j o ur nal ho me pag e: www.elsev ier .com/ locate / jv i romet
roteolytic enzymes in embryonated chicken eggs sustain theeplication of egg-grown low-pathogenicity avian influenza viruses inells in the absence of exogenous proteases
hmed Kandeil a, Ola Bagatoa, Hassan Zaraketb, Jennifer Debeauchampb, Scott Kraussb,abeh El-Sheshenya, Richard J. Webbyb, Mohamed A. Alia, Ghazi Kayalib,∗
Environmental Research Division, National Research Centre, Dokki 12311, Giza, EgyptDepartment of Infectious Diseases, St. Jude Children’s Research Hospital, Memphis, TN 38105, United States
rticle history:eceived 9 August 2013eceived in revised form 25 February 2014ccepted 27 February 2014vailable online 10 March 2014
a b s t r a c t
Low pathogenic influenza viruses grow readily in embryonated chicken eggs but require the additionof exogenous proteases to grow in MDCK cell culture. In this study, we found that the influenza virusespropagated previously in eggs, can grow for up to two passages in cell culture without the additionof exogenous proteolytic enzymes. These results indicate that the reason for virus propagation in cells
eywords:nfluenzaroteaseell culture
during the first two passages may be due to proteases from egg allantoic fluid carried over from eggculture. The ability of influenza viruses to grow in cells in the absence of trypsin is currently consideredas a hallmark of highly pathogenic influenza viruses. Our data indicate that differentiating between highand low pathogenicity using cell culture only is not appropriate and other indicators such as sequenceanalysis and in vitro pathogenicity index should be performed.
Influenza A virus is a single-stranded negative sense RNA virusf the orthomyxoviridae family. Its genome is made of 8 segmentsncoding at least 11 known proteins (Nelson and Holmes, 2007).nfluenza A virus is subtyped based on its surface glycoproteins; theemagglutinin (HA) and the neuraminidase (NA) proteins. Aquaticirds are considered the main natural reservoir of influenza Airuses, of which there are currently 16 HA subtypes and 9 NA sub-ypes (Salomon and Webster, 2009; Webster et al., 1992). Only twoubtypes, the H3N2 and the H1N1, currently circulate in humans.ore recently, influenza A viral genetic material was isolated from
ats, leading to the designation of additional antigenically distinctubtypes, H17N10; however, the N10 neuraminidase has not beenhown to possess actual neuraminidase activity (Garcia-Sastre,012; Tong et al., 2012).
The HA protein is a class I membrane fusion protein. It is cleavedost-translationally into mature metastable protein made of twoubunits; HA1 and HA2 (Skehel and Wiley, 2000; Wilson and Cox,
1990). The HA1 subunit harbors the receptor binding pocket whichregulates the �(2, 3) versus �(2, 6)-linked sialic acid receptor bind-ing preference. The HA2 subunit harbors a fusion peptide at itsN-terminal that mediates virus–cell membrane fusion. Cleavage ofthe HA protein occurs at a basic amino acid linker between theHA1 and HA2 subunits (Bertram et al., 2010; Klenk and Garten,1994). In low pathogenic influenza viruses (LPIV), this linker con-sists of a single amino acid residue that is recognized by a limitednumber of serine-like proteases that are present in the respira-tory (in mammalian and avian species) and intestinal (in avianspecies) tracts (Klenk and Garten, 1994; Webster and Rott, 1987).In case of highly pathogenic avian influenza viruses (HPAIV), thecleavage site is polybasic and is recognized readily by ubiquitoussubtilisin-like proteases enabling systemic replication of the virus(Stieneke-Grober et al., 1992; Webster and Rott, 1987). In humans,host cells and bacteria in the airway epithelium play a role in HAcleavage (Bottcher-Friebertshauser et al., 2013). In tissue culture,replication of LPIVs but not HPAIVs requires the addition of exoge-nous proteases such as tosyl phenylalanyl chloromethyl ketone(TPCK)-treated trypsin to cleave the HA protein. Replication of LPIVs
and HPAIVs in embryonated chicken eggs is supported by proteasespresent in allantoic fluid (Horimoto and Kawaoka, 1997). Sinceavian LPIVs are grown commonly in eggs, the aim of this studywas to investigate the effect of residual proteases present in the
llantoic fluid on replication of egg-grown influenza viruses repre-enting a broad array of subtypes in cell culture in the absence ofxogenous trypsin.
