Articles

https://doi.org/10.1038/s41556-020-00615-4

Asparagine enhances LCK signalling to potentiate

CD8+ T-cell activation and anti-tumour responses
Jun Wu1,2,4, Gen Li1,2,4, Le Li1,2, Dan Li3, Zhongjun Dong3 and Peng Jiang   1,2 ✉

Nutrient availability is central for T-cell functions and immune responses. Here we report that CD8+ T-cell activation and
anti-tumour responses are strongly potentiated by the non-essential amino acid Asn. Increased Asn levels enhance CD8+ T-cell
activation and effector functions against tumour cells in vitro and in vivo. Conversely, restriction of dietary Asn, ASNase admin-
istration or inhibition of the Asn transporter SLC1A5 impairs the activity and responses of CD8+ T cells. Mechanistically, Asn
does not directly alter cellular metabolic fluxes; it instead binds the SRC-family protein tyrosine kinase LCK and orchestrates
LCK phosphorylation at Tyr 394 and 505, thereby leading to enhanced LCK activity and T-cell-receptor signalling. Thus, our
findings reveal a critical and metabolism-independent role for Asn in the direct modulation of the adaptive immune response
by controlling T-cell activation and efficacy, and further uncover that LCK is a natural Asn sensor signalling Asn sufficiency to
T-cell functions.

T
lymphocytes are key components of the adaptive immune of Asp (Fig. 1a,b); this was unlikely to be due to a defect in Asp
system and protect the host from infection and cancer. T-cell uptake, as a significant accumulation of cellular 15N-Asp and reduc-
survival and expansion following antigen stimulation and tion in media 15N-Asp were found when T cells were cultured with
inflammatory signals require both extrinsic and intrinsic incen- 15
N-Asp during activation (Extended Data Fig. 1a). Thus, these data
tives; and metabolic reprogramming has recently emerged as a indicate a specific role for Asn in modulating CD8+ T-cell activa-
critical factor for determining T-cell fate and memory formation1–3. tion. Unsurprisingly, Asn itself could not trigger CD8+ T-cell activa-
Metabolic switch to glycolytic flux and glutaminolysis triggered tion when antigen was absent (Extended Data Fig. 1e–g, column 1
by T-cell antigen receptors (TCR) occurs early after activation to versus column 6).
provide energy and building blocks for T-cell expansion and effec- Activated T cells transited from a naive state are able to prolif-
tor functions4–6. Furthermore, elevated mitochondrial capacity erate rapidly and produce inflammatory cytokines15,16. Consistent
and fatty-acid oxidation are fundamental for memory CD8+ T-cell with this, Asn supplementation resulted in an increased percent-
development7–9. Interestingly, the availability of certain nutrients, age of CD8+ T cells expressing interferon γ (IFN-γ), tumour necro-
including Ser, kynurenine and l-Arg, in the micro-environment sis factor α (TNF-α) and granzyme B (GzmB; Fig. 1c). A lack of
can impact T-cell survival, proliferation and effector responsive- l-Arg, Gln or Ser is known to limit T-cell proliferation12,13. Similar
nesses10–14. Although the effect of metabolic alteration on T-cell to these molecules, depletion of Asn noticeably reduced CD8+ T-cell
expansion and survival becomes increasingly clear, the mechanisms expansion (Fig. 1d). This pro-proliferative effect of Asn was further
underlying the metabolic manipulation of CD8+ T-cell activation verified by carboxyfluorescein diacetate succinimidyl ester (CFSE),
remain poorly understood. Ki67 and BrdU staining (Extended Data Fig. 1h–j). Moreover,
Here we report that Asn strongly enhances CD8+ T-cell activa- TCR-transgenic OT-I CD8+ T cells (specific for OVA257–264)17 that
tion both in vitro and in vivo. Dietary restriction of Asn or inhi- were supplied with Asn were markedly activated in response to
bition of Asn uptake impairs T-cell activation and differentiation OVA-peptide stimulation (Extended Data Fig. 1k,l).
into memory-like cells. Mechanistically, Asn can directly bind to Following antigen-specific and inflammatory signals, naive
LCK and induce its Tyr394 phosphorylation, consequently activat- T cells soon undergo rapid division and subsequent differentia-
ing LCK signalling to promote T-cell activation and improve T-cell tion15,18,19. Notably, Asn treatment resulted in a significant reduc-
responses to pathogens or malignant cells. tion in the naive CD8+ T-cell population and increased percentages
of CD62LlowCD44highCD8+ and CD62LhighCD44highCD8+ T cells
Results (Fig. 1e). Moreover, treatment with Asn, but not Asp, increased the
Asn promotes CD8+ T-cell activation. We sought to determine population of CD8+ T cells with stemness properties, suggesting
whether Asn influences immune functions by measuring its effect that Asn regulates the activation and differentiation of CD8+ T cells
on CD8+ T-cell responses. Strikingly, Asn depletion resulted in a (Fig. 1f). Parenthetically, Asn-enhanced CD8+ T-cell activation was
sharp reduction in CD8+ T-cell activation triggered by anti-CD3 not reliant on serum components (Extended Data Fig. 1m–q).
and -CD28 antibodies (Fig. 1a). Conversely, the addition of Asn, In line with the above findings, Asn supplementation increased
possibly 100 µM at most, could sufficiently potentiate CD8+ T-cell the overall number and size of the CD8+ T cells (Fig. 1g,h).
activation (Fig. 1b and Extended Data Fig. 1a–d). Cellular Asn is Conversely, removal of Asn by treating cells with ASNase (Fig. 1i),
equivalent to Asp, and asparagine synthetase (ASNS) catalyses the an enzyme that hydrolyzes Asn into Asp and ammonia, reversed the
only route to generate Asn from Asp. However, unlike Asn, the effect of Asn on the CD8+ T cells (Fig. 1g,h,j). Similarly, treatment of
activity of CD8+ T cells was unaffected by the depletion or addition human CD8+ T cells with Asn augmented the cell-activation capacity

1
Tsinghua-Peking Center for Life Sciences, Beijing, China. 2School of Life Sciences, Tsinghua University, Beijing, China. 3School of Medicine and Institute for
Immunology, Tsinghua University, Beijing, China. 4These authors contributed equally: Jun Wu, Gen Li. ✉e-mail: [email protected]

Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology 75

Articles NATuRE CELL BIoLogy

a b c d 1.5
NS NS
1.2 ** 1.2 ** 60 40 20 20 *

TNF-α+CD8+ T cells (%)

80

CD25+CD8+ T cells (%)

GzmB+CD8+ T cells (%)

**

IFN-γ+CD8+ T cells (%)

CD44+CD8+ T cells (%)
*** **

Relative number of
CD69+CD8+ T cells
Relative number of

CD44+CD8+ T cells
Relative number of
1.0 1.0 **** NS

CD8+ T cells
60 30 15 15 1.0
0.8 0.8 40 **
0.6 0.6 10 *** ***
40 20 10 ****
0.4 0.4 0.5 ****
20
20 10 5 5
0.2 0.2
0 0 0 0 0 0 0
0
–A p
–A n
sp

–A p
–A n
sp

trl
n
p

trl
n
p

Asrl
n

As l
n

As l
n

–L p
–A u
–A n
er – sp
d g
–G y
ln
tr

tr

l
om

om

As
As

As
As

om

e
s

an Ar
G
C

C
C

–S
e 13.8% 16%
f g h
80 **** **** NS *
40 60 *** 80 ****

CD62L+ CD8+ T cells (%)

30 * 10

CD27+CD8+ T cells (%)

20 ***

Relative CD8 T-cell size

**
CD62LhighCD44highCD8+

**** ****
CD62LhighCD44lowCD8+
CD62LlowCD44highCD8+

Ctrl

Cell number (×105)

30 60 60
15

(FSC counts)
8
CD44–APC
T cells (%)

40
T cells (%)

20
T cells (%)

18.5% 51.7%
40 40

+
20 10
21.7% 30% 6
10 20 20
10 20 5 4

Asn
2
0 0 0 0 0 0 0
12.2% 36.1%
As l
n

As l
n

As l
n

As l
n
p

trl
n
p

trl

N n
e

trl

N n
e
tr

tr

tr

tr

As

As
As

AS As
as

AS As
as
C

C
CD62L–PB

+
n

n
As

As
i j k

Human TNF-α+CD8+ T cells (%)

*

Human IFN-γ+CD8+ T cells (%)

2.5 100 ** 60 *
Extracellular Asn (µg ml–1)

5 **** NS ** 50 20 **
**** 100 15
CD69+CD8+ T cells (%)

CD25+CD8+ T cells (%)

CD25+CD8+ T cells (%)

* *
CD44+CD8+ T cells (%)

* *
****
(ng per 106 cells)

2.0 4 80 **
Intracellular Asn

80 40
15
40 10
1.5 3 60 30

Human
60
10
1.0 2 40 40 20
20 5
0.5 1 20 10 5
20
0 0 0 0 0 0 0 0
trl
n AS sn

N e
e

n AS A trl
+ N sn
N se
e

trl

N n
e

trl

N n
e

trl

N n
e

trl

N n
e

AS As l
N n
e

AS As l
N n
e
tr

tr
AS s
as

as

AS As
as

AS As
as

AS As
as

AS As
as

as

as
C

C
A
+ Na

AS a

+
n

n
As

As

As

As

As

As

As

As
l m n o p Asn
*** 20 **** *** ****
40 ** 50 *** 20 40 *** 40
*** 60 **
GzmB+CD8+ T cells (%)

Medium concentrations
CD25+CD8+ T cells (%)

CD44+CD8+ T cells (%)

CD44+CD8+ T cells (%)

CD25+CD8+ T cells (%)

* 600
IFN-γ+CD8+ T cells (%)

* * * *** * 10
B16-OVA cells (%)

40 50
30 15 15 30 30 5
Annexin V+

40

(ng ml–1)
30 0
20 10 10 20 20 30 800 Asp
20
20 600
10 5 5 10 10 400
10 10
200
0 0 0 0 0 0 0 0
trl
AS Asn

trl

N n
e

trl

N n
e

trl
AS Asn

trl

N n
e

+ 7C l
AS M

+ 7C l
AS M

+ 7C l
AS M

e
M G- Ctr

M G- Ctr

M G- Ctr
as

AS As
as

AS As
as

as

AS As
as

as

as

as
C

C
N

N
E

E
C

C
-7

-7

-7
EG

EG

EG

Fig. 1 | Identification of Asn as a promotor of CD8+ T-cell activation. a–c,e,f, Naive CD8+ T cells from mice were activated by anti-CD3 and -CD28
antibodies (unless otherwise indicated) in complete medium (Comp) and medium depleted (−) of either Asn or Asp for 24 h (to detect CD69 expression)
or 40 h (to detect CD44 expression; a), or medium lacking Asn and Asp (Ctrl, control) and medium containing either Asn or Asp for 40 h (b,c,e,f), and
analysed for expression of the indicated proteins or cytokines using FACS. d, Proliferation of mouse naive CD8+ T cells activated in Comp medium or
medium depleted of the indicated amino acids for 4 d. e, Representative flow cytometry plots are shown with the percentage of cells in each quadrant
indicated (right). In f, The cell culture medium was supplied with 10% dialyzed FBS. g–j, CD8+ T cells from mice were activated in Ctrl medium or medium
containing Asn and/or ASNase (10 µg ml−1) for 40 h. The number (g) and size (h) of the cells, levels of intracellular and extracellular Asn (i) and surface
expression of the indicated proteins (j) were analysed. k, CD25 expression and cytokine production by human CD8+ T cells (n = 3 independent human
donors) activated in Ctrl medium or Asn medium with or without ASNase for 40 h. l–n, OT-I CD8+ T cells were cultured with B16-OVA melanoma cells
in the absence or presence of Asn or ASNase for 40 h. The surface expression of CD25 and CD44 (l) and cytokine production (m) by OT-I CD8+ T cells
as well as the percentage of annexin V+ B16-OVA cells (relative to total B16-OVA cells; n) were measured by FACS analysis. o,p, Naive CD8+ T cells from
mice were cultured in Ctrl or EG-7 cell-conditioned medium (CM) with or without ASNase (10 μg ml−1) for 40 h. The surface expression of CD25 and CD44
on CD8+ T cells (o), and Asn (p, top) and Asp levels (p, bottom) in the media were measured. All data are the mean ± s.e.m.; n = 3 independent wells per
experiment. Two-tailed Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and NS, not significant. Representative FACS plots for a–c,e,f,j–o
(Supplementary Figs. 1 and 2) and numerical source data are provided.

