RESEARCH ARTICLE | IMMUNOLOGY AND INFLAMMATION OPEN ACCESS

Infection-elicited microbiota promotes host adaptation to

nutrient restriction
Mirian Krystel De Siqueiraa,1,2, Vinicius Andrade-Oliveirab,c,2,3 , Apollo Stacyb,c,2,4,5, João Pedro Tôrres Guimarãesa, Ricardo Wesley Alberca-Custodioa ,
Angela Castoldia, Jaqueline Marques Santosa, Marcela Davoli-Ferreiraa, Luísa Menezes-Silvaa, Walter Miguel Turatod, Seong-Ji Hanb,c,6, Arielle Glatman Zaretskyb,c,7,
Timothy Wesley Handb,8, Niels Olsen Saraiva Câmaraa, Momtchilo Russoa, Sonia Jancara, Denise Morais da Fonsecaa,2,5, and Yasmine Belkaidb,c,2,5

Contributed by Yasmine Belkaid; received October 21, 2022; accepted December 14, 2022; reviewed by Wendy S. Garrett and Manuela Raffatellu

The microbiota performs multiple functions vital to host fitness, including defense
against pathogens and adaptation to dietary changes. Yet, how environmental chal- Significance
lenges shape microbiota resilience to nutrient fluctuation remains largely unexplored.
Here, we show that transient gut infection can optimize host metabolism toward the The long-term impacts of infection
usage of carbohydrates. Following acute infection and clearance of the pathogen, mice on the microbiota and its
gained more weight as a result of white adipose tissue expansion. Concomitantly, pre- regulation of host physiology are
viously infected mice exhibited enhanced carbohydrate (glucose) disposal and insulin poorly understood. Here, we
sensitivity. This metabolic remodeling depended on alterations to the gut microbiota, report that long-term alterations
with infection-elicited Betaproteobacteria being sufficient to enhance host carbohydrate to the gut microbiota following a
metabolism. Further, infection-induced metabolic alteration protected mice against single, acute episode of bacterial
stunting in the context of limited nutrient availability. Together, these results propose
or protozoan gut infection can
that alterations to the microbiota imposed by acute infection may enhance host fitness
remodel host metabolism to
and survival in the face of nutrient restriction, a phenomenon that may be adaptive in
settings where both infection burden and food precarity are prevalent. preferentially and more efficiently
consume carbohydrates. This
host metabolism | white adipose tissue | carbohydrate | malnutrition | Yersinia infection-triggered metabolic
remodeling ultimately results in
The survival of living organisms relies on their ability to adapt to environmental fluc- white adipose tissue expansion
tuations and challenges. In this regard, the symbiotic alliance between the host and its and host weight gain.
microbiota is essential for optimal host physiology and fitness in the face of environ-
Furthermore, in the context of low
mental stressors (1–3). A vast array of biodiverse ecosystems and complex interaction
networks, the gut microbiota intimately shapes human physiology through its regulation limited nutrient availability,
of host immunity and metabolism (4). A principal trait of the gut microbiota is its infection-triggered carbohydrate
versatility in extracting energy from the biochemically diverse components (carbohy- metabolism benefits host fitness
drates, protein, and fat) of the omnivorous human diet (5). This versatility extends from by preventing host stunting. Our
the microbiota’s remarkable capacity to adapt both functionally and rapidly to host study suggests a new perspective
dietary changes (6). Because of this, one of the greatest environmental drivers of micro- in which infection (pathogen-
biota diversity and metabolic output is host diet, not only its components but also induced stress) can be co-opted as
fluctuations in its availability. For example, the malnutrition widespread in under-re- a cue to prime host adaptation to
sourced settings can compromise the gut microbiota’s development and diversity (7). nutritional stress.
In contrast, a high-fat Western diet can dramatically remodel the microbiota (6) and
promote metabolic syndrome, a cluster of conditions including high blood glucose, 2
M.K.D.S., V.A.-O., A.S., D.M.d.F., and Y.B. contributed
triglycerides, and body fat [white adipose tissue (WAT)] that increases risk for diabetes equally to this work.
and other health conditions (8). 3
Present Address: Center for Natural and Human Sciences,
While host diet plays an outsized role in regulating the microbiota, the microbiota can Federal University of ABC, Santo André SP 09210-580, Brazil.
in turn regulate host usage and storage of diet-derived energy. Illustrating this, germfree 4
Present Address: Department of Cardiovascular and
Metabolic Sciences, Lerner Research Institute, Cleveland
mice devoid of microbiota are protected against diet-induced obesity (9), accumulating Clinic, Cleveland, OH 44195.
less WAT than conventional mice (10). Furthermore, host metabolism can be regulated, 5
To whom correspondence may be addressed. Email:
either favorably or detrimentally, by defined taxa within the microbiota. For example, the [email protected], [email protected], or ybelkaid@
niaid.nih.gov.
mucus-degrading bacterium Akkermansia muciniphila protects against obesity and diabetes 6
Present Address: Jill Roberts Institute for Research in
(11), while the deltaproteobacterium Bilophila wadsworthia, which expands in response Inflammatory Bowel Disease, Joan and Sanford I. Weill
to fat-induced bile acids (12), can aggravate metabolic syndrome (13). Department of Medicine, Department of Microbiology and
Immunology, Weill Cornell Medicine, Cornell University,
Besides host diet, environmental stressors such as infection and antibiotics can also impact New York, NY 10021.
the microbiota. Indeed, in industrialized countries, changes in the infectious landscape combined 7
Present Address: Regeneron Pharmaceuticals, Tarrytown,
with nutritional alterations and the overuse of antibiotics are strongly associated with a decline NY 10591.
in gut microbiota diversity and the recent, concomitant rise in various inflammatory and met- 8
Present Address: R.K. Mellon Institute for Pediatric
Research, Pediatrics Department, Infectious Disease
abolic diseases (14). Accordingly, some degree of pathogen exposure may benefit the host by Section, University of Pittsburgh Medical Center Children’s
helping to maintain a diverse gut microbiota and disease resistance (15). Further, exposure to Hospital of Pittsburgh, University of Pittsburgh, Pittsburgh,
PA 15224.
pathogens may in fact be necessary for the microbiota to optimally develop and maintain host
This article contains supporting information online at
fitness. Supporting this idea, the microbiota of wild mice, which encounter more pathogens https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.​
than laboratory mice, limits influenza infection and colon cancer, while also protecting against 2214484120/-/DCSupplemental.
obesity and metabolic syndrome (16, 17). Similarly, self-limited gut infection can induce changes Published January 18, 2023.

