1
Distinct SagA from hospital-associated clade A1 Enterococcus faecium strains 1
contributes to biofilm formation. 2
3
F. L. Paganelli1, M. de Been1, J.C. Braat1, T. Hoogenboezem2,3, C. Vink2,3, J. Bayjanov1, 4
M.R.C. Rogers1, J. Huebner4, M. J. M. Bonten1, R. J. L. Willems1, H. L. Leavis1#. 5
6
1. Department of Medical Microbiology, University Medical Center Utrecht, Utrecht, 7
The Netherlands. 8
2. Laboratory of Pediatrics, Pediatric Infectious Diseases and Immunity, Erasmus MC-9
Sophia Children's Hospital, Rotterdam, 10
3. The Netherlands Erasmus University College, Department Life Sciences, Erasmus 11
University Rotterdam, The Netherlands. 12
4. Division of Pediatric Infectious Diseases, Hauner Children´s Hospital Ludwigs-13
Maximilian Universität München, München, Germany. 14
15
Running title: SagA in Enterococcus faecium biofilms 16
17
18
19
20
21
#Correspondence and requests for materials should be addressed to Helen L. Leavis 22
email: [email protected] 23
24
25
AEM Accepted Manuscript Posted Online 24 July 2015Appl. Environ. Microbiol. doi:10.1128/AEM.01716-15Copyright © 2015, American Society for Microbiology. All Rights Reserved.
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
2
Abstract 26
Enterococcus faecium is an important nosocomial pathogen causing biofilm-mediated 27
infections. Elucidating E. faecium biofilm pathogenesis is pivotal for development of new 28
strategies to treat these infections. In several bacteria extracellular DNA (eDNA) and proteins 29
act as matrix components contributing to biofilm development. In this study, we investigated 30
biofilm formation capacity and the role of eDNA and secreted proteins in 83 E. faecium 31
strains with different phylogenetic origin that clustered in clade A1 and clade B. Although 32
there was no significant difference in biofilm formation between E. faecium strains from both 33
clades, addition of DNase I or proteinase K to biofilms demonstrated that eDNA is essential 34
for biofilm formation in most E. faecium strains, whereas proteolysis primarily impacted 35
biofilms of clade A1 E. faecium strains. Secreted antigen A (SagA) was the most abundant 36
protein in biofilms from clade A1 and B E. faecium strains, although localization differed 37
between the two groups. sagA was present in all sequenced E. faecium strains, with a 38
consistent difference in the repeat region between the clades, which correlated with proteinase 39
K susceptibility in biofilms. This indicates an association between the SagA variable repeat 40
profile and the localization and contribution of SagA in E. faecium biofilms. 41
42
43
44
45
46
47
48
49
50
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
3
Introduction 51
Enterococci, specifically Enterococcus faecium and Enterococcus faecalis, are the third most 52
common cause of nosocomial infections (1). The increase in the number of hospital-53
associated infections caused by E. faecium in the last decades is thought to be, at least partly, 54
driven by the cumulative acquisition of novel adaptive traits, such as antibiotic resistant 55
determinants and virulence factors, particularly in a defined subpopulation of E. faecium, 56
enriched in hospital isolates (2, 3). Different molecular technologies and increasingly 57
sophisticated phylogenetic models have been used to study various aspects of the dynamics of 58
E. faecium evolution. This resulted in that the clinical E. faecium subpopulation was initially 59
designated as lineage C1 (4) and later renamed to clonal complex 17 (CC17) (2, 5). Using 60
Bayesian-based population genetic modeling of MLST data of a large set of isolates 61
demonstrated that nosocomial E. faecium strains clustered into distinct subgroups, suggesting 62
different evolutionary trajectories for clinical isolates (6, 7). Based on whole genome 63
sequencing, subsequently Lebreton et al. (8), described three clades: (i) clade A1, including 64
the majority of clinical E. faecium isolates, which were previously classified as lineage C1 65
and CC17, (ii) clade A2, including the majority of animal-derived isolates; and (iii) clade B, 66
containing human commensal isolates (8). We will use an adaptation of the latter 67
nomenclature for our sequenced based phylogenetic analyses in the present study. 68
Many E. faecium infections in hospitalized patients are biofilm-mediated and associated with 69
the use of indwelling medical devices, such as (central) venous and urinary catheters, 70
orthopedic implants, and prosthetic cardiac valves (9). A critical step in the pathogenesis of 71
these infections is the adherence of enterococci to implanted medical devices and the 72
colonization of these surfaces by the formation of biofilm (10). 73
The formation of multilayer biofilms involves a complex process, from attachment of single 74
cells to development of a three dimensional bacterial community (11). Under optimal 75
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
4
conditions, mature biofilm can develop into multilayered microcolonies held together with a 76
matrix and interspersed with water channels, through which nutrients reach deeper parts of the 77
biofilm. The matrix or extracellular polymeric substances (EPS) is an important biofilm 78
component stabilizing the biofilm and protecting it against antimicrobials and immune cells. It 79
is mainly composed of polysaccharides, proteins, and extracellular DNA (eDNA) (12). 80
Autolysis is the common mechanism by which eDNA is released in Gram positive bacteria 81
(13-19). DNA can act as an adhesive, implicated in biofilm attachment and stability. In E. 82
faecium clinical strain E1162, eDNA is mainly generated by lysis of a bacterial subpopulation 83
by the major autolysin AtlAEfm (19). 84
Structural proteins present in the matrix have been characterized in several bacterial species 85
(20). These biofilm matrix-associated proteins include extracellular carbohydrate-binding 86
proteins, such as LecA, LecB and CdrA in P. aeruginosa (21-24) or TasA in B. subtilis, and 87
high-molecular mass proteins like the biofilm-associated surface protein Bap present in S. 88
aureus and Bap-like proteins in others species (25). So far, biofilm-associated matrix proteins 89
have not been studied in E. faecium. 90
In the present study, we analyzed the role of eDNA and extracellular proteins in biofilm 91
formation of E. faecium hospital-associated and community strains, that clustered in clade A1 92
and clade B, respectively. We demonstrated that eDNA is an essential structural component of 93
extracellular matrix of all E. faecium strains, irrespective of their origin or phylogenetic 94
background. Moreover, we identified that the Secreted antigen A (SagA) is the most abundant 95
