Article
Distinct hypothalamic control of same- and
opposite-sex mounting behaviour in mice
In the format provided by the
authors and unedited
Nature | www.nature.com/nature
Supplementary information
https://doi.org/10.1038/s41586-020-2995-0
Article
Distinct hypothalamic control of same- and
opposite-sex mounting behaviour in mice
In the format provided by the
authors and unedited
Nature | www.nature.com/nature
Supplementary information
https://doi.org/10.1038/s41586-020-2995-0
1
Supplementary Notes
SUPPLEMENTARY NOTE 1
–
Related to Fig. 1b, c and Extended Data Fig. 1a
-
g
We trained two supervised classifiers (decoders) to distinguish female
-
and
male
-
directed
mounting, using the same set of videos, but different sets of frames to extract mouse pose features
(
Supplementary Table 1
). One classifier was trained using pose features derived exclusively from
frames in which actual mounting bouts occurr
ed (“frame
-
by
-
frame decoder”;
Fig. 1b, c, ED Fig. 1b
).
The other was trained using pose features extracted from video frames spanning 3 seconds before to 1
second after mount onset (“temporal decoder”;
ED Fig. 1e
-
g
). Performance was higher using the
tempor
al decoder (78%) than the frame
-
by
-
frame decoder (63%). This result suggests that that pose
features extracted from frames prior to the initiation of mounting make a large contribution to decoder
performance.
We examined four pose features (inter
-
mouse dis
tance, resident axis ratio
, resident acceleration,
and resident nose speed) which made a large contribution to distinguishing female
-
vs. male
-
directed
mounting in the temporal decoder. Feature histograms (
ED Fig.1g
, bottom row) for female
-
vs. male
-
direc
ted mounting were better separated for the temporal decoder, than for the frame
-
by
-
frame decoder
(
ED Fig. 1b
).
Closer inspection of individual features revealed that inter
-
mouse distance during the pre
-
mounting period was larger with female than male intr
uders. This may reflect the tendency of resident
males to engage in relatively longer periods of close investigation of male intruders before mount
initiation, whereas they approached and quickly initiated mounting to female intruders (
ED Fig.
1
d
). A
difference in resident axis ratio was also evident during residents’ pre
-
mounting
behaviour
s towards
male vs. female intruders. With female intruders, the axis ratio increased immediately before mount
initiation since
residents walked
over to females in a stretched posture and then mounted. With male
intruders, by contrast, the
axis ratio was smaller since residents remained close to intruders and
investigated them, resulting in a more contracted body posture.
Other feature differenc
es may reflect distinctions in the intruders’ peri
-
mount
behaviour
s. Male
intruders attempted to walk away from the residents when they were being intensively investigated,
causing residents to follow them around; this
behaviour
is reflected in a higher re
sident acceleration for
male
than for
female intruders. Similarly, male intruders tended to initiate escape when mounted,
whereas female intruders remained in place and were receptive to mounting. This difference is reflected
in a higher resident nose s
peed (locomotion) for male
than for
female intruders, after mount initiation.
Thus, examination of different feature distributions during male
-
versus
female
-
directed mounting
provides some insights into how these
behaviour
s
differ, and the reasons for the difference in decoder
performance.
SUPPLEMENTARY NOTE 2
–
Related to Extended Data Fig. 4
Although the majority of male
-
directed mounting events were USV
-
(
Fig. 1g, k
), we observed
rare cases of animals that exhibited USV
+
male
-
male mounting. These cases included resident males
with relatively less sexual and social experience (
Fig. 1g
, Day 1, green), or with castrated male intruders
(
Fig. 4k
, green). Of ten animals implanted for dual
-
site fiber photometry and used for anal
ysis, two
showed USV
+
mounting toward male as well as female intruders (
Extended Data Fig. 4
). While they
are rare examples, they provide useful cases of “exceptions that prove the rule.”
One sexually and socially experienced male mouse
(
no.
