Sei sulla pagina 1di 8

Proceedings

The antimicrobial properties of C-reactive protein (CRP)

Sandra S.H. Tan1, Patricia M.L. Ng1, Bow Ho2, Jeak Ling Ding1

Departments of 1Biological Sciences and 2Microbiology, National University of Singapore, Singapore

C-reactive protein, CRP, is a predominant pattern-recognition receptor (PRR) in the plasma of the
horseshoe crab, which recognizes lipopolysaccharide (LPS). Native CRP2 has previously been
shown to exhibit agglutination activity against the polysialic capsule of Escherichia coli K1 but its
role in bacterial clearance is not well characterized. In this work, the antimicrobial activity of a
recombinant CRP2 isoform (rCRP2) was tested against E. coli, Pseudomonas aeruginosa and
Staphylococcus aureus. rCRP2 agglutinates bacteria and exhibits bactericidal activity against Gram-
negative bacteria. In addition, the antimicrobial activity of rCRP2 is calcium-independent. GST
pulldown experiments suggest that in the naïve physiological state, CRP2 interacts with
hemocyanin, native CRPs, a 35-kDa plasma lectin and an as yet unidentified 40-kDa protein. This
interaction was enhanced upon Pseudomonas infection. We propose that rCRP2 is a PRR with
potent antimicrobial activity and its interacting partners contribute to effective bacterial clearance.

Keywords: C-reactive protein, horseshoe crab, endotoxin, lipopolysaccharide, antimicrobial activity

INTRODUCTION is a formidable obstacle to its prevention and treatment.


The ability to detect and neutralize LPS before it
Infections by Gram-negative bacteria are the predomi- exerts detrimental effects on host systems is an attractive
nant cause of clinical sepsis.1 The key to the primacy of anti-sepsis strategy. Infection by pathogens involves
Gram-negative bacteria in causing sepsis lies in their crossing the blood barrier and systemic infection is char-
unique pathogen-associated molecular pattern (PAMP) – acterized by invasion of pathogens into the plasma.
lipopolysaccharide (LPS). Structural analysis indicates Rapid neutralization is thus dependent upon fast-acting
that LPS protects Gram-negative bacteria against exter- and readily-accessible immune molecules. Owing to
nal assaults by forming a protective barrier.2,3 In humans, their extracellular location, plasma lectins, such as CRP,
LPS triggers a series of physiological responses that that recognize PAMPs represent frontline pattern recog-
result in increased biosynthesis of both pro-and anti- nition receptors (PRRs) that can be envisaged to play a
inflammatory cytokines such as TNF-α and inter- pivotal role in innate immunosurveillance to suppress
leukins.4 LPS also activates the complement system, the development of septicaemia upon pathogen entry.
fuelling the inflammatory response. Despite attempts at We have previously isolated CRP2 as a member of the
ameliorating the effects of sepsis by novel applications repertoire of LPS-binding proteins.6 In an earlier study,
of cytokine antagonists,5 the complexity of septicaemia three types of tCRPs were purified from the plasma of
Tachypleus tridentatus.17 These are classified based on
their different affinities to fetuin–agarose and phospho-
Received 17 December 2004
Revised 21 February 2005
rylethanolamine–agarose. CRP2 from Carcinoscorpius
Accepted 21 February 2005 is so named because it exhibits the highest sequence
homology to tCRP2.
Correspondence to: Jeak Ling Ding, Department of Biological
Sciences, 14 Science Drive 4, National University of Singapore,
Here, we demonstrate that a single isoform of CRP2
Singapore 117543 from the horseshoe crab, Carcinoscorpius rotundicauda,
Tel: +65 68742776; Fax: +65 67792486; E-mail: dbsdjl@nus.edu.sg exhibits antimicrobial activity. In vivo infection experi-
URL: http://www.dbs.nus.edu.sg/Staff/ding.html ments suggest that CRP expression was significantly up-

Journal of Endotoxin Research, Vol. 11, No. 4, 2005 © W. S. Maney & Son Ltd
DOI 10.1179/096805105X37402
250 Tan, Ng, Ho, Ding

regulated following infection.6 In vitro, recombinant

Units Purification
CRP2 (rCRP2) causes agglutination and is bactericidal

(-fold)

16.33
14.42
1.02

15.7
towards Gram-negative bacteria. The bactericidal activ-

16
1
ity of rCRP2 is Ca2+-independent. Additionally, rCRP2
recruits plasma proteins and this interaction is enhanced

(total)
during infection. This complex probably boosts bacterial

0.052
0.053

0.832
0.849
0.75
0.82
clearance by mediating other antimicrobial mechanisms.