PNAS 2023 Vol. 120 No. 4 e2214484120 https://doi.org/10.1073/pnas.2214484120 1 of 8

 to the microbiota that enhance its resistance to subsequent pathogen Yersinia pseudotuberculosis (Yptb) model of transient gut infection
encounters (18). In contrast, dysregulated host metabolism can disrupt (21). Yptb is a food-borne bacterial pathogen that causes transient
or alter the microbiota’s resistance to pathogens (19, 20), but to date, weight loss in mice (Fig. 1A) before being cleared from the gut
whether infection can in turn impact the microbiota’s regulation of and peripheral tissues within 4 weeks (wk) post-infection (21).
host metabolism warrants further exploration. At >15 wk post-infection, previously Yptb-infected (post-Yptb)
Here, we show that prior gut infection, long after clearance of mice began to gain more weight than naïve control mice (Fig. 1A).
the pathogen, can promote WAT expansion and host weight gain, This increase was not associated with increased food intake
while at the same time optimizing host metabolism for carbohy- (SI Appendix, Fig. S1A). X-ray imaging revealed significant expansion
drates. This shift toward carbohydrates, characterized by enhanced of peripheral body fat (Fig. 1B), with the weight of three major WAT
glucose disposal and insulin sensitivity, occurs in a microbiota-de- depots–mesenteric, perigonadal, and subcutaneous–significantly
pendent manner, with infection-elicited Betaproteobacteria being increased in post-Yptb mice compared to naïve mice (Fig. 1C). In
sufficient to enhance carbohydrate usage. Furthermore, infec- line with this observation, post-Yptb mice produced higher circulating
tion-optimized host carbohydrate metabolism can prevent stunt- levels of adiponectin, a hormone secreted primarily by WAT (Fig. 1D).
ing in the context of nutrient restriction, a phenomenon that we WAT expansion can be driven by an increase in the size (hyper-
speculate may sustain host fitness in settings where infection and trophy) of adipocytes and/or proliferation (hyperplasia) of progen-
food precarity often co-occur. itors (22). In all three depots, there was a greater frequency of smaller
adipocytes in post-Yptb mice than naïve mice, ruling out a role for
Results hypertrophy (Fig. 1E and SI Appendix, Fig. S1 B and C). We next
assessed WAT adipocyte progenitors for expression of the prolifer-
Prior Gut Infection can Remodel WAT Physiology. To explore the ation marker Ki-67 (gating strategy in SI Appendix, Fig. S1D). We
impact of infection on host metabolism, we took advantage of the chose to perform this analysis at a relatively early time point (4 wk

A B D
X-ray imaging 2.5
50 ctrl 35 *
Weight gain (% initial)

Yptb *** 30 *

(mg/dl serum)
2.0

(% total area)

Adiponectin
30 25

Body fat
20 1.5
10 15 1.0
10
0.5
-10 5
*** 0 0.0
rl b rl b
0 2 4 6 8 10 12 14 16 18 20
ct Ypt ct Ypt
Weeks post-Yptb infection
C F mesWAT adipocyte progenitors
White adipose tissue (WAT)
mes pg sc 1000 6 * 0.6

Ki-67+ (x103/g tissue)

ctrl **
Tissue weight (mg)

800 Yptb
Ki-67+ (%)

*
ctrl

4 0.4
600
*
400
** 2 0.2
Yptb

200

0 0 0.0
mes pg sc rl b rl b
ct Ypt ct Ypt
E WAT
mesWAT 35
Frequency (% total)

ctrl Yptb 30 ctrl

* Yptb
25 *
20
15
10 *
5
*
0
2000
1200
1600

2400
2800
3200
3600
4000
4400
4800
0
400
800

Adipocyte cell size (µm2)

Fig. 1. Prior gut infection can remodel WAT physiology. (A–F) Analyses were performed at (B–E) >15 or (F) 4 wk post-Y. pseudotuberculosis (Yptb) infection. Naïve
mice were used as a control (ctrl). (A) Weight gain (% initial) over time. (B) Representative X-ray images (Left) and quantification (Right) of peripheral body fat (%
total area). (C) Representative images (Left) and quantified weights (Right) of WAT depots (mes, mesenteric; pg, perigonadal; sc, subcutaneous). (D) Adiponectin
levels in the serum of fasting mice. (E) Representative images of H&E-stained mesWAT tissue sections (Left) and quantification of adipocyte cell size distribution
(Right). Each x-axis tick mark corresponds to a bin size of 200 µm2. (Scale bar: 100 µm.) (F) Frequency (Left) and number (Right) of proliferating (Ki-67+) mesWAT
adipocyte progenitors. Data represent mean ± SEM (n ≥ 4 mice per group). *P < 0.05; **P < 0.01; ***P < 0.001 by two-tailed Student’s t test. Data are representative
of at least two experiments.

2 of 8 https://doi.org/10.1073/pnas.2214484120 pnas.org
post-infection), reasoning that heightened proliferation would respectively). In the GTT, blood glucose levels are monitored at
precede the expansion of WAT first observed at >15 wk post-infec- regular intervals after administering glucose, while in the ITT,
tion. Adipocyte progenitors in the mesenteric and perigonadal but blood glucose levels are monitored at regular intervals after admin-
not the subcutaneous WAT of post-Yptb mice were more prolifer- istering insulin (23). Post-Yptb mice disposed glucose more effi-
ative (as assessed by Ki-67+ cell frequency and number) than those ciently, according to the GTT (Fig. 2C), and responded to insulin
of naïve mice, supporting a role for increased adipocyte hyperplasia more sensitively, according to the ITT (Fig. 2D). Thus, despite
(Fig. 1F and SI Appendix, Fig. S1 E and F). Thus, prior gut infection expansion of WAT, prior gut infection can actually improve rather
can induce physiological remodeling (associated with expansion than negatively impact host metabolic parameters (triglyceride lev-
and increased cell proliferation) of WAT and increased weight gain els, glucose disposal, and insulin sensitivity) for the long-term.
long-term after pathogen clearance.
Locally Constrained Gut Infection can Improve Host Metabolism.
Prior Gut Infection can Improve Host Metabolism. To assess the Previously, we showed that infection with Yptb permanently damages
physiological consequences of post-infection WAT expansion, we gut-associated lymphatic structures (21). Because these structures are
first assessed circulating triglyceride levels. Post-Yptb mice had lower crucial for the transport of lipids (21), we assessed if post-infection
levels of triglycerides than naïve mice, suggesting that triglycerides “scarring” of lymphatics could contribute to host metabolic alterations.
were preferentially stored in WAT (Fig. 2A). In line with this, post- To this end, we employed an attenuated form of Yptb, ΔyopM, that
Yptb mice also had lower circulating levels of free glycerol, indicating lacks the virulence factor YopM (24) and as such is rapidly cleared
that following infection, mice also reduce triglyceride breakdown from the gut and does not cause lymphatic scarring (24). Similarly
into free glycerol and fatty acids (Fig. 2B). Thus, prior infection to what we observed with Yptb, mice previously infected with
may increase host energy storage in the form of WAT triglycerides. ΔyopM also exhibited increased WAT depot expansion and lower
We next performed two common assessments of glucose metab- circulating levels of both triglycerides and free glycerol as compared
olism: the glucose and insulin tolerance tests (GTT and ITT, to control naïve mice (SI Appendix, Fig. S2 A–C). Furthermore, post-

