Surviving Serum: the Escherichia coli iss Gene of Extraintestinal Pathogenic E. coli Is Required for the Synthesis of Group 4 Capsule

ABSTRACT

Extraintestinal pathogenic Escherichia coli (ExPEC) strains constitute a serious and emerging clinical problem, as they cause a variety of infections and are usually highly antibiotic resistant. Many ExPEC strains are capable of evading the bactericidal effects of serum and causing sepsis. One critical factor for the development of septicemia is the increased serum survival (iss) gene, which is highly correlated with complement resistance and lethality. Although it is very important, the function of the iss gene has not been elucidated so far. We have been studying the serum survival of a septicemic strain of E. coli serotype O78, which has a group 4 capsule. Here, we show that the iss gene is required for the synthesis of capsules, which protect the bacteria from the bactericidal effect of complement. Moreover, we show that the deletion of the iss gene results in significantly increased binding of the complement proteins that constitute the membrane attack complex to the bacterial surface.

INTRODUCTION

Septicemic strains of Escherichia coli constitute a clinical problem of increasing importance, as most of them are highly antibiotic resistant. These strains, termed ExPEC (extraintestinal pathogenic E. coli), are the major problem in health care-associated infections and are responsible for high morbidity and mortality in immunocompromised patients. ExPEC strains are involved in a large spectrum of infections, such as urinary tract infections and newborn meningitis, which often lead to sepsis with a low survival rate (1–7).

To cause sepsis, the infecting bacteria have to survive the bactericidal effects of the complement system. Therefore, serum resistance is critical for bacterial sepsis and bacteremia. Septicemic bacteria produce several factors, which confer serum resistance. These include capsules and specific lipopolysaccharide types (8–15). In septicemic E. coli, an additional gene increased serum survival (iss) (16–18) was identified and has been extensively studied. In chickens, the presence of this gene is associated with embryo lethality and complement resistance (19–21). Moreover, the Iss proteins have also been extensively studied as a potential target for protective immunization (22).

The iss gene is homologous to the bor gene of E. coli K-12, which originated from bacteriophage lambda (23) and encodes an outer membrane lipoprotein. In septicemic strains, the iss gene is present in one or more copies, one of which is usually located on a plasmid. Transcriptome analyses indicated that the iss gene is expressed constitutively (24). It was suggested that the Iss protein may have a role in formation of functional complexes in the junctions between the inner and outer membranes (25). However, so far, the molecular mechanism by which iss provides serum resistance is not understood.

We have been studying the iss gene in a septicemic E. coli strain, serotype O78 (E. coli O78-9) (26–32), which is highly serum resistant (30, 33). We previously showed that these bacteria have O-antigen (group 4) capsules (34, 35). The group 4 capsule contains a polysaccharide that is usually identical to that of the O-antigen and is therefore called O-antigen capsule. Seven genes are involved in the biosynthesis of group 4 capsule form an operon, as described in Fig. 1 (35). In E. coli O78-9, the group 4 capsule is required for serum resistance (36). Here, we show that deletion of the iss gene results in reduced production of the capsule and changes the pattern of complement proteins that bind to the bacteria.

FIG 1.

The operon coding for the synthesis of group 4 capsule. The 7-gene operon and adjacent genes are shown, as described by Peleg et al. (35).

RESULTS

The iss gene confers serum resistance.

E. coli O78-9 is highly serum resistant, and its growth is not inhibited even by 75% serum (Fig. 2). E. coli O78-9 also carries a ColV plasmid, which encodes an iss gene (25, 30, 37–39). Previous experiments indicated that serum resistance is associated with cloned fragments of the plasmid ColV, I-K94, which contains a gene named iss (16). Indeed, when E. coli K-12, which is serum sensitive—inhibited by 3% serum—was transformed with a plasmid which carries a cloned iss gene from E. coli O78-9, its serum resistance was significantly improved. The results presented in Fig. 3 indicate that bacteria carrying the recombinant plasmid, pACYC184/iss, but not the native pACYC184, can grow in the presence of 15% serum.

FIG 2.

Response of E. coli O78-9 to serum. Cultures of E. coli O78-9 were grown as described in Materials and Methods. Human serum (Sigma) was added at OD₆₀₀ of 0.04, and turbidity was measured.

FIG 3.

Effect of cloned iss DNA fragment on serum survival of E. coli K-12. Cultures of E. coli K-12 MG1655 pACYC184 (empty circles) and its transformant with a plasmid containing the iss gene from E. coli O78-9 (full circles) were grown and treated with serum as described in Fig. 2. (A) No serum. (B) Fifteen percent serum.

