-containing reporter constructs compared with the promoterlessEGFP plasmid.

transcription. While the 2 kb construct obtained from the U937-ECcell line showed a somewhat higher level of activity than that ofthe U937-PM cell line, this was not statistically significant (Fig. 1),which suggests that there is no detectable difference in the tran-scriptional activity between the two promoter regions cloned.

3.2. Sequences important for transcription of the C5aR arepresent in the 200 bp region upstream of the start codon

To narrow down the regions of importance for the transcrip-tion of the C5aR, deletion constructs were made. Stable transfectionof the reporter constructs in either the U937-EC or the U937-PMcell line showed that the majority of the 2 kb region upstream ofthe start codon is dispensable for basal expression (Fig. 2A and B).Removing the region from −2 kb to −500 bp, increased the pro-moter activity, suggesting the presence of suppressor elements inthese regions. Further deletions resulted in decreased expression,suggesting the presence of important sequences between −100 bpand −1 bp.

3.3. Putative NF�B, CCAAT/NF-Y and NFAT binding sites areimportant for transcription of the C5aR

To identify regions of promoter activity, the −2 kb promotersequence from both cell lines was entered into Genomatix MatIn-spector Software (URL http://www.genomatix.de/), which can beused to predict potential transcription factor binding sites. A totalof 292 putative sites were identified (data not shown). In the first250 bp, in which most of the promoter activity was found, MatIn-spector identified only 24 motifs. Many of these motifs showed alower core and matrix similarity with the consensus sequence, orwere tissue specific to a non-relevent tissue and therefore were notinvestigated further. However, three potential cis-acting elements,NF�B (−238 bp to −232 bp), CCAAT/NF-Y (−123 bp to −119 bp) andNFAT (−93 bp to −87 bp) were selected for mutation. The putativeNF�B site was mutated as this transcription factor regulates

expression of genes whose proteins encode pro-inflammatorymolecules or are involved in the immune responses (Hoffmannet al., 2006). The CCAAT site had previously been shown to be animportant site in regulation of the mouse C5aR expression and

E. Palmer et al. / Molecular Immunology 52 (2012) 88– 95 91

Table 1Comparison of cloned U937-EC and U937-PM 2 kb 5′UTR construct sequences with GenBank sequences. Position number is based on the U937-EC sequence numbering.Previously published SNPs were identified using blastn suite-SNP (URL http://blast.ncbi.nlm.nih.gov/Blast). The U937-EC and U937-PM cell lines were found to be identicaland heterozygous for the nucleotide differences at −464, −396 and −253 bp.

Nucleotideposition

U937-EC (cloned) U937-PM (cloned) Human GenBankID: AC099491.2

Chimpanzee GenBank ID:AC193188.4

SNP

−1857 T T C C rs13967284−1842 A A – – Undocumented

−1663-1 – AAA – AAAA Undocumented−1183 A A G G rs149879338

GA

C

r2sirran

pirobiatw

Fa5bsc

−464 A G−396 G A

−253 C T

egulates expression by binding of NF-YA (Hunt et al., 2005; Martin,007). The putative NFAT site was mutated because NFAT has beenhown to regulate expression of several cell surface GPCRs, includ-ng CCR2, and the thrombin receptor (Rao et al., 1997), however itsole in C5aR expression has not been investigated. New sequenceseplacing the above putative sites were checked in MatInspectornd shown to abolish the putative transcription factor sites and noew transcription factor binding sites were introduced.

Site directed mutagenesis was carried out and the resultantlasmids were stably transfected into the U937-EC cell line. Mutat-

ng the putative CCAAT/NF-Y in the −355 bp promoter fragmentesulted in a large reduction of EGFP fluorescence, while mutationf NFAT motif reduced the C5aR promoter EGFP reporter activity toackground levels (Fig. 3), suggesting that both sites play keys roles

n transcription of the C5aR gene. Some promoter activity was alsottributed to the putative NF�B site, as site directed mutagenesis of

his site significantly reduced EGFP fluorescence compared to theildtype −355 bp fragment (Fig. 3).