A C Glucose tolerance test

80 350 25 n.s.
** ctrl
(mg/dl serum)

20
(mg/dl blood)

Yptb
Triglycerides

60

AUC (x103)
Glucose

250
15
40
10
150 * *
20
5
0 50 0
rl b rl b
ct Ypt 0 15 30 60 90 120
ct Ypt
Min post-glucose injection
B D Insulin tolerance test
80 ** 150 15
(% initial in blood)

ctrl
(mg/dl serum)
Free glycerol

60 Yptb
AUC (x103)
Glucose

100 10
* *
40
50
* 5
20

0 0 0
rl rl b
ct ptb 0 30 60 90 120
ct Ypt
Y Min post-insulin injection
E Dark cycle Light cycle **
1.1 1.00 n.s.
ctrl
Yptb
exchange ratio

exchange ratio

0.95
Respiratory

Respiratory

1.0
0.90
0.9
0.85

0.8 0.80
18 21 24 3 6 9 12 15 18 Dark Light
Time of day
Fig. 2. Prior gut infection can enhance host carbohydrate metabolism. (A–E) Analyses were performed at >15 wk post-Y. pseudotuberculosis (Yptb) infection. Naïve
mice were used as a control (ctrl). (A and B) Triglyceride (A) and free glycerol (B) levels in the serum of fasting mice. (C) Glucose tolerance test results (Left) with
area under the curve (AUC, Right). (D) Insulin tolerance test results (Left) with AUC (Right). (E) Respiratory exchange ratio (RER) at 18 min intervals (Left) and average
RER during the dark and light cycles (Right). Data represent mean ± SEM (n ≥ 4 mice per group). *P < 0.05; **P < 0.01; n.s., not significant by two-tailed Student’s
t test or (C and D, Left) two-way ANOVA followed by Bonferroni post hoc test. Unless otherwise indicated, data are representative of at least two experiments.

PNAS 2023 Vol. 120 No. 4 e2214484120 https://doi.org/10.1073/pnas.2214484120 3 of 8

 A Juvenile SPF mice B Glucose tolerance test C Insulin tolerance test
100 Microbiota * 500 35 p=0.05 150 10
Weight gain (% initial)
* *

(% initial in blood)
ctrl
* * * * 30
Yptb * *
400 8

(mg/dl blood)
75

AUC (x103)

AUC (x103)
25

Glucose
Glucose
p=0.06 100
300 20 6
50
* *
* 200 15 4
25 * 10 50
100 2
0 5
0 0 0 0
rl b rl b
0 1 2 3 4 0 15 30 60 90 120 ct Ypt 0 30 60 90 120 ct Ypt
Weeks post-microbiota transfer Min post-glucose injection Min post-insulin injection

D Proteobacteria Proteobacteria genera

E Yptb microbiota-enriched genera
1.0 1.0 1.0 1.0 g_Parasutterella

rel. abundance (%)

rel. abundance (%)
rel. abundance (%)

*
abundance (%)

Rhodospirillales
0.8 0.8 0.8 0.8 f_Atopobiaceae
Parasutterella

Desulfovibrio
n.s.
Relative

0.6 0.6 0.6 0.6 g_Ruminococcus

f_Lachnospiraceae
0.4
0.4 **** 0.4 0.4
g_Parvibacter
0.2 0.2 0.2 0.2 n.s. f_Lachnospiraceae
0.0 0.0 0.0 0.0
rl b rl b rl b rl b
ct Ypt 0.0 0.2 0.4 0.6 0.8
ct Ypt ct Ypt ct Ypt
Correlation ) with
)
weight gain AUC

F Ex-GF mice G H Insulin tolerance test

0.3 120 * 85 160
ctrl
15
**
rel. abundance (%)

Sutt *
Parasutterella

(mg/dl serum)

(mg/dl serum)

12
Triglycerides

Free glycerol

**
(mg/dl blood)
0.2 90 80

AUC (x103)
120
Glucose

9
60 75 *
0.1 80 * 6
30 70
3
0.0 0 65 40 0
rl t rl t rl t rl t
ct Sut ct Sut ct Sut 0 15 30 60 90 120 ct Sut
Min post-insulin injection
Fig. 3. Infection-elicited microbiota can enhance host carbohydrate metabolism. (A–E) Juvenile specific pathogen-free (SPF) mice were engrafted with microbiota
from naïve (control, ctrl) or previously Y. pseudotuberculosis (Yptb)-infected mice (collected at 17 to 18 wk post-infection). Analyses were performed at >5 wk
post-microbiota transfer. Data represent (B and C) one or (A, D, and E) two pooled experiments (n = 4 to 5 or 9 mice per group, respectively). (A) Weight gain
(% initial) over time. (B) Glucose tolerance test results (Left) with AUC (Right). (C) Insulin tolerance test results (Left) with AUC (Right). (D) Relative abundance of the
Proteobacteria phylum (Left) and genera within the Proteobacteria (Right). (E) Correlation (Spearman’s ρ) of genera abundances with weight gain AUC. Only shown
are genera enriched in juvenile SPF mice engrafted with Yptb relative to naïve microbiota (log2 fold change > 0.5, P < 0.05 by two-tailed Mann–Whitney test). Genera
are named according to their lowest taxonomic classification (f, family; g, genus). (F–H) Ex-germfree (GF) mice were mono-colonized with S. wadsworthensis (Sutt)
or treated with culture medium as a control (ctrl) prior to housing under SPF conditions. Analyses were performed at 11 wk post-S. wadsworthensis colonization.
Data represent one experiment (n = 5 mice per group). (F) Relative abundance of the Parasutterella genus. (G) Triglyceride (Left) and free glycerol levels (Right)
in the serum of fasting mice. (H) Insulin tolerance test results (Left) with AUC (Right). Data represent mean ± SEM (n ≥ 4 mice per group). *P < 0.05; **P < 0.01;
****P < 0.0001; n.s., not significant by two-tailed Student’s t test, (B, C, and H, Left) two-way ANOVA followed by Bonferroni post hoc test, or (D and F) two-tailed
Mann–Whitney test.