protein in the supernatant of biofilm-forming cells that it is part of the E. faecium biofilm 96
matrix. However, SagA present in the biofilm matrix of clade A1 strains and contains a 97
distinct repeat motif that correlates with proteinase K susceptibility. 98
99
100
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
5
Materials and Methods 101
Bacterial strains and growth conditions 102
The 85 E. faecium strains used in this study are listed in Table S1. Unless otherwise 103
mentioned, E. faecium was grown in brain heart infusion broth (BHI; Oxoid) at 37°C. For 104
biofilm assays, tryptic soy broth media (TSB; Oxoid) with 1% glucose (TSBg) was used. 105
Growth was determined by measuring the optical density at 660 nm (OD660). 106
107
Genome sequencing 108
For genome sequencing, 21 E. faecium strains were grown on 4 ml BHI for 24 h at 37°C and 109
genomic DNA was extracted using Wizard® Genomic DNA Purification Kit (Promega). 110
Samples were prepared and sequenced employing Nextera XT DNA Sample Prep Kit and 111
MiSeq Reagent Kit V2 with 2 x 250 bp reads (Illumina Inc.). Reads were first quality filtered 112
using Nesoni Clip version 0.109. Reads with an average quality score of less than 10 were 113
removed, as well as reads with a length shorter than 150 bp. The remaining reads were de 114
novo assembled into contigs using SPAdes assembler version 3.0.0 with default parameters 115
(26). Subsequently, contigs with a length of less than 500 bp and minimum nucleotide 116
coverage of 6 were discarded. 117
Functional gene annotation was performed using Prokka (27). All protein sequences were 118
aligned using BLAST and based on the alignment results, orthologous groups (OG) of 119
proteins were identified using orthAgogue (28) and MCL (29). orthAgogue and MCL were 120
run using "-u -o 50" and "-I 1.5" parameters, respectively. Based on the orthology relationship 121
of 1186 core OGs, including the 64 previous sequenced E. faecium strains, multiple sequence 122
alignment of nucleotide sequences of genes in an OG was performed using MUSCLE (30). 123
Gaps in multiple sequence alignment were removed using trimal (31), which results in an 124
equal sequence length for every gene in an OG. Subsequently, core genomes of the 85 strains 125
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
6
were used for building a phylogenetic tree in a FastTree 2 with 1000 bootstrap samples and 126
GTR model (generalized time-reversible) as a maximum-likelihood model (32). The genome 127
sequences have been submitted to the European Nucleotide Archive (ENA) 128
(http://www.ebi.ac.uk/ena/data/view/PRJEB9441). 129
The number of SNPs between isolates was extracted from the core genome alignment 130
described above. For each pairwise strain comparison, we only considered SNPs at positions 131
in the alignment where both nucleotides were A, T, C, or G (i.e. ambiguous positions were 132
ignored). The number of SNPs between all pairwise strain comparisons is displayed in Table 133
S2. 134
135
Biofilm polystyrene assay 136
The biofilm polystyrene assay was performed as described previously with some 137
modifications (19). In brief, overnight bacterial suspensions were diluted to OD660 0.01 in 138
TSBg and incubated for 24 h. When mentioned, 1.5 µg µl-1 of DNase I (Roche) or 1.0 µg µl-1 139
of proteinase K (Sigma) was added to the bacterial suspension before starting the biofilm. The 140
plates were washed and stained as described before (19). The experiments were performed in 141
triplicate and statistical analysis of the data was performed using a two-tailed Student’s t-test. 142
143
Proteomics of biofilm supernatant 144
To determine the proteins present in the supernatant of the E1162 clade A1 strain growing in 145
biofilm, 6 ml of the supernatant of E1162 cells, after 24 h growth in biofilm using the semi-146
static model as described below, was filtered using a 0.2 µm filter (Corning) and concentrated 147
with a 10 k column (Amicon Ultra – Merck Millipore). Proteins were loaded on a 12.5% 148
SDS-polyacrylamide gel electrophoresis (SDS-PAGE), electrophoresed, and stained with 149
Coomassie brilliant blue. The 3 predominant proteins with sizes of approximately 75-kDa, 50-150
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
7
kDa and 25-kDa, were quantified with Image J (33) and excised from the gel. Mass 151
spectrometry of the excised proteins was performed by matrix-assisted laser 152
desorption/ionisation-time of flight (MALDI-TOF) mass spectrometry (MS), using an 153
Ultraflex MALDI-TOF/TOF mass spectrometer (Bruker Daltonics), as described previously 154
(34). 155
156
Biofilm semi-static model and confocal laser scanning microscopy (CLSM) 157
To visualize E. faecium biofilms and detect SagA, E. faecium biofilms were grown in a semi-158
static model as previously described by Paganelli et al. (19) with some modifications. In brief, 159
biofilms were grown in TSBg on a cover slip coated with poly-L-lysine (0.45 µm; diameter, 160
12 mm; Becton Dickinson) inside a well of a six-well polystyrene plate (Corning Inc.) at 37ºC 161
for 24 h at 120 rpm. To study the effect of proteolysis, 1.0 µg µl-1 of proteinase K was added 162
at the start of biofilm formation (t=0). After 24 h, biofilms were washed with 3 ml PBS and 163
fixed with 3% para-formaldehyde for 15 min at room temperature. After removal of the 164
fixative, cells were washed with PBS. To detect SagA, bacteria were incubated with 165
polyclonal SagA antibodies, α-SagA, (1:250 in PBS with 1%BSA) (35) for 1 h on ice and 166
subsequently washed with PBS. Polyclonal α-SagA antiserum was raised in rabbit using 167
purified SagA from E. faecium strain E155 (35). Secondary antibody, i.e. Alexa Fluor 488 168
goat α-rabbit IgG (life technology) (1:500 in PBS with 1%BSA) was added, and incubated for 169
an additional 1 h on ice. Cells were washed once more, and incubated with the FM 5-95 170
(Invitrogen) dye (5 µg ml-1) for 1 min on ice to visualize cell membranes. The FM 5-95 dye 171
was removed and the cover slips were transferred to microscope slides. Fluorescence was 172
analyzed using a confocal laser microscope (Leica SP5) equipped with an oil plan-neofluor 173
63x/1.4 objective. Alexa 488 and FM 5-95 were excited at 488 nm. Pictures were analyzed 174
with LAS AF software (Leica) and the level of biofilm formation was quantified using 175
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
8
Comstat (36)/Matlab R2010b software (The MathWorks). The average thickness and biomass 176
of the biofilms was measured at five randomly chosen positions. Statistical analysis of the 177
data was performed using a two-tailed Student’s t-test. 3D images were generated with Image 178
