Supplemental Data Real-Time Analysis of Effector Translocation by the

Document Sample
Supplemental Data Real-Time Analysis of Effector Translocation by the Powered By Docstoc
					Cell Host & Microbe, Volume 3

Supplemental Data

Real-Time Analysis of Effector Translocation

by the Type III Secretion System of

Enteropathogenic Escherichia coli
Erez Mills, Kobi Baruch, Xavier Charpentier, Simi Kobi, and Ilan Rosenshine

Supp. 1. Construction of chromosomal BlaM fusions
Two representative examples of construction of blaM translational fusions are shown in
Figure S1A, that of tir-blaM and that of espH-blaM. With tir, map, and espZ, the native
as well as the blaM-fused alleles are expressed from the native promoters, and with espF,
espG, and espH, the blaM-fused alleles are expressed from the native promoters, and the
native alleles are promoter-less. We used the merodiploidic forms in cases where we
suspected that the lack of a native effector might influence the translocation rate. For
instance, native Tir is required for intimate attachment and thus its absence might affect
the translocation kinetics.

Supp. 2. Formation of actin pedestals by strains containing chromosomal effector-
blaM fusions.
Strains carrying chromosomal effector-blaM fusion were grown under LEE-inducing
conditions and were used to infect HeLa cells for one hour. Cells were then washed,
fixed, and stained by phalloidin-rhodamine to visualize the formation of actin pedestals.
As a positive control, we used wild-type EPEC and as a negative control we used an
EPEC escV::Tn10kan mutant. All the effector-blaM-containing strains induced the
formation of actin pedestals at the same efficiency as that of the wild-type EPEC (Figure
S1 C and data not shown).
Figure S1. Construction and characterization of chromosomal effector-blaM strains.
(A) The different effector-blaM fusions were inserted into pCX391, with or without their corresponding
native promoters. For example, espH-blaM was inserted to create pCX445 (left) and tir-blaM, accompanied
by its native promoter (PLEE5), was inserted to create pCX442 (right). The plasmids were then integrated
with the chromosome via homologues recombination between the insert and the corresponding
chromosomal regions (indicated in dark blue). Integration with pCX445 created the strain CX2138 (left),
which contains espH-blaM under the control of the native chromosomal promoter (PLEE3) and a second,
promoterless espH allele. Integration with pCX442 created the strain CX2135 (right), which contains tir-
blaM under the control of the native chromosomal promoter (PLEE5) and a second tir allele controlled by the
plasmid-supplied PLEE5. To maintain wild-type regulation via the cloned promoters, when cloning the
promoter regions, we made sure that we included the regulatory regions related to the Ler and H-NS
binding sites, which are located within the coding regions of the flanking genes (Bustamante et al., 2001;
Haack et al., 2003; Sanchez-SanMartin et al., 2001; Sperandio et al., 2000; Umanski et al., 2002).
(B) Strain CX2135, carrying the tir-blaM fusion, was grown under LEE-inducing conditions. Whole cell
extracts were analyzed by western blotting with anti-Tir antibodies. The two forms of Tir are indicated on
the left: native tir (Tir) and fusion protein (Tir-BlaM).
(C) Strain CX2135, carrying the tir-blaM fusion, was grown under LEE-inducing conditions and was used
to infect HeLa cells for one hour. Cells were then washed, fixed, and stained by phalloidin-rhodamine to
visualize the formation of actin pedestals. Representative phase contrast (left) and the corresponding
stained actin (right) images are shown. We found that the formation of actin pedestals by this strain, as well
as by all the other strains containing effector-blaM fusions (data not shown) is indistinguishable from
pedestal formation by wild-type EPEC.

Supp 3. Calculation of Effector-BlaM concentration in EPEC
Proteins were extracted from strains expressing the chromosomal effector-blaM genes
and specific activity (activity per culture density determined at OD600) was measured
using nitrocefin as a BlaM substrate. This was compared to a calibration curve that was
obtained by measuring the activity of known concentrations of purified BlaM (a gift from
G. Schreiber, the Weizmann Institute). We calculated the bacterial volume, assuming that
the average bacteria have a cylindrical shape 1 µm long and a diameter of 0.5 µm. The
number of bacteria was deduced from the OD: we determined that 1 OD of EPEC equals
5x108 bacteria/ml.