ΔyopM mice outperformed naïve mice in both the GTT and ITT Prior Gut Infection can Remodel Host Metabolism toward
(SI Appendix, Fig. S2 D and E). Thus, gut-restricted pathogens can Carbohydrate Usage. The improved glucose disposal that
confer long-term metabolic improvements. we observed in previously Yptb, ΔyopM, and C. tyzzeri-
We next sought to explore whether infection-induced metabolic infected mice suggested, among other possibilities, that prior
improvement was pathogen-specific. To this end, we used the gut infection could shift host metabolism toward the usage
protozoan parasite Cryptosporidium tyzzeri. C. tyzzeri is a natural of carbohydrates, a major source of glucose. To explore this
murine pathogen that causes infection similar to water-borne possibility, we subjected post-Yptb mice to indirect calorimetry,
cryptosporidium infection in humans, a leading cause of child- a method that can infer energy usage (carbohydrates vs. fat)
hood diarrhea among under-resourced populations (25). Like based on the respiratory exchange ratio (RER), the ratio of
Yptb, C. tyzzeri is cleared within 4 wk post-infection (25). Prior carbon dioxide produced to oxygen consumed (26). A higher
C. tyzzeri infection triggered pronounced weight gain, WAT depot RER indicates preferential usage of carbohydrates (26), and
expansion, and lower circulating levels of triglycerides and free indeed, the RER of post-Yptb mice was higher than that of
glycerol (SI Appendix, Fig. S3 A–C). Furthermore, post-C. tyzzeri naïve mice during the dark cycle (at night, when mice are most
mice outperformed naïve mice in both the GTT and ITT active). Thus, prior gut infection can remodel host metabolism
(SI Appendix, Fig. S3 D and E). Thus, remodeling of host metab- to preferentially utilize carbohydrates over other macronutrients
olism can occur in the context of diverse gut infections. such as fat (Fig. 2E).

4 of 8 https://doi.org/10.1073/pnas.2214484120 pnas.org
 A B Host Diet
malnourished naïve control

Weight gain (% initial)

50
naïve naïve Yptb naïve malnourished
Yptb malnourished
30

10
* * * *
-10

0 1 2 3 4 5 6 7 8 9 10 11
Weeks post-diet
C malnourished
D
1.5 *
naïve naïve Yptb **
*

Tissue weight (g)

1.0 **
* Host Diet
0.5 ** naïve control
naïve malnourished
0.0 Yptb malnourished
mesWAT pgWAT scWAT

Fig. 4. Infection-induced carbohydrate metabolism can sustain host fitness during nutrient restriction. (A–D) Naïve or previously Y. pseudotuberculosis (Yptb)-
infected mice (at 4 wk post-infection) were placed onto a control or low-protein/fat (malnourished). Analyses were performed at 11 wk post-diet. Data are
representative of two experiments. (A) Representative images of malnourished diet-induced stunting. (B) Weight gain (% initial) over time. The Yptb control diet
group and error bars are not shown for visual clarity. All statistical comparisons were made against the naïve control diet group. The Yptb control diet group was
not significantly different from the naïve control diet group at any time point. (C) Representative X-ray images of peripheral body fat. (D) Weights of WAT depots
(mes, mesenteric; pg, perigonadal; sc, subcutaneous). Data represent mean ± SEM (n ≥ 4 mice per group). *P < 0.05; **P < 0.01 by two-tailed Student’s t test.

Infection-Elicited Microbiota can Enhance Host Carbohydrate the Proteobacteria, the only enriched genera were Parasutterella
Metabolism. Previously, we showed that prior Yptb infection can (Betaproteobacteria class), Desulfovibrio (Deltaproteobacteria
remodel the gut microbiota long-term (18). This remodeling, which class), and an unclassified genus within the Rhodospirillales order
can persist for >15 wk post-infection, is characterized by a pronounced (Alphaproteobacteria class) (SI Appendix, Fig. S4E and Dataset S1).
expansion of Proteobacteria (18). Accordingly, we next sought to assess Of these three genera, Parasutterella correlated most strongly with
whether Yptb-elicited microbiota, such as Proteobacteria, contribute insulin sensitivity, following only two Clostridia (two-tailed P =
to the improvement of host metabolism following infection. To 0.029, SI Appendix, Fig. S4G and Dataset S1).
this end, we first transferred gut microbiota from naïve or post- We next repeated these analyses for our experiment with juvenile
Yptb mice into germfree (GF) recipient mice, and after 12 wk, SPF mice. There, the most enriched phylum was again the
assessed metabolic parameters. Remarkably, post-Yptb microbiota Proteobacteria (Fig. 3D and Dataset S1), though to a lesser extent
was sufficient to impose most of the metabolic alterations observed (1.8-fold) than in our experiment with GF mice (11-fold). However,
in post-Yptb donor mice, including WAT expansion, decreased within the Proteobacteria, the only enriched genus was Parasutterella
circulating triglycerides/glycerol, and increased insulin sensitivity, (Fig. 3D), with Parasutterella being the most enriched genus (60-fold)
despite not increasing weight gain (SI Appendix, Fig. S4 A–D). among all genera (Dataset S1). Consequently, Parasutterella corre-
Because GF mice have been shown to have altered metabolism lated most strongly with weight gain (two-tailed P = 0.002, Fig. 3E
(10), we next assessed whether post-Yptb microbiota could improve and Dataset S1) and insulin sensitivity (Spearman’s ρ = −0.52,
host metabolism when transferred into mice with intact microbiota. Dataset S1). Collectively, these results suggest a prominent role for
To promote microbiota engraftment, we transferred microbiota Parasutterella in improving host carbohydrate metabolism.
into juvenile (3-wk-old) specific pathogen-free (SPF) mice.
At >5 wk post-engraftment, post-Yptb microbiota was sufficient to Betaproteobacteria can Enhance Host Carbohydrate Metabolism.
enhance weight gain, glucose disposal, and insulin sensitivity in To assess whether Parasutterella is sufficient to improve host
recipient juvenile SPF mice (Fig. 3 A–C). Together, these results carbohydrate metabolism, we colonized mice with a cultivable,
support the idea that remodeling of the gut microbiota following human-derived isolate of the closely related betaproteobacterium
prior infection can be sufficient to improve host metabolism for Sutterella wadsworthensis (classifiable as Parasutterella based on
the long-term. 16S V4 sequence), after multiple attempts to isolate the murine
strain proved unsuccessful. To promote engraftment with the
Infection-Elicited Betaproteobacteria Correlate with Enhanced human isolate, we first introduced the isolate into GF mice,
Host Carbohydrate Metabolism. To identify specific taxa within where it would not face competition from resident microbiota.
the post-Yptb microbiota that may modulate host metabolism, Following this, we fully conventionalized these mice via housing
we performed correlation analyses between taxa abundances and under SPF conditions. As expected, this procedure increased the
metabolic parameters. In our experiment with GF mice, the most abundance of S. wadsworthensis (detected as Parasutterella) within
enriched phylum, the Proteobacteria (SI Appendix, Fig. S4E and the gut microbiota (Fig. 3F and Dataset S1). Colonization of
Dataset S1), correlated most strongly with insulin sensitivity (lower ex-GF mice with S. wadsworthensis recapitulated the metabolic
area under the curve [AUC] in the insulin tolerance test [ITT]) (two- improvements observed in post-Yptb mice, including decreased
tailed P = 0.048, SI Appendix, Fig. S4F and Dataset S1). Within circulating triglycerides/glycerol and increased insulin sensitivity,