J (33) and Orthoview images by LAS AF software (Leica). 179
180
Proteolytic stability of SagA 181
To determine SagA degradation during biofilm development, biofilm supernatants of three 182
clade A1 strains (E470, E1162, E1904) and three clade B strains (E980, E3548, E1590) were 183
analyzed. The supernatants originated from biofilms grown in a semi-static model for 4 h 184
(early biofilm) and 24 h (mature biofilm), were filtered and precipitated at -20ºC overnight 185
with 1/10 (v/v) 100% Trichloroacetic acid (TCA). After incubation, supernatants were spun 186
down for 10 min at 4000 g at 4ºC. Protein pellets were washed with 500 µL 100% acetone 187
and again centrifuged in the same condition. Final protein pellets were resuspended in sample 188
buffer (100 mM Tris-HCl, 5% dithiothreitol, 2% SDS, 0.004% bromophenol blue, and 20% 189
glycerol) and analyzed by SDS-PAGE and Western blot as described below. 190
For testing SagA stability in biofilm, supernatants of 24 h biofilms of E1162 and E980, 191
processed as described above, were incubated at 37ºC for 30 min, 1 h, 4 h, 8 h and 24 h. 192
Degradation of SagA was analyzed over time by SDS-PAGE gel and Western blot as 193
described above. 194
To analyze the susceptibility of SagA to proteinase K, filtered supernatants of E1162 and 195
E980 24 h biofilm were incubated with 0.1 µg µl-1 proteinase K at 37ºC for 1, 5, 15, 30 and 60 196
min. After proteinase K challenge, supernatants were precipitated at 4ºC for 10 min with 1/5 197
(v/v) 100% TCA. After incubation, supernatants were spun down for 10 min at 4000 g at 4ºC. 198
Protein pellets were washed with 500 µL 100% acetone and again centrifuged in the same 199
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
9
condition. Final protein pellets were resuspended in sample buffer and analyzed by SDS-200
PAGE and Western blot as described below. 201
202
SDS-PAGE and Western blot 203
Protein samples were equally mixed with sample buffer and boiled for 5 min. Western 204
blotting was carried out as described previously (37). Membranes were blocked with 4% skim 205
milk (Campina Holland, Alkmaar, The Netherlands) in PBS-0.1% Tween 20 for 24 h at 4°C. 206
Incubation with primary antibody (α-SagA) (35) was carried out for 1 h in 1% BSA in PBS-207
1% Tween 20 at 37°C, followed by two washes 10 min (each) in PBS-0.1% Tween 20 at 208
37°C. Subsequently, membranes were incubated for 1 h with anti-rabbit IgG (H+L) 209
horseradish peroxidase (HRP) (Bio-Rad Laboratories, Veenendaal, The Netherlands) in 1% 210
BSA in PBS-1% Tween 20 at 37°C. Membranes were washed twice with PBS-0,1% Tween 211
20, and proteins were visualized using the ECL Plus Western blotting detection system (GE 212
Healthcare, Diegem, Belgium) and ImageQuant LAS 4000 biomolecular imager GE 213
Healthcare. 214
215
sagA sequences 216
The sagA complete gene sequence was determined for 33 E. faecium strains by Sanger 217
sequencing. To this end, the sagA gene was amplified using AccuPrime™ Taq DNA 218
Polymerase High Fidelity (Life technology) and primers sagA-F (5′ATGACCTT-219
GACTGCCGTAGCAT-3′) and sagA-R (5′TTACATGCTGACAGCAAAGTCAG -3′), which 220
are specific for E. faecium sagA, in a total volume of 50 µl. The calculated sagA amplicon 221
size was 1500 bp. The conditions for PCR amplification were as follows: initial denaturation 222
was conducted for 3 minutes at 95°C, and 30 cycles of denaturation, annealing, and extension 223
were conducted at 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 2 minutes, 224
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
10
respectively. PCR products were purified with GeneJET PCR Purification Kit (Thermo 225
Scientific) and sequenced by Sanger methodology (38). The sagA gene sequence of the 226
residual 51 strains was extracted from whole genome sequence data. In these strains the sagA 227
was assembled on one contig. These sequence data have been submitted to the European 228
Nucleotide Archive (ENA) and can be accessed by the accession numbers LN714742 to 229
LN714774. 230
231
Phylogenetic tree based on SagA Variable Repeat Region VRR 232
Variable Repeat Region (VRR) sequences of SagA of all 85 E. faecium strains used in this 233
study were first aligned using MUSCLE (30) and a Maximum Likelihood tree was built from 234
the alignment data using Jones-Taylor-Thornton (JTT) model using in Mega version 6 (39). 235
236
SagA secondary structure prediction 237
Based on the SagA primary amino acid sequence, the SagA secondary structure was predicted 238
using the Chou & Fassman Secondary Structure Prediction Server at 239
http://www.biogem.org/tool/chou-fasman/ (40, 41). 240
241
Results 242
Phylogenetic analysis of E. faecium strains 243
A phylogenetic reconstruction of the 85 E. faecium strains included in this study was 244
performed. Of the 85 genome sequences used in this analysis, 64 were publicly available (3, 245
8, 42-44), whereas the genomes of the other 21 strains were newly sequenced for the purpose 246
of this study (Table S1). A total of 1186 single-copy core orthogroups (OGs) were identified 247
using OrthAgogue and MCL. Individual core OG alignments were built and concatenated to 248
generate a core genome sequence alignment. The resulting phylogenetic tree (Fig. 1a) 249
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
11
revealed the presence of two out of the three previously described clades: clade A1, which 250
includes the majority of hospital-associated strains, and clade B, including mainly human 251
commensal strains. Strains that formed a clear monophyletic clade A2 in previous analyses 252
(8) now formed a polyphyletic group located in-between clades B and A1, which we 253
designated intermediate group (IG). This IG should be considered as a multi-clonal group of 254
clinical, animal-derived and community strains with a topology, which is very different from 255
the ladder-like topology of clade A1 (Fig. 1a). However, as observed before in clade A2, 256
strains present in IG are closely related to clade A1, which makes its separation from clade A1 257
in some cases uncertain. 258
In total, 41 of the analysed strains clustered in clade A1, 13 in clade B and 31 strains were 259
assigned to IG. Clonal relatedness of the strains was considered based on pairwise SNP 260
comparisons (Table S2). The lowest number of SNPs found between two strains was 6 261
(between strains E1574 and E120), followed by 10 SNPs between strains E2560 and U317. 262
To avoid overrepresentation of these strains in clade A1, one strain from each described pair 263
was excluded, for the subsequent analysis, in this case E120 and U317 strains. 264
265
Biofilm formation among E. faecium strains 266