. Methods
.1. Viruses and cell culture
A panel of egg-grown LPIVs H4–H16 was used in this studyTable 1). The H5N1 and H7N7 viruses were highly pathogeniciruses that were rendered low pathogenic by reverse geneticsWebby et al., 2004). This was accomplished by deleting the polyba-ic cleavage site from the HA gene and creating reassortant virusesontaining the altered HA and NA of the wildtype virus and the
internal genes from the A/Puerto Rico/38 virus. Viruses wererown initially in the allantoic fluids of 10-day-old embryonatedPF chicken eggs following standard procedures (WHO, 2002). Eachgg-propagated virus was inoculated into 6-well tissue culturelates (100 �l/well) (Greiner, Kremsmunster, Austria) containing0–90% confluent MDCK cells with and without the addition ofPCK-treated trypsin in the infection media and grown for 72 hWHO, 2002). Confluent cells were washed twice with PBS beforehe addition of the virus inoculum. After an 1 h incubation at 37 ◦C,he inoculum was removed and the cells were washed once withBS and then incubated for 72 h. Following that, the viruses wereropagated twice more in MDCK cells by adding 100 �l of therevious passage per well. We recorded cytopathic effect (CPE)ubjectively, and conducted a hemagglutination assay (HA) usinghicken RBCs, HA titers were recorded as the reciprocal of the virusilution that caused agglutination to RBCs (WHO, 2002).
.2. Plaque assay
In order to quantify virus replication, plaque assays were usedor virus titration. Six-well tissue culture plates were seeded with
DCK cells (105 cells/well). At 90–100% confluence (one day post-eeding), the cells were washed twice with PBS. Viruses wereiluted 10-fold in DMEM (Lonza, Basel, Switzerland) and 100 �l ofndiluted virus and each dilution were mixed with 400 �l DMEMnd inoculated into MDCK cells. The plates were incubated at7 ◦C for 1 h. The wells were then aspirated to remove residual
noculum. Each well was then immediately covered with 2 ml ofMEM overlay medium containing 1% agarose type 1 (Lonza), 1%ntibiotic-antimycotic mixture (Lonza), and 1ug/ml TPCK-treatedrypsin (Worthington, Lakewood, NJ, USA). Plates were then incu-ated at 37 ◦C with 5% CO2 for 2 days. The formation, number, and
rowth rate of the plaques were microscopically observed daily.nce clear plaques could be visualized, 1 ml of 10% formaldehydeas added to each well for 1 h for cell fixation and virus inactiva-
ion. The formaldehyde was then discarded and the plates rinsed
al Methods 202 (2014) 28–33 29
with water and dried. For visualization of the plaques, 1 ml of thestaining solution, consisting of 1% crystal violet and 20% methanolin distilled water, was added to each well and incubated at roomtemperature for 5 min, the dye was then discarded and the wellswere rinsed with water and dried. Viral plaques were then countedand virus titer was calculated using the Reed and Muench method(Reed and Muench, 1938).
2.3. Western blotting
In order to qualify cleavage of HA in the presence and absenceof trypsin, 2 consecutive passages of avian influenza H9N2 viruswere subjected to western blotting. Propagated viruses were ana-lyzed by SDS-PAGE as described previously (Ruppel et al., 1985);the only modification was that 1% BSA in PBS–0.3% Tween20 wasused to block the protein-free binding sites on the nitrocellulosemembrane. Immunorecognition was performed on cut membranestrips carrying chicken anti-H9N2 sera (dilution 1:100). Immunedetection was carried out with peroxidase-conjugated goat anti-chicken IgG (KPL, Gaithersburg, MD) diluted 1:2000 in PBS–0.3%Tween20.