Supp 4. The translocation system and possible concerns.
The translocation measurement system and possible pit falls in using it are shown in
Figure S2. Concerns that we addressed include the stability of the CCF2 concentration in
the loaded cells (section S5 and Figure S3 A), stability of the CCF2 product (coumarin)
in the infected cells (section S6 and Figure S3 B), and the effects of possible leakage of
CCF2 from the cells and possible leakage of Eeffector-BlaM from the bacteria (section
S5 and Figures S3, S4, S5, S7 and S8).
Figure S2. Schematics of the translocation system.
CCF2/AM is an acetylated form of CCF2 and the hydrophobic CCF2/AM freely diffuses in and out of the
host cell. CCF2/AM does not fluoresce. Once inside the cell, CCF2/AM is modified by endogenic esterases
to CCF2; then CCF2 i) is charged and thus trapped and accumulates inside the host cell and ii) fluoresces
upon excitation. CCF2 consists of a cephalosporin core linking two fluorophores: 7-hydroxycoumarin and
fluorescein. Excitation of the coumarin at 409 nm results in fluorescence resonance energy transfer (FRET)
to the fluorescein moiety, which emits at 520 nm (green). Cleavage of the intracellular CCF2 by
translocated Effector-BlaM results in the disruption of FRET and excitation of the coumarin at 409 nm,
which results in emission at 447 nm (blue). Possible concerns are denoted by fragmented arrows. These
include leakage to the medium of CCF2 (from the host cells) and of the Effector-BlaM (from the bacteria).
Such leakages might lead to cleavage of extracellular CCF2 by extracellular Effector-BlaM, thus distorting
the assay. To minimize leakage of CCF2 from the cell, we included probenecid, a non-specific inhibitor of
anion transport, in the infection medium. Also, the addition of CCF2/AM to the infection medium increases
the stability of CCF2 levels within the HeLa cells (Figure S3A). To block activity of possible extracellular
Effecor-blaM, we included the beta-lactamase-inhibitory-protein (BLIP) in the infection medium when
needed. BLIP can not inhibit intracellular BlaM activity (Fig S4). We verified that BLIP and probenecid
affected neither bacterial growth nor the rate of effector translocation (data not shown). Note that similar
hierarchy of effector translocation efficiencies was obtained when BLIP and CCF2/AM were eliminated
from the infection medium (compare Fig 2 to supplemental Figure S8A). Therefore, although including
these reagents in the infection medium improves the assay’s accuracy, they are not needed to reveal the
relative translocation efficiencies.

Supp. 5. Optimizing the conditions to retain stable CCF2 concentrations in HeLa
To determine the stability of CCF2 concentrations in the cells, we preloaded HeLa cells
with CCF2/AM as described in the Experimental Procedures, but we did not infect them
with bacteria. Instead, we replaced the loading medium with bacteria-free infection
medium supplemented or not supplemented with 1.0 µM CCF2/AM (reload-protocol or
regular-protocol, respectively). Fluorescence was recorded for the next 100 minutes. In
parallel experiments, every 10 minutes the medium was removed to eliminate CCF2 that
leaked out of the cells, and the level of fluorescence retained by the HeLa cells after the
wash was recorded. Importantly, we noticed low, but significant CCF2 leakage from the
cells. The ratio between before and after the wash signifies the percent of retained
substrate (Figure S3 A). The results show that using the regular protocol, translocation
measurements will not be reliable after about 80 min post loading. We also demonstrated
that the reload-protocol enhances the accuracy of the measurements and allows reading
the translocation for longer periods. Nevertheless, in this study we tested translocation up
to 100 min post infection because after more than 100 min post infection, the host cells
become intoxicated, followed by cell death and detachment from the substratum (data not

Supp. 6. The stability of coumarin, the CCF2 product, in infected HeLa cells.
We estimated the stability of coumarin, ([P]), in the infected HeLa cells. Since infection
with EPEC tir-blaM results in a massive translocation of enzyme and depletion of
substrate (see Figures 2A and B), we used this strain to determine the stability of the
coumarin. Briefly, HeLa cells were loaded with CCF2, and then infected with EPEC tir-
blaM. After three hours of infection, we began reading the levels of fluorescence emitted
by coumarin, as described in the Experimental Procedures. The results indicate that
coumarin stably fluoresces for at least three hours (Figure S3 B), and that coumarin ([P])
decay is not a concern in this system.
Figure S3. CCF2 substrate and product dynamics.
(A) Substrate stability. Preloaded HeLa cells were washed and covered with 200 µL cDMEM (Regular
protocol, Pink line) or with cDMEM supplemented with 1.0 µM CCF2/AM (Reload protocol, Dark Purple
line). The level of the substrate was monitored for 100 minutes (excitation at 405 nm, and emission
detection using a 535-nm filter set). In parallel assays, every 10 minutes the medium was exchanged with
fresh medium and we determined how much of the substrate-signal was retained in the HeLa cell layer. The
level of substrate retained is plotted by the empty circles. Reload protocol: Dark Purple circles and regular
protocol: Pink circles. On the right are bars representing the amplitudes of change in the retained substrate
levels during these 100 minutes. Also on the right is the average level of retained substrate and the standard
deviation. The results show that CCF2 leaks out of the cells, thus reducing the CCF2 intracellular
concentration, but the reload protocol significantly improved the stability of the CCF2 intracellular
(B) Product stability. Preloaded HeLa cells were infected with EPEC tir::blaM for three hours so that all
the substrate was converted to product. The level of the product was then monitored for an additional three
hours (excitation at 405 nm, and emission detection using a 465-nm filter set). A trend line’s formula was
generated by Excel. The results show that the product signal is very stable
(C) Linearity between fluorescence and CCF2 concentration. CCF2/FM (a preactivated form of CCF2) was
diluted progressively and then fluorescence was recorded (excitation at 405 nm, and emission detection
using a 535-nm filter set). Excel was used to calculate the conversion formula contained in the graph and
the R2 is shown.

