Torben Heise1, Christian Büll2, Daniëlle M Beurskens3, Emiel Rossing1, Marien I de Jonge4, Gosse J Adema2, Thomas J Boltje1, Jeroen D Langereis4. 1. Institute for Molecules and Materials, Radboud University Nijmegen , Heyendaalseweg 135, 6525 AJ Nijmegen, The Netherlands. 2. Department of Radiation Oncology, Radiotherapy & OncoImmunology Laboratory, Radboudumc , Geert Grooteplein Zuid 32, 6525 GA Nijmegen, The Netherlands. 3. Department of Biochemistry, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University , Universiteitssingel 50, 6229 ER Maastricht, The Netherlands. 4. Laboratory of Pediatric Infectious Diseases, Radboud Center for Infectious Diseases, Radboudumc , Geert Grooteplein Zuid 32, 6525 GA Nijmegen, The Netherlands.
Abstract
Metabolic incorporation of azide- or alkyne-modified sialic acids into the cellular glycosylation pathway enables the study of sialoglycan expression, localization, and trafficking via bioorthogonal chemistry. Herein, we report that such modifications of the sialic acid sugar can have a profound influence on their hydrolysis by neuraminidases (sialidase). Azidoacetyl (Az)-modified sialic acids were prone to neuraminidase cleavage, whereas propargyloxycarbonyl (Poc)-modified sialic acids were largely resistant to cleavage. Because the influenza virus infection cycle depends on the hydrolysis of host-cell-surface sialic acids, influenza cell-to-cell transmission was strongly reduced in Poc sialic acid glycoengineered host cells. The use of Poc sialic acids may disturb biological processes involving neuraminidase cleavage but also provides perspective for use in applications in which sialic acid hydrolysis is not desired, such as antibody modification, viral infection, etc.
Metabolic incorporation of azide- or alkyne-modifiedsialic acids into the cellular glycosylation pathway enables the study of sialoglycan expression, localization, and trafficking via bioorthogonal chemistry. Herein, we report that such modifications of the sialic acidsugar can have a profound influence on their hydrolysis by neuraminidases (sialidase). Azidoacetyl (Az)-modifiedsialic acids were prone to neuraminidase cleavage, whereas propargyloxycarbonyl (Poc)-modifiedsialic acids were largely resistant to cleavage. Because the influenza virus infection cycle depends on the hydrolysis of host-cell-surface sialic acids, influenza cell-to-cell transmission was strongly reduced in Poc sialic acid glycoengineered host cells. The use of Poc sialic acids may disturb biological processes involving neuraminidase cleavage but also provides perspective for use in applications in which sialic acid hydrolysis is not desired, such as antibody modification, viral infection, etc.
Sialic acids
(also known as
neuraminic acids) are nine-carbonsugars abundantly expressed at the
termini of cell surface glycoconjugates.[1] Sialylated glycans are recognized by endogenous and exogenous receptors
and play an important role in physiology and pathology, respectively.[2] An important tool for the study of the function
and localization of glycans in living cells and organisms is metabolic
oligosaccharide engineering (MOE).[3] In
this process, monosaccharides carrying an unnatural modification are
supplied to cells and incorporated into glycans at the cell surface.
In a second step, the unnatural modification, also called a chemical
reporter, can be visualized using bioorthogonal reactions with, for
instance, fluorescent molecules.[4] This
powerful technique can be used to study and to manipulate glycans
at the cell surface in various cells and living organisms. Studying
biological processes without disturbing them is a major challenge,
and hence, the chemical reporter carried by the monosaccharides must
preferably be small and biologically inert. This approach is based
on earlier work showing that small modifications on the N-acyl chain of N-acetylmannosamines were well-tolerated
by the glycosylation machinery and ultimately expressed at the cell
surface in the form of modifiedsialic acids.[5] However, these modifications were intended to actively perturb biological
processes involving sialic acids, such as pathogenic infection, immune-cell
dampening, and neurite outgrowth.[6,7] The acyl chain
of sialic acid is also a site of natural modifications, which has
probably arisen to avoid pathogen recognition.[8] Chemical modifications at this site may therefore affect the biological
function of sialic acids. Here, we report that azidoacetyl (Az) sialic
acids are sensitive to bacterial and viral neuraminidases (also called
sialidases), whereas propargyloxycarbonyl (Poc) sialic acids are resistant.