2.4. Zymogram
The presence of serine proteases in egg uninfected allantoic fluidand cultured viruses was determined by zymography (Heussen andDowdle, 1980). In this assay, serine proteases will degrade gelatinat their specific molecular weights and the gel at that site willnot absorb the dye, thus the protease activity will be visualizedas an unstained band on a stained background. Gelatin was usedas a protease substrate and was added to the separating gel beforepolymerization of acrylamide. A volume of 10 �l of infected or unin-fected allantoic fluid was added to 10 �l of sample buffer containing0.1 M Tris–HCl, 4% SDS, 10% glycerol, and 50 mg bromophenol blueat pH 6.8. The sample was then loaded to wells of an acrylamidestacking gel. A volume of 5 �l of 10–200 kDa protein marker (Lonza)was applied to one of the gel wells and allowed to run parallel to thesamples for detection of the molecular weight(s) corresponding tothe proteases. Electrophoresis was conducted at 60 V for 2 h. Afterelectrophoresis, the marker was cut and Coomassie-stained thendestained while the rest of the gel was soaked in renaturing (2.5%triton-X-100) solution with gentle shaking for 30 min at room tem-perature to remove SDS from the gel. The gel was then incubatedovernight at 37 ◦C with gentle shaking in substrate buffer (30 mMTris–HCl, 60 mM NaCl, CaCl2 (Serva, Heidelberg, Germany)) at pH 5,7 and 10 then stained with 0.5% Coomassie blue dye (in 10% aceticacid, 5% methanol (Sigma, St. Louis, MO, USA)) and de-stained using60% methanol. The molecular weight of the proteases present wasdetermined by comparison with the molecular weight marker.
3. Results
At passage 1, all viruses produced an HA titer at 48 h with orwithout using trypsin. At 72 h (Fig. 1), titers increased to 32 or 64for the viruses propagated without trypsin while the viruses thatgrew with trypsin had HA titers between 8 and 64. At passage 2,viruses propagated with trypsin continued to grow. However, only6 viruses, mostly having higher titers in passage 1, were able togrow without the presence of trypsin and had low titers (HA 2–16).This trend continued into passage 3 where all viruses except H16grew with trypsin while none propagated in its absence. To confirm
these results, 2 consecutive passages of H4, H9, and H10 viruseswere repeated in the presence or absence of trypsin. HA titers mea-sured were within 1-fold difference of the measurements obtainedearlier.
30 A. Kandeil et al. / Journal of Virological Methods 202 (2014) 28–33
0
16
32
48
64
80
96
112
128
144
P1 +T 72 P2+T 72 P3+T 72 P1 -T 72 P2-T 72 P3-T 72
HA
Tite
r
Passage
H4
H5
H6
H7
H8
H9
H10
H11
H12
H13
H14
H15
H16
F MDCp of tryp
t(fvCpCI
Ff+
ig. 1. Chicken RBCs hemagglutination assay titers of influenza viruses grown inropagated in cell culture for 3 consecutive passages in the presence or absence
resence of trypsin, 72 indicates 72 h post inoculation.
We also observed CPE at 48 h and 72 h post-infection (Fig. 2). Inhe presence of trypsin, all viruses produced high CPE in passage 1>80%). In passage 2, high CPE was observed for all viruses exceptor H6, H13, and H16 that had lower CPE (30–50%). The same obser-ation was made in passage 3 and H6, H13, and H16 had even lower
PE (0–20%). In the absence of trypsin, all viruses had high CPE inassage 1. In passage 2, viruses H4, H5, H7, H8, H9, and H11 had aPE on cells (40–100%), while the other viruses did not show CPE.
n passage 3, none of the viruses had CPE.
0
20
40
60
80
100
120
P1 +T 72 P2+T 72 P3+T 72
% C
PE
Passag
ig. 2. Cytopathic effect (%) exerted by influenza viruses on MDCK cells at 72 h post-inocor 3 consecutive passages in the presence or absence of trypsin. Cytopathic effect was quT indicates presence of trypsin, 72 indicates 72 h post inoculation.