Supp. 7. Leakage of Effector-BlaM from the bacteria is very low
The experiments, shown in Figures S4 and S5, demonstrate that under our experimental
conditions secretion (or leakage from wild-type EPEC) of BlaM fused to the six LEE
effectors (EspF, EspF, EspG, EspZ, Map, or Tir) is very low. Thus, the addition of BLIP
(to inhibit the activity of secreted Effector-BlaM) is not always required. In these
experiments we used as a positive control a strain containing a plasmid expressing espB-
BlaM. EspB is a translocon component that is known to be both secreted to the medium
and translocated into cells (Wolff et al., 1998). In contrast to the other effectors, strong
secretion to the medium of EspB-BlaM was recorded.

Figure S4. The presence of BLIP in the infection media eliminates background owing to the activity of
secreted Effector-BlaM fusions.
The six effector-blaM strains were grown under LEE-inducing conditions and used for a translocation
experiment, as detailed in the Experimental Procedures section. BLIP was either added during the infection,
or not (green and blue lines, respectively). As a negative control, cells were infected with wild-type EPEC
that did not carry an effector-blaM gene. As a positive control, we used EPEC-expressing espB-blaM. The
tested effector fusion is indicated within each graph. The results show that in the case of EspB-BlaM, BLIP
was needed to prevent activity of secreted EspB-BlaM. In all other cases, similar results were obtained
whether or not BLIP was added. These results indicate that under our experimental conditions, effector
secretion to the medium is negligible and thus the presence of BLIP in the infection medium is not

Figure S5. Low leakage of Effector-BlaMs, from strains containing plasmid-borne effector-blaM genes.
EPEC strains carrying plasmids with the different effector-blaM fusions, as well as espB-blaM (as a
positive control) and a wild-type EPEC that did not carry an effector-blaM gene (as a negative control),
were grown under LEE-inducing conditions and used to infect HeLa cells. After 80 minutes, supernatant
containing possible secreted Effector-BlaM proteins was harvested and cleared by centrifugation. To assess
the amounts of Effector-BlaM in the cleared supernatants, we determined the β-lactamase activity in these
cleared supernatant using nitroceffin as substrate. In all cases, excluding EspB-BlaM, secretion of Effector-
BlaM to the medium was very low. Each column is the average value of three independent assays and the
bars represent the standard error of the mean.

Supp. 8. The influence of the multiplicity of infection (MOI) on the translocation
In Fig. 2A the cells were infected at MOI 700. Under these conditions, about 1% of the
bacteria attach within 30 min. Increasing the MOI accelerated the attachment rate as well
as the translocation; in contrast, reducing the MOI reduced the attachment and
translocation rates. However, in all cases the translocation order, as shown in Fig 2, was
maintained (data not shown). In the case of Fig. 2, an MOI of 700 was chosen because it
allowed the presentation of data generated for all the effectors in a single experiment.
Upon using higher MOIs, the Tir-BlaM signal overshoots, and upon using lower MOIs,
more than 30 min are needed to obtain a clear signal from Map-BlaM.

Supp. 9. The translocation assay is not affected by the rate of BlaM refolding.
We tested whether the translocation results might be distorted due to differences in BlaM
refolding kinetics between the different Effector-BlaM fusions. The effectors are
translocated via TTSS in a partially unfolded state and refold upon entry into the host
cells (Akeda and Galan, 2005). Thus, variations in the amount of accumulated CCF2
product in cells infected with different strains (Figure 2A) might reflect differences in the
refolding kinetics of the BlaM domains. To test this possibility, we denatured the
Effector-BlaM fusions in 6 M urea. Next, to allow renaturation, they were diluted into a
reaction buffer (final urea concentration: 0.02 M) and activity was monitored. All the
Effector-BlaM fusions gained activity within less than 100 seconds upon dilution and
their activities were similar to those of the native proteins (Figure S6). This rapid
refolding might reflect the need of native BlaM, a periplasmic protein, to refold upon
entering the periplasm via the Sec translocon. The effect of urea treatment is probably
more drastic than the protein partial unfolding during translocation. Thus, refolding of
translocated BlaM in the host cell cytoplasm is possibly even faster than that seen upon
urea treatment. In addition, like in the case of urea treatment, the refolding of the
translocated proteins is expected to be independent of the attached effector. In
conclusion, possible differences in the efficiency of BlaM refolding would have a
negligible effect on the translocation assay.