Subsequently, incorporation of a propargyloxycarbonyl (Ac5SiaNPoc) into host cells, but not an azidoacetyl
(Ac5SiaNAz), markedly reduced neuraminidase-dependent influenza cell-to-cell transmission.
Results and Discussion
We have previously shown that sialic acidsmodified at the C-5
position with Ac5SiaNAz or Ac5SiaNPoc are utilized by the cellular sialylation
pathway with high selectivity resulting in their incorporation into
cell surface sialoglycans (Figure ).[9,10] Both sialic acid analogues are
well-tolerated by the cellular sialylation pathway, and these small
modifications to the C-5 position are generally thought to be nonintrusive.
There is evidence, however, that the modification of sialic acids
at the C-5 position affects their recognition and cleavage by bacterial
neuraminidases.[11] Therefore, we assessed
whether cell surface Az and Poc sialic acids can be cleaved by bacterial
neuraminidase. Human monocytic THP-1 cells were cultured for 3 days
with Ac5SiaNAz or Ac5SiaNPoc and reacted to biotin–alkyne or biotin–azide,
respectively, using copper-catalyzed alkyne azide cycloaddition (CuAAc).[12−14] The biotin-tagged surface sialoglycans were visualized with fluorescent
streptavidin. Confocal images show the dense expression of Az and
Poc sialic acids at the cell surface of THP-1 cells (Figure a). Quantification of the fluorescence
with flow cytometry showed that both sialic acid analogues were incorporated
into surface glycans with high efficiency. Ac5SiaNAz (MFI 7979) showed slightly lower incorporation compared
to Ac5SiaNPoc (MFI 9345) (Figure b).
Figure 1
Schematic presentation
of metabolic labeling with Ac5SiaNAz or
Ac5SiaNPoc
and the resistance of Ac5SiaNPoc against
influenza neuraminidase.
Figure 2
Metabolic labeling of THP-1 cells. (a) The cells were incubated
for 3 days with PBS, 100 μM Ac5SiaNAz, or 100 μM Ac5SiaNPoc and reacted
to biotin–alkyne or biotin–azide, respectively, using
CuAAC. Biotin-containing sialoglycans were stained with PE–streptavidin
and visualized by confocal microscopy. (b) Fluorescence was quantified
by flow cytometry, and the mean fluorescence intensity ± SEM
is shown (n = 3).
Schematic presentation
of metabolic labeling with Ac5SiaNAz or
Ac5SiaNPoc
and the resistance of Ac5SiaNPoc against
influenzaneuraminidase.Metabolic labeling of THP-1 cells. (a) The cells were incubated
for 3 days with PBS, 100 μM Ac5SiaNAz, or 100 μM Ac5SiaNPoc and reacted
to biotin–alkyne or biotin–azide, respectively, using
CuAAC. Biotin-containing sialoglycans were stained with PE–streptavidin
and visualized by confocal microscopy. (b) Fluorescence was quantified
by flow cytometry, and the mean fluorescence intensity ± SEM
is shown (n = 3).The difference in labeling intensity between Ac5SiaNAz and Ac5SiaNPoc
is linked
to their different incorporation efficiency and not the labeling step
because the CuAAc reaction is faster for Ac5SiaNAz compared to Ac5SiaNPoc.[16]a Next, it was determined
whether surface SiaNAz and SiaNPoc
containing glycans can be cleaved by bacterial neuramindase (Figure ). THP-1 cells cultured
for 3 days with Ac5SiaNAz and Ac5SiaNPoc were treated with Clostridium perfringensneuraminidase and then reacted with biotin–alkyne and biotin–azide,
respectively, followed by staining with streptavidin–phycoerythrin
(PE).[17] Flow cytometry analysis revealed
that SiaNAz was cleaved from the cells (>64% removal), whereas
SiaNPoc
could not be removed from the cell surface (<8% removal) (Figure a). To evaluate the
cleavage selectivity with respect to the linkage type (α-2,3
or α-2,6), the cells were analyzed with the lectins Maackia amurensis lectin (MALII, binding to α-2,3-sialoglycans), Sambucus nigra lectin (SNA-I, binding to α-2,6-sialoglycans),
and peanut agglutinin lectin (PNA, binding to terminal β-galactose).