PNAS 2023 Vol. 120 No. 4 e2214484120 https://doi.org/10.1073/pnas.2214484120 5 of 8

 without promoting gain weight (Fig. 3 G–H and SI Appendix, Fig. In modern human settings, food precarity primarily occurs
S5 A and B). Likewise, S. wadsworthensis colonization of juvenile among under-resourced populations, where the staple foods (e.g.,
SPF mice also enhanced glucose disposal and insulin sensitivity corn, potato, and rice) are often high in carbohydrates (38). Because
(SI Appendix, Fig. S5 C–E). Thus, these results support the idea of this, food precarity can preferentially restrict access to protein and
that defined taxa, such as Parasutterella, that expand following prior fat. Such skews in nutrient intake, through diets high in carbohy-
gut infection can directly improve host carbohydrate metabolism. drates but low in protein and fat, potentially contribute to the high
rates of malnutrition among under-resourced populations (39). This
Infection-induced Carbohydrate Metabolism can Sustain Host widespread condition induces a range of developmental defects, such
Fitness during Nutrient Restriction. Throughout evolution, as growth stunting (7), and it often afflicts regions simultaneously
mammals and their associated microbiota have faced periodic burdened with high rates of infectious disease (31). Our finding that
(e.g., seasonal) fluctuations in both infection burden and nutrient infection can promote resistance to malnutrition, particularly mal-
availability (27–30). Because of the regular co-occurrence of these nutrition induced by limited access to protein and fat, raises the
and other environmental stressors, the presence of one stressor intriguing possibility that this phenomenon may help to support
could serve as a reliable cue for the host to mount preparative human fitness and survival in under-resourced settings.
defenses against other stressors. Accordingly, infection could Though we firmly established a role for the gut microbiota in
signal an impending food shortage, in which some macronutrients our observed phenotypes, we presently lack a mechanistic under-
(carbohydrates, protein, or fat) become more available than others. standing of how infection-elicited microbiota can remodel distal
In this context, an adaptive host strategy could be to specialize tissues (i.e., WAT) and systemic physiology (i.e., carbohydrate
in utilization of the macronutrients that are most consistently metabolism). We hypothesize a prominent role for microbiota-as-
available or that require less energy to acquire. sociated molecular patterns (MAMPs) and/or metabolites, since
In the wild, many rodent species primarily consume carbohy- these molecules are known to mediate much of the microbiota’s
drate-rich, plant-derived foods (e.g., seeds), though as omnivores, control over host metabolism, for example by modulating the activ-
they can also complement their diets with animal-derived foods ity of enteroendocrine cells (40). These specialized gut epithelial
(e.g., insects) that are richer in protein and fat (27, 28). As such, cells secrete systemically acting hormones that stimulate production
some rodents may be required to subsist largely on carbohydrates of the glucose-disposing signal insulin (41). Furthermore, enteroen-
during periods when food is scarce. Because food scarcity has, docrine hormones can regulate distal tissues such as WAT (42, 43),
throughout evolution, often co-occurred with disease outbreaks potentially leading to WAT expansion and weight gain (43, 44).
(31–33), we hypothesized that infection would prompt the host to Supporting a role for microbiota-derived MAMPs/metabolites in
optimize their metabolism toward the usage of carbohydrates. modulating the enteroendocrine cell activity of previously infected
To explore this possibility, we utilized a low-protein/fat mice, Parasutterella, which we showed are elicited by prior infection
(high-carbohydrate) diet previously shown to induce a state of and sufficient to enhance glucose homeostasis, have been shown to
malnutrition (34, 35). As previously shown, the low-protein/fat directly adhere to gut epithelial cells (45) and to localize along the
diet stunted the growth, weight gain, body fat, and WAT depots entire length of the gastrointestinal tract, both the large intestine
of mice, even when food was provided ad libitum (Fig. 4 A–D). and the upper (46) and lower (45) small intestine (duodenum and
In contrast, under similar dietary conditions, post-Yptb mice ileum). These two sites are, respectively, where two major types of
remained impervious to signs of malnutrition and developed com- enteroendocrine cells, K and L cells, are primarily found (47).
parably to naïve mice on the control diet (Fig. 4 A–D). Thus, prior Furthermore, Parasutterella produce the MAMP lipopolysaccharide
infection can promote host fitness in the context of limiting dietary (45) and enhance host production of bile acid metabolites (46),
settings in which the main source of nutrients is carbohydrates. both of which have been shown to modulate enteroendocrine cell
activity (48, 49). In ongoing studies, we are exploring how
Discussion Parasutterella-associated MAMPs/metabolites potentially synergize
to enhance host metabolism long-term after infection.
The long-term consequences of environmental stressors such as
infection on host physiology remain largely unexplored. Here, we Study Limitations. Extrapolation of our results to humans is
show that infection-elicited gut microbiota can remodel host limited by the fact that our study took place in mice and was
metabolism to preferentially utilize carbohydrates, resulting in highly controlled (used defined pathogens and diets). Further
heightened glucose disposal, WAT expansion, and weight gain. studies of heterogeneous human cohorts are warranted to assess the
Furthermore, infection-optimized carbohydrate metabolism could impact of prior infection and/or other environmental exposures
sustain host fitness in the context of limited protein and fat avail- on host metabolic status later in life.
ability, preventing the growth stunting (malnutrition) that other-
wise occurs in infection-naïve mice. Conclusion
Our observations add to a growing body of evidence that
environmental stressors are in fact necessary for the full devel- Together, our work highlights the fundamental role of infection
opment and optimization of host physiology (18, 36). For in mediating host adaptation to nutrient precarity. By probing the
example, we previously uncovered that nutrient restriction can long-term impacts of infection on host metabolism, we discovered
enhance the formation of immunological memory and thus that infection-induced gut microbiota optimizes host metabolism
resistance to subsequent infection (36), while in the present toward the usage of carbohydrates. Because carbohydrates are
work, we identify essentially the reverse phenomenon that prior often a more available nutrient source than protein or fat, we
infection can trigger host adaptation to nutrient restriction. speculate that infection-optimized carbohydrate metabolism may
This reciprocal regulatory structure (i.e., infection priming for be adaptive in under-resourced settings where infection and nutri-
nutrient restriction, and nutrient restriction priming for infec- ent precarity often co-occur. In contrast, in settings where carbo-
tion) is likely adaptive and may have arisen as a defense against hydrates are overly accessible (such as with a high-sugar Western
the frequent co-occurrence of food precarity with infection and diet) or restricted (ketogenic diet), infection-induced carbohydrate
other stressors throughout mammalian evolution (31, 33, 37). metabolism may instead be maladaptive.