Figure S6. Effector-BlaM refolding efficiencies and rates.
Overnight cultures of EPEC strains containing plasmids encoding effector-blaM fusions (pCX392,
pCX394, pCX395, pCX396, pCX397, and pME21) were diluted 1:100 into 100 ml LB, grown under
aerobic conditions for 2 hours at 37°C, reaching an OD of ~0.4. Cultures were then supplemented with
IPTG (0.5 mM), grown for an additional 2 hours, washed 3 times with PBS, resuspended in 1 ml of PBS
with 1% Triton X100 and sonicated after the addition of PMSF (0.1 mM final concentration). Cleared
lysates were then diluted 1:10 in a solution of 6.6M Urea, 10 mM Tris pH 7.5 (final urea concentration 6M)
and incubated for 10 min at room temperature. Under these denaturing conditions, the BlaM is completely
inactive (data not shown). To allow protein refolding, we diluted the denatured proteins 1:300 in 10 mM
Tris pH 7.5, reducing the urea concentration to 0.02 M. The diluted solution (40 ul) was added immediately
into 5 µl of 100 mM CCF2/FA substrate (Invitrogen) in a 96-well plate (black, clear bottom) and activity
(fluorescence data at 465 and 535 nm) was recorded in 30-sec intervals (green line). As a "native protein"
control, we used cleared lysates, which were not subjected to the denaturing step, but directly diluted
(1:3000) in reaction buffer (10 mM Tris pH 7.5, 0.02 M urea) (blue lines). The identity of the fused
effectors is indicated within each of the graphs. The shown results are of a typical experiment out of two
independent experiments. In all cases the monitoring system was too slow to record the refolding process,
which is rapid (measured in seconds), and the activity of the refolded proteins was similar to that of the
native proteins.

Supp. 10. Calculating the CCF2 product concentration in the infected HeLa cells
Infection was carried out in a plate reader and the data was collected with Magelan5
software and processed using the MATLAB computing platform. Background
fluorescence was subtracted from the whole data set. This includes background at 465-
nm (the average of the three first time points in which a product has not yet been created)
and background at 535 nm (fluorescence of uninfected wells). Next, a smoothing process
was used (moving average in a 5 time point window). Finally, fluorescence at 465 nm
was divided by the initial reading (before infection) of fluorescence at 535 nm. This latter
step normalized the data for well-to-well variations in HeLa cell number and/or CCF2
loading efficiency. This last step was omitted when calculating [E] in molar units (see

Supp. 11. Calculating the number of Effector-BlaM molecules in single HeLa cells:
The BlaM β-lactamase (TEM-1) follows simple Michaelis-Menten enzymatic kinetics
(Zlokarnik, 2000) and its Kcat and Km to CCF2 were determined (29 s                          and 23 µM,
respectively; (Zlokarnik, 2000)). According to the Michaelis-Menten model:

         V * ([ s ] + Km)
[E] =
           [ s ] * Kcat
The rate of product formation, (V), values were extracted from the data by subtracting
each [P](t) with its predecessor, [P](t-1), divided by 150 seconds, which is the gap between
each two measurements. The substrate concentration [s] was taken from our
measurements. The [s] fluorescent units were converted to molar units using a calibration
curve of known substrate (CCF2/FA) concentrations (Figure S3 C). To assess the
effectors' concentration in the HeLa cells, we estimated the combined volume of all the
HeLa cells in a well. To determine the average volume of HeLa cells, we assumed that
the volume of flat cells, which are attached to the substrate, would be similar to that of
trypsinized cells, which assume a spherical shape. Attached HeLa cells were trypsinized,
became spherical, and their average diameter was determined by microscopy. We found
that the HeLa cell diameter is 22.65 µmeter (n=34, SD= +/- 5.85) and thus the average
volume of HeLa cells is 6.086*10-6 µL. The number of HeLa cells was determined by
counting in a cell-counting chamber. To arrive at [E] in molar units, we omitted the
normalization step when calculating [P] (see section 7 above). Knowing the molarity, cell
volume, and Avogadro’s number (6.022*1023 molecules/mole), we calculated the number
of effector molecules per cell. Unfortunately, upon maximal loading of the HeLa cells
with CCF2, our starting CCF2 concentration, ([S]), allowed an initial velocity (Vi) of the
enzymatic reaction that is lower than the Vmax value (using fluorescence to CCF2-
concentration calibration curve, Supplemental Figure S3C, the substrate concentration
within loaded HeLa cells was determined to be ~100 µM and at this substrate
concentration the reaction velocity is estimated to be only about 70-85% of the Vmax
value, Figure S7 and data not shown). Given all these considerations, the calculated
effectors' concentrations (indicated in brackets within Figure 2A) are underestimates.
Figure S7. The CCF2 concentration in the cells limits the velocity of the enzymatic reaction.
Strains containing tir-blaM (upper panels) or espG-blaM (lower panels) were used for a translocation assay,
as described above. At 40 min post infection, the infection medium was replaced with fresh medium
supplemented with 1 µM CCF2/AM (blue line), or it remained untreated (green line). As a negative
control, cells were infected with wild-type EPEC that did not carry an effector-blaM gene and at 40 min
post infection, the infection medium was replaced with new medium containing 1 µM CCF2/AM (red line).
For both effectors, adding fresh CCF2/AM at 40 min post infection increased the reaction velocity
(compare the blue with the green lines). Exchanging the infection medium without adding fresh CCF2/AM
had no effect on the reaction velocity (data not shown). These results indicate that the reaction velocity is
limited by the CCF2 concentration in the infected HeLa cells.
Figure S8. The translocation hierarchy into Caco2 cells is similar to that found for HeLa cells.
(A) Translocation into HeLa cells was determined as described in Figure 2A, except that the infection
buffer did not contain BLIP and CCF2/AM. The results show that the translocation hierarchy was
maintained even in the absence of BLIP and CCF2/AM in the infection medium. A typical experiment is
(B) An experiment identical to that shown in (A) was carried out except that Caco2 cells were used instead
of HeLa cells. The results show that translocation hierarchy into Caco2 and HeLa cells is very similar, but
not identical. In Caco2 the clustering of the effectors into two groups was more pronounced. These include
the early translocated effectors group (Tir, EspF, and EspZ) and the late translocated effectors group (Map,
EspG, and EspH). A typical experiment out of three is shown.