Biofilm formation of 83 E. faecium strains (Table S1) in TSBg was tested in a polystyrene (96 267
wells) plate assay. We observed biofilm formation in 59% (49 strains) of E. faecium strains 268
with no significant difference in biofilm formation between the different phylogenetic clades 269
(Fig. 2). 270
271
eDNA, not secreted proteins, is essential for E. faecium biofilm formation 272
eDNA and secreted proteins are extracellular matrix components important for biofilm 273
attachment and stability in different bacterial species (20). In polystyrene plates, the impact 274
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
12
of DNase I and proteinase K on initial biofilm formation in TSBg was tested in 49 E. faecium 275
strains belonging to different phylogenetic clades. Biofilm formation reduced (from 10 to 276
77% reduction) in 41 of 49 strains (83%) from the different phylogenetic clades when DNase 277
I was added to the biofilms (Fig. 3a). A statistically significant (p<0.05) decrease in biofilm 278
formation was observed among E. faecium clade A1 strains when proteinase K was added to 279
growing biofilms, but this was not observed for clade B strains (Fig. 3b). These results 280
indicate that both eDNA and proteins are essential for biofilm formation of clade A1 E. 281
faecium strains, and eDNA for clade B strains. The fact that we did not observe a reduction in 282
biofilm formation of clade B strains after proteinase K treatment may suggest that proteins are 283
less important than eDNA for biofilm formation of clade B E. faecium strain. Alternatively, it 284
is possible that proteins in clade B biofilms are less accessible to proteinase K or that in the 285
context of a biofilm clade B strains are able to degrade the exogenously added protease. 286
287
SagA is the major secreted protein in E. faecium biofilms 288
Since proteolytic treatment reduced biofilm formation, at least in clade A1 E. faecium strains, 289
we investigated the proteins secreted in the supernatant of biofilm-forming cells of 290
representative E. faecium strain from clade A1 (E1162) and clade B (E980) (Fig. 4). 291
Differences were not observed between the protein profiles in the supernatants of these strains 292
by SDS-PAGE gel. Using MALDI-TOF MS we identified the three dominant protein bands, 293
highlighted in Fig. 4, as SagA (Locus tag EfmE1162_2437 or EFF33872), Sulfatase domain 294
protein (Locus tag EfmE1162_1520 or EFF34609) and LysM domain protein 295
(EfmE1162_2234 or EFF34034) (Table S3). The three most predominant bands present in the 296
SDS-PAGE gel were also quantified by Image J (33). Since SagA was the most abundant 297
protein quantified in the supernatant of the biofilm-forming E. faecium cells (60% relative to 298
the other two proteins), we subsequently focused on the role of SagA in E. faecium biofilms. 299
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
13
SagA is susceptible to proteinase K and localization in the biofilms differs between clade 300
A1 and clade B E. faecium strains 301
To confirm the presence of SagA in the extracellular biofilm matrix and its susceptibility to 302
proteinase K, we selected six E. faecium strains, three clade A1 (E470, E1162, E1904) and 303
three clade B (E980, E3548, E1590), and analyzed biofilm formation using a semi-static 304
biofilm model in the presence or absence of proteinase K. As in the polystyrene assay, all E. 305
faecium strains, irrespective of their ecological or phylogenetic origin, were able to form 306
biofilms. After proteinase K treatment, SagA was detected neither in any of the E. faecium 307
biofilms (Fig. 5a), nor in the supernatant of biofilm growing cells (Fig. 5b). Biofilm thickness 308
of clade A1 strains was significantly decreased in the presence of proteinase K, in contrast to 309
clade B strains, in which the biofilm thickness was not significantly affected (Fig. 5c). This 310
confirmed the previous results in the polystyrene assay (Fig. 3b). 311
312
SagA was detected in biofilms of both clade A1 and clade B strains, but the localization of 313
SagA in biofilm was different in the two clades. In clade A1 strains, SagA was localized in 314
the middle of the biofilm with points of co-localization of bacteria (red) and SagA (green), 315
visualized in yellow (Fig. S1a), which suggests that it acts as an adhesive keeping bacterial 316
cells tightly connected. In contrast, in biofilms of clade B strains SagA seemed more surface 317
localized, which may explain why proteolytic removal of SagA does not have a major impact 318
on biofilm integrity in these strains as it does in clade A1 strains (Fig. S1b). 319
To confirm, the different roles for SagA in clade A1 and clade B we attempted to construct a 320
markerless sagA mutant and to select an insertion mutant in sagA from a transposon library in 321
E1162 (45). Both approaches unfortunately failed most probably because SagA was 322
previously described to be essential for E. faecium, probably due to its role in cell division 323
(46). 324
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
14
The susceptibility of SagA to proteolytic degradation is similar for E. faecium clade A1 325
and clade B strains 326
The observation that proteinase K treatment impacted biofilm formation of E. faecium clade 327
A1 strains more than clade B strains could, in addition to a difference in localization as 328
described above, also be explained by the fact that SagA of clade B strains is less susceptible 329
to proteolytic degradation by proteinase K. Therefore, we analyzed degradation of SagA, 330
present in 24 h biofilm supernatant, over time by proteinase K. Using E1162 (clade A1) and 331
E980 (clade B) strains, no significant difference was observed in SagA degradation with 332
proteinase K. In both E1162 and E980 SagA was not detected after 30 min of treatment (Fig. 333
6). This indicates that the degradation rate of SagA from E1162 and E980 by proteinase K is 334
comparable. 335
336
SagA is present in early and mature biofilms without detectable endopeptidase activity 337
Another explanation for the difference in susceptibility of clade A1 and clade B biofilms for 338
proteinase K treatment could rely on differences in proteolysis of SagA by Enterococcal 339
proteases produced by clade A1 or clade B strains or in SagA endopeptidase activity. 340
Proteolysis of SagA during biofilm development was studied by SDS-PAGE and Western blot 341
of early (4 h) and mature (24 h) biofilm supernatants of all six strains that were also analyzed 342
by confocal microscopy. In all strains, similar amounts of SagA were detected by SDS-PAGE 343
and Western blot in early biofilms and the mature biofilms, without any indications of 344
degradation of SagA over time (Fig. 7). Alternatively, constitutive production of SagA during 345