6 of 8 https://doi.org/10.1073/pnas.2214484120 pnas.org
Methods Microbiota Preservation and Transfer. Ceca were harvested from 3 to 4
mice per group, and the contents of the harvested ceca were collected in an
Mice. Adult (6 to 8-wk-old) and juvenile (3-wk-old) SPF female mice were acquired anaerobic chamber. Cecal contents were suspended in reduced PBS (0.4 g/mL),
through Taconic Biosciences (C57BL/6NTac) or the University of Campinas pooled, mixed 1:1 with a reduced PBS + 20% glycerol solution, passed through
Multidisciplinary Center for Biological Research (CEMIB) (C57BL/6JUnib). GF a 70-µm filter, and preserved at −80 °C as aliquots. Prior to transfer into GF or
C57BL/6 mice were bred and maintained in the National Institute of Allergy and juvenile mice, microbiota aliquots were thawed, mixed 1:1 with reduced PBS
Infectious Diseases (NIAID) Gnotobiotic Animal Facility. All mice were housed at (for final concentration of 0.1 g/mL), and administered via oral (200 µL) and/or
a maximum of five animals per cage in temperature-controlled rooms under a rectal gavage (100 µL per mouse). Gavages were performed twice with up to 3
12-h light/dark cycle and provided water and chow ad libitum (Teklad TD.160179 d between administrations.
or Nuvilab CR-1 for SPF mice and NIH 31M for GF mice). All mouse procedures
were performed under animal study proposals approved by the NIAID Animal Care Microbiota Analysis. 16S data were generated and analyzed as previously
and Use Committee (LHIM-2E) or the University of Sao Paulo Ethics Committee described (18). Mann–Whitney tests were performed in R using a custom script.
for the Use of Animals (49/2016). Correlation analyses were performed in Excel. The significance of correlations was
determined in Prism (GraphPad).
Y. pseudotuberculosis. Y. pseudotuberculosis strains 32777 and 32777 ΔyopM
were revived from freezer stocks on MacConkey agar. Individual colonies were S. wadsworthensis Colonization. S. wadsworthensis ATCC 51579 was revived
cultured overnight in 2× YT at 25 °C with shaking. Mice were fasted for at least from freezer stocks on YCFAC-B plates (Anaerobe Systems). Colonies were
12 h prior to gavage with 1 × 107 colony forming units in 200 μL phosphate picked into 30 mL cooked meat broth (CMB) with glucose, hemin, and vitamin
buffered saline (PBS). K (Thermo) + 3% formate + 3% fumarate (CMBFF) and cultured overnight at
37 °C in an anaerobic chamber. After centrifugation at 2,500 × g for 10 min,
−/−
C. tyzzeri. C. tyzzeri was propagated in Ifng
mice as described (25). Mice were cultures were resuspended in 2 mL fresh CMBFF, and 200 µL was gavaged per
4
gavaged with 1 × 10 cysts in 100 μL PBS. Mice were not fasted beforehand. GF mouse immediately after housing under SPF conditions. Sterile CMBFF was
used as a control. Additional gavages were performed every other day for 1 wk
Food Intake. Food intake was calculated by dividing the daily change in food
(four total gavages). When colonizing juvenile SPF mice, mice were fasted for
weight by the number of mice per cage for 5 d.
several hours (from the morning until the afternoon) prior to the first two of
Peripheral Body Fat. Peripheral body fat was assessed using an In-Vivo MS FX four total gavages.
PRO (Bruker) in the University of Sao Paulo Center for Research Facilities and Diet Studies. To induce malnutrition, mice were placed onto a diet (Teklad
Support (CEFAP-USP) using the following settings: high energy X-ray, 0.8 mm TD.160180) with low levels of protein and fat (7% and 5% w/w, respectively).
filter; total X-ray, no filter. Body fat, lean, and bone area were quantified using Control mice were placed onto an isocaloric diet (Teklad TD.160179) with standard
Bruker MI software. Peripheral body fat area (%) was calculated using the fol- levels of protein and fat (20% and 15% w/w, respectively).
lowing formula: ([fat + lean + bone area] – [lean + bone area]) / (fat + lean +
bone area) * 100. Data, Materials, and Software Availability. 16S data were deposited into the
Sequence Read Archive (PRJNA819152) (51).
Metabolic Markers. Mice were fasted for 12 h prior to collecting serum for
assessing the following markers using commercially available kits: adiponectin ACKNOWLEDGMENTS. This work was supported by the NIAID Intramural
(R&D Systems), triglycerides (Pointe Scientific), and free glycerol (Sigma). Research Program (ZIAAI001115 and ZIA-AI001132 to Y.B.), the Sao Paulo Research
Foundation (FAPESP) (2015/25364-0, 2016/08385-7, and 2021/06881-5 to
Adipose Tissue Histology. WAT depots were fixed overnight in 10% paraform- D.M.d.F.; 2017/02298-8 and 2018/14562-4 to M.K.D.S.; 2018/00458-0 to
aldehyde, after which they were washed with 70% ethanol, sectioned in paraffin J.M.S.; 2017/14209-0 to M.D.-F.; 2018/15981-0 to L.M.-S.), Coordination for
(5 µm thickness), and stained with hematoxylin and eosin (H&E). Adipocyte cell the Improvement of Higher Education Personnel/National Council for Scientific
area was assessed using ImageJ. and Technological Development (CAPES/CNPq) (D.M.d.F.), the Pew Latin American
Fellows Program (V.A.-O.), the NIGMS Postdoctoral Research Associate Training
Flow Cytometry. Adipocyte progenitors were isolated as previously described
(PRAT) Program (FI2GM128736 to A.S.), NIDCR (K99DE31372 to A.S.), and the
(50). Briefly, WAT depots were cut into small pieces, suspended in DMEM + 50 mM
Helen Hay Whitney Foundation (A.G.Z.). We thank Dr. Boris Striepen (University
HEPES + 0.05 mg/mL Liberase TL (Roche) + 0.25 mg/mL DNase I (Sigma) + 1%
of Pennsylvania) for providing Cryptosporidium tyzzeri. Dana Trageser-Cesler
BSA (low fatty acid), digested for 25 min at 37 °C with agitation, and passed through
and Cesar Acevedo (NIAID Gnotobiotic Animal Facility), Eliane Mello Gomes for
70-µm filters. The resulting single-cell suspensions were then stained in Hank’s
technical assistance, the University of Sao Paulo Center for Research Facilities
Balanced Salt Solution (HBSS) with Invitrogen fixable LIVE/DEAD (Thermo Fisher
and Support (CEFAP-USP) staff, the NIAID animal facility staff, and the NIAID
Scientific), stained in 2% fetal bovine serum (FBS) with fluorochrome-conjugated
Microbiome Program staff.
antibodies for identifying adipocyte progenitors, fixed for 30 min using the Foxp3/
Transcription Factor Staining Buffer Set (eBioscience), permeabilized, and stained
for at least 1 h with a fluorochrome-conjugated antibody against Ki-67 (clone:
Author affiliations: aDepartment of Immunology, Institute of Biomedical Sciences,
SolA15). Stained cells were acquired on a BD LSRFortessa with FACSDiva software University of Sao Paulo, Sao Paulo, SP 05508-000, Brazil; bMetaorganism Immunity
and analyzed using FlowJo v10. Adipocyte progenitors were identified as follows: Section, Laboratory of Host Immunity and Microbiome, National Institute of Allergy
and Infectious Diseases, NIH, Bethesda, MD 20892; cNational Institute of Allergy and
live+, lineage- (CD45- [30-F11], CD31- [MEC13.3]), CD29+ (eBioHMb1-1), CD34+ Infectious Diseases Microbiome Program, National Institute of Allergy and Infectious
(RAM34), PDGFRα+ (APA5), Sca-1+ (D7), and CD24+ (M1/69). Diseases, NIH, Bethesda, MD 20892; and dDepartment of Clinical and Toxicological
Analysis, Faculty of Pharmaceutical Sciences, University of Sao Paulo, Sao Paulo, SP
Glucose and Insulin Tolerance Tests. For glucose tolerance tests, mice 05508-000, Brazil