The iss gene is essential for serum resistance.

The results presented in Fig. 4 indicate that deletion of the iss gene of E. coli O78-9 results in serum sensitivity. The phenotype of serum sensitivity could be complemented by introducing the iss gene on a plasmid (Fig. 4).

FIG 4.

Effect of a deletion of the iss gene. Effect of serum on growth was determined as described in Fig. 2. The bacteria used were E. coli O78-9 and its mutant deleted for the iss gene. Both strains were transformed with a pBR322 plasmid or a pBR322 carrying the iss gene. (A) Growth without serum. (B) Growth with 50% human serum.

The iss gene is required for synthesis of the capsule.

In order to understand the effect of iss deletion, we performed proteomic studies comparing the mutant with the wild-type strain. The results presented in Fig. 5A show the effect of the iss deletion on protein abundance. The results indicate that there was a significant decrease in the abundance of tyrosine-protein kinase (Etk), which, together with Etp, is required for the synthesis of the group 4 capsule (35). The proteomic results were validated by showing that antibodies against Etk could not detect this protein in iss deletion mutants (Fig. 5B).

FIG 5.

Effect of deletion of the iss gene on Etk levels. (A) Volcano plot displaying protein abundance differences between wild-type O78-9 and its iss deletion mutant. Proteins were considered differentially expressed proteins if the fold change is greater than 2 and the P value is smaller than 0.01. These borders are indicated as well. Location of the Etk protein is marked by the arrow. (B) Detection of the Etk protein by anti-Etk antibodies was carried out as previously described (35). The position of the Etk protein is indicated by an arrow.

The reduction in Etk level should lead to inhibition of capsule formation and, consequently, should result in increased sensitivity to polymyxin (40). Indeed, the iss deletion mutants show increased sensitivity to polymyxin in a way similar to the sensitivity conferred by mutations in capsule synthesis due to deletion of the etp gene (Fig. 6).

FIG 6.

Effect of deletion of the iss gene and the etp gene on growth with polymyxin B. Overnight cultures of E. coli O78-9 and its iss or etp deletion mutants grown in LB medium were diluted 1:20 into fresh medium containing 4 μg/ml polymyxin B. Growth was determined by turbidity at OD₆₀₀.

The inhibition in capsule synthesis in iss deletion mutants was further demonstrated by analyses of the O-antigen. In bacteria that produce group 4 capsule, also called O-antigen capsules, the newly synthesized O-antigen is distributed between the lipopolysaccharide and the capsule under physiological conditions. However, when capsule synthesis is inhibited, the O-antigen level goes up (35). Such an increase in O-antigen was observed in the iss deletion mutants (Fig. 7). The results presented in Fig. 8 compare the levels of O-antigen in wild-type bacteria, bacteria deleted for iss, and bacteria deleted for etp. No O-antigen was visible in mutants deleted for cap (cps), a mutant in the expression of the O-antigen at the cell surface.

FIG 7.

Effect of an iss deletion on production of LPS. LPS from E. coli O78-9 and its iss deletion mutant was purified, separated by SDS-PAGE, and stained with Pro-Q Emerald 300. Three biological samples (I, II, and II) of wild type and of the iss deletion mutant (Δiss) are shown.

FIG 8.

Production of LPS in E. coli O78-9-carrying mutations in genes coding for capsule formation. LPS was purified from wild-type E. coli O78-9 and its deletion mutants for iss, etp, and cap (csp) genes.

So far, our results indicated that a deletion in the iss gene results in a reduced level of the Etk protein in reduced synthesis of capsule and in serum sensitivity. To validate our hypothesis that the serum sensitivity in iss deletion mutants is due to the reduction in Etk expression, we cloned the etp-etk genes on a plasmid under the araC promoter (41) and transformed into the iss deletion mutant. The results presented in Fig. 9 indicate that the overexpression of etp-etk abolished the serum sensitivity conferred by the iss deletion.

FIG 9.

Complementation of serum sensitivity. Effect of serum on growth was determined as described in Fig. 2. The bacteria used were E. coli O78-9 and its mutant deleted for the iss gene. Both strains were transformed with a pBAD24 plasmid or a pBAD24 carrying the etp-etk genes and exposed to 50% serum.

Binding of serum components by iss mutants.

Analysis of the proteins bound to the bacteria after exposure to serum was carried out in the wild-type cells and in iss and etp deletion mutants. The results indicate that several proteins of the complement system were bound significantly more to the mutants, and some were bound only to the mutants (Fig. 10 and Table S1 in the supplemental material). These proteins constitute the components of the complement, which are essential for the formation of the membrane attack complex (8, 42–44) (Fig. 10). The finding that proteins of the membrane attack complex are bound to the iss deletion mutants much more than to the wild-type bacteria may explain the high serum sensitivity of the mutants.