K cells at 72 h post-inoculation. 13 low pathogenic avian influenza viruses werepsin. P# indicates passage number, −T indicates absence of trypsin, +T indicates
The results of the plaque assays are shown in Fig. 3. All virusesexcept H13 provided high titers in the plaque assay in all passageswith trypsin. The H13 virus only provided a titer by the third pas-sage. For the other viruses titers remained the same or increasedby the 3rd passage. In the absence of trypsin, all viruses except for
H13 and H16 yielded a titer at passage 1 (4–7 log10 PFU/ml). Titersdropped by the second passage but H13 and H16 did not plaque(3–6 log 10 PFU/ml). Only three viruses (H4, H9, and H10) yieldeda titer by the third passage and these titers were ≤4 log10 PFU/ml.
P1 -T 72 P2-T 72 P3-T 72
e
H4
H5
H6
H7
H8
H9
H10
H11
H12
H13
H14
H15
H16
ulation. 13 low pathogenic avian influenza viruses were propagated in cell culturealitatively recorded. P# indicates passage number, −T indicates absence of trypsin,
A. Kandeil et al. / Journal of Virological Methods 202 (2014) 28–33 31
0
1
2
3
4
5
6
7
8
9
10
P1 +T 72 P2+ T 72 P3+ T 72 P1 -T 72 P2-T 72 P3-T 72
Log1
0 PF
U/m
l
Pas sage
H4
H5
H6
H7
H8
H9
H10
H11
H12
H13
H14
H15
H16
F oculat3 ductet
cTt
Fet
ig. 3. Plaque assay titers of influenza viruses grown on MDCK cells at 72 h post in consecutive passages in the presence or absence of trypsin. Plaque assay was conrypsin, +T indicates presence of trypsin, 72 indicates 72 h post inoculation.
Western blot data (Fig. 4) showed that the H9N2 HA was beingleaved into HA1 and HA2 efficiently in the presence of trypsin.his was also evident in the first MDCK passage in the absence ofrypsin.
ig. 4. Western blot showing HA cleavage of H9N2 at different passages in the pres-nce or absence of trypsin. P# indicates passage number, −T indicates absence ofrypsin, +T indicates presence of trypsin.
ion. 13 low pathogenic avian influenza viruses were propagated in cell culture ford in the presence of trypsin. P# indicates passage number, −T indicates absence of
Zymogram data of uninfected allantoic fluid showed severalegg protease activities at neutral and alkaline pH. Clear proteolyticbands were visualized at 155, 133, 69, 62, 54, 39, and 36 kDa at pH 7and 10. An additional proteolytic band at MW of about 107 kDa wasdetected at pH 7 (Fig. 5). Zymogram of MDCK-grown H4, H5, H6,and H7 viruses was conducted. Bands at molecular weight around130 kDa for all passages of these viruses were visualized, the inten-sity of the band decreased with increased passaging. Trypsin isvisualized at around 24 kDa in passages where trypsin was used.Fig. 6 illustrates this finding for MDCK-grown H5 virus.
In order to verify the potential of egg allantoic fluid as a protease-replacement, the first cell passage (without trypsin) of H4, H5, H10,H14, and H15 were passaged in MDCK cells without trypsin butwith the addition of 10% to 50% allantoic fluid or trypsin in the infec-tion media. All these viruses propagated successfully with trypsinas verified by HA titers between 128 and 256. However, only twoviruses grew in the presence of allantoic fluid (H5 and H10) but had
low titers of 4 (Fig. 6).
Fig. 5. Visualization of various proteolytic enzymes in allantoic fluid of un-infectedembryonated chicken eggs by Zymography. Clear bands in the gel indicate the pres-ence of proteolytic enzymes. M denotes protein marker, pH# indicates pH at whichthe assay was carried out.
32 A. Kandeil et al. / Journal of Virologic
Fig. 6. Visualization of proteolytic enzymes in different passages of MDCK-culturedH5 virus by Zymography. Clear bands in the gel indicate the presence of proteolyticenzymes. Egg proteases appear in all passages at molecular weight around 130 kDawwi
4
eitpsatzHwepsiMme
cHdckcIfiHahifHwa2
dHbnti
hile trypsin appears in passages that included the addition of trypsin at moleculareight 24 kDa. P# indicates passage number, −T indicates absence of trypsin, +T
ndicates presence of trypsin.