Supp. 12. The stability of the translocated effectors.
We predicted that if the effectors are targeted by the proteasome system, then treatment
with the proteasome inhibitor MG132 should increase the effectors' concentration in the
cells and thus the respective translocation signal. In contrast to this prediction, we
demonstrated that MG132 does not affect the translocation rate of any of the effectors
(Figure S9A). To verify that MG132 was active, we demonstrated its capacity to inhibit
TNFα-induced IκB degradation (Figure S9B). To corroborate these findings, we used an
immunoblot analysis of translocated Tir and Tir-BlaM to show that both are equally
translocated and equally stable in the infected HeLa cells (Figure S9C). Finally, we
directly tested the stability of the six BlaM fusions in the host cells. Bacteria were
allowed to inject the BlaM fusions into HeLa cells. Next, chloramphenicol was added to
inhibit translation and thus eventually to block further effector translocation. We then
compared the intracellular BlaM activities at 10 and 80 min post chloramphenicol
addition and found them to be similar (Figure S9D). These results indicate that within the
experiment time frame, degradation of the Effectors-BlaM is not significant and thus
should not affect the translocation assay.
Figure S9. The stability of translocated effectors.
    A. Inhibition of the proteasome does not affect effector-BlaM translocation kinetics. The
         translocation assay was conducted as described in figure 2A, but the HeLa cells were treated or
         not treated with the proteasome inhibitor (MG132, 20 µM). MG132 was added to the host cell
         medium 90 minutes prior to infection and was present throughout the experiment. A typical
         experiment is shown. The fused effectors are indicated within each graph. A wild-type EPEC
         strain that does not carry an effector-blaM fusion was used as a negative control (red line). Note
         that blue and green overlap for all effectors except espZ.
    B. MG132 inhibits TNFα-induced IκB degradation. To verify MG132 activity, we applied it to HeLa
         cells (90 minutes, 20µM) before treating it with TNFα (200U/ml, 30 minutes). Cellular proteins
        were extracted and analyzed by western blotting using anti-IκB antibodies (cell signaling
   C. Tir and Tir-BlaM are equally stable in the host cell. Overnight cultures of EPEC ∆tir mutant
        (strain CX2167) and an EPEC strain expressing both Tir and Tir-BlaM (strain CX2135) were
        diluted 1:50 in DMEM and immediately used to infect HeLa cells. After 3.5 hours of infection, the
        media in the infection plates was replaced with DMEM supplemented with 100 µg/ml of
        chloramphenicol to block translation in the bacteria and thus to eventually block further
        translocation. HeLa cell proteins were extracted at 0, 30, 60, and 90 minutes post chloramphenicol
        addition and used for western blotting analysis with anti-Tir antibody. kDa size is indicated by the
        numbers on the right. The major Tir and Tir-BlaM forms are indicated by arrows. As a protein
        concentration control, we used non-specific bands that reacted with anti-Tir, denoted by an
        arrowhead. As an additional control, we used an extract generated from EPEC ∆tir mutant.
   D. HeLa cells grown in 96 well plates were washed 3 times with 200 µl of warm CDMEM and
        infected with 200 µl of the pre-activated bacteria. Pre-activated EPEC were generated by dilution
        (1:100) of over night cultures containing the different blaM fusions, in CDMEM supplemented
        with 2.5 mM probenecid followed by 2:45 h growth at 37°C. After 40 minutes of infection, the
        supernatants were removed (with the unattached bacteria) and replaced with CDMEM
        supplemented with 100 µg/ml chloramphnicol to block translation and thus to eventually block
        further translocation. Ten minutes after adding chloramphnicol, a set of wells was washed and
        supplemented with media containing CCF2-AM loading buffer, 1 µM CCF2-AM, 100 µg/ml
        chloramphenicol and probenecid. The plate was immediately placed in the pre-warmed reader for
        measurements (as in the real time assays). This procedure was repeated with a second set of wells
        at 80 minutes after adding chloramphnicol. Assays were conducted in triplicates and average
        slopes (representing the BlaM activity) were calculated. The ratios between the slops at 10
        minutes (representing 100% activity), and at 70 minutes post chloramphnicol addition are shown
        (T=Tir-BlaM, Z=EspZ-BlaM, F=EspF-BlaM, H=EspH-BlaM, M=Map-BlaM, G=EspG-BlaM).
        Standard deviation was less than 10% in all cases.