76 Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
and increased the population of IFN-γ+CD8+ and TNF-α+CD8+ Asn in the culture media of tumour cells (Extended Data Fig. 3q).
T cells; however, these effects were reduced when the cells were Consistent with this, the culture medium from glutamine-starved
then supplied with ASNase (Fig. 1k and Extended Data Fig. 1r,s). EG-7 cells had a markedly reduced ability to promote CD8+ T-cell
These results were strengthened by the observations that ASNase activation (Extended Data Fig. 3r). Together, these data dem-
had a minimal effect on the levels of Gln and Glu (Extended Data onstrate that increased extracellular Asn levels enhance CD8+
Fig. 2a–d), and Glu did not impact CD8+ T-cell activation (Extended T-cell activation. Notably, the finding of CD8+ T-cell sensing of
Data Fig. 2e,f). tumour-cell-secreted Asn reveals a potential mechanism underly-
In addition, Asn enhanced human T lymphocyte Jurkat ing the detection of de novo tumorigenesis by the immune system.
cell activation in a dose dependent manner (Extended Data
Fig. 2g–j). Moreover, a noticeable increase in the activation of Asn potentiates CD8+ T-cell responses in vivo. We next investi-
OT-I mouse CD8+ T cells stimulated by B16-OVA melanoma cells gated whether environmental Asn could influence T-cell responses
was observed when Asn was present (Fig. 1l,m). However, these in vivo. Mice maintained on a diet lacking Asn were subjected to
effects were completely reversed following the addition of ASNase intraperitoneal (i.p.) injection of Asn or PBS (control). Compared
(Fig. 1l,m). Co-culture with activated CD8+ T cells triggered apop- with the control animals, the Asn-treated mice displayed increased
tosis of B16-OVA cells, and this phenomenon was more obvious in levels of blood Asn (Fig. 2a), spleen weights (Extended Data Fig. 4a)
Asn-treated cells and largely suppressed by ASNase supplementa- and numbers of cells in the spleen and lymph nodes (Extended Data
tion (Fig. 1n). Fig. 4b). However, the weight (Extended Data Fig. 4c) as well as the
We next examined the cellular response to antigen stimulation homeostatic status of the CD4+ and CD8+ T cells in the spleens and
following inhibition of de novo Asn synthesis. Expression of ASNS lymph nodes of the Asn-treated mice were not significantly changed
remarkably increased during activation but was almost undetectable (Extended Data Fig. 4d,e). Notably, we observed an increased popu-
in the naive state (Extended Data Fig. 3a,b). Intriguingly, the levels lation of CD25+CD8+ T cells as well as central memory-like CD8+
of cellular Asn declined during the early period of antigen stimula- T cells in mice injected with Asn (Fig. 2b,c and Extended Data
tion, whereas both the extracellular and intracellular levels of Asn Fig. 4f,g). In addition, the Asn-fed mice had a relatively higher per-
increased sharply at 48 h (Extended Data Fig. 3c). These findings centage of activated CD4+ T cells and memory-like CD4+ T cells
suggest that Asn may be critical for the onset and/or maintenance of (Extended Data Fig. 4h–j), suggesting that Asn not only affects
CD8+ T-cell activation and therefore heavily consumed during acti- the activity of CD8+ T cells but may also have a broad impact on
vation. In support of this idea, Asn did not affect the activated CD8+ T-cell functions.
T-cell state further (Extended Data Fig. 3d). Intriguingly, inhibition In agreement with these findings (Fig. 2a–c and Extended Data
of ASNS by l-albizziine moderately decreased CD8+ T-cell activa- Fig. 4a–j), Asn deprivation by ASNase in mice maintained on
tion (Extended Data Fig. 3e–h), indicating that, newly synthesised a diet containing Asn led to a marked decrease in the Asn levels
Asn has a lower impact on CD8+ T-cell activation compared with in the blood (Fig. 2d), smaller spleens and reduced numbers of
extracellular Asn. cells in the spleen and lymph nodes of the mice (Extended Data
The cytokine interleukin-2 (IL-2) promotes CD8+ T-cell activa- Fig. 4k–l). However, the overall percentage of CD4+ and CD8+
tion and expansion, and increases the cytotoxicity of CD8+ T cells20. T cells was unchanged (Extended Data Fig. 4m). Notably, the mice
Following antigenic stimulation, Asn enhanced the survival and that were injected with ASNase had reduced numbers of activated
reduced apoptosis of CD8+ T cells in the presence of IL-2 (Extended CD8+ (Fig. 2e,f), central memory-like CD8+ and, in a tissue-specific
Data Fig. 3i–k). In contrast, Asn failed to affect the survival, apop- manner, effector memory-liked CD8+ T cells (Extended Data
tosis and proliferation of activated CD8+ T cells that were cultured Fig. 4n,o). Similar findings were obtained for CD4+ T cells (Extended
without IL-2 (Extended Data Fig. 3l–n), further supporting the idea Data Fig. 4p–r). Consistent with this, culturing with serum from
that Asn regulates the transition of naive CD8+ T cells into the active ASNase-administered mice significantly impeded CD8+ T-cell acti-
state but not the functions of the activated cells. vation (Extended Data Fig. 4s).
Many tumour cells produce and secrete a large amount of Asn. We further examined T-cell responses to infection with
Notably, naive CD8+ T cells cultured in medium from mouse EG-7 OVA-expressing Listeria monocytogenes (LmOVA) in vivo (Fig. 2g).
lymphoma cells and human lung cancer H1299 cells displayed an Notably, a dietary restriction of Asn reduced the number of OT-I
increased capacity for activation, and ASNase supplementation CD8+CD45.2+ T cells in both the spleens and lymph nodes of
completely reversed this effect, thereby providing an argument for congenic CD45.1 mice adoptively transferred with naive OT-I
the involvement of Asn in this event (Fig. 1o,p and Extended Data CD45.2+CD8+ T cells (Fig. 2h,i). We also examined whether Asn
Fig. 3o,p). Gln starvation resulted in a reduction in the levels of affects endogenous T-cell responses to Listeria infection (Extended

Fig. 2 | Asn enhances CD8+ T-cell activity and responses in vivo. a–c, C57BL/6 mice maintained on a diet lacking Asn and Asp were i.p. injected with PBS
(Ctrl) or Asn every 2 d, for 30 d. The blood Asn levels (a), and percentage of CD25+CD8+ T cells in the spleen (b) and lymph nodes (c) were measured (Ctrl,
n = 5; Asn, n = 6). d–f, Levels of blood Asn (d), and percentage of CD25+CD8+ T cells in the spleens (e) and lymph nodes (f) of mice on a normal diet, which
were i.p. injected with PBS (Ctrl) or ASNase (0.2 mg per mouse) every 2 d, for 30 d (n = 6 or 7 mice per group). g, Schematic of the OT-I adoptive transfer and
LmOVA infection experiment. OVA-specific (CD45.2+) CD8+ naive T cells were transferred into CD45.1+ host mice fed with control or Asn-supplemented
water and the mice were then infected with LmOVA for another 7 d. h,i, The percentage and number of OVA-specific (CD45.2+) CD8+ T cells in the spleens
(h) and lymph nodes (i) of the mice from g were measured (n = 5 mice per group). j, C57BL/6 mice that had been orally administered Asn-free (Ctrl) or
-supplemented water were subjected to LmOVA infection and then rechallenged with a lethal dose of virulent LmOVA on the normal diet, as indicated. k,l,
The percentage and number of OVA-TET+CD8+ T cells (k), and percentage of TNF-α+ and IFN-γ+ CD8+ T cells (l) in the spleens of the infected mice from
j were measured 7 d after virulent LmOVA rechallenge (n = 5 mice per group). m,n, CD45.1+ mice on an Asn-free diet were i.p. injected with PBS (Ctrl) or
Asn, adoptively transferred with OT-I CD45.2+CD8+ naive T cells and injected with B16-OVA cells, as illustrated in Extended Data Fig. 5j. The percentage
and number of CD44+CD8+ T cells (m), and percentage of central (CD62LhighCD44high) and effector (CD62LlowCD44high) memory-like CD8+ T cells (n) in the
spleens of the mice were analysed 16 d later (n = 5 mice per group). b,c,e,f,k–n, Cell percentages are expressed as a percentage of the total CD8+ T cells. Blood
samples in a and d were 5 times diluted before assaying for Asn levels. h,i,k,l,m, Representative FACS plots are shown (h,i,k,l, left and m, right); the percentages
of cells in each quadrant are indicated. All data are the mean ± s.e.m. Two-tailed Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and NS, not
significant. Numerical source data for a–f,h,i,k–n and representative FACS plots (Supplementary Fig. 3) are provided.

Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology 77

Articles NATuRE CELL BIoLogy

Data Fig. 5a). Similarly, the weight of the spleens and the overall Asn revealed a significant elevation in the percentage and number
percentage of CD8+ and CD4+ T cells in the infected mice were of OT-I CD8+ T cells 30 d post rechallenge (Fig. 2k). Consistent with
unchanged (Extended Data Fig. 5b–g). However, the percentages this, the mice that were administered Asn for the duration of the
of endogenous OT-I CD8+ T cells were reduced in mice on the primary LmOVA response displayed increased IFN-γ+CD8+ and
Asn-free diet compared with the animals maintained on the control TNF-α+CD8+ T-cell populations, demonstrating that Asn can pro-
diet (Extended Data Fig. 5h,i). mote secondary T-cell responses to LmOVA (Fig. 2l).
Enhancement of CD8+ T-cell activation and the memory T-cell Tumour-cell recognition by T cells induces effector and mem-
pool may potentiate the T-cell response and pathogen clearance fol- ory T-cell responses. Following malignancy, Asn-treated congenic
lowing rechallenge12,21. We thus investigated whether availability of CD45.1 mice adoptively transferred with naive OT-I CD45.2+CD8+
Asn during T-cell priming influences protective immunity (Fig. 2j). T cells had elevated numbers of OT-I CD8+CD45.2+ T cells (Fig. 2m
The analysis of the spleens of the mice that were initially fed with and Extended Data Fig. 5j,k). Moreover, higher percentages of

a b c d e f **
3,500 * 8 25 ** 2,000 **** 15
Lymph-node CD25+CD8+
10

Lymph-node CD25+CD8+
* *
Blood Asn (ng ml–1)

2,800
Spleen CD25+CD8+

Blood Asn (ng ml–1)

Spleen CD25+CD8+
20 8
6 1,500
10
T cells (%)

T cells (%)
T cells (%)

2,100

T cells (%)
15 6
4 1,000
1,400 10 4
5
700 2 500
5 2
0 0 0 0 0 0
trl

trl

trl

trl

trl

trl

e
As

As

As

as

as

as
C

C
N

N
AS

AS

AS
g LmOVA
h
CD8+ OT-I
50 *** 2.0
(CD45.2) **

Spleen CD45.2+CD8+

Number of spleen
40 1.5

CD45.2+CD8+
T cells (×107)
–14 –1 0 7

T cells (%)
OT-I cell 30
CD45.1

–Asn feed 1.0

response
(CD45.2) 20
–14 –1 0 7 0.5
10
Ctrl 12.9% Asn 38.8%
+Asn feed 0 0
CD45.1 (1 mM in water) CD45.2

trl

trl

n
As

As
C

C
i j LmOVA
CD45.2+CD8+ T cells (×104)

** 4
Lymph-node CD45.2+CD8+

5 ***
Number of lymph-node

A Normal
4 3 –14 0 7 feed (+Asn)
for 30 d
T cells (%)

3
CD45.1

2 –Asn feed
Rechallenge
2
1 –14 0 7
1
Ctrl 2.17% Asn 4.44%
0 0 +Asn feed (1 mM in water)
CD45.2
trl

trl

n
As

As
C

k l 25
20 ** *
25
8 ** ***
Spleen OVA-TET+CD8+

Spleen TNF-α+CD8+
spleen OVA-TET+CD8+

Spleen INF-γ+CD8+

20
2.83% 13.9% 15
6 20
T cells (%)

T cells (%)
T cells (×106)
T cells (%)

15
Number of
OVA-TET

10 4 10
15
5 2 5
Ctrl Asn
0 0 10 0
CD44
trl

trl

trl

trl

n
As

As

As

As
C

m n
** Ctrl Asn
Spleen effector memory-like
Spleen central memory-like

6 ***
Spleen CD44+CD8+ T cells

45 25 NS
** 20
spleen CD44+CD8+ T

51.6% 22%
CD62L–APC-cy7

40 18 64.1% 13.3%
CD8+ T cells (%)

CD8+ T cells (%)

cells (×105)

4 20
Number of

16
35
(%)

14
30
2 15 12
25 10 11.7% 10.8% 8.46% 18%
0 20 10 8 CD44–APC
trl

trl

trl

trl

n
As

As

As

As
C

78 Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
a b c
**** ****

N-Asn level in CD8+ T cells

C-Ser level in CD8+ T cells

25 15 3 8 80 **** **** 100 80
NS * *** ***

C-Ser level in medium

****
N-Asn level in medium

CD25+CD8+ T cells (%)

CD69+CD8+ T cells (%)

CD44+CD8+ T cells (%)

NS
20 6 80
60 60
10 2
15 60
4 40 40
10 40
5 1
2 20 20
5 20

13
0 0 0 0 0 0 0
15

13
15

– + – + – + – + Asn – + – + Asn – + – + Asn – + – +

Ser + Gly Ser + Gly Asn Asn Ser + Gly – + Ser + Gly – + Ser + Gly – +

d e f

A2
5
1A

38
C
C
****

SL
SL
Relative expression (RPKM)
100 80

Relative mRNA expression

20 60 Spleen Lymph node 1.5
*** ** * **
CD69+CD8+ T cells (%)
CD25+CD8+ T cells (%)

CD44+CD8+ T cells (%)

80
15 60
40 1.0
60
10 40
40
20 0.5
5 20
20

0 0 0 0 0
Asn – + Asn – + Asn – +
SLC1A5
SLC38A2
SLC38A3
SLC38A5
SLC6A19
SLC7A10
SLC7A1
SLC7A5
SLC1A4

SLC1A5
SLC38A2
SLC38A3
SLC38A5
SLC6A19
SLC7A10
SLC7A1
SLC7A5
SLC1A4

C rl
si s 5
C trl
A2
SL iCt
1A
SL iC
38
si s
g * h
*
5 10 *** 3 1,500 2,000 ** ****
**** **** *** **

Relative abundance in
Cellular 15N-Asn level

Relative abundance in

Jurkat T cells (a.u.)