biofilm development could explain detection of apparently stable amounts of SagA in early 346
and late biofilm. Nevertheless, even after incubation of mature biofilm supernatant from 347
E1162 and E980 that contains high amounts of SagA for 24 h at 37ºC, no SagA degradation 348
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
15
could be detected (Fig. S3). This virtually excludes endopeptidase activity of SagA, present in 349
early and mature biofilm. 350
351
SagA is present in all E. faecium sequenced strains, but has different amino acid motifs 352
in clade A1 and clade B strains 353
We investigated the differences in sagA sequence among all E. faecium strains used in this 354
study as a possible explanation for differences in localization of SagA in biofilms grown from 355
E. faecium clade A1 or clade B strains. The sagA gene is present in all sequenced E. faecium 356
strains and the size of the SagA proteins ranged from 479 to 572 amino acids (aa) (average = 357
520 ± 12 AA) (Fig. S2). In a multiple sequence alignment of the 85 SagA orthologous protein 358
sequences (built using MUSCLE v.3.8.31) the SagA N- and C-terminal regions were highly 359
conserved (data not shown). The SagA N-terminal region was composed of a COG3883 360
domain, which is an uncharacterized conserved domain in bacteria, yet the C-terminal region 361
harbored a NLPC_60 domain, which has been described as a cell wall-associated hydrolase 362
(Fig. S2). However, the SagA central region was highly variable, mainly due to the presence 363
of imperfect 9-aa repeats. The number of these repeats was highly variable across the different 364
SagA orthologues. The variable repeat region (VRR) was flanked at the N-terminal side by 365
the amino acid sequence “AQATQASS”, which was 100% conserved among all analyzed 366
strains. Similarly, at the C-terminal side, the VRR was flanked by the amino acid sequence 367
“TTPSTDQSVD”, which was also 100% conserved among all strains. Using these regions as 368
boundaries for the VRR, the starting position for the VRR was found between SagA amino 369
acid positions 250 – 256, and the average size of the VRR was found to be 109 ± 12-aa, 370
including the 100% conserved boundaries. We further characterized the 9-aa repeated element 371
in SagA, and identified three types of VRR compositions (Fig. 1b). VRR-1 is highly 372
conserved in clade A1 strains, and it is composed of 7 to 9 repeat motifs (T/A-A/T-Q-S-S-373
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
16
A/M-T-E-E, in yellow and A/V-P-E-S-S-A/T-T-E-E, in green), which are interrupted by an 8-374
aa repeat (T-P-E-S-S-T-E-E, in blue) after the first three repeats. VRR-2 has a similar VRR 375
composition to type 1; however, it has an extra 8-aa repeat (V-P-E-S-S-T-E-E, in blue) at the 376
end of the VRR. VRR-3 is only present in clade B strains and has a different composition 377
compared to VRR-1. It is composed of 7 or 8 repeats (A-A-Q-S-S-A/T-T-E-E, in yellow and 378
A/V-P-E-S-S-T-T-E-E, in green) and one 8-aa repeat (V-P-E-S/T-S-T-E-E) at end of the 379
VRR. 380
Seventeen of 21 E. faecium strains with SagA VRR-1 (Fig. 8a) formed biofilms, which were 381
susceptible to proteinase K and represent clade A1 strains. Strains with VRR-2 are from 382
mixed origins and are found in the polyphyletic IG. Biofilms produced by these strains are 383
predominantly (13 of 20 strains) proteinase K-susceptible (Fig. 8b). In contrast, strains with 384
VRR-3 produced biofilms that were proteinase K-resistant (7 of 8 strains). All strains with 385
VRR-3 clustered in clade B (Fig. 8a). 386
Based on the variations observed in this SagA repeat region, we built a phylogenic tree (Fig. 387
S4). Phylogenetic inferences based on VRR identified two distinct clusters supported by 388
bootstrap values. Cluster 1 (VRR1 and VRR2) is composed of strains clustering in clade A1 389
and IG, while cluster 2 (VRR3) is composed of clade B strains. These results also show that 390
clade A1 and IG are phylogenetically closer related than clade A1 and clade B, as observed in 391
the genome based phylogenetic analysis. 392
393
Variable SagA VRR corresponds to different predicted secondary protein structures 394
Based on in silico predictions, the structural differences between the SagA proteins from 395
clade A1 and clade B strains were investigated (40, 41). In both clade A1 and clade B strains, 396
the repeat regions may serve as a linker between the N-terminal and the C-terminal domain 397
(Fig. 9a). SagA protein with VRR-1 consists of relatively more alpha-helices in the N-398
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
17
terminal fragment of the protein compared to SagA protein with VRR-3. In contrast, SagA 399
proteins with VRR-3 are composed of more beta-sheets than proteins with VRR-1 (Fig. 9b). 400
These predicted differences in SagA protein structure, might have a distinct effect on SagA 401
functionality and localization in biofilms of hospital-associated and community E. faecium 402
strains. 403
404
Discussion 405
Most difficult-to-treat chronic infections caused by multi-resistant E. faecium are biofilm-406
mediated. In the present study, we demonstrated that the major secreted antigen, SagA, is part 407
of the biofilm matrix produced by E. faecium and as such implicated in biofilm formation in 408
this species. The biofilms produced by hospital-associated (clade A1) strains and community 409
(clade B) strains were both found to be destabilized by DNase I treatment. Interestingly, the 410
biofilms generated by hospital-associated strains were highly sensitive to proteinase K 411
treatment, whereas the community strain-induced biofilms were not. 412
Three proteins, LysM domain protein, Sulfatase domain protein and SagA, were identified in 413
mature biofilm supernatant of clade A1 and clade B E. faecium strains. The LysM domain 414
protein (EFF34034) is a non-covalently binding peptidoglycan protein that has been reported 415
to be involved in erythromycin resistance in E. faecalis (47). Antibodies against the LysM 416
domain protein are protective against E. faecium and E. faecalis infection in a mouse 417
bacteremia model (47). The Sulfatase domain protein (EFF34609) is a predicted 418
transmembrane protein and has a MdoB conserved domain, described to be involved in cell 419
envelope biogenesis (47). The roles of these proteins in E. faecium biofilms remain to be 420
determined. 421
The most abundant protein identified was SagA, which has a COG3883 uncharacterized 422
conserved domain in the N-terminal and a NLPC_60 conserved domain in the C-terminal 423
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
18