Recovery

68.80%

38.70%
32.23%
30.00%
26.25%
100%
rCRP-2
MATERIALS AND METHODS

Collection of horseshoe crab hemolymph

3.303

1.855
1.545
Units
(mg)

1.44
1.26
4.8
The C. rotundicauda were collected from the Kranji
River estuary, Singapore and acclimatised overnight in
minimal levels of 30% sea water. Hemolymph was

(mg/ml)

0.077
0.066

0.124
0.103
0.096
0.086
obtained by cardiac puncture and cells were removed by
centrifugation before the plasma was used for further

The amount of rCRP2 was determined by densitometric measurements on SDS-PAGE and Bradford assay quantification.
analysis. For infection, 106 CFU of Pseudomonas aerug-
inosa was resuspended in 200 µl of 0.9% saline and
Purification
(-fold)

16.33

16.46
injected intracardially. At 1 h post-infection (hpi), the

1.02

16
infected hemolymph was similarly collected and 1
processed.
(total)
0.052
0.053

0.832
0.849

0.856
Units

Cloning, expression and purification of rCRP2

A DNA fragment encoding the mature sequence of


rGST-CRP-2
Recovery

68.80%

38.70%
32.23%

27.10%
100%

CRP2 was cloned into the pGEX-4T-3 plasmid and


transformed into Escherichia coli Top10-competent
cells. Sequence-verified plasmids were then transformed
into the E. coli BL21 strain for expression. To obtain
3.303

1.855
1.545

1.301
Units
(mg)
4.8

rCRP2 protein, E. coli BL21 transformants were inocu-


lated into 2x YTA media containing 100 µg/ml of ampi-
cillin and cultured overnight at 37°C with shaking at 230
(mg/ml)

rpm. An aliquot of 100 µl of overnight culture was added


0.077
0.066

0.124
0.103

0.089

to 1 l of 2x YTA medium.
The bacteria were grown at room temperature with
shaking at 230 rpm until OD600nm ~0.6–0.8 was achieved.
Total
(mg)
92.4

2.23
1.82
1.92
1.52
62

Isopropyl-1-thio-β-D-galactopyranoside (IPTG) was


Protein

added to a final concentration of 0.1 mM. At 3 h post-


induction, bacteria were harvested and washed by centrifu-
(mg/ml)

0.148
0.121
0.128
1.036
1.54
1.24

gation at 5000 g for 5 min at 4°C. The cells were


resuspended in 50 ml ice-cold Tris-buffered saline, pH 7.4
(TBS) and French pressed at 15 kPa. The cell debris was
Volume

14.7
(ml)

removed by centrifuging the lysate at 9000 g for 1 h at 4°C.


60
50

Glutathione–Sepharose 4B 15
15
15
Table 1. Purification of rCRP2

Bacterial supernatants containing GST-CRP2 were repeat-


edly passaged through a 50% glutathione–Sepharose 4B
Thrombin/benzamidine

(Amersham Biosciences) column at 4°C overnight.


Bacterial supernatant

Following washing with TBS, target proteins were eluted


Whole cell lysate

Soluble proteins

with 20 mM reduced glutathione and dialysed overnight


LPS removal

against TBS containing 10 mM EDTA at 4°C. A super-


Purification

natant containing GST was similarly treated as a control.


Dialysis

For purification of untagged CRP2, the GST domain


step

of the fusion protein was cleaved on-column with 2 units


Antimicrobial activity of CRP 251
GST pulldown assay to determine the protein–protein
interactions of rCRP2

Purified GST and GST-CRP2 fusion proteins were incubated


overnight with freshly-prepared glutathione–Sepharose 4B.
Plasma samples were introduced to aliquots of the recombi-
nant protein-bound slurry. Bound proteins were washed with
TBS, 0.35% Tween-20 and eluted with 50 mM Tris-Cl, 10
mM reduced glutathione, pH 8.0. The eluants were analysed
on 12% SDS-PAGE.