Supp. 13. Translocation of all the LEE-encoded effectors is subjected to
We examined whether translocation of all six effectors is subjected to autoinhibition. Our
results show that all six effectors are subjected to translocation autoinhibition (Figure
Figure S10. Translocation of all six effectors is subjected to translocation autoinhibition.
Experiments as in Figure 6A and B were repeated using all Effector-BlaM-expressing strains as the second
infection wave (wild-type EPEC or a mutant EPEC escV::Tn5kan were used for the first infection wave).
The caption on each graph notes the strain used for the second wave of infection. As a positive control, we
used cells subjected only to the second infection wave. As a negative control, cells were infected with wild-
type EPEC that did not carry an effector-blaM gene. The results shown are for an MOI of 1:2500. Results
in an MOI of 1:625 were similar (data not shown).

Supp. 14. Direct correlation between attachment and translocation efficiencies
The BlaM translocation assay used in this study likely reflects both the rate of bacterial
attachment to the host cells as well as the rate of effector translocation. Bundle forming
pili (BFP) promote strongly the attachment of EPEC to host cells. To test whether BFP
production affects translocation efficiency we constructed the six effector-blaM fusions in
the background of a bfpA::TnphoA EPEC mutant, which cannot assemble functional BFP
(Stone et al., 1996). As expected, bfpA inactivation was associated with strongly reduced
EPEC attachment to host cells (Figure S11C) and translocation of all six BlaM fusions
was attenuated (Figure S11A). The order of effector translocation in the bfpA mutant
remained similar to that seen in wild type EPEC (Figure S11B), except EspH translocated
very inefficiently such that it became the least efficiently translocated effector. These
results confirm that BFP promotes attachment and enhances strongly translocation
efficiency and demonstrates that BFP have little effect on translocation order.
Figure S11. BFP enhance translocation efficiency
A. The translocation efficiencies of wild type EPEC (blue line) and a bfpA::TnphoA mutant (green line)
were compared. The different strains carrying effector-blaM fusions were subjected to real time
translocation analysis; typical experiments are shown. The tested effector is indicated within each graph. A
wild type strain that does not carry an effector-blaM fusion was used as the negative control (red line).
B. The translocation order by the EPEC bfpA mutant.
C. BFP promotes attachment to host cells. Wild-type EPEC or an bfpA::TnphoA mutant, both carrying a
GFP expressing plasmid, were grown under LEE activating conditions and used to infect HeLa cells grown
on a 96 well plate. At the shown time points the medium was removed, the well was washed 5 times with
buffer and fluorescence emission at 535 nm was quantified using a plate reader. Each data point is the
average of three wells (the bars represent the standard error of the mean). Shown is a representative
experiment out of two. Similar results were obtained by using microscopy to count the number of bacteria
attached to HeLa cell (data not shown). Both bacterial strains grow similarly (data not shown).