region. Proteins containing a NLPC_60 domain have been characterized as lipoproteins or cell 424
hydrolases involved in cell division, cell wall maturation, and also virulence (48). SagA was 425
first identified by Teng et al. (46) by screening an E. faecium genomic expression library with 426
serum from a patient, who suffered from E. faecium endocarditis. The SagA protein appeared 427
essential for growth, to bind to a broad spectrum of extracellular matrix (ECM) proteins and 428
to be sensitive to proteinase K (46). Subsequently, Kropec et al. (35) demonstrated that serum 429
raised against purified SagA is opsonic and inhibits E. faecium infections in an experimental 430
animal model. While these data suggest that SagA is an important virulence factor in E. 431
faecium, its function has not previously been linked to biofilm formation. In the closely 432
related species E. faecalis, however, two secreted antigens, SalA and SalB, which displayed 433
similarity to SagA, have been found to play a role in biofilm formation (49). 434
The present study demonstrates that SagA is the major protein secreted during biofilm 435
formation in E. faecium. Furthermore, SagA is localized within the biofilm matrix and 436
proteolytic degradation, which also degrades SagA, markedly reduced E. faecium biofilms of 437
only clade A1, but not of clade B. The potential difference between clade A1 and clade B in 438
the impact of proteolytic degradation on biofilm formation correlates with differences 439
between these clades in repeat profiles of their SagA proteins. We hypothesize that 440
differences in repeat profiles of SagA can lead to changes in the secondary structure of the 441
protein, which may determine the localization as well as the role of SagA in E. faecium 442
biofilm matrices, possibly as a result of diverse interactions of SagA variants with other E. 443
faecium cell surface determinants. A difference in SagA localization within biofilms may 444
explain the difference in proteolytic susceptibility of biofilms that was observed between 445
clade A1 strains, all of which express SagA proteins with VRR-1, and clade B strains, which 446
express SagA proteins with VRR-3. 447
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
19
Our findings also demonstrate that eDNA is an essential part of biofilms produced by all 448
tested E. faecium strains. This is consistent with previous findings that the major autolysin 449
AtlAEfm, which is present in all sequenced E. faecium strains, is responsible for eDNA release 450
in E. faecium E1162 biofilms (19). Our observation that clade A1 and clade B E. faecium 451
strains form biofilms with distinct properties, suggests that under different ecological 452
conditions different types of biofilms are produced, possibly contributing to adaptation to 453
different niches. 454
455
Acknowledgments 456
The authors wish to thank Ellen Brouwer for her expert technical assistance during the 457
genomic DNA isolation for whole genome sequence and Prof. Jukka Corander for the helpful 458
discussion on the interpretation of the phylogenetic analysis. Part of this work was supported 459
by ZonMW VENI grant 91610058 to H.L.L. from The Netherlands Organization for Health 460
Research and Development and European Union Seventh Framework Programme (FP7-461
HEALTH-2011-single-stage) under grant agreement no. 282004, EvoTAR to R.J.L.W. 462
463
References 464
1. Hidron AI, Edwards JR, Patel J, Horan TC, Sievert DM, Pollock DA, Fridkin 465
SK. 2008. NHSN annual update: antimicrobial-resistant pathogens associated with 466
healthcare-associated infections: annual summary of data reported to the National 467
Healthcare Safety Network at the Centers for Disease Control and Prevention, 2006-468
2007. Infect Control Hosp Epidemiol 29:996-1011. 469
2. Heikens E, Singh KV, Jacques-Palaz KD, van Luit-Asbroek M, Oostdijk EA, 470
Bonten MJ, Murray BE, Willems RJ. 2011. Contribution of the enterococcal surface 471
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
20
protein Esp to pathogenesis of Enterococcus faecium endocarditis. Microbes Infect 472
13:1185-1190. 473
3. de Been M, van Schaik W, Cheng L, Corander J, Willems RJ. 2013. Recent 474
recombination events in the core genome are associated with adaptive evolution in 475
Enterococcus faecium. Genome Biol Evol 5:1524-1535. 476
4. Homan WL, Tribe D, Poznanski S, Li M, Hogg G, Spalburg E, Van Embden JD, 477
Willems RJ. 2002. Multilocus sequence typing scheme for Enterococcus faecium. J 478
Clin Microbiol 40:1963-1971. 479
5. Willems RJ, Top J, van Santen M, Robinson DA, Coque TM, Baquero F, 480
Grundmann H, Bonten MJ. 2005. Global spread of vancomycin-resistant 481
Enterococcus faecium from distinct nosocomial genetic complex. Emerg Infect Dis 482
11:821-828. 483
6. Willems RJ, Top J, van Schaik W, Leavis H, Bonten M, Siren J, Hanage WP, 484
Corander J. 2012. Restricted gene flow among hospital subpopulations of 485
Enterococcus faecium. MBio 3:e00151-00112. 486
7. Tedim AP, Ruiz-Garbajosa P, Corander J, Rodriguez CM, Canton R, Willems 487
RJ, Baquero F, Coque TM. 2015. Population biology of intestinal Enterococcus 488
isolates from hospitalized and nonhospitalized individuals in different age groups. 489
Appl Environ Microbiol 81:1820-1831. 490
8. Lebreton F, van Schaik W, McGuire AM, Godfrey P, Griggs A, Mazumdar V, 491
Corander J, Cheng L, Saif S, Young S, Zeng Q, Wortman J, Birren B, Willems 492
RJ, Earl AM, Gilmore MS. 2013. Emergence of epidemic multidrug-resistant 493
Enterococcus faecium from animal and commensal strains. MBio 4:e00534-00513. 494
9. Donlan RM, Costerton JW. 2002. Biofilms: survival mechanisms of clinically 495
relevant microorganisms. Clin Microbiol Rev 15:167-193. 496
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
21
10. Sandoe JA, Witherden IR, Cove JH, Heritage J, Wilcox MH. 2003. Correlation 497
between enterococcal biofilm formation in vitro and medical-device-related infection 498
potential in vivo. J Med Microbiol 52:547-550. 499
11. O'Toole G, Kaplan HB, Kolter R. 2000. Biofilm formation as microbial 500
development. Annu Rev Microbiol 54:49-79. 501
12. Abee T, Kovacs AT, Kuipers OP, van der Veen S. 2011. Biofilm formation and 502
dispersal in Gram-positive bacteria. Curr Opin Biotechnol 22:172-179. 503
13. Ahn SJ, Burne RA. 2006. The atlA operon of Streptococcus mutans: role in autolysin 504
maturation and cell surface biogenesis. J Bacteriol 188:6877-6888. 505
14. Allesen-Holm M, Barken KB, Yang L, Klausen M, Webb JS, Kjelleberg S, Molin 506
S, Givskov M, Tolker-Nielsen T. 2006. A characterization of DNA release in 507
Pseudomonas aeruginosa cultures and biofilms. Mol Microbiol 59:1114-1128. 508
15. Petersen FC, Tao L, Scheie AA. 2005. DNA binding-uptake system: a link between 509