FIG 10.

Heat map showing binding of serum proteins. Relative abundance of selected human proteins bound to the bacterial surface of E. coli O78-9 and its iss and etp deletion mutants.

DISCUSSION

The iss gene contributes to serum survival in many virulent septicemic E. coli strains, but its function is not fully understood. In virulent strains involved in extraintestinal infections (ExPEC), iss is carried on the ColV plasmid, but many strains have also several chromosomal copies. In E. coli K-12, there is only the chromosomal copy of bacteriophage origin, called bor. The proteins products of bor and iss are very similar, with only few amino acid differences (17). We have been working with a septicemic E. coli O78-9 strain, which has a group 4 capsule essential for serum resistance (36). In this strain, deletion of the plasmid iss gene renders the bacteria completely serum sensitive (Fig. 4), indicating that the chromosomal allele is not sufficient to complement the deletion of the plasmid gene. Moreover, we did not succeed to complement the iss deletion with the cloned bor gene (D. Biran, S. Navok, and E. Z. Ron, unpublished data). This result is unexpected, as the chromosomal and plasmid genes are very similar. However, although the proteins coded by the iss gene are similar, there are extensive differences in the upstream and downstream regions of the genes, which probably bring about regulatory changes.

Here, we show that an iss deletion results in a significant reduction of Etk. Etk is a protein-tyrosine kinase, a member of the bacterial PTK family, which, together with Etp (protein-tyrosine-phosphatase) carries out the synthesis of group 4 capsules (34). We further show that the reduced level of Etk results in depletion of the capsule and in serum sensitivity. Moreover, we could complement the serum sensitivity of iss deletion mutants by overexpression of the etp-etk genes. Taken together, these findings indicate that the serum sensitivity of iss mutants is due to the decreased level of the enzymes involved in capsule synthesis.

Etk is located in the inner membrane and has two predicted transmembrane domains (45). Iss is an outer membrane protein (46), and its depletion leads to changes in the bacterial envelope (D. Biran, S. Navok, and E. Z. Ron, unpublished data). These changes affect other membrane proteins and are probably responsible for the reduction in the level of Etk.

Although we established the effect of Iss on production of group 4 capsule and on serum resistance in E. coli O78-9, there are still several open questions regarding the physiological function of Iss. Thus, Iss is highly conserved in ExPEC strains. Yet many such strains do not produce a group 4 capsule. We could not detect a reduction in capsule formation consequently to introducing an iss deletion in strains which produce K1 (sialic acid) capsule (D. Biran, S. Navok, and E. Z. Ron, unpublished data). These results suggest that Iss has additional functions which contribute to the fact that it is highly conserved. It is also not clear how the presence of an iss gene from an ExPEC strain increases serum resistance in E. coli K-12, as shown in Fig. 3 and previously by Binn et al. (16). K-12 strains have the operon for producing group 4 capsule, which is probably not expressed due to the presence of an insertion element 1 (IS1) sequence in the promoter region. However, there is evidence showing that the etk gene is expressed in K-12 (40), a finding which may explain the interaction with Iss. We assume that Iss, being a membrane protein, influences not only capsule production but also changes the structure of the membrane. These changes could contribute to increased serum survival in addition to the presence of Etk.

The effect of Iss on production of the group 4 capsule in E. coli O78-9 is dramatic, yet this protein cannot be considered a general regulator of this pathway, as many E. coli strains, usually intestinal pathogens, do not carry a ColV plasmid or an iss gene. As the other regulators of the pathway are probably present in E. coli O78-9, it should be possible to investigate the interaction between Iss and the other regulators and understand the complicated systems which control capsule formation.

Upon exposure to serum, the iss deletion results in a significantly increased level of complement proteins bound to the bacteria (Fig. 10). This binding is probably facilitated by the absence of a protective capsule. Interestingly, the proteins bound to the membrane of iss deletion mutants constitute the complement membrane attack complex (Fig. 11) (42–44, 47). Therefore, it is probably this excessive binding of complement proteins which renders the iss deletion mutants sensitive to serum. Thus, the findings presented here can explain the protective effect of group 4 capsules in the presence of serum and the role of the iss gene in serum survival.

FIG 11.

Identified proteins of the membrane attack complex (MAC). This is a schematic of the formation and activity of the membrane attack complex. Identified proteins are marked in green.