Table S1: primers used in this study.
Name Used For                               Sequence 5’ to 3’
Primers used for the creation of deletion mutants.
103    cesT Forward                         ATGTCATCAAGATCTGAACTTTTATTAGATAGGTTTGC
104    cesT Reverse                         TTATCTTCCGGCGTAATAATGTTTATTATCGCTTGAGC
442    cesF LFH-PCR (P4)                    CACTGCAGCTACAGCCGAGTATCCTGC
Primers used for cloning.
       espZ-blaM) Forward
       espZ-blaM) Reverse
       blaM) Forward
       blaM) Reverse
       blaM) Forward
       blaM) Reverse
       blaM) Forward
       blaM) Reverse
       map-blaM) Forward
       map-blaM) Reverse
       blaM) Forward
       blaM) Reverse
Primers used for sequence analysis
226    Verification of BlaM fusion GATAATACCGCGCCACATAG
397    Verification of BlaM fusion CGGATAACAATTTCACACAG
Enzymatic restriction sites are underlined.
Table S2: strains used in this study
Strain name (our Description                                                Source
stock #)
DH5α (#1293)          E. coli, F-, φ80dlacZ∆M15, ∆(lacZYA-argF)U169,
                      deoR, recA1, endA1, hsdR17(rk-, mk+), phoA,
                      supE44, λ-, thi-1, gyrA96, relA1
SY327, λpir (#50) E. coli, araD, ∆(lac-pro)XIII, Rifr,nalA,                 MS.
                      argE(Am), recA56, λpir                                Donnenberg
MC1061,        λpir E. coli, ∆(araA-leu)7697, araD139, ∆(codB-              KY. Leung
(#2312)               lac)3=∆lac74, galK16, galE15, mcrA0, relA1,
                      rpsL150, spoT1, mcrB1, hsdR2, λpir, λ-, F-
SM10,          λpir E. coli, thi-1, th, leu, tonA, lacY, supE, recA::RP4-   KY. Leung
(#2313)               2-Tc::Mu, λpir
E2348/69 (#1)         EPEC Wild-Type (strain O127:H6) clinical              J. kaper
                      isolation / isolated from au outbreak in Taunton,
SN9 (#1961)           E2348/69 escV::miniTn5Kn                           (Nadler et
                                                                         al., 2006)
ME2018 (#2018)        E2348/69 ∆cesT::Kn                                 (Li et al.,
ME2849 (#2849)        E2348/69 ∆cesF::Kn                                 This study
31-6-1(1) (#41)       E2348/69 bfpA::TnphoA                              MS.
ME3083 (#3083)        E2348/69 /pSA11                                    This study
ME3084 (#3084)        E2348/69 bfpA::TnphoA /pSA11                       This study
CX2135 (#2135)        E2348/69 PLEE5-tir-blaM (:pCX442)                  This study
CX2137 (#2137)        E2348/69 Pmap-map-blaM (:pCX444)                   This study
CX2138 (#2138)        E2348/69 espH-blaM (:pCX445)                       This study
CX2139 (#2139)        E2348/69 espF-blaM (:pCX446)                       This study
CX2140 (#2140)        E2348/69 espG-blaM (:pCX447)                       This study
ME2383 (#2383)        E2348/69 PLEE2-espZ-blaM (:pME22)                  This study
CX2141 (#2141)        E2348/69 ∆cesT::Kn, PLEE5-tir-blaM (:pCX442)       This study
CX2143 (#2143)        E2348/69 ∆cesT::Kn, Pmap-map-blaM (:pCX444) This study
CX2144 (#2144)        E2348/69 ∆cesT::Kn, espH-blaM (:pCX445)            This study
CX2145 (#2145)        E2348/69 ∆cesT::Kn, espF-blaM (:pCX446)            This study
CX2146 (#2146)        E2348/69 ∆cesT::Kn, espG-blaM (:pCX447)            This study
ME2541 (#2541)        E2348/69 ∆cesT::Kn, PLEE2-espZ-blaM (:pME22)       This study
ME2853 (#2853)        E2348/69 ∆cesF::Kn, PLEE5-tir-blaM (:pCX442)       This study
ME2854 (#2854)        E2348/69 ∆cesF::Kn, Pmap-map-blaM (:pCX444) This study
ME2855 (#2855)        E2348/69 ∆cesF::Kn, espH-blaM (:pCX445)            This study
ME2856 (#2856)        E2348/69 ∆cesF::Kn, espF-blaM (:pCX446)            This study
ME2857 (#2857)        E2348/69 ∆cesF::Kn, espG-blaM (:pCX447)            This study
ME2858 (#2858)        E2348/69 ∆cesF::Kn, PLEE2-espZ-blaM (:pME22)       This study
ME2545 (#2545)        E2348/69       bfpA::TnphoA,        PLEE5-tir-blaM This study
ME2546 (#2546)    E2348/69     bfpA::TnphoA,     Pmap-map-blaM     This study
ME2547 (#2547)    E2348/69 bfpA::TnphoA, espH-blaM (:pCX445)       This study
ME2548 (#2548)    E2348/69 bfpA::TnphoA, espF-blaM (:pCX446)       This study
ME2549 (#2549)    E2348/69 bfpA::TnphoA, espG-blaM (:pCX447)       This study
ME2550 (#2550)    E2348/69    bfpA::TnphoA,      PLEE2-espZ-blaM   This study
LM2511 (#2511)    E2348/69 /pCX-espB                               This study
ME2761 (#2761)    E2348/69 /pCX392                                 This study
ME2889 (#2889)    E2348/69 /pCX394                                 This study
KB2847 (#2847)    E2348/69 /pCX395                                 This study
ME2797 (#2797)    E2348/69 /pCX396                                 This study
ME2890 (#2890)    E2348/69 /pCX397                                 This study
ME2891 (#2891)    E2348/69 /pME21                                  This study
ME1925 (#1925)    E2348/69 /pME19                                  This study