were fasted for 12 h prior to challenge with glucose (2 g/kg mouse) via intra- Author contributions: M.K.D.S., V.A.-O., A.S., D.M.d.F., and Y.B. designed research;
M.K.D.S., V.A.-O., A.S., J.P.T.G., R.W.A.-C., S.-J.H., A.G.Z., T.W.H., and D.M.d.F. performed
peritoneal injection or oral gavage. For insulin tolerance tests, mice were research; A.C., J.M.S., M.D.-F., L.M.-S., W.M.T., N.O.S.C., M.R., and S.J. contributed new
fasted for 4 to 6 h prior to intraperitoneal injection with insulin (0.75 U/ reagents/analytic tools; M.K.D.S., V.A.-O., A.S., D.M.d.F., and Y.B. analyzed data; and
M.K.D.S., V.A.-O., A.S., D.M.d.F., and Y.B. wrote the paper.
kg mouse; Novo Nordisk). Blood glucose levels were assessed by tail vein
Reviewers: W.S.G., Harvard T.H. Chan School of Public Health; and M.R., University of
bleeding using a glucometer (Accu-Chek). AUC were determined using Prism California San Diego.
software (GraphPad).
The authors declare no competing interest.

Indirect Calorimetry. RER (the ratio of CO2 produced to O2 consumed) was Copyright © 2023 the Author(s). Published by PNAS. This open access article is distributed
under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0
assessed using Oxymax small-animal metabolic cages (Columbus Instruments). (CC BY-NC-ND).
Mice were individually housed and acclimated to cages for 24 h prior to calibration 1
Present Address: Department of Integrative Biology and Physiology, University of
and measurement of RER over a 24-h period. California, Los Angeles, CA 90095.

PNAS 2023 Vol. 120 No. 4 e2214484120 https://doi.org/10.1073/pnas.2214484120 7 of 8