MATERIALS AND METHODS

Bacteria, plasmids, and growth conditions.

The strains and plasmids used are described in Table 1. Unless stated otherwise, all E. coli strains were grown with shaking at 37°C in Davis and Mingioli minimal medium, with 0.2% glucose (48). When required, antibiotics were added (300 μg/ml ampicillin, 50 μg/ml kanamycin). For proteomic analyses, all strains were grown in Davis minimal medium with 0.2% glucose at 37°C in shaking flasks.

TABLE 1.

Strains and plasmids used in this work

Plasmid or strain	Description	Reference no. or source
Plasmids
pKD4	Template for kanamycin resistance cassette	49
pACYC184	Tetr plasmid	55
pACYCiss	Tetr plasmid containing the iss gene	This study
pBR322	Ampr plasmid	56
pBRiss	Ampr plasmid containing iss gene	This study
pBAD24	Ampr plasmid with araC promoter	41
pBAD24etp-etk	pBAD24 with cloned etp-etk genes
E. coli strains
MG1655	Wild-type K-12 strain
MG1655 pACYCiss		This study
O78-9	Wild-type ExPEC O78 strain	33
O78-9 Δetp		36
O78-9 Δcps		This study
O78-9 pBR	O78-9-carrying pBR322 plasmid	This study
O78-9 pBRiss		This study
O78-9 pBAD
O78-9 pBADetp-etk
O78-9 Δiss		30
O78-9 Δiss pBRiss		This study
O78-9 Δiss pBAD		This study
O78-9 Δiss pBADetp-etk		This study

Growth curves and serum survival assays.

All the results were obtained from three biological samples. Turbidity results were confirmed by viable counts. Exponentially growing cultures were diluted in the same medium to an optical density at 600 nm (OD₆₀₀) of 0.04, as determined with a BioTek Eon plate reader, and incubation was continued with shaking at 37°C. Turbidity at OD₆₀₀ was measured every 20 min. For determining the effect of serum, human serum (Sigma) was added at OD₆₀₀ of 0.04, and turbidity was measured for several hours.

Construction of deletion mutants and plasmids.

All site-specific gene knockouts were performed by the method of Datsenko et al. (49). Briefly, cultures of competent wild-type E. coli O78:H19 ST88 isolate 789 (O78-9) were transformed with pKD46 plasmid. The transformants were grown in LB medium containing ampicillin, induced with arabinose, made competent for electroporation, and stored at −70°C until used. Linear PCR products were made, according to the region to be deleted, on the template of the kanamycin resistance cassette flanked by the FLP recognition target (FRT) sequences from the pKD4 plasmid. The primers (36 nucleotides each) were designed from the flanking region of the sequence to be deleted in strain O78-9 (Table 2). Kanamycin-resistant recombinants were screened by means of colony PCR. The pKD46 plasmid was cured by growth on LB medium at 42°C.

TABLE 2.

Primers used for specific gene knockouts

Primer	Sequence (5′–3′)
pACYCissF	ACAGGAGGGACAGCTGATAGAAACAGAAGCAATAGTGGCTTTTGTGGC
pACYCissR	AGAACATATCCATCGCGTCCGCCATCTCCAGCAAAAATATTGGGGATG
pBRissF	TATGGATCCAATAGTGGCTTTTGTGGC
pBRissR	TATAAGCTTGCAAAAATATTGGGGATG
issP1	ATGCAGGATAATAAGATGAAAAAAATGTTATTTTCTGCCGGTGTAGGCTGGAGCTGCTTC
issP2	AGCGGAGTATAGATGCCAAAAGTGATAAAACCGAGCAATCCATATGAATATCCTCCTTAG
cpsP1	GCAAATAAAATTAAAAAACTACAAGAAATCAACAATGTGTAGGCTGGAGCTGCTTC
cpsP2	AGCTTGATCGGATATGACGGCTGCGAAAAATTGGAACATATGAATATCCTCCTTAG
etpP1	ATGGCCCAACTAAAATTTAACTCAATCCTGGTGGTTGTGTAGGCTGGAGCTGCTTC
etpP2	TACCGTAGACATGTTCAAATGCGTCCTGACTTTTACCATATGAATATCCTCCTTAG
pBADetpF	GTTTTTTTGGGCTAGCAGGAGGAATTCACCATGGCCCAACTAAAATTTAA
pBADetkR	CTTCTCTCATCCGCCAAAACAGCCAAGCTTTTTACCTTCAGTTTTACTCT

Recombinant plasmids were constructed using one of the following methods: (i) cloning PCR-amplified DNA fragments (primers listed in Table 2) into a pACYC vector or pBAD24 vector using the HiFi DNA assembly cloning kit (NEBuilder), or (ii) digesting PCR-amplified DNA fragments (primers listed in Table 2) with the appropriate restriction enzyme, ligated into a chosen vector, and digested with the same restriction enzymes, using Fast-Link DNA ligase (Epicentre).