. Discussion
These results demonstrated that most LPIVs propagated inmbryonated chicken eggs can grow readily for up to 2 passagesn MDCK cells without the need for the addition of TPCK-treatedrypsin albeit to titers that were lower than those obtained in theresence of trypsin. We hypothesized that endogenous trypsin-likeerine proteases found in the eggs’ allantoic fluids sustain the prop-gation of LPIVs in MDCK cells and once depleted, TPCK-treatedrypsin becomes essential for MDCK culture of these viruses. Byymography, several egg proteases that may be involved in viralA activation were visualized. Furthermore, one of these proteasesas visualized in the MDCK-grown viruses further supporting the
gg-protease involvement in virus propagation on cells. However,ropagating viruses by replacing trypsin with egg protease was notuccessful as only 2 viruses grew but to very low titers. It remainsmportant to verify the involvement of such proteases in virus
DCK cultures and explain the means by which such proteasesay aid HA cleavage, and further research is required if allantoic
nzymes are to be used as a replacement for trypsin.Other researchers were not able to detect protease-mediated
leavage for the H13 and H16 Has (Galloway et al., 2013). In the16 HA, the cleavage site that would create the HA1/HA2 complexisplays an �-helical structure and hides in the negatively chargedavity, in contrast to the flexible loop conformations found in othernown structures, and more importantly, this unusual structure isorrelated with its inefficient cleavage by trypsin (Lu et al., 2012).n our data, H16 only yielded a high titer in the plaque assay in therst passage in the presence of trypsin (Fig. 3.) and had the lowestA titer. This might be explained by Lu et al. findings. The HA titerfter the first cell passage in the absence of trypsin was one-foldigher than that in the presence of trypsin. Since the difference
s minimal, we do not believe there is an underlying mechanismor this finding and, most likely, it is due experimental error. The13 virus had higher HA and plaque assay titer in the third passageith trypsin. This surprising finding may be due to the H13 virus
dapting to cell culture as suggested by others (Keawcharoen et al.,010).
One weakness of this study was that different HA titers forifferent viruses prior to MDCK infection was not adjusted for.owever, since the aim was not to compare viruses to each other
ut rather compare the different passages of the same virus, we doot believe that adjusting titers would have altered the findings. Ifhis had an effect, it will be mostly affecting our subjective CPE read-ng as higher content of viral particles should provide more CPE.
al Methods 202 (2014) 28–33
However, all assays that we conducted led to the same conclu-sion. On another hand, we did not wash the cells extensively afterthe initial 1 hr incubation to allow the inoculum to adsorb on thecells. Extensive washing of the cells post this incubation period mayreduce carried-over allantoic fluid and hence inhibit the propaga-tions of LPIVs in the absence of exogenous proteases.
Current influenza protocols state that only few avian influenzaviruses of the H5 or H7 subtypes are able to grow on MDCK cells inthe absence of exogenous trypsin (Krauss et al., 2012). The inabil-ity of some AI viruses, especially reassortants, to grow in MDCKcells in the absence of trypsin is commonly used as a differentia-tion method to determine their pathogenicity. In light of our results,this cannot be used as a single method to determine pathogenicityin case the viruses were passaged in eggs. Other assays, such aspathogenicity testing in chickens and HA sequence data (APHIS,2011), are needed to determine whether viruses have low or highpathogenicity.
TPCK has been shown to induce mammalian cell death throughvarious mechanisms (Fabian et al., 2009). The addition of TPCK-treated trypsin to MDCK cells has a cytotoxic effect that maydecrease viral titers in cell culture. Thus, investigating alternativesto TPCK-treated trypsin such as serine proteases from chicken eggsis warranted.
Acknowledgements
This work was funded by the National Institute of Allergyand Infectious Diseases, National Institutes of Health, Depart-ment of Health and Human Services, under contract numberHHSN266200700005C, and supported by the American LebaneseSyrian Associated Charities (ALSAC).
References
APHIS, U., 2011. Guidelines for Avian Influenza Viruses.Bertram, S., Glowacka, I., Steffen, I., Kuhl, A., Pohlmann, S., 2010. Novel insights
into proteolytic cleavage of influenza virus hemagglutinin. Reviews in MedicalVirology 20, 298–310.