In-gel digestion and protein identification by mass


spectrometry

Protein bands of interest were excised from SDS-PAGE


and dehydrated with acetonitrile. Proteins were then
reduced with 10 mM DTT at 57°C for 1 h, and alkylated
by 55 mM iodoacetamide at room temperature for 1 h.
In-gel digestion was carried out with 12.5 ng/µl trypsin
at 37°C for 15 h. The resultant peptide fragments were
identified by MALDI-TOF MS/MS analysis (4700
Proteomics Analysis, ABI). Peptide mass fingerprints
(pmfs) of the digested proteins were analysed by Mascot
(<http://www.matrixscience.com>) against the Mass

Fig. 1. rCRP2 exerts its effects on Gram-negative bacteria. Bacterial


suspensions (105 CFU/ml) were incubated with a final concentration of 10
µM of rCRP2 (test) or with TBS (control). OD600nm readings of each
sample were plotted as a function of time. In the presence of rCRP2, both
P. aeruginosa (A) and E. coli (B) test cultures did not exhibit increases in
density, relative to controls. In contrast, S. aureus cultures were not
affected by the presence of rCRP2 (C). Data points represent the mean
values of three independent experiments and flags indicate SD.

of thrombin (Amersham) to release the rCRP2.


Following 24 h incubation at 4°C, target proteins were
removed by eluting with TBS. The eluant was incubated
with benzamidine–Sepharose (Amersham) at 4°C to
remove excess thrombin. Protein solutions containing
either GST-tagged or untagged rCRP2 were subjected to
LPS removal by Triton X-114, according to the method
developed by Petsch and Anspech.7
Total protein obtained from various purification steps
was quantified using the Bradford assay.8 To determine the
expression level of rCRP2, different fractions were elec- Fig. 2. The effects of rCRP2 are calcium-independent. The antimicrobial
activity of rCRP2 on both P. aeruginosa and E. coli is calcium-independent.
trophoresed using SDS-PAGE. CRP-specific proteins OD600nm measurements of 105 CFU/ml bacteria were taken following 1 h of
were verified by Western blotting and the relative amounts culture with CRP, in the presence and absence of Ca2+. The results show stasis
were quantified densitometrically using Quantity One soft- in OD600nm measurement regardless of the presence or absence of Ca2+. A
ware (Biorad). control was set up by replacing the CRP with TBS buffer.
252 Tan, Ng, Ho, Ding

A B

Fig. 3. rCRP2 causes bacterial agglutination and is bactericidal. rCRP2


(10 µM) causes visible bacterial agglutination of 5 x 108 CFU/ml of P.
aeruginosa (A, bottom panel). Bacterial clumps are indicated by C
brackets/arrows. This effect was immediately observable. In contrast,
bacteria suspended in TBS did not agglutinate (A, top panel). Both
photomicrographs were taken at x400 magnification. (B) Residual colony
counts reveal that 2.5 µM rCRP2 is capable of reducing viable bacteria
densities by up to 109-fold. This result was confirmed by (C) the Miles and
Misra method10 of serial dilutions to monitor viable bacteria after the
antimicrobial assault by rCRP2.

Spectrometry protein sequence Data Base (MSDB).


Peaks of pmfs were also matched to known proteins fol-
lowing in-silico analysis.9