Plasmids integrated in the chromosome are indicated in brackets as (:pXXXX)
Table S3: plasmids used in this study
Plasmid      Description                                      Source
pCX340                                                        (Charpentier and
             Vector for formation of fusions with the blaM reporter.
                                                              Oswald, 2004)
pCX341     pCX340 with ori from pBR322, lower copy number.    This study
pGP704     Suicide, conjugative plasmid                       (Miller        and
pCX391     Derivate of pGP704. The blaM gene was replaced by This study
           tetRA. Contains the T1 terminator of rrnB (from
           pQE30) between tetR and the MCS.
pCX392     pCX341 containing PLEE5-tir-blaM.                  This study
pCX394     pCX341 containing Pmap-map-blaM.                   This study
pCX395     pCX341 containing espH-blaM.                       This study
pCX396     pCX341 containing espF-blaM.                       This study
pCX397     pCX341 containing espG-blaM.                       This study
pME21      pCX341 containing PLEE2-espZ-blaM.                 This study
pCX442     pCX391 containing PLEE5-tir-blaM.                  This study
pCX444     pCX391 containing Pmap-map-blaM.                   This study
pCX445     pCX391 containing espH-blaM                        This study
pCX446     pCX391 containing espF-blaM                        This study
pCX447     pCX391 containing espG-blaM                        This study
pME22      pCX391 containing PLEE2-espZ-blaM.                 This study
pQE30lacIq A derivative of Qiagen’s pQE30 6His fusion cloning This study
           plasmid containing lacIq.
pME19      cesT cloned in pQE30lacIq.                         This study
pKD46      Contains the λ red genes.                          (Datsenko and
                                                              Wanner, 2000)
pKD4       Template for the kanamycin resistance cassette.    (Datsenko and
                                                              Wanner, 2000)
pSA11      Expresing gfp-mut3 via regulated tac promoter      (Schlosser-
                                                              Silverman et al.,
pCXespB    pCX341 containing espB-blaM.                       KY. Leung
pEspH-Flag Encoding espH-FLAG from E2348/69                   (Tu et al., 2003)

All the genes cloned in this study were amplified from genomic DNA of EPEC E2348/69

Akeda, Y., and Galan, J. E. (2005). Chaperone release and unfolding of substrates in type
III secretion. Nature 437, 911-915.

Bustamante, V. H., Santana, F. J., Calva, E., and Puente, J. L. (2001). Transcriptional
regulation of type III secretion genes in enteropathogenic Escherichia coli: Ler
antagonizes H-NS-dependent repression. Mol Microbiol 39, 664-678.

Charpentier, X., and Oswald, E. (2004). Identification of the secretion and translocation
domain of the enteropathogenic and enterohemorrhagic Escherichia coli effector Cif,
using TEM-1 beta-lactamase as a new fluorescence-based reporter. J Bacteriol 186, 5486-

Datsenko, K. A., and Wanner, B. L. (2000). One-step inactivation of chromosomal genes
in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 97, 6640-6645.

Haack, K. R., Robinson, C. L., Miller, K. J., Fowlkes, J. W., and Mellies, J. L. (2003).
Interaction of Ler at the LEE5 (tir) operon of enteropathogenic Escherichia coli. Infect
Immun 71, 384-392.

Li, M., Rosenshine, I., Yu, H. B., Nadler, C., Mills, E., Hew, C. L., and Leung, K. Y.
(2006). Identification and characterization of NleI, a new non-LEE-encoded effector of
enteropathogenic Escherichia coli (EPEC). Microbes Infect 8, 2890-2898.

Miller, V. L., and Mekalanos, J. J. (1988). A novel suicide vector and its use in
construction of insertion mutations: osmoregulation of outer membrane proteins and
virulence determinants in Vibrio cholerae requires toxR. J Bacteriol 170, 2575-2583.

Nadler, C., Shifrin, Y., Nov, S., Kobi, S., and Rosenshine, I. (2006). Characterization of
enteropathogenic Escherichia coli mutants that fail to disrupt host cell spreading and
attachment to substratum. Infect Immun 74, 839-849.

Sanchez-SanMartin, C., Bustamante, V. H., Calva, E., and Puente, J. L. (2001).
Transcriptional regulation of the orf19 gene and the tir-cesT-eae operon of
enteropathogenic Escherichia coli. J Bacteriol 183, 2823-2833.

Schlosser-Silverman, E., Elgrably-Weiss, M., Rosenshine, I., Kohen, R., and Altuvia, S.
(2000). Characterization of Escherichia coli DNA lesions generated within J774
macrophages. J Bacteriol 182, 5225-5230.

Sperandio, V., Mellies, J. L., Delahay, R. M., Frankel, G., Crawford, J. A., Nguyen, W.,
and Kaper, J. B. (2000). Activation of enteropathogenic Escherichia coli (EPEC) LEE2
and LEE3 operons by Ler. Mol Microbiol 38, 781-793.
Stone, K.D., Zhang, H.Z., Carlson, L.K., and Donnenberg, M.S. (1996). A cluster of
fourteen genes from enteropathogenic Escherichia coli is sufficient for the biogenesis of a
type IV pilus. Mol. Microbiol. 20, 325–337.</jrn>

Tu, X., Nisan, I., Yona, C., Hanski, E., and Rosenshine, I. (2003). EspH, a new
cytoskeleton-modulating effector of enterohaemorrhagic and enteropathogenic
Escherichia coli. Mol Microbiol 47, 595-606.

Umanski, T., Rosenshine, I., and Friedberg, D. (2002). Thermoregulated expression of
virulence genes in enteropathogenic Escherichia coli. Microbiology 148, 2735-2744.

Wolff, C., Nisan, I., Hanski, E., Frankel, G., and Rosenshine, I. (1998). Protein
translocation into host epithelial cells by infecting enteropathogenic Escherichia coli.
Mol Microbiol 28, 143-155.

Zlokarnik, G. (2000). Fusions to beta-lactamase as a reporter for gene expression in live
mammalian cells. Methods Enzymol 326, 221-244.