4 8 **
CD8+ T cells (a.u.)
Cellular Gln level

Cellular Glu level

* 1,500
2 1,000
3 6
1,000
2 4 NS
1 500
2 NS 500
1

0 0 0 0 0
V9302 (µM) – 20 40 – 20 40 – 20 40 V9302 (µM) 0 20 40 0 20 40 0 20 40
si LC trl
C 5
A2

si LC rl
C 5
A2

si LC1 l
C 5
A2
tr
SL 1A

t
SL 1A

SL A
S C

si siC

si siC
38

38

38

15 15
N-Asn Gln Tyr N-Asn Gln Tyr
si

S
si

i NS – V9032 Asn Asn + V9032

40 NS
40 **** 20.5% 15.5% 29.2% 14.9%
CD69+CD8+ T cells (%)

CD69–PB
CD25 CD8 T cells (%)

****
30 30 ***
*
20 20
+

10 10
+

24.3% 12.8% 35.6% 11.3%

CD25–APC

0 0
Asn – – + + Asn – – + +
V9032 (17 µM) – + – + V9032 (17 µM) – + – +

FSC-A FSC-A FSC-A FSC-A

Fig. 3 | CD8+ T-cell activation by Asn relies on its uptake through SLC1A5 and not its amino-acid exchange function. a, Extracellular (left) and intracellular
(right) quantities of 15N-Asn of naive CD8+ T cells from mice activated in complete medium (+) or medium depleted (−) of Ser and Gly in the presence
of 0.5 mM 15N-Asn for 24 h (n = 3 independent wells). b, Extracellular (left) and intracellular (right) quantities of 13C-Ser of naive CD8+ T cells from mice
cultured for 24 h in 0.5 mM 13C-Ser-supplemented medium containing no (−) or 0.4 mM Asn (n = 3 independent wells). a,b, Values were calculated as the
peak area × 107 for 3 × 106 cells. c,d, Surface expression of CD69, CD25 and CD44 on CD8+ T cells stimulated in medium lacking or supplemented with
Ser and Gly (c), or in amino-acid-free medium containing no or 0.2 mM Asn (d), as indicated (n = 3 independent wells). e, Relative mRNA expression of
potential transporters for Asn uptake in tissues from human spleens (n = 4) and lymph nodes (n = 5). RPKM, reads per kilobase of transcript per million
mapped reads. f,g, Jurkat cells treated with control siRNA (siCtrl) or siRNA against SLC1A5 (siSLC1A5) or SLC38A2 (siSLC38A2) were cultured in medium
containing 15N-Asn for 2 h. The expression of SLC1A5 and SLC38A2 mRNA (f; determined using quantitative PCR with reverse transcription), and the cellular
levels of 15N-Asn, Gln and Glu (g) were measured (n = 3 independent wells). Data are the normalized peak area × 104 for 15N-Asn and normalized peak
area × 105 for Gln and Glu. h, Levels of 15N-Asn, Gln and Tyr in naive CD8+ T cells from mice (left) and Jurkat cells activated in medium containing 0.5 mM
15
N-Asn for 2 h in the presence of increasing concentrations of V9302; a.u., arbitrary units. i, Mouse CD8+ T cells activated (3.5 μg ml−1 anti-CD3 and
1 μg ml−1 anti-CD28) in Asn or Asn-free medium in the presence or absence of V9302 for 12 h. The surface expression of CD25 and CD69 relative to the
total CD8+ T cell populations was quantified (n = 3 independent wells). Representative flow cytometry plots are shown and the percentages of cells in the
boxed areas are indicated (right). All data are the mean ± s.e.m. Two-tailed Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and NS, not
significant. Representative FACS plots for c,d,i (Supplementary Fig. 4) and numerical source data for a–i are provided.

Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology 79

Articles NATuRE CELL BIoLogy

effector and central memory-like CD8+ T cells were found in the bated by Asn supplementation (Extended Data Fig. 6a,b). However,
Asn-treated mice (Fig. 2n and Extended Data Fig. 5l). Collectively, although the mTORC1 inhibitor rapamycin could reduce T-cell
these findings suggest that Asn influences the activity and responses activation, we still observed a substantial elevation of CD8+ T-cell
of CD8+ T cells in vivo. activation induced by Asn (Extended Data Fig. 6c), indicating that
upregulation of mTORC1 activity may contribute to, but is not suf-
Cellular uptake, not the exchange factor role, is required for Asn ficient for, Asn-enhanced CD8+ T-cell activation.
to promote CD8+ T-cell activation. Asn is an amino-acid exchange When Gln is limited, Asn is used for protein synthesis in many
factor in proliferating cells22. Similarly, Asn starvation significantly tumour cells25. Similarly, the addition of Asn during the activation of
lowered Ser uptake in CD8+ T cells, despite the fact that Asn uptake T cells, promoted protein synthesis and significantly restored mes-
was unaffected by the depletion of Ser and Gly (Ser + Gly; Fig. 3a,b). senger RNA translation when Gln was absent or scarce (Extended
However, Ser + Gly deficiency failed to influence Asn-enhanced Data Fig. 6d). Moreover, Gln limitation reduced the proteasome
CD8+ T-cell activation (Fig. 3c) and Asn supplementation was still inhibition-induced accumulation of short-lived proteins (MYC
able to improve CD8+ cell activation even when the cells were cul- and NRF2), and Asn supplementation resulted in a noticeable res-
tured in amino-acid-free medium (Fig. 3d), suggesting that the role toration of this accumulation (Extended Data Fig. 6e). Thus, these
of Asn as an amino-acid exchange factor is not required for its pro- data suggest a role for Asn in restoring translation in the absence of
moting effect on CD8+ T-cell activation. Gln. In keeping with this, Gln depletion or scarcity did not suffi-
We then investigated whether cellular uptake is essential for the ciently block the enhancement of T-cell activation mediated by Asn
Asn-promoted activation of CD8+ T cells. As it remains unclear (Extended Data Fig. 6f).
which transporters are involved in Asn import and exchange, we
analysed the expression of several potential Asn transporters22 in LCK signalling mediates Asn-promoted CD8+ T-cell activation
human lymph-node and spleen tissues. Notably, high expression and responses. T-cell activation is accompanied by metabolic repro-
levels of the transporters SLC1A5 and SLC38A2 as well as three gramming13 (Extended Data Fig. 7a,b). Although Asn supplementa-
antiporters (SLC7A1, SLC7A5 and SLC1A4) potentially involved tion changed the cellular levels of many metabolites (Extended Data
in Asn exchange function were found in these immune organs Fig. 7a,b), isotope tracing using 15N-Asn demonstrated that the inte-
(Fig. 3e). Similar findings were also obtained in Jurkat cells grated Asn was not metabolized further (Fig. 4a,b), suggesting that
(Extended Data Fig. 5m). As examined above, Asn augmentation Asn modulation of T-cell activation may not be attributable to its
of T-cell activation does not rely on its exchange function; we thus direct metabolic participation despite the fact that it can impinge on
focused on the transporters SLC1A5 and SLC38A2. Activation of activation-induced T-cell metabolic reprograming.
CD8+ T cells increased the expression of SLC1A5, but not SLC38A2, T-cell activation requires the Src family tyrosine kinase LCK to
to some extent (Extended Data Fig. 5n,o). Consistent with this, initiate antigen-specific TCR signalling26 (Extended Data Fig. 7c).
ablation of SLC1A5, but not SLC38A2, reduced the Asn uptake Notably, when CD8+ T cells were supplied with Asn, we observed a
(Fig. 3f,g). Interestingly, the cellular levels of Gln, but not Glu, remarkable enhancement in the phosphorylation of LCK (Y394) and
were significantly reduced following SLC1A5 silencing (Fig. 3f,g). the LCK substrates ZAP70 and PI3K during activation (Fig. 4c,d).
Similarly, pharmacological inhibition of SLC1A5 dramatically Similar findings were also observed in vivo. Following OVA-peptide
reduced the levels of intracellular 15N-Asn and Gln but not Tyr treatment, adoptively transferred CD45.1 mice displayed signifi-
(Fig. 3h). Thus, both Asn and Gln can use SLC1A5 for their impor- cantly elevated LCK phosphorylation in spleen OT-I CD8+ T cells
tation into cells. Next, we examined the effect of Asn uptake on when Asn was provided (Fig. 4e and Extended Data Fig. 7d).
CD8+ T-cell activation by inhibiting SLC1A5 in the absence of Conversely, treatment with ASNase reduced the phosphorylation of
Gln. SLC1A5 inhibition was sufficient to block the Asn-mediated LCK, ZAP70 and PI3K (Fig. 4f). Moreover, using high-resolution
enhancement of CD8+ T-cell activation (Fig. 3i and Extended Data imaging, we found elevated probable microclusters of phosphory-
Fig. 5p). Collectively, these findings suggest that Asn uptake medi- lated TCR signalling proteins in Asn-treated CD8+ T cells (Fig. 4g,h).
ated by SLC1A5 is critical for Asn-promoted CD8+ T-cell activation. Together, these findings suggest a previously unappreciated role for
Asn in the regulation of TCR signalling.
Asn-enhanced CD8+ T-cell activation is independent of mTORC Addition of the LCK inhibitor PP2 totally blocked LCK and PI3K
and Gln. Next, we explored the mechanism(s) underlying T-cell phosphorylation triggered by Asn. The PI3K inhibitor GDC-0032
activation by Asn. Upregulation of mTORC1 activity metabolically affected LCK phosphorylation slightly but reduced PI3K phosphor-
supports T-cell activation23, and Asn is able to stimulate mTORC1 ylation in cells cultured with Asn (Fig. 5a and Extended Data Fig. 7e).
activity indirectly22,24. Activation of CD8+ T cells indeed displayed In keeping with these findings, although the Asn feed enhanced
increased mTORC1 activity, and the effect was strongly exacer- CD8+ T-cell activation, supplementation with PP2 and GDC-0032

Fig. 4 | Asn promotes LCK activation. a,b, Naive CD8+ T cells from mice were activated in medium containing 400 µM 15N-Asn for 0, 6 or 12 h, as indicated.
a, The cellular abundances of metabolites labelled with the stable isotope 15N for the indicated time periods were analysed by liquid chromatography with
tandem mass spectrometry (LC–MS/MS). The heat map shows the differences in metabolite abundances. b, Mass isotopomer distribution (MID) of
cellular amino acids at 12 h. The incorporation of 15N atoms from Asn are denoted as M + n, where n is the number of 15N atoms, M means mass isotopomer.
c, Mouse CD8+ T cells left unstimulated (naive) or stimulated (10 µg ml−1 anti-CD3 and 1 µg ml−1 anti-CD28 antibodies) for 12 min in the presence of
increasing quantities of Asn were analysed by western blotting. d, Mouse CD8+ T cells in medium lacking Asn and Asp (vehicle) or Asn medium were
stimulated for the indicated times and protein expression was analysed by western blotting. e, CD45.1 C57BL/6 mice on an Asn-free diet and treated
with PBS (Ctrl) or Asn were adoptively transferred with naive OT-I CD45.2+CD8+ T cells and i.p. injected with 50 µg OVA peptide in complete Freund’s
adjuvant (CFA) as indicated in Extended Data Fig. 7d. The expression levels of LCK phosphorylation in spleen OT-I CD8+ T cells at 12 h post treatment
with the OVA peptide were analysed by FACS (relative to the total CD8+ cell population; n = 5 mice per group). f, Immunoblot analysis of naive CD8+ T cells
from mice activated in medium containing Asn and/or ASNase for the indicated time periods. g,h, Confocal images of phosphorylated (p-) LCK, ZAP70 and
PI3K (g), as well as AKT and LAT171 (h) clustering in naive CD8+ T cells stimulated with or without Asn for 10 min (left). The total fluorescence intensity
(tFI) was quantified (right; n = 48 cells). Scale bars, 4 µm. All data are the mean ± s.e.m. Two-tailed Student’s t-test; *P < 0.05, ***P < 0.001 and
****P < 0.0001. Western blot data are representative of at least three independent experiments. Uncropped blots for c,d,f, representative FACS plots
for e (Supplementary Fig. 5) and numerical source data for a,b,e,g,h are provided.

80 Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
remarkably reduced (PP2 caused a total reversal) Asn-promoted Asn-treated mice (Fig. 5e–k). We further confirmed these findings
T-cell activation (Fig. 5b,c). Consistent with this, the introduction using LmOVA infections (Extended Data Fig. 7g). The administra-
of the mutant LCK(Y394F) reduced T-lymphocyte activation and tion of Asn increased the population of activated, effector and cen-
blunted the susceptibility of these cells to Asn availability (Fig. 5d). tral memory-like T cells, and treatment with PP2 totally reversed
In addition, OT-I adoptively transferred mice injected with Asn had these effects (Fig. 5l–o). Together, these findings suggest that Asn
an increased population of activated CD8+ T cells as well as effector enhances T-cell activation through LCK.
and central memory-like CD8+ T cells (Fig. 5e–g,i–k and Extended
Data Fig. 7f). Apparently, the administration of PP2 or GDC-0032 Asn binds to LCK directly and promotes its activity. Whereas
reduced T-cell activation and the number of spleen OT-I+ CD8+ Tyr505 phosphorylation is inactivated, Tyr394 autophosphorylation
T cells, thereby diminishing the difference between the control and activates LCK26. Interestingly, treatment with Asn dose-dependently

a b c Asn (µM)
15
N-Asn treatment (h)
M+2 M+1 M+0

0 ve

0
0 6 12 Size

00
01
10
00

0
ai

10
10
100

1,
0.
0.
1.
(kDa)

N
Ala

MID from (U-15N)Asn (%)

Asn p-LCK 55
Arg 80
Asp LCK 55
Gln 60
Gly p-CD3ζ 25
His 80
40
Ile CD3ζ 25
Leu 20
Lys p-Zap70 70
Met 0
Phe Zap70 70

Lys

Val
Ser
Pro
Ala

Gly

Leu

Arg
His

Tyr
Gln

Ile

Thr
Met
Asn

Trp
Asp
Gln

Phe
Pro 100
Ser 60 p-PI3K
Thr 100
Trp PI3K
Tyr
Val p-AKT
55
ADP/dGDP AKT
AMP d Vehicle Asn 55
40
CMP Actin
Size 40
Creatine CD3 + CD28 (min)
10
20