Sample preparation for liquid chromatography-tandem mass spectrometry.

To access serum proteins that bind to the cellular surface of E. coli O78-9, the wild-type and the iss deletion mutants were grown in Davis minimal medium and supplemented with 0.2% glucose at 37°C. From cultures grown to an OD₆₀₀ of 0.4, 25 ml of the culture were transferred to 25 ml of prewarmed Davis minimal medium and supplemented with 0.2% glucose and 20% normal human serum. After an incubation of 30 min at 37°C with shaking, the bacteria were harvested by centrifugation at 10,000 × g for 15 min at 4°C. Pellets were washed twice with 5 ml ice-cold Dulbecco's phosphate-buffered saline (DPBS). For removal of unbound serum protein, the pellets were resuspended in 25 ml ice-cold DPBS and placed onto 15 ml 45% (wt/vol) sucrose followed by centrifugation for 40 min at 10,000 × g and 4°C. After repeating the previous step, samples were washed five times with 5 ml ice-cold DPBS for sufficient sucrose removal. The pellets were resuspended in 400 μl 50 mM triethylammonium bicarbonate (TEAB), and the cells were disrupted by sonication (3× 1 min, 20% amplitude, 0.1-s pulse, 0.5-s cycle; Bandelin SonoPuls 3100, type MS 72). Lysates were cleared by centrifugation for 15 min at 20,000 × g and 4°C. Protein concentration was determined using the bicinchoninic acid (BCA) assay (Thermo Fisher Scientific) according to the manufacturer’s protocol. One hundred micrograms of protein were subsequently reduced, alkylated, and digested with trypsin (5 h, 37°C). Resulting peptides were desalted by C₁₈ ZipTip (Merck Millipore) purification. Peptides were dried in a vacuum centrifuge and reconstituted in 10 μl solvent A (0.1% acetic acid in water).

Liquid chromatography-tandem mass spectrometry.

Peptide composition was analyzed by liquid chromatography with an Easy-nLC II coupled to an Orbitrap Velos Pro mass spectrometer. Peptides were loaded on in-house packed reversed-phase columns. Afterward, peptides were eluted by a nonlinear binary gradient of 156 min from 1% to 99% solvent B (0.1% acetic acid in acetonitrile) in solvent A. Eluting peptides were recorded by a full scan in the Orbitrap from 300 to 2,000 m/z at a resolution of 30,000. The 20 most abundant precursor ions were isolated and fragmented by collision-induced dissociation at a normalized collision energy of 35V. Ions with an undetermined charge state, as well as a charge of 1, were excluded from fragmentation. Dynamic exclusion was set to 30 s, and lock mass correction was enabled.

Database search and label-free quantification.

The database search was performed by the Andromeda search engine, which is implemented in MaxQuant (version 1.5.5.1; Max Planck Institute of Biochemistry) (50–52) using databases for E. coli O78-9 and the human reference proteome, downloaded from GenBank (12 February 2016) and UniProt (15 July 2016), respectively. Precursor mass deviation was set to 4.5 ppm, and fragment mass tolerance was set to 0.5 Da. A false-discovery rate (FDR) of 0.01 on protein, peptide, and spectra levels was applied. Oxidation on methionine was set as a variable modification, and carbamidomethylation on cysteine was set as a fixed modification. The MaxLFQ (52) quantification was enabled, as well as “match between runs.” For data analyses, Perseus software (version 1.5.5.3) (53) was used. Listed proteins were filtered for contaminants and reverse and only identified by site hits. Statistical testing was done by a two-class unpaired t test for proteins which have at least two valid values per strain. Proteins were considered differentially expressed proteins if the fold change is greater than 2 and the P value is smaller than 0.01.

Quantified human proteins were analyzed separately. Briefly, proteins were filtered for at least 3 valid values in at least one group. Statistical testing was done if possible by a two-class unpaired t test. Heat maps, as well as volcano plots, were created in Perseus and Inkscape (v0.92.2).

Extraction and analysis of LPS.

Bacterial lipopolysaccharide (LPS) was obtained by hot aqueous-phenol extraction and separated by SDS-PAGE, as described elsewhere (54). Separated LPS was stained with Pro-Q Emerald 300 (Thermo Fisher Scientific) according to the manufacturer’s protocol.