Antimicrobial assays

E. coli ATCC25922, P. aeruginosa ATCC27853 and


Staphylococcus aureus ATCC25923 were used for antimi-
crobial activity assays. Bacteria were cultured, washed twice
with TBS and adjusted to final concentrations of 1 × 105–9
cells/ml with half-strength Muller-Hinton broth (MHB;
Becton Dickinson). Aliquots of culture were separately
incubated with 1.25–10 µM of rCRP2 in 96-well ELISA Misra.10 As a control, TBS was added to the bacterial
plates at 37°C, with shaking at 180 rpm. OD600nm readings suspensions and the mixture was similarly treated.
were taken at 0, 2, 3, 6, 12 and 24 h to monitor bacterial Bacterial agglutination was performed using P. aerug-
density. Additionally, cultures were serially diluted with inosa according to the method of Lanyi and Bergan.11
TBS, plated onto tryptone soy agar (TSA; Oxoid) and The culture was grown to mid-log phase and the concen-
incubated overnight at 37°C for colony enumeration. tration adjusted to 5 x 108 CFU/ml. Varying concentra-
The density of the bacteria following incubation with tions of rCRP2 were separately added to aliquots of
rCRP2 was also visualized by the method of Miles and bacteria. The results were observed on glass slides after
Antimicrobial activity of CRP 253
3–5 min. Agglutination of live cells is characterized by a Contrary to earlier reports,12,13 the bactericidal activity
coarse granular bacterial clumping. of rCRP2 appears to be calcium-independent (Fig. 2).
This observation is supported by our recent demonstra-
tion that Carcinoscorpius CRPs bind LPS in a calcium-
RESULTS AND DISCUSSION independent manner.6 This calcium independence is not
due to high pre-existing levels of Ca2+ in the medium
Purified rCRP2 (Table 1) was separately incubated with (MHB). The background level of Ca2+ in MHB has been
1 x 105 CFU/ml of P. aeruginosa, E. coli and S. aureus. investigated by atomic absorption spectrometry and
At 3 h, the E. coli and P. aeruginosa cultures showed no shown to be much lower than the naturally occurring
increase in OD600nm (Fig. 1A,B). In contrast, S. aureus levels in eukaryotic hosts.14 The calcium independence
cultures incubated with rCRP2 continued to exhibit of rCRP2 is in contrast to the activities of known amino-
increases in bacterial density (Fig. 1C). The results sug- glycosides; these become progressively less effective
gest that rCRP2 possesses antibacterial activity selec- with increasing concentrations of calcium.15,16 Previous
tively against Gram-negative bacterial species. studies have established that interaction between Ca2+

Fig. 4. rCRP2 interacts with plasma proteins. rCRP2 does not act in isolation but interacts with plasma proteins, forming a formidable antimicrobial
complex. (A) Members of this complex include hemocyanin (solid arrow), native CRPs (dashed arrow) and a mixture of other plasma lectins, including
CLs (dotted arrows). (B) Peaks of pmfs derived from 28-kDa and 70-kDa protein bands from different samples identified the proteins to be CRPs and
hemocyanin (flagged in white and black), respectively. That rCRP2 is able to recruit a complex from naïve plasma suggests that it is a pivotal pre-existing
PRR that mediates pathogen-recognition and recruitment of downstream immune molecules (continued on next page).
254 Tan, Ng, Ho, Ding

Bi

Fig. 4. rCRP2 interacts with plasma proteins. rCRP2 does not act in isolation but interacts with plasma proteins, forming a formidable antimicrobial
complex. (A) (see previous page) Members of this complex include hemocyanin (solid arrow), native CRPs (dashed arrow) and a mixture of other plasma
lectins, including CLs (dotted arrows). Peaks of pmfs derived from 28-kDa (Bi) and 75-kDa (Bii) protein bands from different samples identified the
proteins to be CRPs and hemocyanin (flagged in white and black), respectively. That rCRP2 is able to recruit a complex from naïve plasma suggests that it
is a pivotal pre-existing PRR that mediates pathogen-recognition and recruitment of downstream immune molecules.

ions and the bacterial membrane is responsible for the pathogen, and appears to circumvent the problem of cal-
bacteria becoming less susceptible, although the exact cium-mediated bacterial resistance.
mechanism of this resistance remains unknown.16 rCRP2 To determine the minimum antimicrobial concentration
is thus capable of potent antimicrobial activity, even as of rCRP2, bacterial suspensions were incubated with
calcium is recruited to strengthen the survival of the decreasing concentrations from 10 µM to 1.25 µM rCRP.
Antimicrobial activity of CRP 255