A

200

300

*****

***** ***

ge

EG

FP

ch

an

ge

B

-200 0 -150 0 -100 0 -500 -35 5 -20 0 -10 0 pEGFP0

100 ** ** ** ***

Reporter construct

Pe

rce

nta

150

200

250

300

350

**

***

**

*

EG

FP

ch

an

ge

BRepo rter con struct

-2000 -150 0 -100 0 -500 -35 5 -20 0 -10 0 pEGFP0

50

100

150

Pe

rce

nta

ge

Repo rter cons truc t

ig. 2. Effect of 5′ region deletions on reporter construct expression in U937-ECnd U937-PM. U937-EC (A) and U937-PM (B) cells were stably transfected with′ promoter deletion EGFP reporter constructs. EGFP fluorescence was monitoredy flow cytometry. One-way ANOVA followed by Tukey’s multiple comparison testhowed significant differences represented by *p < 0.05, **p < 0.01 and ***p < 0.001ompared to promoterless pEGFP.

G rs60925053G rs59259352C rs10404456

3.4. Role of 3′UTR in transcription of the C5aR

To determine if the C5aR expression is regulated at the mRNAlevel, the 3′UTR of both the U937-EC and the U937-PM werecloned. The cloned sequences were highly homologous with theGenBank sequence M62505.1, which differed by a few bases withanother GenBank sequence NM 001736.3 (Table 2). The sequenceof the U937-EC differed by only one nucleotide from the sequencesin Genbank, while the sequence of the U937-PM differed byfour nucleotides, which could represent allelic variants in link-age disequilibrium, but this was not explored further. None ofthe differences between the cloned 3′UTRs and the Genbanksequences altered the putative ARE sequences (ARE1: 372–384;ARE2: 474–483). The 3′UTR cloned from the U937-EC cell line,which was most identical to the published sequences, was inserteddownstream of the EGFP gene in the EF1�-pEGFP vector (Grayet al., 2010). The core sequences of the two AU-rich elements weremutated using site directed mutagenesis. The C5aR wildtype 3′UTRtransfectants showed lower EGFP fluorescence compared to thepEGFP vector SV40 3′UTR after stable transfection (Fig. 4), whichwas expected as the SV40 3′UTR is very stable. Mutating either AREhad no significant effect on EGFP fluorescence compared to the wildtype (WT), suggesting that these AREs are not of importance forC5aR expression (Fig. 4).

3.5. Effect of differentiation on C5aR and reporter constructexpression

To investigate the role of the promoter regions in inducedexpression of the C5aR, first the expression of the C5aR proteinwas analysed. Numerous agents (e.g. PMA, Bt2cAMP, IFN�, IL-6,

100

120

*

han

ge

20

40

60

80*

****** ** *

rcen

tag

e E

GF

P c

WT NFkB CCAAT NFAT pEGFP0P

er

Fig. 3. Site directed mutagenesis of putative CCAAT/NF-Y or NFAT cis-acting ele-ments abolished transcription control in −355 bp + pEGFP reporter construct.Mutant −355 bp promoter constructs were stably transfected into the U937-EC cellline and EGFP fluorescence was monitored by flow cytometry. One-way ANOVA fol-lowed by Tukey’s multiple comparison test show significantdifference, with the WT−355 bp construct (*p < 0.05, ***p < 0.001).

92 E. Palmer et al. / Molecular Immunology 52 (2012) 88– 95

Table 2Sequence comparison of the 3′UTR cloned from U937-EC and U937-PM with GenBank sequences.

Nucleotideposition

U937-EC U937-PM Human GenBankID: M62505.1

Human GenBankID: NM 001736.3

Chimpanzee Genbank ID:AC193188.4

SNP

67 T C T T T Undocumented237 C A C (C-rich) CA CC rs17218697: A/C

rs56349709: –/A/C364 – – A (A-rich) – – rs34285755:–/A571 A G G G G Undocumented584 – G – – – Undocumented678 – G – – – Undocumented701 C C C – C rs35042977: –/C

T

C

A

LCGflnoP(

ttFEtsEImsaoc

4

ogU

Fc(aAA

971 T A T

1089 C T C

1219 A T A

PS or TNF�) have been reported to induce the expression of the5aR in monocytic cells including U937 (Burg et al., 1995, 1996;asque et al., 1995, 1998; Gerard and Gerard, 1991). Analysis byow-cytometry and western blotting showed that the U937-EC didot respond to the above mentioned agents and that U937-PM cellsnly showed an increased C5aR expression after incubation withMA and Bt2cAMP (Fig. 5B), but not with IFN�, IL-6, LPS or TNF�not shown).