10
20

(kDa)
0
5

0
5

Cyclic AMP
Dihydrooratic acid p-LCK 55 g
Glu p-LCK DAPI Merge
p-LCK/LCK ratio 1 1.7 1.9 2.1 1 1.7 2.5 4.2 15
Glutathione
LCK 55 ***

tFI of p-LCK(Y394)
GMP
20
IMP Ctrl
Inosine p-Zap70 70 10

(×105)
NAD
Orotic acid Zap70 70
Taurine 5
p-PI3K
UMP 70 Asn
Xanthine
0 PI3K 0
70
Cellular 15N-labelling abundance
Actin 40

trl

n
As
C
p-ZAP70 DAPI Merge
e f Asn Asn + ASNase 3 ****

tFI of p-ZAP70(Y319)
Size Ctrl
80 * CD3 + CD28 (min) (kDa)
10
20

10
20

2
0
5

0
5

(×105)
Spleen p-LCK(Y394)

p-LCK 55
CD8+ T cells (%)

60
LCK 1
55
40 p-Zap70 70 Asn
0
20 Zap70 70
trl

n
As
C

p-PI3K p-PI3K DAPI Merge

0 70
Ctrl Asn PI3K
6 ****
tFI of p-PI3K (×105)

70
Actin 40 Ctrl
4

Asn
0
trl

n
As
C

h
p-AKT DAPI Merge p-LAT171 DAPI Merge

6 2.5 ****
tFI of p-LAT1(Y171)

***
tFI of p-AKT(S473)

Ctrl Ctrl
2.0
4 1.5
(×105)
(×105)

1.0
2
Asn 0.5
Asn
0 0
trl

n
As
trl

C
As
C

Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology 81

Articles NATuRE CELL BIoLogy

a GDC- b c
PP2 0032
Size ****
Asn – + – + – +
**** **** ****
(kDa) 50 **** 80 50 50

CTLA-4+CD8+ T cells (%)

*** *

PD-1+CD8+ T cells (%)

CD44+CD8+ T cells (%)
CD25+CD8+ T cells (%)
p-LCK 55
40
*** 40 40
60
p-LCK/LCK: 1 2.9 0.1 0.1 0.4 0.7 30 30
30 NS
LCK 40
55
20 ** NS ** 20 NS NS 20
NS
p-PI3K 20 10
70 10 NS 10

0 0 0 0
PI3K 70 Asn – + – + – +
Asn – + – + – + Asn – + – + – + Asn – + – + – +
Actin PP2 GDC- PP2 GDC- PP2 GDC- PP2 GDC-
40
0032 0032 0032 0032

d Vehicle Asn e f g
* *

Spleen CD44+CD8+ T cells (%)

50 *
40
pEYFP–LCK

39.6% 46.0% 70 ** 30
CD25+CD8+ T cells (%)

CD62LhighCD44highCD8+

CD62LlowCD44highCD8+
40 60 NS NS
NS NS NS NS
35 NS
30 20

T cells (%)
50

T cells (%)
Spleen

Spleen
40 30
20
pEYFP–LCK

25.8% 21.2% 30 10
(Y394F)

10 20 25
10
0 0 20 0
Asn – + – + Asn – + – + – + Asn – + – + – + Asn – + – + – +
CD25–PE
pEYFP– LCK LCK PP2 GDC- PP2 GDC- PP2 GDC-
(Y394F) 0032 0032 0032

h i j k
*
Number of spleen OT-I+CD8+

3 70 40 40
10
* *** ** ***
Lymph-node CD44+CD8+

CD62LhighCD44highCD8+

CD62LlowCD44highCD8+
60
Spleen OT-I+CD8+

8 30 30
50 NS
T cells (×106)

NS

Lymph-node
Lymph-node
2 NS NS

T cells (%)
T cells (%)
T cells (%)
T cells (%)

6 NS NS NS 40 NS NS
NS 20
20
4 30
1 20 10 10
2 10
0 0 0 0 0
Asn – + – + – + Asn – + – + – + Asn – + – + – + Asn – + – + – + Asn – + – + – +
PP2 GDC- PP2 GDC- PP2 GDC- PP2 GDC- PP2 GDC-
0032 0032 0032 0032 0032

l m n o
*** * *
90
*** 60 35 * 30 20
CD62L CD44highCD8+

CD62LlowCD44highCD8+

GzmB CD8 T cells (%)

IFN-r+CD8+ T cells (%)

55 30
Spleen CD44+CD8+

80 NS NS
50 NS 25 NS 15
70 NS 20
T cells (%)

T cells (%)
T cells (%)

45
Spleen

20
Spleen

60
40 10
+

50 15
high

35 10
40 10 5
+

15
20 5
0 0 0 0 0
Asn – + – + Asn – + – + Asn – + – + Asn – + – + Asn – + – +
PP2 PP2 PP2 PP2 PP2

Fig. 5 | Asn promotes T-cell activation through LCK. a, Immunoblot of the LCK, p-LCK, PI3K and p-PI3K proteins in mouse CD8+ T cells treated with Asn and/or PP2
or GDC-0032, as indicated, during stimulation. Data are representative of three independent experiments. b,c, Naive CD8+ T cells from mice were treated as in a.
The percentage of CD25+CD8+ T cells and CD44+CD8+ T cells relative to total CD8+ T cells (b), and surface expression of PDL-1 and CTLA-4 on the CD8+ T cells
(c) were measured (n = 3). d, Jurkat T cells expressing yellow fluorescent protein (YFP)–LCK or the YFP–LCK(Y394F) mutant were stimulated in control medium
lacking Asn and Asp (Ctrl) or medium containing Asn. The percentage of CD25+CD8+ T cells relative to total CD8+ T cells was quantified (left; n = 3 independent
experiments). Representative FACS plots are shown (right). e–k, CD45.1 C57BL/6 mice on an Asn-free diet and i.p. injected with Asn or PBS (−), or with/without PP2
(30 µg per mouse) or GDC-0032 (100 µg per mouse) were adoptively transferred with naive OT-I CD45.2+CD8+ T cells and injected with OVA peptide as indicated
in Extended Data Fig. 7f. The percentage of CD44+CD8+ (e,i), CD62LhighCD44highCD8+ (f,j) and CD62LlowCD44highCD8+ (g,k) T cells relative to the total CD8+ T-cell
population in the spleens (e–h) and lymph nodes (i–k) of the mice were measured. The overall percentage and number of OT-I CD8+ T cells in spleen were also
determined (h, right; n = 5 mice per group). l–o, CD45.1 C57BL/6 mice on an Asn-free diet and i.p. injected with Asn and/or PP2 were adoptively transferred with
naive OT-I CD45.2+CD8+ T cells and infected with LmOVA as indicated in Extended Data Fig. 7g. The percentage of CD44+CD8+ (l), CD62LhighCD44highCD8+ (m)
and CD62LlowCD44highCD8+ (n) T cells (as a percentage of the total CD8+ T-cell population), and production of cytokines by CD8+ T cells (o) in the spleens of the
mice were measured at day 11 (n = 5 mice per group). All data are the mean ± s.e.m. Two-tailed Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and
NS, not significant. Uncropped blots for a, representative FACS plots for b–o (Supplementary Figs. 6,7) and numerical source data for b–o are provided.

increased Tyr394 phosphorylation and decreased Tyr505 phos- following Asp treatment (Fig. 6a,b). We next performed an equi-
phorylation (Fig. 6a,b), and directly activated LCK in vitro (Fig. 6c librium binding assay using radioactive amino acids to investi-
and Extended Data Fig. 8a). However, these effects were not evident gate the possibility of Asn–LCK binding27. Purified recombinant

82 Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
a Asn (µM)
b c
Asp (µM) 0.4 0.4
Asn Ctrl

p-LCK(Y394) level (OD450)

****
Asp Asn

0
Size

LCK activity (OD450)

02

02
20

20
.0

.0
00

00
trl

trl
0.3 **** ****

20

20
0.3

0.

0.
0.

0.
2.

2.
C

C
(kDa)

0
Asp
p-LCK(Y394) 55 ****
Ratio of p-LCK(Y394) 0.2
**** 0.2
and LCK 1 1.4 1.6 2.1 2.8 1 0.9 1 1.1 1.2
****
p-LCK(Y505) 55 0.1 * 0.1
p-LCK(Y505)/LCK 1 0.7 0.3 0.3 0.3 1 0.9 0.8 1.3 1.4
0 0
LCK 55 0 0.2 2 20 50 2 10 15 30 50
Concentration (µM) Reaction time (min)

d e 2 µCi (3H)Asn 2 µCi (3H)Asp

2 µCi (3H)-Asn 2 µCi (3H)Asn + 10 mM Asn 2 µCi (3H)Asp + 10 mM Asp
Vector

LCK

ZAP70
SLP76
Size 2 µCi (3H)-Asn + 10 mM Asn
MW

LCK
MW
(kDa) Size
100 1,500 (kDa) *** 2,000
1,800
**
(3H)-Asn bound (c.p.m.)

(3H)-Asn bound (c.p.m.)

75 100

(3H)Asp bound (cpm)

1,500
75 1,500
55 1,000 1,200
55
40 ** 900 1,000
500 45
37 600
37 500
300
25 0 25 0 0
Vector

K
0

6
P7

P7

P7

P7
LC

LC
LCK

ZA

SL

ZA

SL
f g
Size
45 20
K

200
W
LC

(kDa)
M

100 40 150

Asn concentration
15
Response units
Response units

75 100
35 50
55
(µM)
10
40
30
45 30
5 20 Kd of LCK
25
37 10 for Asn
Kd = 8.142 × 10–6 M
20 0 0
25 0 50 100 150 200 0 50 100 150 200 CD3 + CD28 (h) – – – + + + +
Asn (µM) Asp (µM) Time (min) 0 5 30 5 30 5 30
Asn Asn

h i j
n)
m -
e

do 3
n) s

ai
2 (SH
ai ina
m (k

SH CK

Size
do CK

L
W

500 Size 5 30
M

(kDa)
L
W

400
M

(kDa)
100 25
Asn concentration (µM)

300 100 4
200 75
Response units
Response units

50 75 20
3 55
40 55
15
30 2 45
Kd of LCK for 45 10
20 Asn
1 37 5
10 37
Kd = 3.438 × 10–6 M
0 25 0 25 0
CD3 + CD28 (h) – – 4 4 12 12 0 100 200 0 100 200
ASNase – + – + – + Asn (µM) Asn (µM)
Asn (400 µM) – – + + + +

Fig. 6 | Asn binds directly to LCK and modulates its activity. a, In vitro kinase assays of recombinant LCK proteins left untreated (Ctrl) or incubated with
ATP and increasing concentrations of Asn or Asp for 15 min. LCK phosphorylation was analysed by western blotting (a) and enzyme-linked immunosorbent
assay (ELISA) using anti-p-LCK (Y394) antibody (b). The results show the average values of the optical density at 450 nm (OD450) from triplicates. c, The
activity of recombinant LCK proteins left untreated (Ctrl) or treated with either Asn or Asp for the indicated times in the presence of ATP was investigated
by ELISA using biotin-poly (Glu4: Tyr, 4:1) as described in Methods (n = 3 biological replicates). d,e, 3H-radiolabelled Asn bound to purified recombinant
LCK, but not ZAP70 and SLP76 (n = 3 biological replicates). Unlabelled amino acid was added where indicated (see Methods for details). e, 3H-radiolabelled
Asp did not bind LCK, ZAP70 or SLP76. f, BIAcore measurements of the interaction between purified LCK and Asn or Asp. Graphs of equilibrium response
units and compound concentrations are shown (middle and right). g,h, Absolute concentrations of Asn in mouse CD8+ T cells left unstimulated (naive) or
stimulated in medium containing 0 or 400 µM Asn and/or 10 µg ml−1 ASNase for the indicated times. The Kd value of LCK for Asn is indicated by the dashed
line (n = 3 independent wells). i,j, BIAcore measurements of the interaction between Asn and the purified LCK kinase domain (amino acids 231–509; i) and
SH3-SH2 domain (amino acids 60–240; j). Graphs of the equilibrium response units and compound concentrations are shown (right). b–e,g,h, Statistical
significance was determined using a two-way analysis of variance (ANOVA), followed by Tukey’s multiple comparisons test (b,c) or a two-tailed Student’s
t-test (d,e,g,h); *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. d–f,i,j, Purified proteins were analysed by SDS–PAGE, followed by Coomassie blue staining
(left). a,f,i,j, Data are representative of at least three independent experiments. All data are the mean ± s.e.m. MW, molecular weight markers. Uncropped
blots for a and numerical source data for b–j are provided.

Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology 83

Articles NATuRE CELL BIoLogy

a b c
10 ** ****
Ctrl 50 **
100 100 Ctrl (n = 10)

Annexin V+ EG-7 cells

Relative cytotoxicity

Tumour multiplicity
8 40 Asn (n = 10)
80

Survival (%)
6 1 cm 30
60
***

(%)
50
4 40 Asn 20

2 20 10

0 0 0 0
Asn – + – + Asn – + 1 cm
Ctrl Asn 0 5 10 15 20 25 30 35 40 45 50 55
+ +
CD8 :EG-7 1:1 1:2 CD8 :EG-7 1:2 Time post tumour inoculation (d)

d *** e
30 **
Ctrl –Arg –Asn PBS or Asn
Tumour multiplicity intravenously
Tumour
20 Day 0 every 2 d 20 21 Normal diet and
survival
10 Asn and Asp-free diet analysis
1 cm 1 cm 1 cm
8 × 105
B16-OVA cells injected
0 subcutaneously
trl

rg

sn
C

–A

f g –A h
250 Ctrl 200
**** Ctrl (n = 8) Ctrl
100 ****

Tumour size (mm2)

Asn
Tumour size (mm2)

200 n=3
Asn (n = 8) Asn
150
Survival (%)

150 *** n=4 100 ****

50
100 *** n=6
****
50
50

0 0 0
11 15 19 23 27 29 31 0 5 10 15 20 25 30 35 40 9 14 17 20 23
Time post B16-OVA cell injection (d) Time post tumour inoculation (d) Time (d)

i j
Ctrl (n = 10) 80 ** 80 60
** 40 **
***
CD44+CD8+ T cells (%)

100
GzmB CD8 T cells

Asn (n = 10)
IFN-γ+CD8+ T cells

60 60 30
Survival (%)

40

Ki67 (%)
+

(%)

(%)

40 40 20
50
+

20
20 20 10

0 0 0 0 0
0 5 10 15 20 25 30 35 40 45 50 55 Ctrl Asn Ctrl Asn Ctrl Asn Ctrl Asn
Time post tumour inoculation (d)

Fig. 7 | Asn improves CD8+ T-cell anti-tumour activity. a, OVA-specific OT-I CD8+ T cells pre-activated in medium lacking both Asn and Asp (−) or with
Asn (+) were cultured with EG-7 cells for 18 h. The relative cytotoxicity (left) and percentage of apoptotic CD8+ T cells (right) were determined (n = 3
independent wells). b,c, C57BL/6 mice were intravenously injected with B16-OVA cells, immunodepleted by whole-body irradiation13,33 and injected with
naive OT-I CD8+ T cells pre-activated in medium lacking Asn and Asp (Ctrl) or Asn-supplemented medium (Asn), as illustrated in Extended Data Fig. 9q.
The tumour multiplicity on day 17 (n = 7 mice per group; b, right) and mouse survival (n = 10 mice per group; c) were measured. Representative images of
lung tumour multiplicity are shown (b, left). d, C57BL/6 mice were tail-vein injected with 4 × 105 B16-OVA cells and immunodepleted by irradiation twice
(2 Gy each time). After 12 h, the mice were tail-vein injected with 8 × 105 CD8+ T cells pre-activated in the presence (Ctrl) or absence (−) of Arg and Asn.
The tumour multiplicity in the lungs of the mice was measured on day 12 (n = 5; right) and the representative images of lungs are given. e–g, C57BL/6
mice on an Asn-free diet and intravenously injected with PBS (Ctrl) or Asn were subcutaneously (s.c.) administered B16-OVA cells as indicated (e), and
the tumour burden (f) and survival (g) over time were assessed (n = 8 mice per group). h–j, C57BL/6 mice were subjected to B16-OVA-cell injection,
whole-body irradiation, naive OT-I CD8+ T-cell transfer and treatment of with Asn or PBS (Ctrl), as illustrated in Extended Data Fig. 9s. h,i, The tumour
burden (h) and survival (i) over time were assessed. Tumour-infiltrating CD8+ T-cell proliferation was determined by Ki67 staining (n = 8 mice per group).
j, Percentage of CD44+CD8+ T cells and production of cytokines by CD8+ T cells in tumours relative to the total CD8+ cell population (n = 10 mice per
group). a,b,d,f,h,j, Statistical significance was determined using a two-tailed Student’s t-test (a,b,d,j), or a two-way ANOVA, followed by Tukey’s multiple
comparisons test (f,h); **P < 0.01, ***P < 0.001 and ****P < 0.0001. All data are the mean ± s.e.m. Representative FACS plots for a,j (Supplementary Fig. 8)
and numerical source data for a–d,f–j are provided.

LCK protein immobilized on beads was incubated with 3H-Asn, only bound to LCK; it did not bind the ZAP70 and SLP76 proteins
followed by washing and quantification. Notably, LCK bound to (Fig. 6e). In contrast, Asp exhibited no binding affinity towards
3
H-Asn in a dose-dependent manner, which could be fully com- LCK and the other proteins (Fig. 6e). These findings were further
peted by non-radiolabelled Asn (Fig. 6d). Specifically, 3H-Asn confirmed using surface plasmon resonance (BIAcore; Fig. 6f);

84 Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Asn also bound to LCK mutants (Y394F and Y505F; Extended tumour-infiltrated T cells with a higher capacity for activation,
Data Fig. 8b,c). cytokine generation and even proliferation (Fig. 7j). Collectively,
In addition, we assessed the binding of LCK to other amino acids. these data demonstrate that increased Asn levels enhance the acti-
Unlike Asn, other amino acids did not bind to LCK, with the excep- vation of CD8+ T cells and their anti-tumour activity in vivo.
tion of Gly and His, which interacted weakly with LCK (Extended
Data Fig. 8d–u). Neither Asn nor Asp bound to ZAP70, as measured Discussion
using BIAcore (Extended Data Fig. 8v). Consistent with the in vitro The availability of specific nutrient(s) dramatically influences
BIAcore and radioactive-labelling data, cellular thermal shift and innate and/or adaptive immunity through manipulation of the
drug affinity responsive targets stability assays revealed direct bind- antigen-specific T-cell expansion capacity and efficacy28. However,
ing between LCK and Asn inside cells (Extended Data Fig. 8w,x). it remains unclear whether T-cell activation requires the direct
Because Asn depletion reduced the LCK activity and T-cell activa- involvement of certain nutrient(s), and in this work we sought to
tion, the cellular Asn concentrations should be sufficient for Asn determine the role of Asn in T-cell differentiation29. We found that
binding to LCK. The Asn concentrations in both naive and activated Asn can strongly enhance T-cell activation and the memory-like
cells were indeed higher than the dissociation constant (Kd) of LCK potential of T cells, demonstrating the importance of Asn as a cru-
for Asn, making it probable that Asn is a physiologically relevant cial metabolic factor that can cause a sharp increase in T-cell immu-
modulator of LCK activity (Fig. 6g). In keeping with this, when the nity. More importantly, Asn serves as a signalling molecule in CD8+
cells were supplied with Asn elevated the cellular Asn abundances T cells, which can be directly sensed by LCK.
quickly and substantially resulting in increased LCK activity and Tumour cells outcompete T cells for nutrients (for example,
T-cell functions (Fig. 6g). Conversely, the Asn concentrations in glucose), leading to impeded T-cell effector function29. Similarly,
CD8+ T cells decreased following ASNase treatment, with the Kd T and tumour cells may also compete for Asn in the tumour
decreasing from a value above that of LCK for Asn to below this micro-environment for proliferation and survival, particularly
value (Fig. 6h). when Asn is scarce or tumour cells are defective in the de novo
Phosphorylation controls LCK activity26. Blocking phosphoryla- synthesis of Asn. However, many tumour cells usually have a high
tion on Tyr394 (Y394F) largely reduced LCK activity, whereas inhi- biogenetic activity that can enable them to produce and release
bition of Tyr505 phosphorylation (Y505F) stimulated LCK activity Asn25,30,31 (Fig. 1p and Extended Data Fig. 3p), which can fuel
(Extended Data Fig. 8y,z). Asn addition failed to significantly acti- the bioenergetic demands of both T and tumour cells for growth
vate LCK(Y394F) and LCK(Y505F) (Extended Data Fig. 8y), and or preserve the anti-tumour potential of T cells. In this situation,
the Y394F mutation did not affect the Asn-mediated reduction tumour cells may be able to manipulate T-cell responses by chang-
of Tyr505 phosphorylation, whereas the Y505F mutation boosted ing the levels of Asn in the micro-environment. Given that Asn also
Tyr394 phosphorylation and minimized the effect of Asn (Extended positively regulates breast tumour-cell metastasis32, increasing the
Data Fig. 8z). Thus, Asn might inhibit Tyr505 phosphorylation to levels of Asn in the tumour micro-environment may have a pro-
promote Tyr394 autophosphorylation. To identify the region of tective role in preventing tumorigenesis by enhancing the T-cell
binding, we compared the ability of various LCK deletion mutants responses but would also benefit tumour cells if they have escaped
to bind Asn and found that the binding site for Asn is the region the immunosurveillance system. Under this condition, it is possible
between amino acids 240 and 320 on LCK within the kinase domain that Asn contributes to tumour-cell-mediated T-cell exhaustion in
(Fig. 6I,j and Extended Data Fig. 9a–c). the tumour micro-environment. Nevertheless, the availability of
Asn in vivo may exhibit complex and time-dependent functions in
Asn enhances the anti-tumour T-cell response. Next we investi- cancer immunosurveillance but would have pathophysiological and
gate whether Asn potentiates anti-tumour T-cell responses. Naive therapeutic implications for immunotherapy.
OT-I CD8+ T cells pre-stimulated by the OVA257–264 peptide in con-
trol or Asn-supplemented medium were co-cultured with EG-7 or Online content
B16-OVA cells. Notably, Asn endowed OT-I CD8+ T cells with a Any methods, additional references, Nature Research report-
higher cytotoxicity and capacity to kill tumour cells (Fig. 7a and ing summaries, source data, extended data, supplementary infor-
Extended Data Fig. 9d). Furthermore, following intravenous injec- mation, acknowledgements, peer review information; details of
tion of B16 cells, the Asn-treated mice had an increased percent- author contributions and competing interests; and statements of
age of activated CD8+ T cells, and effector and central memory-like data and code availability are available at https://doi.org/10.1038/
T cells (Extended Data Fig. 9e–j). Conversely, ASNase treatment s41556-020-00615-4.
reduced CD8+ T-cell activation, the memory-like T-cell population
and cytokine production (Extended Data Fig. 9k–p). Received: 2 April 2020; Accepted: 26 November 2020;
Remarkably, when adoptively transferred into immunode- Published online: 8 January 2021
pleted mice injected with B16-OVA cells, Asn-treated OT-I CD8+
T cells exhibited a superior anti-tumour efficacy, as measured by References
the reduced tumour-cell metastasis and increased survival of the 1. Pearce, E. L., Poffenberger, M. C., Chang, C. H. & Jones, R. G. Fueling
immunity: insights into metabolism and lymphocyte function. Science 342,
mice (Fig. 7b,c and Extended Data Fig. 9q). Conversely, Asn star- 1242454 (2013).
vation endowed OT-I CD8+ T cells with reduced anti-tumour 2. MacIver, N. J., Michalek, R. D. & Rathmell, J. C. Metabolic regulation of T
activity (Fig. 7d). Similarly, Asn rendered OT-I T cells with an lymphocytes. Annu Rev. Immunol. 31, 259–283 (2013).
increased capacity to suppress the growth of subcutaneous xeno- 3. Wang, R. & Green, D. R. Metabolic checkpoints in activated T cells.
grafted tumours (Extended Data Fig. 9r). To assess the effect of Asn Nat. Immunol. 13, 907–915 (2012).
4. Wang, R. et al. The transcription factor Myc controls metabolic
on the endogenous anti-tumour immune response, we subcutane- reprogramming upon T lymphocyte activation. Immunity 35, 871–882 (2011).
ously injected B16-OVA cells into mice pretreated with PBS or Asn. 5. Gubser, P. M. et al. Rapid effector function of memory CD8+ T cells requires
The mice that were treated with Asn displayed strong anti-tumour an immediate-early glycolytic switch. Nat. Immunol. 14, 1064–1072 (2013).
immune responses and increased survival (Fig. 7e–g). Moreover, 6. Rathmell, J. C., Vander Heiden, M. G., Harris, M. H., Frauwirth, K. A. &
Thompson, C. B. In the absence of extrinsic signals, nutrient utilization by
naive OT-I CD8+ T cells primed by OVA in immunodepleted and
lymphocytes is insufficient to maintain either cell size or viability. Mol. Cell 6,
tumour-bearing mice that were intraperitoneally injected with 683–692 (2000).
Asn also mounted an effective anti-tumour response (Fig. 7h,l 7. van der Windt, G. J. et al. Mitochondrial respiratory capacity is a critical
and Extended Data Fig. 9s). Specifically, Asn endowed these regulator of CD8+ T cell memory development. Immunity 36, 68–78 (2012).

Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology 85

Articles NATuRE CELL BIoLogy
8. Pearce, E. L. et al. Enhancing CD8 T-cell memory by modulating fatty acid 22. Krall, A. S., Xu, S., Graeber, T. G., Braas, D. & Christofk, H. R. Asparagine
metabolism. Nature 460, 103–107 (2009). promotes cancer cell proliferation through use as an amino acid exchange
9. O’Sullivan, D. et al. Memory CD8+ T cells use cell-intrinsic lipolysis to factor. Nat. Commun. 7, 11457 (2016).
support the metabolic programming necessary for development. Immunity 23. Chi, H. Regulation and function of mTOR signalling in T cell fate decisions.
41, 75–88 (2014). Nat. Rev. Immunol. 12, 325–338 (2012).
10. O’Sullivan, D., Sanin, D. E., Pearce, E. J. & Pearce, E. L. Metabolic 24. Meng, D. et al. Glutamine and asparagine activate mTORC1 independently of
interventions in the immune response to cancer. Nat. Rev. Immunol. 19, Rag GTPases. J. Biol. Chem. 295, 2890–2899 (2020).
324–335 (2019). 25. Pavlova, N. N. et al. As extracellular glutamine levels decline, asparagine
11. Wei, J., Raynor, J., Nguyen, T. L. & Chi, H. Nutrient and metabolic sensing in becomes an essential amino acid. Cell Metab. 27, 428–438 (2018).
T cell responses. Front Immunol. 8, 247 (2017). 26. Brownlie, R. J. & Zamoyska, R. T cell receptor signalling networks: branched,
12. Ma, E. H. et al. Serine is an essential metabolite for effector T cell expansion. diversified and bounded. Nat. Rev. Immunol. 13, 257–269 (2013).
Cell Metab. 25, 482 (2017). 27. Wolfson, R. L. et al. Sestrin2 is a leucine sensor for the mTORC1 pathway.
13. Geiger, R. et al. l-Arginine modulates T cell metabolism and enhances Science 351, 43–48 (2016).
survival and anti-tumor activity. Cell 167, 829–842 (2016). 28. Kelly, B. & Pearce, E. L. Amino assets: how amino acids support immunity.
14. Munn, D. H. & Mellor, A. L. Indoleamine 2,3 dioxygenase and metabolic Cell Metab. 32, 154–175 (2020).
control of immune responses. Trends Immunol. 34, 137–143 (2013). 29. Chang, C. H. et al. Metabolic competition in the tumor microenvironment is
15. Schluns, K. S. & Lefrancois, L. Cytokine control of memory T-cell a driver of cancer progression. Cell 162, 1229–1241 (2015).
development and survival. Nat. Rev. Immunol. 3, 269–279 (2003). 30. Richards, N. G. & Kilberg, M. S. Asparagine synthetase chemotherapy.
16. Shiow, L. R. et al. CD69 acts downstream of interferon-α/β to inhibit S1P1 Annu. Rev. Biochem. 75, 629–654 (2006).
and lymphocyte egress from lymphoid organs. Nature 440, 540–544 (2006). 31. Deng, L. et al. p53-mediated control of aspartate-asparagine homeostasis
17. Best, J. A. et al. Transcriptional insights into the CD8+ T cell response to dictates LKB1 activity and modulates cell survival. Nat. Commun. 11,
infection and memory T cell formation. Nat. Immunol. 14, 404–412 (2013). 1755 (2020).
18. Surh, C. D., Boyman, O., Purton, J. F. & Sprent, J. Homeostasis of memory 32. Knott, S. R. V. et al. Asparagine bioavailability governs metastasis in a model
T cells. Immunol. Rev. 211, 154–163 (2006). of breast cancer. Nature 554, 378–381 (2018).
19. Kaech, S. M. & Cui, W. Transcriptional control of effector and memory CD8+ 33. Vodnala, S. K. et al. T cell stemness and dysfunction in tumors are triggered
T cell differentiation. Nat. Rev. Immunol. 12, 749–761 (2012). by a common mechanism. Science 363, eaau0135 (2019).
20. Boyman, O. & Sprent, J. The role of interleukin-2 during homeostasis and
activation of the immune system. Nat. Rev. Immunol. 12, 180–190 (2012). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
21. Harty, J. T. & Badovinac, V. P. Shaping and reshaping CD8+ T-cell memory. published maps and institutional affiliations.
Nat. Rev. Immunol. 8, 107–119 (2008). © The Author(s), under exclusive licence to Springer Nature Limited 2021