Figure 3
Enzymatic
removal of Az and Poc sialic acids from the cell surface
of THP-1 cells. Cells incubated for 3 days with PBS, 100 μM
Ac5SiaNAz, or 100 μM Ac5SiaNPoc were treated for 1 h with 250mU/mL Clostridium perfringens neuraminidase. Az and Poc sialoglycans
were reacted to fluorescent biotin using CuAAC (a), α2,3-sialoglycans
were detected with MALII lectin (b), α2,6-sialoglycans were
detected with SNA-I lectin (c), and terminal β-galactose was
detected with PNA lectin (d). Bar diagrams show mean fluorescence
intensity or mean lectin binding normalized to control ± SEM
of three independent experiments. MALII: Maackia amurensis lectin; PE: phycoerythrin, PNA: Peanut agglutinin lectin; SEM: standard
error of the mean; SNA-I: Sambucus nigra lectin.
Enzymatic
removal of Az and Poc sialic acids from the cell surface
of THP-1 cells. Cells incubated for 3 days with PBS, 100 μM
Ac5SiaNAz, or 100 μM Ac5SiaNPoc were treated for 1 h with 250mU/mL Clostridium perfringensneuraminidase. Az and Poc sialoglycans
were reacted to fluorescent biotin using CuAAC (a), α2,3-sialoglycans
were detected with MALII lectin (b), α2,6-sialoglycans were
detected with SNA-I lectin (c), and terminal β-galactose was
detected with PNA lectin (d). Bar diagrams show mean fluorescence
intensity or mean lectin binding normalized to control ± SEM
of three independent experiments. MALII: Maackia amurensis lectin; PE: phycoerythrin, PNA: Peanut agglutinin lectin; SEM: standard
error of the mean; SNA-I: Sambucus nigra lectin.Cells incubated with Ac5SiaNAz showed
also a strong reduction in sialic acid expression following treatment
with neuraminidase (MALII, ↓93%; SNA-I, ↓72%). In contrast,
the sialylation of THP-1 cells cultured with Ac5SiaNPoc was barely altered following neuraminidase treatment
(MALII, ↓18%; SNA-I, ↓13%). The low sensitivity of Poc
sialic acids toward neuraminidase activity was also confirmed by the
minor change in PNA binding. Similar findings were obtained when THP-1
cells were treated with neuraminidase derived from Vibrio
cholerae or Arthrobacter ureafaciens (Figure S1). Together, these data show that sialic
acids modified with an Az group at the C-5 position are substrates
for neuraminidases, whereas Poc-modifiedsialic acids are not. Whether
this effect is due to inhibited recognition or hydrolysis of sialic
acids by neuraminidases remains
to be investigated.Several pathogens make use of neuraminidases
to infect host cells
or release host cell sialic acids for their own metabolism.[18,19] In particular, influenza virus makes use of the host cell sialic
acids to enter and leave the cell. Influenza virus has two major glycoproteins
on the outside of the viral particle, hemagglutinin (HA) and neuraminidase
(NA).[20] Hemagglutinin is a lectin that
mediates binding to sialic acids on host cells, whereas neuraminidase
is involved in the release of progeny virus from infected cells by
cleaving the sialic acid that binds the mature viral particle to the
cell surface. Neuraminidase inhibitors including Oseltamivir (Tamiflu)
and Zanamivir (Relenza) block neuraminidase enzyme activity and prevent
influenza virus reproduction.[21−23] Therefore, we assessed whether
neuraminidase resistant Poc sialic acid alters influenza reproduction
in an A549 cell infection model. In this model, a GFP-reporter influenza
virus strain A/PR8-GFP/8/34 was used that expresses GFP in infected
cells.[24] A549 cells were cultured for 3
days with Ac5SiaNAz or Ac5SiaNPoc, and incorporation of Az and Poc sialic acids as well
as cleavage-resistance of Poc sialic acid was confirmed (Figure S2). Next, confluent A549 cells were infected
with different dilutions of influenza virus strain A/PR8-GFP/8/34
(IVA) for 1 h and the percentage of GFP-positive A549 cells was determined
by flow cytometry 6 and 24 h post-infection, representing the initial
infection and the infection of viral progeny, respectively. The initial
infection after 6 h did not show a significant difference in the percentage
of GFP-positive A549 cells between cells cultured with Ac5SiaNAz, Ac5SiaNPoc,
or a solvent control (333× dilution Figure a; 1000× dilution, Figure S3).
Figure 4
Effect of Az and Poc sialic acid on influenza reproduction
(333×
dilution). A549 cells cultured with either Ac5SiaNAz or Ac5SiaNPoc were treated
with influenza virus strain A/PR8-GFP/8/34, and the GFP expression
was measured as a readout for successful viral infection after 6 h
and reinfection after 24 h (a). Az and Poc sialoglycans were detected
24 h post-infection by reaction with fluorescent biotin using the
CuAAC (b). Bar diagrams show mean fluorescence intensity ± SEM
(n = 4).