The effects of PMA and Bt2cAMP on reporter gene expression inhe 2 kb promoter constructs, which had been transfected into thewo U937 cell lines were subsequently investigated. As shown inig. 5C, although some increase in fluorescence was observed in theC reporter construct in both cell lines after incubation with PMA,his was statistically not significant. Incubation with Bt2cAMP onlyhowed a slight but not significant increase in the expression of theC promoter construct but only in the U937-EC cell line (Fig. 5D).ncubation with Bt2cAMP reduced the expression of the PM pro-

oter construct, rather than increasing it as expected. These resultsuggest that the increase in expression of the C5aR induced by PMAnd Bt2cAMP is not mediated via the promoter sequences. No effectf the differentiating agents was observed on the 3′UTR reporteronstructs (not shown).

. Discussion

The aim of this study was to investigate the presence and role

f motifs within the promoter region and 3′UTR of the human C5aRene that may regulate expression of the C5aR in the monocytic937 cell line. We found three motifs in the promoter region that

750

1000

1250

1500

1750

MF

I

WT ARE 1 ARE 2 SV400

250

500

ig. 4. Effects of 3′UTR and ARE mutations on EGFP fluorescence. EGFP reporteronstructs containing the 3′UTR of the C5aR (WT) or the mutated AU-rich regionsARE1 and ARE2) or the SV40 3′UTR were stably transfected into the U937-EC cellsnd changes in EGFP fluorescence were monitored by flow cytometry. One-wayNOVA performed comparing the C5aR 3′UTR WT, 3′UTR ARE1 mutant and 3′UTRRE2 mutant showed no significant difference.

T UndocumentedC UndocumentedA Undocumented

are involved in basal C5aR expression, while the AU-rich sequencespresent in the 3′UTR were not involved.

The 2 kb promoter region cloned from two different U937 celllines differed in a few nucleotides, from each other and from thesequence published in GenBank. All differences could be ascribedto documented SNPs and common insertions (Table 1). Analysisof the two cell lines by a combination of restriction digestion andnucleotide sequence analysis, showed that they were actually het-erozygous for the three SNPs closest to the start codon (data notshown), that the SNPs were in linkage disequilibrium, and one ofeach haplotype had been cloned. One of the SNPs at −253 (T/C) hadpreviously been investigated and found not to affect a luciferasereporter construct, transiently transfected into U937 cells (Barneset al., 2004).

The insertions that we detected were all in poly-A tracts; the sin-gle A insertion was present in both clones of the two U937 cell linesin a relatively short tract of 5 As; the AAA insertion was only found inthe U937-PM in a large adenosine repeat stretch (22 uninterruptedAs in Genbank). Interestingly, an AAAA insertion was present inthis C5aR promoter region from Chimpanzees (GenBank accessionnumber AC193188.4). Insertions in A-repeat stretches are com-mon and several such SNP insertions can be found in the sequenceupstream of the start codon of the C5aR. It is also possible that theA insertions we observed were a consequence of ‘slipping’ of thepolymerase during the cloning procedure.

The obtained sequences were cloned into an EGFP reporter vec-tor (Gray et al., 2010). U937 cells are difficult to transfect and inorder to obtain sufficient cells for analysis, stable transfectantswere generated using the G418 selection medium. The use of stabletransfectants has the advantage that the actual process of transfec-tion does not affect the expression of the reporter construct. This isin contrast to transient transfection, where transfection-inducedcell activation can occur. Another advantage of using EGFP as areporter is that it can directly be measured by flow-cytometry anddoes not require cell counting and cell extraction and non-viablecells can be discarded as they have a different forward/side scat-ter pattern than live cells. Both U937 cell lines after transfectionand selection, expressed the EGFP reporter under control of thepromoter of the C5aR. No difference was found between the twocloned sequences in levels of expression (Fig. 1), suggesting that theSNPs in the two constructs had no effect on expression of the C5aR.The U937-EC cells seemed to express both constructs at a slightlyhigher level than the U937-PM cells; however this difference wasnot statistically significant (Fig. 1).