Bottcher-Friebertshauser, E., Klenk, H.D., Garten, W., 2013. Activation of influenzaviruses by proteases from host cells and bacteria in the human airway epithe-lium. Pathogens and Disease 69, 87–100.
Fabian, Z., O’Brien, P., Pajecka, K., Fearnhead, H.O., 2009. TPCK-induced apoptosis andlabelling of the largest subunit of RNA polymerase II in Jurkat cells. Apoptosis:An International Journal on Programmed Cell Death 14, 1154–1164.
Galloway, S.E., Reed, M.L., Russell, C.J., Steinhauer, D.A., 2013. Influenza HA subtypesdemonstrate divergent phenotypes for cleavage activation and pH of fusion:implications for host range and adaptation. PLoS Pathogens 9, e1003151.
Garcia-Sastre, A., 2012. The neuraminidase of bat influenza viruses is not a neu-raminidase. Proceedings of the National Academy of Sciences of the UnitedStates of America 109, 18635–18636.
Horimoto, T., Kawaoka, Y., 1997. Biologic effects of introducing additional basicamino acid residues into the hemagglutinin cleavage site of a virulent avianinfluenza virus. Virus Research 50, 35–40.
Klenk, H.D., Garten, W., 1994. Host cell proteases controlling virus pathogenicity.Trends in Microbiology 2, 39–43.
Krauss, S., Walker, D., Webster, R.G., 2012. Influenza virus isolation. In: Kawaoka,Y., Neumann, G. (Eds.), Influenza Virus Methods and Protocols. Humana Press,London, UK, pp. 11–24.
Lu, X., Shi, Y., Gao, F., Xiao, H., Wang, M., Qi, J., Gao, G.F., 2012. Insights into avianinfluenza virus pathogenicity: the hemagglutinin precursor HA0 of subtype H16has an alpha-helix structure in its cleavage site with inefficient HA1/HA2 cleav-
age. Journal of Virology 86, 12861–12870.
Nelson, M.I., Holmes, E.C., 2007. The evolution of epidemic influenza. Nature ReviewsGenetics 8, 196–205.
Reed, L.J., Muench, H., 1938. A simple method of estimating fifty percent endpoints.The American Journal of Hygiene 27, 493–497.
uppel, A., Rother, U., Vongerichten, H., Lucius, R., Diesfeld, H.J., 1985. Schistosomamansoni: immunoblot analysis of adult worm proteins. Experimental Parasito-logy 60, 195–206.
alomon, R., Webster, R.G., 2009. The influenza virus enigma. Cell 136, 402–410.kehel, J.J., Wiley, D.C., 2000. Receptor binding and membrane fusion in virus entry:
the influenza hemagglutinin. Annual Review of Biochemistry 69, 531–569.tieneke-Grober, A., Vey, M., Angliker, H., Shaw, E., Thomas, G., Roberts, C., Klenk,
H.D., Garten, W., 1992. Influenza virus hemagglutinin with multibasic cleav-age site is activated by furin, a subtilisin-like endoprotease. EMBO Journal 11,
2407–2414.
ong, S., Li, Y., Rivailler, P., Conrardy, C., Castillo, D.A., Chen, L.M., Recuenco, S., Ellison,J.A., Davis, C.T., York, I.A., Turmelle, A.S., Moran, D., Rogers, S., Shi, M., Tao, Y., Weil,M.R., Tang, K., Rowe, L.A., Sammons, S., Xu, X., Frace, M., Lindblade, K.A., Cox, N.J.,Anderson, L.J., Rupprecht, C.E., Donis, R.O., 2012. A distinct lineage of influenza A
al Methods 202 (2014) 28–33 33
virus from bats. Proceedings of the National Academy of Sciences of the UnitedStates of America 109, 4269–4274.
Webby, R.J., Perez, D.R., Coleman, J.S., Guan, Y., Knight, J.H., Govorkova, E.A.,McClain-Moss, L.R., Peiris, J.S., Rehg, J.E., Tuomanen, E.I., Webster, R.G., 2004.Responsiveness to a pandemic alert: use of reverse genetics for rapid develop-ment of influenza vaccines. Lancet 363, 1099–1103.