cell-to-cell communication and biofilm formation. J Bacteriol 187:4392-4400. 510
16. Qin Z, Ou Y, Yang L, Zhu Y, Tolker-Nielsen T, Molin S, Qu D. 2007. Role of 511
autolysin-mediated DNA release in biofilm formation of Staphylococcus epidermidis. 512
Microbiology 153:2083-2092. 513
17. Steinberger RE, Holden PA. 2005. Extracellular DNA in single- and multiple-514
species unsaturated biofilms. Appl Environ Microbiol 71:5404-5410. 515
18. Vilain S, Pretorius JM, Theron J, Brozel VS. 2009. DNA as an adhesin: Bacillus 516
cereus requires extracellular DNA to form biofilms. Appl Environ Microbiol 75:2861-517
2868. 518
19. Paganelli FL, Willems RJ, Jansen P, Hendrickx A, Zhang X, Bonten MJ, Leavis 519
HL. 2013. Enterococcus faecium biofilm formation: identification of major autolysin 520
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
22
AtlAEfm, associated Acm surface localization, and AtlAEfm-independent extracellular 521
DNA Release. MBio 4:e00154. 522
20. Flemming HC, Wingender J. 2010. The biofilm matrix. Nat Rev Microbiol 8:623-523
633. 524
21. Tielker D, Hacker S, Loris R, Strathmann M, Wingender J, Wilhelm S, Rosenau 525
F, Jaeger KE. 2005. Pseudomonas aeruginosa lectin LecB is located in the outer 526
membrane and is involved in biofilm formation. Microbiology 151:1313-1323. 527
22. Diggle SP, Stacey RE, Dodd C, Camara M, Williams P, Winzer K. 2006. The 528
galactophilic lectin, LecA, contributes to biofilm development in Pseudomonas 529
aeruginosa. Environ Microbiol 8:1095-1104. 530
23. Johansson EM, Crusz SA, Kolomiets E, Buts L, Kadam RU, Cacciarini M, 531
Bartels KM, Diggle SP, Camara M, Williams P, Loris R, Nativi C, Rosenau F, 532
Jaeger KE, Darbre T, Reymond JL. 2008. Inhibition and dispersion of 533
Pseudomonas aeruginosa biofilms by glycopeptide dendrimers targeting the fucose-534
specific lectin LecB. Chem Biol 15:1249-1257. 535
24. Branda SS, Chu F, Kearns DB, Losick R, Kolter R. 2006. A major protein 536
component of the Bacillus subtilis biofilm matrix. Mol Microbiol 59:1229-1238. 537
25. Lasa I, Penades JR. 2006. Bap: a family of surface proteins involved in biofilm 538
formation. Res Microbiol 157:99-107. 539
26. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, Lesin 540
VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV, Vyahhi N, 541
Tesler G, Alekseyev MA, Pevzner PA. 2012. SPAdes: a new genome assembly 542
algorithm and its applications to single-cell sequencing. J Comput Biol 19:455-477. 543
27. Seemann T. 2014. Prokka: rapid prokaryotic genome annotation. Bioinformatics 544
30:2068-2069. 545
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
23
28. Ekseth OK, Kuiper M, Mironov V. 2014. orthAgogue: an agile tool for the rapid 546
prediction of orthology relations. Bioinformatics 30:734-736. 547
29. Enright AJ, Van Dongen S, Ouzounis CA. 2002. An efficient algorithm for large-548
scale detection of protein families. Nucleic Acids Res 30:1575-1584. 549
30. Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accuracy and 550
high throughput. Nucleic Acids Res 32:1792-1797. 551
31. Capella-Gutierrez S, Silla-Martinez JM, Gabaldon T. 2009. trimAl: a tool for 552
automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 553
25:1972-1973. 554
32. Price MN, Dehal PS, Arkin AP. 2010. FastTree 2-approximately maximum-555
likelihood trees for large alignments. PLoS One 5:e9490. 556
33. Schneider CA, Rasband WS, Eliceiri KW. 2012. NIH Image to ImageJ: 25 years of 557
image analysis. Nat Methods 9:671-675. 558
34. Verkaik NJ, de Vogel CP, Boelens HA, Grumann D, Hoogenboezem T, Vink C, 559
Hooijkaas H, Foster TJ, Verbrugh HA, van Belkum A, van Wamel WJ. 2009. 560
Anti-staphylococcal humoral immune response in persistent nasal carriers and 561
noncarriers of Staphylococcus aureus. J Infect Dis 199:625-632. 562
35. Kropec A, Sava IG, Vonend C, Sakinc T, Grohmann E, Huebner J. 2011. 563
Identification of SagA as a novel vaccine target for the prevention of Enterococcus 564
faecium infections. Microbiology 157:3429-3434. 565
36. Heydorn A, Nielsen AT, Hentzer M, Sternberg C, Givskov M, Ersboll BK, Molin 566
S. 2000. Quantification of biofilm structures by the novel computer program 567
COMSTAT. Microbiology 146 ( Pt 10):2395-2407. 568
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
24
37. Hendrickx AP, van Wamel WJ, Posthuma G, Bonten MJ, Willems RJ. 2007. Five 569
genes encoding surface-exposed LPXTG proteins are enriched in hospital-adapted 570
Enterococcus faecium clonal complex 17 isolates. J Bacteriol 189:8321-8332. 571
38. Sanger F, Coulson AR. 1975. A rapid method for determining sequences in DNA by 572
primed synthesis with DNA polymerase. J Mol Biol 94:441-448. 573
39. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013. MEGA6: Molecular 574
Evolutionary Genetics Analysis version 6.0. Mol Biol Evol 30:2725-2729. 575
40. Chou PY, Fasman GD. 1974. Conformational parameters for amino acids in helical, 576
beta-sheet, and random coil regions calculated from proteins. Biochemistry 13:211-577
222. 578
41. Chou PY, Fasman GD. 1974. Prediction of protein conformation. Biochemistry 579
13:222-245. 580
42. Lam MM, Seemann T, Bulach DM, Gladman SL, Chen H, Haring V, Moore RJ, 581
Ballard S, Grayson ML, Johnson PD, Howden BP, Stinear TP. 2012. Comparative 582
analysis of the first complete Enterococcus faecium genome. J Bacteriol 194:2334-583
2341. 584
43. de Regt MJ, van Schaik W, van Luit-Asbroek M, Dekker HA, van Duijkeren E, 585
Koning CJ, Bonten MJ, Willems RJ. 2012. Hospital and community ampicillin-586
resistant Enterococcus faecium are evolutionarily closely linked but have diversified 587
through niche adaptation. PLoS One 7:e30319. 588
44. van Schaik W, Top J, Riley DR, Boekhorst J, Vrijenhoek JE, Schapendonk CM, 589
Hendrickx AP, Nijman IJ, Bonten MJ, Tettelin H, Willems RJ. 2010. 590
Pyrosequencing-based comparative genome analysis of the nosocomial pathogen 591
Enterococcus faecium and identification of a large transferable pathogenicity island. 592
BMC Genomics 11:239. 593
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
25
45. Zhang X, Paganelli FL, Bierschenk D, Kuipers A, Bonten MJ, Willems RJ, van 594
Schaik W. 2012. Genome-wide identification of ampicillin resistance determinants in 595
Enterococcus faecium. PLoS Genet 8:e1002804. 596
46. Teng F, Kawalec M, Weinstock GM, Hryniewicz W, Murray BE. 2003. An 597
Enterococcus faecium secreted antigen, SagA, exhibits broad-spectrum binding to 598
extracellular matrix proteins and appears essential for E. faecium growth. Infect 599
Immun 71:5033-5041. 600
47. Aakra A, Vebo H, Snipen L, Hirt H, Aastveit A, Kapur V, Dunny G, Murray BE, 601
Nes IF. 2005. Transcriptional response of Enterococcus faecalis V583 to 602