Bii

Inspection by light microscopy revealed that at con- bacteria, but can nevertheless act on other Gram-nega-
centrations of 2.5 µM or higher, rCRP2 is able to agglu- tive bacteria at concentrations of above 2.5 µM.
tinate Gram-negative bacteria (Fig. 3A). Previously, We thus continued to investigate the bactericidal activity
Tachypleus CRP2 (tCRP2) was reported to be unable to of rCRP2 at 2.5 µM by enumeration of viable colonies fol-
agglutinate or show antibacterial activity against a range lowing incubation. Results show that at 2.5 µM, rCRP2 is
of Gram-negative and Gram-positive bacteria with the capable of causing up to 109-fold reduction in bacterial den-
exception of E. coli K117. However, the highest protein sity rapidly within 1 h (Fig. 3B), which is 4-fold lower than
concentration tested was apparently only up to 1.64 that initially tested (Fig. 1). Visualization of bacterial
µM.17,18 Here, we show that rCRP2 can agglutinate P. colonies following their incubation with rCRP2 (Fig. 3C)
aeruginosa at a concentration of 2.5 µM. Taken together, confirms that rCRP2 possesses potent bactericidal activity.
these results suggest that horseshoe crab CRP2 exhibits While rCRP2 on its own is a potent PRR, it does not
higher affinity for E. coli K1 than other Gram-negative act in isolation. GST pulldown experiments show that
256 Tan, Ng, Ho, Ding
GST-CRP2 is able to recruit hemocyanin, native CRP 4. Wenzel RP, Pinsky MR, Ulevitch RJ, Young L. Current understanding
isotypes and a mixture of plasma lectins of 35 kDa and of sepsis. Clin Infect Dis 1995; 22: 407–413.
5. Wenzel R, Bone R, Fein A. Results of a second double-blind,
40 kDa (Fig. 4) that includes homologues of tachylectins randomized, controlled trial of antiendotoxin antibody E5 in Gram-
(data not shown). The ability of rCRP2 to form homo- negative sepsis. In: Programme and abstracts of the 31st
and hetero-oligomers with other CRP isotypes and other Interscientific Conference on Antimicrobial agents and Chemotherapy
plasma lectins serves to enhance the avidity of CRPs to (Chicago). Washington, DC: American Society for Microbiology,
the PAMP-bearing surfaces. Such ensemble of plasma 1991.
6. Ng PML, Jin Z, Tan SSH, Ho B, Ding JL. C-reactive protein: a
lectins may enhance frontline pathogen recognition and predominant LPS-binding protein responsive to Pseudomonas
binding, as well as bacterial agglutination as an early infection. J Endotoxin Res 2004; 10: 163–174.
pathogen-neutralization strategy. 7. Petsch D, Anspach FB. Endotoxin removal from protein solutions. J
Hemocyanin is known principally as an oxygen car- Biotechnol 2000; 76: 97–119.
rier, and the abundance of hemocyanin in the plasma has 8. Bradford MM. A rapid and sensitive method for the quantification of
microgram quantities of protein using the principle of protein-dye
long been reported.19 However, more recently, hemo- binding. Anal Biochem 1976; 72: 248–254.
cyanin has been demonstrated to exhibit phenol oxidase 9. Wilkins MR, Lindskog I, Gasteiger E et al. Detailed peptide
activity when induced by proteolytic enzymes and deter- characterisation using PEPTIDEMASS – a WorldWide Web
gents such as SDS.20–22 This is necessary for melaniza- accessible tool. Electrophoresis 1997; 18: 403–408.
tion of pathogens and is a key innate immune defence 10. Miles AA, Misra SS. The estimation of the bacteriocidal power of
blood. J Hygiene (Camb) 1938; 38: 732–749.
mechanism.23 The high level of circulating hemocyanin 11. Lanyi B, Bergan T. Serological characterization of Pseudomonas
suggests that tight control must be maintained to prevent aeruginosa. Methods Microbiol 1978; 10: 93–168.
automelanization of the blood during infection. The 12. Kaplan MH, Volanalis JE. Interaction of C-reactive protein complexes