86 Nature Cell Biology | VOL 23 | January 2021 | 75–86 | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Methods supplemented with the same type and concentration of amino acids as RMPI-1640
Antibodies and reagents. The antibodies that were used in this work are listed (Macgene, CM10041), 10% FBS and/or Asn (100 µM, unless otherwise indicated).
in Supplementary Table 1. The following reagents were purchased from the Specific assays used 10% dialysed FBS-supplemented media containing the
indicated sources: mouse IL-2 (R&D, 402-ML), EasySep mouse biotin selection indicated concentrations of amino acids. Moreover, T-cell special culture medium
kit (Stem Cell, 18556), CD8+ T isolation kit (Stem Cell, 19853), l-Gln (Corning, was supplemented with 50 mg ml−1 streptomycin, 50 U ml−1 penicillin and, unless
25-005-CIR), l-Asn (ProSpec, ENZ-287), l-Albizziine (Goldbio, A-230-250), CFSE otherwise indicated, 10 ng ml−1 IL-2. CD8+ T cells were stimulated with anti-CD3
(Invitrogen, C34554), Percoll (GE Healthcare, 17089101), Ficoll (GE Healthcare, (3.5 µg ml−1) and anti-CD28 (1 µg ml−1) antibodies for 40 h. H1299 and EG-7 cells
17144002), Taselisib (GDC-0032; Selleck, 1282512-48-4), PP2 (Selleck, 172889- were cultured in RPMI-1640 medium (Macgene, CM10041) supplemented with
27-9), poly(Glu4-Tyr) peptide biotin conjugate (Millipore, 12-440). Brain heart 10% FBS. B16-F10 and B16-OVA cells were cultured in DMEM medium (Gibco,
infusion medium (Solarbio, A0360) and mouse special food (Asn-and-Asp-free) C11995) supplemented with 10% FBS. All of the cell lines that were used have been
were purchased from Changzhou Shuyishuer Bio-tech Co., Ltd. The following authenticated and were tested for mycoplasma contamination.
reagents were purchased from Sigma: l-Asn monohydrate (A8381), l-Asp (A9256),
l-Ser (S4311), l-Thr (T8441), l-Trp (T8941), l-Pro (P5607), l-Arg (A8094), l-Tyr Isolation and stimulation of T cells. Mouse T cells were isolated from the spleens
disodium salt hydrate (T1145), l-Ala (A7469), l-Cys (C7352), l-His HCl (H5659), and lymph nodes of the mice using a biotin selection kit (Stem Cell). Human
l-Ile (I7403), l-Leu (L8912), l-Lys HCl (L8662), l-Met (M5308), l-Phe (P5482), T cells were isolated from peripheral blood mononuclear cells by Ficoll gradient
l-Val (V4638), N-acetyl-l-Glu (855642), Gly (V900144), l-Asn-(amide-15N) centrifugation (Ficoll-Paque PLUS, GE Healthcare). The CD8+ T cells were,
monohydrate (485896), anti-FLAG M2 affinity gel (A2220), FLAG peptide (F3290), unless otherwise indicated, stimulated with 3.5 µg ml−1 anti-CD3 and 1 µg ml−1
collagenase IV (C5138) and Freund’s complete Adjuvant (F5581). anti-CD28 antibodies for 40 h in the presence of 100 µM Asn, 1 µg ml−1 ASNase
or PBS. The OT-I CD8+ T cells from OT-I mice were stimulated with 1 µg ml−1
Semi-quantitative PCR with reverse transcription. Total RNA was isolated OVA257–264(SIINFEKL) peptide.
using a total RNA purification kit (GeneMark, TR01). The RNA (2 µg) was reverse
transcribed to complementary DNA using the First-strand cDNA synthesis system Western blot analysis of TCR signalling in CD8+ T cells. Generally, CD8+ T cells
(Thermo Scientific, K1621). For each sample, 0.2 µg cDNA product was used as a were activated in medium (10% dialysed FBS) lacking Asn and Asp, supplemented
template to conduct quantitative PCR. The primer pairs that were used are listed in with or without Asn at 37 °C for 30 min, and further stimulated with 10 μg ml−1
Supplementary Table 2. Quantitative PCR was performed using SYBR Green PCR anti-CD3 and 1 μg ml−1 anti-CD28 antibodies for 15 min at 37 °C. The cells were
master mix (GenStar, A301-10) on a CFX96 real-time PCR system (Bio-Rad). Gene then cooled on ice, followed by centrifugation at 3,000 r.p.m. for 4 min at 4 °C. The
expression of the gene of interest was normalized to that of β-actin. samples were analysed by western blotting.

Animals and human samples. Six- to eight-week-old C57BL/6 mice were purchased Metabolic flux and LC–MS/MS analysis. Naive CD8+ T cells were seeded
from Vital River Laboratory Animal Technology for CD8+ T-cell isolation. The mice into 48-well plates in control medium (lacking Asn and Asp) or medium
were maintained under a 12/12 h light/dark cycle at 22–26 °C and were fed with supplemented with 100 µM Asn or 4 mM 15N-Asp. After activation, the CD8+
sterile pellet food and water ad libitum. The animal facilities used in this study have T cells were washed twice with PBS buffer and subjected to cell counting using a
been accredited by the Association for Assessment and Accreditation of Laboratory TC20 automated cell counter (Bio-Rad). Metabolites were extracted using 100%
Animal Care International and Institutional Animal Care and Use Committee acetonitrile and analysed using the Multi-reaction monitoring mode (MRM) of
(IACUC) of Tsinghua University. All of the mice were maintained in pathogen-free UPLC-QQQ-MS/MS (Agilent 1290/6460 tandem mass spectrum). An ACQUITY
facilities and used strictly in accordance with the protocols approved by the IACUC UPLC BEH HILIC, 2.1 mm × 100 mm, 1.7 µm (Waters) column was used for liquid
of Tsinghua University. All of the animals used were matched for age and sex and chromatography separation. Gradient elution was performed using 0.1% formic
randomly allocated to experimental groups. The study is compliant with all of the acid acetonitrile (solvent A) and 0.1% formic acid water (solvent B). The settings
relevant ethical regulations regarding animal research and human participants. were as follows: 0–1 min, 80% A; 1–5 min, 80 to 50% A; 5–7 min, 50% A; 7–7.1 min,
Human peripheral blood mononuclear cells were obtained from healthy volunteers 50 to 80% A; and 7.1–10 min, 80% A. The injection volume was set to 10 µl and the
under approved protocols of the Peking Union Medical College Hospital Ethics flow rate 0.4 ml min−1. Each sample was run for 10 min.
Board and we obtained informed consent from all of the participants. Relevant
information on the human research participants is listed in Supplementary Table 3. Metabolomics. Naive CD8+ T cells were isolated and activated at different time
points in medium lacking Asn and Asp supplemented with or without Asn. The
LCK kinase activity assays by ELISA. To measure the phosphorylation of Tyr394 cells were then washed twice with ice-cold PBS and metabolites were extracted
on LCK, recombinant LCK proteins pre-incubated with a buffer containing 60 mM using ice-cold 80% methanol. The extracts were analysed by LC–MS/MS.
HEPES (pH 7.5), 5 mM MgCl2, 5 mM MnCl2, 3 µM NaVO4, 1 mM dithiothreitol and
with or without 5 µg ml−1 Asn were incubated at 37 °C for 15 min. ATP was then Protein expression and purification. PRK5–FLAG plasmids coding for the
added to the mixtures to a final concentration of 500 µM and incubated for 15 min full-length mouse LCK, LCK(Y505F), LCK(Y394F), ZAP70 or SLP76 were
at 37 °C. The reaction mixtures were boiled for 10 min at 98 °C and transferred into transfected into 293T cells. After 48 h of transfection, the cells were harvested,
a 96-well ELISA plate (100 µl per well) to detect the phosphorylation of Tyr394 lysed using lysis buffer (PBS, 1% Triton X-100 and proteinase inhibitors) on ice for
on LCK. For the ELISA, a 96-well ELISA plate with the mixtures was incubated at 10 min and then centrifuged at 13,000 r.p.m. for 10 min at 4 °C. The supernatants
37 °C for 2 h before the removal of the supernatant; the plate was then blocked with were incubated with FLAG M2 beads, which were washed three times with wash
100 µl 5% BSA for 1 h at 37 °C. After three washes in PBST buffer, 100 µl mouse buffer (PBS, 150 mM NaCl and 1% Triton X-100), followed by competitive elution
anti-LCK p-Tyr394 antibody (1:2,000) was added and incubated for another 1 h with FLAG peptide PBS solution.
at 37 °C. Next, the antibody was washed out with 200 µl PBST, goat anti-mouse For purification of recombinant proteins from the Escherichia coli
IgG-HRP was added (100 µl per well) and incubated for 1 h before the addition BL21 strain, PET-21B-strep-LCK, PET-21B-strep-LCK(Y394F) and PET-
of TMB buffer and subsequently TMB stop buffer according to manufacturer’s 21B-strep-LCK(Y505F) were transformed into competent BL21 cells. The
manual (BioLegend, 421101). The absorbance at 450 nm was measured using a cells were then cultured in medium containing 200 g ml−1 ampicillin at 37 °C
microplate reader (Molecular Devices, SpectraMAX M2e). with shaking at 250 r.p.m. Protein expression was induced by adding IPTG to
For the determination of LCK activity by substrate phosphorylation of a final concentration of 1 mM at 18 °C for 18 h. The bacteria were collected
biotin-poly (Glu4: Tyr, 4:1), LCK proteins (1 µg ml−1) were incubated with 5 µg ml−1 and ultrasonicated in a lysis buffer containing 125 mM NaCl, 25 mM Tris–HCl
poly (Glu4: Tyr, 4:1)-biotin in a buffer containing 60 mM HEPES (pH 7.5), 5 mM (pH 7.5) and 1 mM PMSF. After centrifugation at 16,000 r.p.m. for 1 h at 4 °C, the
MgCl2, 5 mM MnCl2, 3 µM NaVO4 and 1 mM dithiothreitol in the presence of supernatants were incubated with streptavidin-conjugated agarose beads for 4 h at
100 µM Asp, 100 µM Asn or vehicle control. After an incubation of 15 min at 37 °C, 4 °C. The beads were then washed three times with lysis buffer. The bound proteins
ATP was added and incubated for another 15 min, followed by boiling at 98 °C. were eluted using Strep elution buffer (IBA, 2-1042-025).
The ELISA plate was then loaded with 50 µg ml−1 streptavidin, blocked with 5%
BSA and washed three times with PBST. The LCK reaction mixtures were then Radiolabelled amino-acid binding assay. Purified proteins on beads were incubated
transferred into the plate wells (100 µl per well) and incubated for 15 min. Mouse with 3 µCi 3H-labelled amino acids and/or 10 mM cold amino acids for 20 min at
anti-phospho-Tyr antibody 4G10 (100 µl; 1:5,000) was then added to the wells for room temperature. The tubes were gently shaken every 5 min during incubation.
a further incubation at 37 °C, followed by washing with PBST and incubation with The beads were spun down briefly and washed three times with wash buffer (0.1%
100 µl goat anti-mouse IgG-HRP for another 1 h. The LCK activity was measured Triton X-100, 40 mM HEPES pH 7.4 and 150 mM NaCl), followed by drying and
and calculated (the absorbance at 450 nm) using a microplate reader. resuspension in 50 µl wash buffer. The suspension (10 µl) was added into an Isoplate-
96-plate well and quantified using a MicroBETA2 scintillation counter (PerkinElmer).
Cell lines and culture conditions. B16-F10 and 293T cells were purchased from The remaining beads were then boiled in 50 µl sample buffer and resolved by 10%
the American Type Culture Collection. The B16-OVA cells were a gift from the C. SDS–PAGE. The proteins were visualized by Coomassie blue staining.
Xu laboratory (CAS, Shanghai). H1299 cells and EG-7 cells were provided by Y.
Wang (Tsinghua University, Beijing) and Z. Dong (Tsinghua University, Beijing), BIAcore analysis. The purified proteins were immobilized on a Series S Sensor
respectively. CD8+ T cells were cultured in DMEM medium (Macgene, CM15022) chip CM7 (GE Healthcare) in 10 mM sodium acetate buffer (pH 4.5). The amino

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy
acids were dissolved in PBS buffer to generate gradient concentrations in PBS buffer containing 4% paraformaldehyde and 0.01% Triton X-100. The fixed cells
running buffer. The mixtures were run across each sensor surface at a flow rate were subjected to anti-TNF-α, anti-IFN-γ, anti-GzmB and anti-Ki67 staining.
of 10 µl min−1 for a contact phase of 90 s, followed by a dissociation phase of 120 s
in running buffer. The results were calculated using the BIAcore S200 Evaluation Lung-metastatic melanoma model. Mice were treated by γ-irradiation (2 Gy,
software Version 1.0 (GE Healthcare). twice) and received 4 × 105 OT-I cells by tail-vein injection 24 h later. The mice
were immunized with SIINFEKL peptide (OVA257–264; 20 µg per mouse) in Freund’s
CD8+ T-cell cytotoxicity assay. OT-I CD8+ T cells were activated in medium complete adjuvant the next day. Asn (1 mmol per mouse) or PBS was intraperitoneally
lacking Asn and Asp supplemented with or without 100 µM Asn for 4 d. injected into the mice every 2 d, starting one day before immunization for a total of
The cells were then cultured with pre-loaded B16-OVA or EG-7 cells for 18 h. 14 d. During this period, the mice were fed a diet lacking Asn and Asp; otherwise, the
The cytotoxic efficiency of the CD8+ T cells was measured using a lactate animals were fed a normal diet. Animal survival was monitored. To study tumour
dehydrogenase A release assay using a microplate reader (Molecular Devices, growth, the mice were euthanized on day 17 and the metastatic tumours were
SpectraMAX M2e). counted. Lung-infiltrating T cells were isolated and analysed as described earlier.