erythromycin. Antimicrob Agents Chemother 49:2246-2259. 603
48. Tran SL, Guillemet E, Gohar M, Lereclus D, Ramarao N. 2010. CwpFM (EntFM) 604
is a Bacillus cereus potential cell wall peptidase implicated in adhesion, biofilm 605
formation, and virulence. J Bacteriol 192:2638-2642. 606
49. Mohamed JA, Teng F, Nallapareddy SR, Murray BE. 2006. Pleiotrophic effects of 607
2 Enterococcus faecalis sagA-like genes, salA and salB, which encode proteins that 608
are antigenic during human infection, on biofilm formation and binding to collagen 609
type I and fibronectin. J Infect Dis 193:231-240. 610
611
Figure legends 612
Figure 1. Phylogenetic reconstruction of 85 E. faecium strains based on core genome 613
variation. a. Phylogenetic tree (FastTree) built from an alignment of 1186 core genes in 85 E. 614
faecium genomes. The clades A1 and B and the Intermediate Group (IG) are indicated in red, 615
green and blue, respectively. Selected strains for further phenotypic tests are highlighted in 616
red in clade A1 (E470, E1162 and E1904) and in green in clade B (E980, E3548, E1590). b. 617
SagA variable repeat region (VRR) are indicated after each strain and divided into three types 618
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
26
of VRRs named VRR-1 (red), VRR-2 (blue) and VRR-3 (green). The sequences highlighted 619
in yellow, blue, green and red indicate the different repeat variants. 620
Figure 2. Biofilm polystyrene assay. 24 h biofilm formation of 83 E. faecium strains from 621
different origins was studied in polystyrene plates with TSBg. Biofilms were stained with 622
crystal violet and the amount of biofilm formation was measured at 595 nm. The six strains 623
selected for further phenotypic tests are color code in yellow (E470), red (E1162), orange 624
(E1904), green (E980), purple (E3548) and blue (E1590). 625
Figure 3. DNase I and proteinase K sensitivity in initial biofilm formation. 24 h biofilm 626
formation of 49 E. faecium strains from different origins in polystyrene plates with TSBg in 627
the presence of DNase I (1.5 µg µl-1) (a) or proteinase K (1 µg µl-1) (b). Biofilms were stained 628
with crystal violet and the amount of biofilm formation was measured at 595 nm. Asterisks 629
represent significant differences (*p < 0.05, unpaired two-tailed Student's t-test) between 630
clade A1 and clade B or IG and clade B strains. The six strains selected for further phenotypic 631
tests are color code in yellow (E470), red (E1162), orange (E1904), green (E980), purple 632
(E3548) and blue (E1590). 633
Figure 4. Proteins released in the biofilm supernatant. Proteins present in the 24 h biofilm 634
supernatant of E1162 (clade A1) E. faecium strain and E980 (clade B) E. faecium strain were 635
equally loaded and separated using a 12.5% SDS page gel and stained with Coomassie blue. 636
Indicated bands were excised from the gel and identified by mass spectrometry as secreted 637
antigen A (SagA, EfmE1162_2437); Sulfatase (EfmE1162_1520) and LysM 638
(EfmE1162_2234). 639
Figure 5. Confocal microscopy images depicting biofilm thickness and SagA localization 640
in a semi-static biofilm model with or without proteinase K treatment. a. Biofilms of six 641
E. faecium strains, three clade A1 (E470, E1162, E1904) and three clade B strains (E980, 642
E3548, E1590), were grown for 24 h on poly-L-Lysine coated glass, in TSBg with or without 643
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
27
proteinase K, at 120 rpm, at 37ºC. The ecological origin and clade assignment of the strains 644
are listed in table S1. Biofilms were incubated with rabbit polyclonal α-SagA antiserum and 645
goat α-rabbit Alexa 488 (green). Bacterial membranes were stained with FM 95-5 (red). b. 646
Presence of SagA in the biofilm supernatant of E. faecium strains incubated with (+) or 647
without (-) proteinase K was analyzed by Western Blot using rabbit polyclonal α-SagA 648
antiserum. c. The average thickness of biofilms was measured at five random positions and 649
analyzed with Comstat/Matlab software. Asterisks represent significant differences (*p < 650
0.05, unpaired two-tailed Student's t-test) with proteinase K treatment. Pictures were taken at 651
63x magnification with 2.0 optical zoom. 652
Figure 6. Susceptibility of SagA in 24 h biofilm supernatant to proteinase K treatment. 653
a. Proteins present in the 24 h biofilm supernatant of E1162 (clade A1) E. faecium strain and 654
E980 (clade B) E. faecium strain were equally loaded and separated using a 12.5% SDS page 655
gel and stained with Coomassie blue after incubation with 0,1 µg ml-1 proteinase K at 37ºC 656
for 1 min, 5 min, 15 min, 30 min and 60 min. b. Degradation of SagA in the biofilm 657
supernatant of E. faecium strains incubated with proteinase K was analyzed by Western Blot 658
using rabbit polyclonal α-SagA antiserum. 659
Figure 7. Stability of SagA during biofilm development. a. Proteins present in 4 h or 24 h 660
biofilm supernatants of six E. faecium strains, three clade A1 (E470, E1162, E1904) and three 661
clade B strains (E980, E3548, E1590) were equally loaded and separated using a 12.5% SDS 662
page gel and stained with Coomassie blue. b. Presence of SagA in the biofilm supernatants of 663
E. faecium strains after 4 h or 24 h growth was analyzed by Western Blot using rabbit 664
polyclonal α-SagA antiserum. 665
Figure 8. Correlation between proteinase K susceptibility of biofilm formation and SagA 666
variable repeat cluster (VRR) type. Percentage of biofilm reduction in 49 E. faecium strains 667
in the presence of 1.0 µg µl-1 proteinase K is indicated on the y-axis, while the level of biofilm 668
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from
28
formation in the polystyrene biofilm model is indicated on the x-axis. a. E. faecium strains 669
with SagA VRR-1 and VRR-3, b. E. faecium strains with all three SagA VRRs. The six 670
strains selected for further phenotypic tests are color code in yellow (E470), red (E1162), 671
orange (E1904), green (E980), purple (E3548) and blue (E1590). 672
Figure 9. SagA secondary structure prediction. Comparison of the predicted secondary 673
structures of total SagA protein (a) and the variable repeat region (VRR) (b) of the E. faecium 674
strain E1162, which represents SagA protein with VRR-1, strain E1575, which represents 675
SagA with VRR-2, and strain E980, which represents SagA with VRR-3. Secondary 676
structures were predicted using a web tool (CFSSP: Chou & Fassman Secondary Structure 677
Prediction Server). Alpha-helices are indicated in red, beta-sheets in green, beta-turns in blue 678
and random coils in yellow. 679
680
681
682
on March 14, 2018 by guest
http://aem.asm
.org/D
ownloaded from