Deletion constructs made of the 2 kb sequence cloned from theU937-EC cell line, showed that the majority of this region wasdispensable, which is similar to results that have been previously

shown for the mouse C5aR (Hunt et al., 2005). Similar to this studytoo was the finding that the CCAAT site plays a major role in theexpression of the C5aR, which is not surprising as the CCAAT sitesin mouse and human are highly homologous and the CCAAT motif

E. Palmer et al. / Molecular Immunology 52 (2012) 88– 95 93

Fig. 5. Expression of the C5aR and reporter constructs after differentiation. U937-EC (A) and U937-PM (B) cells incubated for 48 h with 10 nM PMA or 0.5 mM Bt2cAMP or withmedium only (control) were stained for the C5aR and analysed by flow cytometry. Results are expressed as median fluorescent intensities (MFI); background = no primaryantibody. Inserted into B: western blots of the cell pellets of U937-PM detected with rabbit anti-C5aR; treatments (as in B): c, control; p, PMA; b, Bt2cAMP. C and D: U937-EC(C) and U937-PM cells (D), stably transfected with either 2 kb (EC) or 2 kb (PM) EGFP reporter construct or the promoterless pEGFP vector, were incubated for 2 days in thea orescc ANOVt

imosCpitfswsomba

ePvdUFeeiP

bsence or presence of 10 nM PMA (C) or 0.5 mM Bt2cAMP (D). Changes in EGFP fluompared to the untreated cells (represented by dotted line) ±SEM (n = 3). One-wayhe different constructs when cells were treated with either PMA or Bt2cAMP.

s one of the most common promoter elements in eukaryotic pro-oters, and is normally found either in the forward or reverse

rientation and is preferentially located −80/−100 bp to translationtart site (Mantovani, 1998). MatInspector analysis of the human5aR promoter sequence predicted that the CCAAT motif present atosition −123 bp to −119 bp relative to the start codon. Consider-

ng an initiation of transcription site at −39 bp (Gerard et al., 1993),he CCAAT motif in the human C5aR corresponds to −84 to −80 bprom the initiation of transcription site at −39 bp relative to thetart codon. This site would likely bind the NF-Y binding protein,hich would correlate with the study by Hunt et al. (2005) who

howed, using EMSA and super shift EMSA, that NF-Y, but not c-Fosr ETS1/2, bound to the homologous CCAAT motif present in theouse C5aR promoter. However, we were unable to detect specific

inding of proteins to this sequence using electrophoretic mobilityssays (data not shown).

Mutation of the putative NFAT site was based on its role inxpression of a variety of molecules and the observation that thatMA can stimulate CCL23 expression in the U937 cell line via acti-ation of NFAT (Shin et al., 2007). Furthermore, PMA has been wellocumented in the literature to increase C5aR expression in the937 cell line (Burg et al., 1996; Rubin et al., 1991), confirmed inig. 5B. Mutation of a putative NFAT binding site also ablated the

xpression of the reporter, demonstrating an important role in thexpression of the C5aR for this sequence too, which is a novel find-ng; this has not been investigated in the mouse C5aR. However,MA had no effect on the expression of the reporter constructs

ence were monitored by flow-cytometry. Results are expressed as relative changeA followed by Tukey’s multi comparison showed no significant difference between

themselves (Fig. 5C), suggesting that the increased expression ofC5aR induced by PMA, is not mediated via NFAT and is not mediatedby interaction of transcription factors anywhere within the 2 kb 5′

region. The remainder of the reporter activity, observed within the−100 bp deletion fragment, may be attributed to the putative NFATsite (−93 bp to −87 bp), (Figs. 2 and 3). However it is important tonote that the −100 bp fragment may also contain the core promoterTATA box, present in about a quarter of all genes, which could con-tribute towards the remaining transcriptional activity. A TATA-likemotif (TTTAAAA) is present at −65 to −59 bp relative to the startcodon, which corresponds to −26 to −20 bp relative to the tran-scription initiation site at −39. This is consistent with the positionexpected of a TATA box, approximately 20–35 bp upstream of atranscription start site. The observation that mutation of the NFATsite ablated the C5aR promoter construct induced EGFP expression,suggests that the putative TATA box may not be of importance.The study by Hunt et al. did not investigate the role of the NFATsite, and attributed the remainder of the transcriptional activity oftheir shortest reporter construct to the putative TATA box motif(TTTAAA). However that construct also contains an NFAT motif,which could have been responsible for the transcriptional activity.

Mutation of the NF�B site, reduced the expression slightly butsignificantly, which is a novel finding and in contrast to the study of

the mouse C5aR promoter region, where mutation of this sequencewas not found to have an effect (Hunt et al., 2005). This differencemay be attributed to the use of different cell types and becausethe mouse putative NF�B is quite different from the human site.