Effect of Az and Poc sialic acid on influenza reproduction
(333×
dilution). A549 cells cultured with either Ac5SiaNAz or Ac5SiaNPoc were treated
with influenza virus strain A/PR8-GFP/8/34, and the GFP expression
was measured as a readout for successful viral infection after 6 h
and reinfection after 24 h (a). Az and Poc sialoglycans were detected
24 h post-infection by reaction with fluorescent biotin using the
CuAAC (b). Bar diagrams show mean fluorescence intensity ± SEM
(n = 4).In contrast, after 24 h, the cells incubated with Ac5SiaNPoc showed a significant reduction in
GFP-positive
A549 cells compared with solvent control or cells incubated with Ac5SiaNAz, indicating that Ac5SiaNPoc, which is resistant to bacterial neuraminidase activity,
has a profound effect on virus reproduction (333× dilution, Figure a; 1000× dilution, Figure S3). Earlier research has shown that modifications
of the N-acyl chain could inhibit influenza infections
and led to the hypothesis that longer site chains would sterically
hinder binding of hemagglutinin.[25] Our
experimental data with Ac5SiaNAz or Ac5SiaNPoc after 24 h are indeed in line with
earlier observations that sialic acid C-5 modifications can lead to
viral resistance. However, the initial infection after 6 h shows no
resistance, therefore suggesting a different mechanism. In contrast
to the current hypothesis, we found that initial infection (and, therefore,
hemagglutinin recognition) is not significantly hindered by longer
C-5 modifications such as Az or Poc. Instead, the gain in resistance
in the case of Ac5SiaNPoc is probably
caused by neuraminidase resistance, thereby inhibiting the release
of the viral progeny. This was further confirmed by labeling the Ac5SiaNAz- and Ac5SiaNPoc- treated cells after 24 h of IVA infection (Figure b). The Ac5SiaNAz-treated cells showed a significant reduction in labeling
signal, indicating the release of SiaNAz after infection.
In contrast, Ac5SiaNPoc-treated cells
showed very similar labeling levels 24 h post-infection compared to
an untreated control. This clearly indicates that SiaNPoc is stable under IVA infection and resists neuraminidase cleavage.
Conclusions
In conclusion, we demonstrate that metabolic engineering with Poc-modifiedsialic acids confers resistance toward neuraminidases. Consequently,
the incorporation of Poc sialic acids into host cells reduced neuraminidase-dependent
influenza cell-to-cell transmission. Furthermore, these findings indicate
that, depending on the chemical reporter used, metabolic oligosaccharide
engineering may not be biologically inert. The neuraminidase resistance
conferred by Poc sialic acids may disturb biological processes involving
neuraminidase cleavage but could be applied in situations in which
sialic acid hydrolysis is not desired, such as therapeutic protein
modification, viral infection, etc.
Authors: David C Kennedy; Craig S McKay; Marc C B Legault; Dana C Danielson; Jessie A Blake; Adrian F Pegoraro; Albert Stolow; Zoltan Mester; John Paul Pezacki Journal: J Am Chem Soc Date: 2011-10-19 Impact factor: 15.419
Authors: Christian Büll; Torben Heise; Daniëlle M H Beurskens; Moniek Riemersma; Angel Ashikov; Floris P J T Rutjes; Toin H van Kuppevelt; Dirk J Lefeber; Martijn H den Brok; Gosse J Adema; Thomas J Boltje Journal: ACS Chem Biol Date: 2015-08-10 Impact factor: 5.100
Authors: Monique van Scherpenzeel; Federica Conte; Christian Büll; Angel Ashikov; Esther Hermans; Anke Willems; Walinka van Tol; Else Kragt; Marek Noga; Ed E Moret; Torben Heise; Jeroen D Langereis; Emiel Rossing; Michael Zimmermann; M Estela Rubio-Gozalbo; Marien I de Jonge; Gosse J Adema; Nicola Zamboni; Thomas Boltje; Dirk J Lefeber Journal: Glycobiology Date: 2022-03-30 Impact factor: 4.313
Authors: Patricia A Blundell; Dongli Lu; Mark Wilkinson; Anne Dell; Stuart Haslam; Richard J Pleass Journal: J Immunol Date: 2019-01-25 Impact factor: 5.422