recruitment and adhesion of hemocyanin to the Gram- with the complement system. I. Consumption of human complement
negative bacteria-bound CRP may provide such a regu- associated with the reaction of C-reactive protein with pnuemococcal
C-polysaccharide and with the choline phosphatides lecithin and
lation. This process ensures that the defence mechanism sphingomyelin. J Immunol 1974; 112: 2135–2147.
is activated only at the vicinity of pathogens. 13. Osmand AP, Mortensen RF, Siegel JE, Gewurz H. Interaction of C-
reactive protein with the complement system. III. Complement
dependent passive hemolysis initiated by CRP. J Exp Med 1975; 142:
CONCLUSIONS 1065–1075.
14. D’amato RF, Thornsberry C, Baker CN, Kirven LA. Effect of calcium
Overall, differential activities of members of the CRP2- and magnesium ions on the susceptibility of Pseudomonas species to
recruited complex contribute towards effective bacterial tetracycline, gentamycin, polymyxin B and carbenicillin. Antimicrob
Agents Chemother 1974; 7: 596–600.
clearance. In the presence of calcium, the formation of 15. Gilbert DN, Kutscher E, Ireland P, Barnett JA, Sanford JP. Effect of
this complex becomes enhanced following infection. the concentrations of magnesium and calcium on the in-vitro
The effect is specific, since addition of divalent magne- susceptibility of Pseudomonas aeruginosa to gentamycin. J Infect Dis
sium does not enhance this complex formation. Taken 1971; 124: S37–S45.
together with other evidence of CRP2 binding to LPS on 16. Zimelis VM, Jackson GG. Activity of aminoglycoside antibiotics
against Pseudomonas aeruginosa: specificity and site of calcium and
Gram-negative bacteria, this finding strongly suggests magnesium antagonism. J Infect Dis 1973; 127: 663–669.
that CRP2 is a crucial frontline defence molecule that 17. Iwaki D, Osaki T, Mizunoe Y, Sun NW, Iwanaga S, Kawabata S.
mediates pathogen binding as well as being the pivotal Functional and structural diversities of C-reactive proteins present in
plasma lectin that recruits downstream immune-related horseshoe crab hemolymph plasma. Eur J Biochem 1999; 264:
molecules to form a formidable protein-complex for 314–326.
18. Saito T, Kawabata S, Hirata M, Iwanaga S. A novel type of Limulus
antimicrobial activity and effective bacterial clearance. lectin – L6. J Biol Chem 1995; 270: 14493–14499.
Work is currently underway to elucidate the structure 19. Fahrenbach WH. The cyanoblast: hemocyanin formation in Limulus
of rCRP2. This will pave the way towards greater under- polyphemus. J Cell Biol 1970; 44: 445–453.
standing of the interactions of CRP2 and LPS and to 20. Decker H, Ryan M, Jaenicke E, Terwilliger N. SDS-induced phenol
determining the antimicrobial centre of the molecule. oxidase activity of hemocyanins from Limulus polyphemus,
Eurypelma californicum and Cancer magister. J Biol Chem 2001; 276:
17796–17799.
21. Nellaiappan K, Sugumaran M. On the presence of prophenol oxidase
REFERENCES in the hemolymph of the horseshoe crab, Limulus. Comp Biochem
Physiol B 1996; 113: 163–168.
1. Bone RC. The sepsis syndrome. Definition and general approach to 22. Kairies N, Beisel HG, Fuentes-Prior P et al. The 2.0-Å crystal
management. Clin Chest Med 1996; 17: 175–181. structure of tachylectin 5A provides evidence for the common origin
2. Rietschel ET, Brade H. Bacterial endotoxins. Sci Am 1992; August: of the innate immunity and the blood coagulation systems. Proc Natl
26–33. Acad Sci USA 2001; 98: 13519–13524.
3. Seydel U, Okiawa M, Fukase K, Kusumoto S, Brandenburg K. 23. Kawabata S, Nagai T. A link between blood coagulation and
Intrinsic conformation of lipid A is responsible for agonistic and prophenol oxidase activation in arthropod host defense. J Biol
antagonistic activity. Eur J Biochem 2000; 267: 3032–3039. Chem 2000; 275: 29264–29267.