9 r Imm

Es

eetsobcsote

fpitCCtIecFwcitelapTNtpiftlafi

BhcBB

romCgcslorhurr

4 E. Palmer et al. / Molecula

lectrophoretic mobility assays were unsuccessful in determiningpecific binding to these three sequences (data not shown).

The 3′UTR was found to contain two AU-rich elements, how-ver, mutating these sequences did not alter reporter constructxpression, suggesting that these sequences are not involved inhe regulation of expression of the C5aR. PMA has been shown totabilise mRNA regulating ARE containing mRNA such as that ofncostatin M in U937 cells (Bandyopadhyay et al., 2008). The num-er and pattern of ATTTA repeats determines the interaction withertain regulatory molecules; but while the oncostatin M ARE con-ists of multiple ATTTA pentamers and belongs to class 2, the C5aRnly has one true ‘ATTTA’ sequence fitting this classification. It ishus maybe not surprising that mutation of this sequence had noffect on the expression of the reporter construct.

While the majority of our experiments were carried out in undif-erentiated cells, we also wanted to investigate the role of theromoter sequences in differentiated cells. Various agents includ-

ng PMA, Bt2cAMP and IFN�, IL-6, LPS or TNF� have been reportedo differentiate U937 cells and/or increase the expression of the5aR. However, we only observed an increase in expression of the5aR after incubation with PMA or Bt2cAMP, and only in one ofhe U937 cell lines used (Fig. 5B). Publications previously reportingFN�, IL-6, LPS or TNF� induced expression of the C5aR are likelyrroneous, as C5aR expression was either assessed using mono-lonal antibodies of the IgG2a isotype, which, without blocking ofc� receptors on monocytic cells, would give artefactual results, orere assessed by C5a-ligand binding, however, the recently dis-

overed alternative receptor for C5a, C5L2 (Cain and Monk, 2002),s most likely responsible for this increase as it has been shownhat, e.g. IFN� is a strong inducer of C5L2 in U937 cells (Johswicht al., 2006). Surprising, however, was that only one of the U937 cellines used in this study responded to the differentiating agents byn increased expression in the C5aR: although both changed mor-hology, only the U937-PM cell line increased its C5aR expression.he U937 cell line is a long established cell line (Sundstrom andilsson, 1976) and it is well known that when cells have been cul-

ured in different labs, that they may change their phenotype. It hasreviously been reported that Bt2cAMP differentiation occurs only

n a subset ∼60–80% of U937 cells (Rubin et al., 1986). It is there-ore possible that over time in culture the U937-EC cell line has losthe subset of cells which respond to Bt2cAMP, while U937-PM hasost the subset of cells which do not respond to Bt2cAMP. This maylso suggest that there may be regulatory proteins or transcriptionactors, which are responsive to Bt2cAMP and PMA that are presentn the U937-PM but not U937-EC cell line.

Investigating the effect of the differentiating agents PMA andt2cAMP on the 2 kb promoter constructs in either cell line,owever, did not show induction of expression of the reporteronstructs (Fig. 5C and D), suggesting that the effects of PMA andt2cAMP are not on motifs in the promoter region. How PMA andt2cAMP increase the C5aR expression remains to be established.

In conclusion, three motifs have been identified in the promoteregion of the C5aR that are of importance in the basal expressionf this important molecule in a myeloid cell line. However, nootif in the promoter seems to be responsible for increasing the

5aR expression in response to differentiating agents. The C5aRene consists of two exons, separated by a 9 kb intron betweenodons 1 and 2, which may contain regulatory elements. To furthertudy regulatory elements in the C5aR gene, it needs to be estab-ished which physiological stimuli actually change the expressionf the C5aR in primary monocytes and macrophages as in the cur-ent literature too many artefactual/poorly controlled experiments

ave been published. The expression of the C5aR is not just reg-lated at the level of transcription or translation, but may also beegulated at the protein level. C5aR expression was shown to beeduced on human neutrophils during sepsis (Seely et al., 2002),

unology 52 (2012) 88– 95

but the mechanism of this alteration in expression is not known.We recently showed that the C5aR on neutrophils and U937 cellscan be cleaved on the cell surface by endogenous metalloprotease,after incubation with Loxosceles spider venom (van den Berg et al.,2012), which induces some symptoms also observed in sepsis, andthe cleavage of C5aR may be an important regulatory step in reduc-ing the C5aR expression. Another regulatory mechanism throughwhich the function of the C5aR can be regulated is by alteration ofexpression of the alternate receptor C5L2: decreased expression ofthis receptor correlated with increased responsiveness to C5a (Rabyet al., 2011). A better understanding of how the C5aR expressionand function is regulated may offer potential insights/mechanismsto limit the destructive effects of excess C5a during pathology.