Proliferation analysis by CFSE staining. The CFSE staining was performed Confocal immunofluorescence imaging. Imaging was performed on a custom
according to the manufacturer’s manual (Invitrogen). Briefly, the isolated CD8+ modified Olympus FV1200 Laser Scanning Microscope equipped with a ×100
T cells were resuspended in 0.1% BSA (in PBS) at a final concentration of oil immersion lens. Briefly, naive CD8+ T cells were seeded in chambers and
1–5 × 106 cells ml−1. The cells were incubated with CFSE solution (1 µM) at 37 °C stimulated with 10 μg anti-CD3 and 1 μg anti-CD28 antibodies for 10 min at 37 °C.
for 10 min; the staining was stopped by adding five volumes of ice-cold culture The cells were then fixed with 4% paraformaldehyde (PFA) and 0.1% Triton X-100
medium and the cells were incubated for another 5 min on ice. The cells were spun for 1 h at 4 °C, followed by 5% BSA blocking for 30 min at 37 °C. The cells were
down and washed at least twice with fresh medium. The cells were then loaded into gently washed three times with PBS and stained overnight with the indicated
a 48-well plate with Asn-free or -supplemented medium. After 3 d, the cells were antibody at 4 °C. After three washes with PBS, the cells were incubated with the
analysed using flow cytometry. secondary antibody for 1 h at 4 °C, and gently washed three times with PBS before
imaging. The images were analysed using Image J.
Drug affinity responsive targets stability. Jurkat T cells were cultured in Asn-free
medium for 12 h. Approximately 5 × 106 cells were lysed in 150 µl M-PER (Thermo Statistics and reproducibility. The experiments in this study were set up using
Fisher, 78501) with 1×protease inhibitor cocktail. The cell lysates (100 µl) were 3–10 samples or animals per independent group, condition or repeat. Each
transferred to a new tube, incubated with various concentration of Asn or PBS for experiment was repeated independently with similar results. Immunoblot
1 h at room temperature and digested with Pronase (6 µg ml−1) for 15 min at room detection, quantitative PCR with reverse transcription, kinase activity
temperature. After 15 min, the digestion was stopped by adding 5×SDS loading measurements, and the immunoprecipitation and immunofluorescence staining
buffer and the samples were boiled at 98 °C for 10 min. Western blotting was used experiments were independently repeated at least three times. The FACS gating
to determine the LCK protein levels. strategies are described in Extended Data Fig. 10. Representative dot plots for
the FACS analysis are provided in the Supplementary figures. All of the data
Cellular thermal shift assay. Jurkat T cells were cultured in Asn-free medium for are presented as the mean ± s.e.m. and exact n values are indicated in the figure
12 h. The cells (5 × 106) cells were then pretreated with or without 400 μM Asn for legends. The statistical analyses were performed using GraphPad Prism 7.
12 h before use in the cellular thermal shift assays assay. The cells were cooled on Comparisons were analysed using an unpaired two-tailed Student’s t-test or a
ice, washed with PBS containing protease inhibitor cocktail and transferred into two-way ANOVA, followed by Tukey’s multiple comparisons test, as indicated in
PCR tubes. The cells were then heat shocked in a Bio-Rad T100 thermal cycler the corresponding figure legends. The P values are indicated in the related figures;
at the indicated temperature (60 to 40 °C) for 3 min to denature the proteins, and *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and NS, not significant (P > 0.05).
immediately cooled on ice. The cells were lysed by three freeze–thaw cycles with
liquid nitrogen and a thermal cycler, centrifuged at 13,000 r.p.m. for 10 min at 4 °C Reporting Summary. Further information on research design is available in the
and the supernatant was boiled with 5×SDS loading buffer for western blotting. Nature Research Reporting Summary linked to this article.
The bands were quantified using the Image J software.

Translation assay by puromycin. Isolated naive CD8+ T cells were stimulated Data availability
in the indicated medium with 3.5 µg ml−1 anti-CD3 and 1 μg ml−1 anti-CD28 All data supporting the findings of this study are available from the corresponding
antibodies for 24 h. Puromycin (40 µg ml−1) was then added 10 min before sample author on reasonable request. Source data are provided with this paper.
collection. The cells were stained by FITC-conjugated anti-puromycin antibody
and puromycin incorporation was measured by FACS analysis. Acknowledgements
We thank X. Lin, L. Yu, X. Guo, L. Zhang, Y. Wang and X. Hu at Tsinghua University;
Adoptive T-cell transfer. LmOVA were cultured in Brain Heart Infusion Medium H. Wu at Peking University; W. Du at Peking Union Medical College and C. Xu at
(BHI). Briefly, the mice were immunized intravenously with recombinant Shanghai Institute of Biochemistry for materials and/or technical assistance. We thank
attenuated LmOVA (2 × 105 colony-forming units (c.f.u.)). OT-I CD8+ T cells H. Wang and W. Wu at Tsinghua University for their great support of this work during
(CD45.2+; 2 × 105) were intravenously injected into the tail-vein of CD45.1+ mice, the COVID-19 pandemic. We thank all of the members of the Jiang laboratory for their
followed by LmOVA infection 1 d later. Splenocytes were isolated 7 d post OT-I technical assistance and/or discussions. We thank X. Liu, L. Xu, X. Wang and W. Wang
CD8+ T-cell injection and analysed for the presence of OVA-specific CD8+ T cells for their help with the LC–MS/MS experiments. This research was supported by the
by MHC class I tetramer (Kb/OVA257–264). National Key R&D Program of China (grant no. 2019YFA0801701), Tsinghua-Peking
For the memory rechallenge experiments, mice maintained on a specific diet Center for Life Sciences and National Natural Science Foundation of China (grant nos
were immunized with attenuated LmOVA, followed by rechallenge with LmOVA 81930082 and 81722035) to P.J.
(2 × 106 c.f.u.) 21 d after the primary immunization. Splenocytes were isolated 5 d
post the rechallenge and analysed as described earlier.
Author contributions
J.W., G.L. and P.J. designed the experiments. J.W. and G.L. performed all of the
In vivo priming of T cells for tumour experiments. Tumour cells (8 × 105) were
experiments. L.L. and D.L. provided technical assistance. Z.D. provided constructive
subcutaneously injected into the dorsal region of C57BL/6 mice. The mice were
comments and supplied reagents. P.J. supervised the research and wrote the manuscript.
immunodepleted by irradiation (2 Gy, twice) 5 d post injection. After another 24 h,
All authors commented on the manuscript.
4 × 105 OT-I cells were adoptively transferred into these mice by tail-vein injection.
The mice were immunized with SIINFEKL peptide (OVA257–264; 20 µg per mouse)
in Freund’s complete adjuvant the following day. Asn (1 mmol per mouse), ASNase Competing interests
(0.2 mg per mouse) or PBS was intraperitoneally injected into the mice every 2 d, The authors declare no competing interests.
starting one day before immunization and lasting for a period of 14 d. During
this period, the mice were maintained on a diet lacking Asn and Asp or normal
diet. Animal survival was recorded daily. Tumour size (s) was calculated using the Additional information
formula s ¼ πr2 , where r is the radius of the tumour. Extended data is available for this paper at https://doi.org/10.1038/s41556-020-00615-4.
I Supplementary information is available for this paper at https://doi.org/10.1038/
Tumour-infiltrating CD8+ T-cell responses. The tumour was digested with s41556-020-00615-4.
collagenase IV and the tumour-infiltrating leukocytes were isolated by 40–80% Correspondence and requests for materials should be addressed to P.J.
Percoll (GE Healthcare) gradient centrifugation. The isolated cells were stimulated
with 50 ng ml−1 phorbol 12-myristate 13-acetate and ionomycin in the presence of Peer review information Nature Cell Biology thanks Vijay Kuchroo and the other,
5 μg ml−1 Brefeldin A (BFA) for 4 h. The activation of CD8+ T cells were analysed by anonymous, reviewer(s) for their contribution to the peer review of this work.
anti-mCD4, anti-mCD8 and anti-mCD3 ζ staining, followed by fixation using PBS Reprints and permissions information is available at www.nature.com/reprints.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles

Extended Data Fig. 1 | See next page for caption.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 1 | In vitro activation of CD8+ T cells by Asn. a, 15N-Asp levels outside or within mouse naive CD8+ T cells activated in medium
containing 4 mM 15N-Asp for 40 h. b-d, Surface expression of CD25 (b), CD44 (c), or CD69 (d) on CD8+ T cells stimulated in the presence of increasing
amounts of Asn. e-g, Percentage of CD25 (e), CD44 (f), or CD69 (g) expression on CD8+ T cells stimulated with CD28 antibody (1 μg/ml) and increasing
amounts of anti-CD3 antibody in the presence or absence of Asn. h, CFSE staining of mouse naive CD8+ T cells activated in Asn/Asp-free medium
(Ctrl) or medium containing Asn for indicated time points. i and j, Proliferation of mouse CD8+ T cells treated as in (b) was determined by Ki67 staining
(i), or BrdU incorporation assay (j). mFI, mean fluorescence intensity. k and l, Surface expression of CD25, CD44 and CD69 on mouse OT-I CD8+ T
cells activated by OVA peptide in Asn/Asp-free medium (Ctrl) or Asn medium (k). Production of cytokines was also determined (l). m and n, Surface
expression of CD69 (m) and CD44 (n) on mouse naive CD8+ T cells activated in Asn/Asp-free (Ctrl) or Asn medium containing 10% dialysed FBS.
o-q, Surface expression of indicated proteins on mouse naive CD8+ T cells activated in Asn/Asp-free (Ctrl) or Asn medium containing no FBS. r and s,
Representative flow cytometry plots of FSC-A versus TNF-α (r), or IFN-γ (s) staining for human CD8+ T cells treated as in Fig. 1k. Data are mean ± SEM,
two-tailed Student’s t-test. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant. All experiments were done at least 3 times per group or condition.
Related representative FACS plots (Supplementary Figs. 9, 10) and numerical source data for a-g, i, k-p are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles

Extended Data Fig. 2 | Effect of ASNase, Glu and Asn on CD8+ T cell responses. a-d, Mouse CD8+ T cells cultured in RPMI 1640 medium, or medium
containing 10 µg/ml ASNase were left untreated (naive) or stimulated with anti-CD3/CD28 antibodies for 4 h or 12 h. Relative levels of Asn, Asp, Gln
and Glu within and outside cells were measured by LC-MS/MS. e and f, Mouse naive CD8+ T cells stimulated with anti-CD3/CD28 antibodies in control
Glu-free medium (Ctrl) or medium containing Glu in the presence of dialysed FBS. Percentage of CD44+CD8+ T cells (e), CD62LlowCD44high CD8+ T cells
and CD62LhighCD44high CD8+ T cells (f) relative to total CD8+ T cells were measured. g-j, Jurkat cells were stimulated with anti-CD3/CD28 antibodies in
the absence or presence of L-Alb, PP2, or increasing amounts of Asn for 24 h (for detecting CD69 expression) or 40 h (for detecting CD25 expression).
Surface expression of CD69 (c) or CD25 (d) was measured. Production of TNF-α (e), and IFN-γ (f) by the activated Jurkat cells was determined by
FACS analysis. All data are mean ± SEM, n = 3 independent wells. Two-tailed Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not
significant. Representative FACS plots for e-j (Supplementary Fig. 11) and numerical source data for a-j are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 3 | See next page for caption.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Extended Data Fig. 3 | Effect of extracellular and de novo Asn on T activation. a, mRNA expression in naive CD8+ T cells. b and c, Mouse naive CD8+ T
cells were left untreated (naive) or activated for different time points. ASNS expression (b) and Asn levels (c) were measured. d, Surface expression of
CD25 and CD44 on pre-activated mouse naive CD8+ T cells cultured in Asn/Asp-free (Ctrl) or Asn medium. e-h, Mouse naive CD8+ T cells were activated
as indicated. Intracellular (e) and extracellular (f) Asn levels, surface expression of CD25 and CD44 (g), and cytokine production (h) were measured. i
and j, Survival (i) and apoptosis (j) of mouse naive CD8+ T cells activated in Asn/Asp-free (Ctrl) or Asn medium containing exogenous IL-2. k, Annexin
V staining of naive CD8+ T cells activated in Asn/Asp-free (Ctrl) or Asn medium without exogenous IL-2. l-n, Naive CD8+ T cells pre-activated in 1640
IL-2-supplemented medium were cultured in IL-2-free medium without (Ctrl) or with Asn. Living cells (l), apoptotic (m) and proliferating (n) T cells were
assayed at day 3 post IL-2 withdrawal. o and p, Mouse naive CD8+ T cells activated in ASNase-supplied or -free conditioned medium from H1299 cells
or Asn/Asp-free medium (Ctrl) were analysed for activation and Asn levels. q, Asn levels in medium cultured from Gln-fed or starved H1299 and EG-7
cells. r, Mouse naive CD8+ T cells were stimulated in medium cultured from Gln-fed or starved EG-7 cells in the presence of 2 mM Gln. CM, conditioned
medium. Data are mean ± SEM, n = 3 independent samples per group/condition. Two-tailed Student’s t-test or a two-way ANOVA followed by Sidak’s
multiple comparisons (i). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Related representative FACS plots (Supplementary Fig. 12), uncropped blots
for b, and numerical source data for a-m, o-r, are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 4 | See next page for caption.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Extended Data Fig. 4 | Increasing Asn promotes T cell activity in vivo. a-j, C57BL/6 mice intraperitoneally injected with Asn or PBS (Ctrl) every two days
for a period lasting one month. Spleens were excised, weighted and pictured (a). Relative numbers of cells in spleens and lymph nodes (b), weight of mice
(c), Percentage of CD4+ T cells and CD8+ T cells relative to total cells in spleens and lymph nodes (d, e), effector (e) and central memory-like (f) CD8+ T
cells relative to total CD8+ T cells in spleens and lymph nodes, CD25+CD4+ T cells (h), and effector (i) and central memory-like (j) CD4+ T cells relative to
total CD4+ T cells in spleens and lymph nodes were assessed. k-r, C57BL/6 mice intraperitoneally injected with ASNase or PBS (Ctrl) every two days for a
period lasting one month. Spleens were excised, weighted and pictured (k). Relative numbers of cells in spleens and lymph nodes (l), percentage of CD4+
T cells and CD8+ T cells relative to total cells in spleens and lymph nodes (m), effector memory-like (n) and central memory-like (o) CD8+ T cells relative
to total CD8+ T cells in spleens and lymph nodes, percentage of CD25+CD4+ T cells (p), effector memory-like (q) and central memory-like (r) CD4+ T
cells relative to total CD4+ T cells in spleens and lymph nodes were assessed. s, Surface expression of CD25 and CD44 on mouse naive CD8+ T cells
activated with serum from normal diet mice intraperitoneally injected with PBS (Ctrl) or ASNase (n = 6 mice per group). All data are mean ± SEM, in a-r,
Ctrl, n = 7 mice; Asn, n = 6 mice. Two-tailed Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant. Related representative
FACS plots (Supplementary Figs. 13, 14) and numerical source data for a-s are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 5 | See next page for caption.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Extended Data Fig. 5 | Asn enhances OVA-induced CD8+ T cell activation in vivo and a role for SLC1A5 in Asn uptake in CD8+ T cells. a-i, C57BL/6
mice orally administrated with normal (Ctrl) or Asn-supplemented water (Asn) were infected with LmOVA for 7 days (a). spleens were excised, weighted
and pictured (b, c). Percentage of CD8+ T cells and CD4+ T cells (d-g), and OVA-TET+ CD8+ T cells in spleens and lymph nodes (h, i) were determined.
n = 5 mice per group. j-l, Related to Fig. 2m,n. Asn-free diet CD45.1 mice intraperitoneally injected with PBS (Ctrl) or Asn were adoptively transferred
with naive CD45.2+ OT-I CD8+ T cells and injected with B16-OVA cells as indicated (j). 16 days later, percentage and number (#) of CD44+CD8+ T cells
(k), and percentage of central and effector memory-liked CD8+ T cells relative to total CD8+ T cells in lymph nodes (l) were analysed (n = 5 mice per
group). m, mRNA levels of potential transporters for Asn uptake and exchange in Jurkat cells. n and o, SLC1A5 and SLC38A2 expression in mouse naive
CD8+ T cells activated in Asn/Asp-free (Ctrl) or Asn medium was analysed by qRT-PCR (n) and western blotting (o). p, Mouse CD8+ T cells activated
by anti-CD3(5 μg/ml) and anti-CD28(1 μg/ml) in Asn/Asp-free (-) or Asn medium supplied with 10% dialysed FBS and 2 mM Gln were treated with or
without 17 μM V9302 for 12 h. Surface expression of CD25 and CD69 relative to total CD8+ T cells were quantified. All data are mean ± SEM. Data in o
are representative of three independent experiments, two-tailed Student’s t-test. *P < 0.05; ***P < 0.001; ***P < 0.001; ***P < 0.0001; ns, not significant.
Related representative FACS plots (Supplementary Fig. 15), uncropped blots for o, and numerical source data for c-i, k-n, p, are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 6 | Effect of mTORC1 and Gln on Asn-mediated CD8+ T cell activation. a, Phosphorylation of the mTORC1 effector S6 (phospho-S6
Ser 235/236) in mouse CD8+ T cells left unstimulated (naive) or activated in the presence of increasing amount of Asn. b, Western blot analysis of
mouse naive CD8+ T cells left untreated (naive), or activated in Ctrl (Asn/Asp-free) medium or Asn medium supplemented with increasing amounts
of rapamycin for 12 h. c, Surface expression of CD69, CD25, and CD44 on mouse naive CD8+ T cells stimulated in Asn/Asp-free medium (Ctrl) or
Asn medium with 10% dialysed FBS and increasing amounts of rapamycin (n = 3 independent wells). d, Mouse naive CD8+ T cells stimulated in Asn/
Asp-free (-) or Asn-supplied (Asn) medium without (0) or with increasing amounts of Gln for 24 h were treated with 40 μg/ml Puromycin for another
10 min. Medium was added with 10% dialysed FBS. Puromycin incorporation was measured by FACS analysis using FITC-conjugated anti-Puromycin
antibody (n = 3 independent wells). Representative flow cytometry plots are shown. e, Mouse naive CD8+ T cells were stimulated in the indicated
medium with 10% dialysed FBS for 24 h. MG132 (10 μM) was added 2 h before harvesting for western blot analysis. f, Surface expression of CD69,
CD25, and CD44 on mouse naive CD8+ T cells stimulated in Asn/Asp-free (-) or Asn medium containing no (0) or increasing amounts of Gln for 24 h
(n = 3 independent wells). All medium was added with 10% dialysed FBS. All data are mean ± SEM, two-tailed Student’s t-test. *P < 0.05; **P < 0.01;
***P < 0.001; ***P < 0.0001; ns, not significant. Data in a, b and e are representative of three independent experiments. Related representative FACS plots
(Supplementary Figs. 16, 17), uncropped blots for a, b, e, and numerical source data for c, d, f, are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles

Extended Data Fig. 7 | Impact of Asn on CD8+ T cell metabolism and TCR signalling. a and b, Mouse naive CD8+ T cells were left untreated (naive) or
activated in Asn/Asp-free medium (Ctrl), or medium supplied with Asn or Asp for indicated times. All medium was added with 10% dialysed FBS. The
abundance of cellular metabolites at different time points were analysed by LC-MS/MS. The heat map shows the relative abundance of metabolites over
time in CD8+ T cells stimulated in Ctrl, Asn- or Asp-medium. Log2 fold changes (FC) are relative to relative to cells stimulated in Asn/Asp-free (Ctrl)
medium at corresponding time points. Data are representative of three independent experiments. Structural isomers were condensed and counted only
once as the employed technology cannot distinguish metabolites with identical molecular weight. c, TCR signalling that initiates T cell activation. d,
Related to Fig. 4e. Scheme of OT-I adoptive transfer and OVA-peptide treatment experiment. e, Western blot analysis of mouse CD8+ T cells activated in
medium containing increasing amounts of Asn in the absence or presence of PP2 (10 µM). Data are representative of three independent experiments. f,
Related to Fig. 5e–k. CD45.1 C57BL/6 mice maintained on an Asn-free diet were intraperitoneally injected with Asn or PBS (Ctrl), or PP2 (30 µg/mouse) or
GDC-0032 (100 µg/mouse) every two days. 8 days later, mice were adoptively transferred with naive CD45.2+ OT-I CD8+ T cells and injected with OVA
peptide as indicated. g, Related to Fig. 5l–o, Scheme of Asn-free diet CD45.1 C57BL/6 mice i.p. with Asn and/or PP2 and then adoptively transferred with
naive CD45.2+ OT-I CD8+ T cells and infected with LmOVA as indicated. Uncropped blots for e and numerical source data for a, b, are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 8 | See next page for caption.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Extended Data Fig. 8 | LCK is a natural sensor for Asn. a, Recombinant LCK proteins were treated with ATP and Asn or Asp. LCK kinase activity was
determined by ELISA. b and c, BIAcore measurement of the interaction between Asn or Asp and purified mutant LCK (Y394F) (b) or LCK (Y505F) (c).
d-u, Screening the binding specificity of recombinant LCK proteins to different amino acids by BIAcore assay. Graphs of equilibrium response units (RU)
and compound concentrations are shown. Gln, glutamine; Thr, threonine; Isoleu, isoleucine; Trp, tryptophan; Met, methionine; Leu, leucine; Ala, alanine;
Val, valine; Ser, serine; Pro, proline; Gly, glycine; His, histidine; Cys, cysteine; Tyr, tyrosine; Phe, phenylalanine; Lys, lysine; Glu, glutamate; Arg, arginine.
v, BIAcore assay of the interaction between purified ZAP70 proteins and Asn or Asp. w, CETSAs exhibit the binding affinity of LCK to Asn, but not Gln,
in Jurkat T cells. Relative LCK band intensities were quantified and plotted against corresponding incubation temperatures. x, Direct binding of LCK to
Asn, but not Gln in Jurkat cells measured by DARTS assays. y, In vitro kinase assay for LCK and its mutants in the presence of increasing amounts of Asn.
Kinase activity is expressed as relative values to phosphorylation of LCK(Y394). z, In vitro kinase assay using purified LCK and its mutants left untreated
(Ctrl), or incubated with ATP and increasing amounts of Asn for 15 min. In b-v, Graphs of equilibrium response units (RU) and compound concentrations
are shown. In b, c and v, Purified proteins were analysed by SDS-PAGE followed by Coomassie Blue staining. Data are mean ± SEM, in a, n = 3 biological
replicates. Two-way ANOVA followed by Sidak’s multiple comparisons. Data are representative of three independent experiments throughout. Uncropped
blots for x, z, and numerical source data for a-w, y, are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 9 | See next page for caption.

Nature Cell Biology | www.nature.com/naturecellbiology

NATuRE CELL BIoLogy Articles
Extended Data Fig. 9 | Asn binds to the LCK kinase domain and enhances CD8+ T cell anti-toumor activity. a and b, BIAcore measurement of the
interaction between Asn and purified region AA220-390 (a) or AA320-509 (b). Graphs of equilibrium response units (RU) and compound concentrations
are shown. c, Mapping the Asn binding site on LCK. The region responsible for Asn binding and results of Asn binding analysis are shown. d, OVA-specific
OT-I CD8+ T cells pre-activated in Asn/Asp-free (Ctrl) or Asn medium were cultured with B16-OVA cells for 18 h. B16-OVA cell apoptosis was analysed
(n = 3 independent wells). e-p, C57BL/6 mice maintained on a normal (Ctrl) or Asn diet (e-j), or on a normal diet yet subcutaneously injected with PBS
(Ctrl) or ASNase (0.2 mg/mouse) were subcutaneously injected with B16-F10 cells near the inguinal lymph node for 12 days. Percentage of CD44+CD8+
T cells (e, k), effector (f, l) and central memory-like (g, m), naive (h, n) CD8+ T cells, and production of cytokines (i, j, o, p) in lymph nodes were measured
(n = 7 or 8 mice per group). q, Related to Fig. 7b and c. Schematic view of the B16-OVA and activated OT-I CD8+ T cell adoptive transfer experiment.
r, Naive OT-I CD8+ T cells primed in vitro in Asn or Asn/Asp-free medium (Ctrl) were injected into C57BL/6 mice subcutaneously injected (sc) with
B16-OVA cells as indicated. Tumour burden were analysed over time. Data are representative of three independent experiments (n = 5 mice per group).
s, Schematic of the experimental approach in Fig. 7h–j. All data are mean ± SEM. In d-p, two-tailed Student’s t-test, and in r, two-way ANOVA followed by
Sidak’s multiple comparisons. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant. Related representative FACS plots (Supplementary
Fig. 18) and numerical source data for a, b, d-p, r, are provided.

Nature Cell Biology | www.nature.com/naturecellbiology

Articles NATuRE CELL BIoLogy

Extended Data Fig. 10 | Gating strategies used in FACS analysis. a, Gating strategy to identify APC+ in mouse spleen isolated CD8+ T cells which were
activation such as Fig. 1a and Supplemental Fig. 1b. b, Gating strategy to analyse CFSE-stained or BrdU-stained activation CD8+ T cells analysis presented
on Extended Data Fig. 1h,1j and 3n. c, Gating strategy to analyse annexin V stained activation CD8+ T cells in apoptosis analysis presented on Extended
Data Fig. 3k,m. d, Gating strategy to analyse the population of CD4+ T cells or CD8+ T cells in mouse lymph nodes or spleens presented on such as
Extended Data Fig. 4m. e, Gating strategy to analyse the population of OT-I CD8+ T cells in mouse lymph nodes or spleens presented on such as Figs. 2m,
2n, 5e–o, and Extended Data Fig. 5k,l.

Nature Cell Biology | www.nature.com/naturecellbiology

 ζ
γ
γ
γ
α
α

ζ
ζ