 1. C. R. Voolstra, M. Ziegler, Adapting with microbial help: Microbiome flexibility facilitates rapid 28. C. H. S. Watts, R. W. Braithwaite, The diet of rattus lutreolus and five other rodents in Southern
responses to environmental change. BioEssays 42, e2000004 (2020). Victoria. Wildlife Res. 5, 47–57 (1978).
2. K. Troha, J. S. Ayres, Metabolic adaptations to infections at the organismal level. Trends Immunol. 29. A. Bajer et al., Medium-term temporal stability of the helminth component community structure in
41, 113–125 (2020). bank voles (Clethrionomys glareolus) from the Mazury Lake District region of Poland. Parasitology
3. N. Collins, Y. Belkaid, Control of immunity via nutritional interventions. Immunity 55, 210–223 130, 213–228 (2005).
(2022). 30. C. Kiffner, T. Vor, P. Hagedorn, M. Niedrig, F. Rühe, Factors affecting patterns of tick parasitism on
4. E. Ansaldo, T. K. Farley, Y. Belkaid, Control of immunity by the microbiota. Annu. Rev. Immunol. 39, forest rodents in tick-borne encephalitis risk areas, Germany. Parasitol. Res. 108, 323–335 (2011).
449–479 (2021). 31. J. L. Walson, J. A. Berkley, The impact of malnutrition on childhood infections. Curr. Opin. Infect. Dis.
5. K. R. Amato, Incorporating the gut microbiota into models of human and non-human primate 31, 231–236 (2018).
ecology and evolution. Am. J. Phys. Anthropol. 159, S196–215 (2016). 32. A. Gwela, E. Mupere, J. A. Berkley, C. Lancioni, Undernutrition, host immunity and vulnerability to
6. P. J. Turnbaugh et al., The effect of diet on the human gut microbiome: A metagenomic analysis in infection among young children. Pediatr. Infect. Dis. J. 38, e175–e177 (2019).
humanized gnotobiotic mice. Sci. Transl. Med. 1, 6ra14 (2009). 33. E. Mertens, J. L. Peñalvo, The burden of malnutrition and fatal COVID-19: A global burden of disease
7. L. V. Blanton, Gut bacteria that prevent growth impairments transmitted by microbiota from analysis. Front. Nutr. 7, 619850 (2021).
malnourished children. Science 351, aad3311 (2016), 10.1126/science.aad3311. 34. E. M. Brown et al., Diet and specific microbial exposure trigger features of environmental
8. D. Festi et al., Gut microbiota and metabolic syndrome. World J. Gastroenterol. 20, 16079–16094 enteropathy in a novel murine model. Nat. Commun. 6, 7806 (2015).
(2014). 35. N. R. Teodósio, E. S. Lago, S. A. Romani, R. C. Guedes, A regional basic diet from northeast Brazil as a
9. F. Bäckhed, J. K. Manchester, C. F. Semenkovich, J. I. Gordon, Mechanisms underlying the resistance dietary model of experimental malnutrition. Arch. Latinoam. Nutr. 40, 533–547 (1990).
to diet-induced obesity in germ-free mice. Proc. Natl. Acad. Sci. U.S.A. 104, 979–984 (2007). 36. N. Collins et al., The bone marrow protects and optimizes immunological memory during dietary
10. F. Bäckhed et al., The gut microbiota as an environmental factor that regulates fat storage. Proc. Natl. restriction. Cell 178, 1088–1101.e1015 (2019).
Acad. Sci. U.S.A. 101, 15718–15723 (2004). 37. P. C. Calder, A. A. Jackson, Undernutrition, infection and immune function. Nutr. Res. Rev. 13, 3–29
11. H. Plovier et al., A purified membrane protein from Akkermansia muciniphila or the pasteurized (2000).
bacterium improves metabolism in obese and diabetic mice. Nat. Med. 23, 107–113 (2017). 38. S. Ruan et al., Staple food and health: A comparative study of physiology and gut microbiota of mice
12. S. Devkota et al., Dietary-fat-induced taurocholic acid promotes pathobiont expansion and colitis in fed with potato and traditional staple foods (corn, wheat and rice). Food & Function 12, 1232–1240
Il10-/- mice. Nature 487, 104–108 (2012). (2021).
13. J. M. Natividad et al., Bilophila wadsworthia aggravates high fat diet induced metabolic 39. Z. A. Bhutta et al., Severe childhood malnutrition. Nat. Rev. Dis. Primers 3, 17067 (2017).
dysfunctions in mice. Nat. Commun. 9, 2802 (2018). 40. Y. Fan, O. Pedersen, Gut microbiota in human metabolic health and disease. Nat. Rev. Microbiol. 19,
14. J. L. Sonnenburg, E. D. Sonnenburg, Vulnerability of the industrialized microbiota. Science 366 55–71 (2021).
(2019). 41. F. Preitner et al., Gluco-incretins control insulin secretion at multiple levels as revealed in mice
15. D. Ramanan et al., Helminth infection promotes colonization resistance via type 2 immunity. Science lacking GLP-1 and GIP receptors. J. Clin. Invest. 113, 635–645 (2004).
352, 608–612 (2016). 42. H. Hauner, G. Glatting, D. Kaminska, E. F. Pfeiffer, Effects of gastric inhibitory polypeptide on glucose
16. S. P. Rosshart et al., Wild mouse gut microbiota promotes host fitness and improves disease and lipid metabolism of isolated rat adipocytes. Ann. Nutr. Metab. 32, 282–288 (1988).
resistance. Cell 171, 1015–1028.e1013 (2017). 43. L. Getty-Kaushik, D. H. Song, M. O. Boylan, B. E. Corkey, M. M. Wolfe, Glucose-dependent
17. B. Hild et al., Neonatal exposure to a wild-derived microbiome protects mice against diet-induced insulinotropic polypeptide modulates adipocyte lipolysis and reesterification. Obesity (Silver Spring)
obesity. Nat. Metab. 3, 1042–1057 (2021). 14, 1124–1131 (2006).
18. A. Stacy et al., Infection trains the host for microbiota-enhanced resistance to pathogens. Cell 184, 44. R. Ugleholdt et al., Transgenic rescue of adipocyte glucose-dependent insulinotropic polypeptide
615–627.e617 (2021). receptor expression restores high fat diet-induced body weight gain. J. Biol. Chem. 286,
19. C. A. Thaiss et al., Hyperglycemia drives intestinal barrier dysfunction and risk for enteric infection. 44632–44645 (2011).
Science 359, 1376–1383 (2018). 45. K. Hiippala, V. Kainulainen, M. Kalliomäki, P. Arkkila, R. Satokari, Mucosal prevalence and
20. K. K. Sanchez et al., Cooperative metabolic adaptations in the host can favor asymptomatic infection interactions with the epithelium indicate commensalism of sutterella spp. Front. Microbiol. 7, 1706
and select for attenuated virulence in an enteric pathogen. Cell 175, 146–158.e115 (2018). (2016).
21. D. M. Fonseca et al., Microbiota-dependent sequelae of acute infection compromise tissue-specific 46. T. Ju, J. Y. Kong, P. Stothard, B. P. Willing, Defining the role of Parasutterella, a previously
immunity. Cell 163, 354–366 (2015). uncharacterized member of the core gut microbiota. Isme J. 13, 1520–1534 (2019).
22. L. Vishvanath, R. K. Gupta, Contribution of adipogenesis to healthy adipose tissue expansion in 47. R. Gutierrez-Aguilar, S. C. Woods, Nutrition and L and K-enteroendocrine cells. Curr. Opin.
obesity. J. Clin. Invest. 129, 4022–4031 (2019). Endocrinol. Diabetes Obes. 18, 35–41 (2011).
23. S. Virtue, A. Vidal-Puig, GTTs and ITTs in mice: Simple tests, complex answers. Nat. Metab. 3, 48. L. J. Lebrun et al., Enteroendocrine L cells sense LPS after gut barrier injury to enhance GLP-1
883–886 (2021). secretion. Cell Rep. 21, 1160–1168 (2017).
24. S. J. Han et al., White adipose tissue is a reservoir for memory T cells and promotes protective 49. S. Katsuma, A. Hirasawa, G. Tsujimoto, Bile acids promote glucagon-like peptide-1 secretion through
memory responses to infection. Immunity 47, 1154–1168.e1156 (2017). TGR5 in a murine enteroendocrine cell line STC-1. Biochem. Biophys. Res. Commun. 329, 386–390
25. A. Sateriale et al., A genetically tractable, natural mouse model of cryptosporidiosis offers insights (2005).
into host protective immunity. Cell Host. Microbe. 26, 135–146.e135 (2019). 50. C. D. Church, R. Berry, M. S. Rodeheffer, Isolation and study of adipocyte precursors. Methods
26. J. Rozman, M. Klingenspor, M. Hrabě de Angelis, A review of standardized metabolic phenotyping Enzymol. 537, 31–46 (2014).
of animal models. Mamm. Genome. 25, 497–507 (2014). 51. M. K. De Siqueira et al., Infection-induced microbiota enhances host adaptation to nutrient
27. O. J. Reichman, Relation of desert rodent diets to available resources. J. Mammal. 56, 731–751 restriction. Sequence Read Archive. https://www.ncbi.nlm.nih.gov/bioproject/PRJNA819152/.
(1975). Deposited 23 March 2022.

8 of 8 https://doi.org/10.1073/pnas.2214484120 pnas.org