References

Bakheet, T., Frevel, M., Williams, B.R., Greer, W., Khabar, K.S., 2001. ARED: humanAU-rich element-containing mRNA database reveals an unexpectedly diversefunctional repertoire of encoded proteins. Nucleic Acids Research 29, 246–254.

Bakheet, T., Williams, B.R., Khabar, K.S., 2003. ARED 2.0: an update of AU-rich ele-ment mRNA database. Nucleic Acids Research 31, 421–423.

Bakheet, T., Williams, B.R., Khabar, K.S., 2006. ARED 3.0: the large and diverse AU-richtranscriptome. Nucleic Acids Research 34, D111–D114.

Bandyopadhyay, S., Sengupta, T.K., Spicer, E.K., 2008. PMA induces stabilization ofoncostatin M mRNA in human lymphoma U937 cells. Biochemical Journal 410,177–186.

Barnes, K.C., Caraballo, L., Munoz, M., Zambelli-Weiner, A., Ehrlich, E., Burki, M.,Jimenez, S., Mathias, R.A., Stockton, M.L., Deindl, P., Mendoza, L., Hershey, G.K.,Nickel, R., Wills-Karp, M., 2004. A novel promoter polymorphism in the geneencoding complement component 5 receptor 1 on chromosome 19q13.3 is notassociated with asthma and atopy in three independent populations. Clinicaland Experimental Allergy 34, 736–744.

Burg, M., Martin, U., Bock, D., Rheinheimer, C., Kohl, J., Bautsch, W., Klos, A., 1996.Differential regulation of the C3a and C5a receptors (CD88) by IFN-gamma andPMA in U937 cells and related myeloblastic cell lines. Journal of Immunology157, 5574–5581.

Burg, M., Martin, U., Rheinheimer, C., Kohl, J., Bautsch, W., Bottger, E.C., Klos, A., 1995.IFN-gamma up-regulates the human C5a receptor (CD88) in myeloblastic U937cells and related cell lines. Journal of Immunology 155, 4419–4426.

Cain, S.A., Monk, P.N., 2002. The orphan receptor C5L2 has high affinity binding sitesfor complement fragments C5a and C5a des-Arg(74). The Journal of BiologicalChemistry 277, 7165–7169.

Danner, S., Frank, M., Lohse, M.J., 1998. Agonist regulation of human beta2-adrenergic receptor mRNA stability occurs via a specific AU-rich element. TheJournal of Biological Chemistry 273, 3223–3229.

Gasque, P., Chan, P., Fontaine, M., Ischenko, A., Lamacz, M., Gotze, O., Morgan, B.P.,1995. Identification and characterization of the complement C5a anaphylatoxinreceptor on human astrocytes. Journal of Immunology 155, 4882–4889.

Gasque, P., Singhrao, S.K., Neal, J.W., Wang, P., Sayah, S., Fontaine, M., Morgan, B.P.,1998. The receptor for complement anaphylatoxin C3a is expressed by myeloidcells and nonmyeloid cells in inflamed human central nervous system: analy-sis in multiple sclerosis and bacterial meningitis. Journal of Immunology 160,3543–3554.

Gerard, N.P., Bao, L., Xiao-Ping, H., Eddy Jr., R.L., Shows, T.B., Gerard, C., 1993. Humanchemotaxis receptor genes cluster at 19q13.3–13.4. Characterization of thehuman C5a receptor gene. Biochemistry 32, 1243–1250.

Gerard, N.P., Gerard, C., 1991. The chemotactic receptor for human C5a anaphyla-toxin. Nature 349, 614–617.

Gray, L.C., Hughes, T.R., van den Berg, C.W., 2010. Binding of human antigen R (HuR)to an AU-rich element (ARE) in the 3′ untranslated region (3′UTR) reducesthe expression of decay accelerating factor (DAF). Molecular Immunology 47,2545–2551.

Hoffmann, A., Natoli, G., Ghosh, G., 2006. Transcriptional regulation via the NF-kappaB signaling module. Oncogene 25, 6706–6716.

Hunt, J.R., Martin, C.B., Martin, B.K., 2005. Transcriptional regulation of the murineC5a receptor gene: NF-Y is required for basal and LPS induced expression inmacrophages and endothelial cells. Molecular Immunology 42, 1405–1415.

Johswich, K., Martin, M., Thalmann, J., Rheinheimer, C., Monk, P.N., Klos, A., 2006. Lig-and specificity of the anaphylatoxin C5L2 receptor and its regulation on myeloidand epithelial cell lines. The Journal of Biological Chemistry 281, 39088–39095.

Klos, A., Tenner, A.J., Johswich, K.O., Ager, R.R., Reis, E.S., Kohl, J., 2009. The role of theanaphylatoxins in health and disease. Molecular immunology 46, 2753–2766.

Kuersten, S., Goodwin, E.B., 2003. The power of the 3′ UTR: translational control anddevelopment. Nature Reviews Genetics 4, 626–637.

Mantovani, R., 1998. A survey of 178 NF-Y binding CCAAT boxes. Nucleic AcidsResearch 26, 1135–1143.

Martin, B.K., 2007. Transcriptional control of complement receptor gene expression.Immunologic Research 39, 146–159.

Monk, P.N., Scola, A.M., Madala, P., Fairlie, D.P., 2007. Function, structure and thera-peutic potential of complement C5a receptors. British Journal of Pharmacology152, 429–448.

r Imm

P

P

R

R

R

E. Palmer et al. / Molecula

ende, A., Giacche, M., Castigliola, L., Contini, L., Passerone, G., Patrone, M., Port,J.D., Lotti, G., 1999. Characterization of the binding of the RNA-binding proteinAUF1 to the human AT(1) receptor mRNA. Biochemical and Biophysical ResearchCommunications 266, 609–614.

owell, M.B., Marchbank, K.J., Rushmere, N.K., van den Berg, C.W., Morgan, B.P., 1997.Molecular cloning, chromosomal localization, expression, and functional char-acterization of the mouse analogue of human CD59. Journal of Immunology 158,1692–1702.

aby, A.C., Holst, B., Davies, J., Colmont, C., Laumonnier, Y., Coles, B., Shah, S., Hall,J., Topley, N., Kohl, J., Morgan, B.P., Labeta, M.O., 2011. TLR activation enhancesC5a-induced pro-inflammatory responses by negatively modulating the secondC5a receptor, C5L2. European Journal of Immunology 41, 2741–2752.

ao, A., Luo, C., Hogan, P.G., 1997. Transcription factors of the NFAT family: regulationand function. Annual Review of Immunology 15, 707–747.

ubin, J., Titus, L., Nanes, M.S., 1991. Regulation of complement 5a receptorexpression in U937 cells by phorbol ester. Journal of Leukocyte Biology 50,502–508.

unology 52 (2012) 88– 95 95

Rubin, J.E., Chenoweth, D.E., Catherwood, B.D., 1986. 1,25-Dihydroxyvitamin D3and adenosine 3′ ,5′-monophosphate synergistically promote differentiation ofa monocytic cell line. Endocrinology 118, 2540–2545.

Seely, A.J., Naud, J.F., Campisi, G., Giannias, B., Liu, S., DiCarlo, A., Ferri, L.E., Pascual,J.L., Tchervenkov, J., Christou, N.V., 2002. Alteration of chemoattractant receptorexpression regulates human neutrophil chemotaxis in vivo. Annals of Surgery235, 550–559.

Shin, Y.H., Lee, G.W., Son, K.N., Lee, S.M., Kang, C.J., Kwon, B.S., Kim, J., 2007. Pro-moter analysis of human CC chemokine CCL23 gene in U937 monocytoid cells.Biochimica et Biophysica Acta 1769, 204–208.

Sundstrom, C., Nilsson, K., 1976. Establishment and characterization of a human his-tiocytic lymphoma cell line (U-937). International Journal of Cancer 17, 565–577.

van den Berg, C.W., Goncalves-de-Andrade, R.M., Okamoto, C.K., Tambourgi, D.V.,2012. C5a receptor is cleaved by metalloproteases induced by sphingomyelinaseD from Loxosceles spider venom. Immunobiology.

Ward, P.A., 2009. Functions of C5a receptors. Journal of Molecular Medicine 87,375–378.