629) showed US
V
+
mounting towards
male as well as female intruders (
Extended Data Fig. 4a, b
), and no attack. Scaled neural activity in
this animal during USV
+
mounting towards males was higher in MPOA
ESR1
than VMHvl
ESR1
neurons,
similar to the typical activity pattern
exhibited towards female intruders (
Extended Data Fig
s
. 3m,
4a
-
2
c
). In a triadic social encounter (i.e., two intruders, one male and one female), this mouse preferred the
female over the male. This suggests that the lack of attack and USV
+
male
-
directed mounting exhibited
by this mouse was not due to a failure to discriminate the sex of intruders, e.g., as in
TrpC2
mutant
mice
14,45
. In another example, a sexually and socially naïve mouse
(
no.
634) initially exhibited USV
+
mounting towards male as well as female intruders (
Extended Data Fig. 4f
-
h
), and (as in the case of
mouse
no.
629) scaled activity was higher in MPOA
ESR1
than in V
MHvl
ESR1
neurons during male
-
directed USV
+
mounting. After gaining social and sexual experience, however, this animal switched to
USV
-
mounting towards male intruders, and exhibited a male intruder
-
typical pattern of activity in
MPOA
ESR1
versus
VMHvl
ESR1
neurons (
Extended Data Fig
s
. 3q,
4i
-
k
). Thus, in both cases, these mice
exhibited a female intruder
-
typical pattern of activity in MPOA
ESR1
v
ersus
VMHvl
ESR1
neurons when
they exhibited USV
+
mounting towards males.
This suggests that the relative level of activity in these
neurons is not simply representing intruder sex, but rather reflects intent or motivational state.
SUPPLEMENTARY NOTE 3
–
Related to Extended Data Fig. 10q
In
Extended Data Figure. 10q
, we present a working hypothesis to reconcile the results of imaging
experiments with the effects of functional manipulations of
ESR1
+
neurons in MPOA and VMHvl. Here
we describe each class of functional manipulation, together with the most parsimonious e
xplanation for
its
behaviour
al effects, in terms of the physiological properties of neuronal subsets detected in our
imaging experiments.
VMHvl:
GOF
: Strong optogenetic activation of
ESR1
+
neurons and chemogenetic activation of
progesterone
receptor (PGR)
positive
PGR
neurons promote
attack
11,12
①
, while weak optogenetic activation of
ESR1
+
neurons promotes mounting without USVs
②
(
Fig. 4a
-
k
). Strong activation of VMHvl
ESR1
/
PGR
neurons also inhibits female
-
directed sexual
behaviour
11
, via likely projections to MPOA
③
(
Fig. 4o
-
r
).
Since virtually all VMHvl
ESR1
neurons are glutamatergic
16
, this inhibition likely recruits local inhibitory
interneurons in MPOA
④
, which is 80% GABAer
gic
15
(although this is not proven). The larger
proportion of male
-
vs. female
-
selective
ESR1
+
neurons in VMHvl (~2.4:1; ref
6
and this study) likely
explains why aggressive
behaviour
s dominate the response to GOF manipulations of VMHvl
ESR1
/
PGR
neurons
11,1
2
.
LOF
: Genetic ablation
10
or chemogenetic inhibition of VMHvl
PGR
neurons
11
weakly inhibits female
-
directed male mounting. Here we show that chemogenetic inhibition of VMHvl
ESR1
neurons strongly
inhibits female
-
directed
mounting
(
ED Fig. 9
j
-
m
), using a longer incubation time between CNO
injection and testing. It is likely (but not directly demonstrated) that this effect is due to inhibition of the
female
-
select
ive subset of neurons in VMHvl
⑤
(refs
6,16,26
and
Fig. 2f, g
). Such chemogenetic
inhibition presumably also silences VMHvl
ESR1
neurons that inhibit mating via projections to MPOA
③
(
Fig. 4o
-
r
), which in theory should increase mating. However, these neurons
are likely found in the
male
-
preferring population, and therefore should not be active in the presence of a female; hence
silencing them has no effect. The female
-
preferring VMHvl
ESR1
neurons that are required for sexual
mounting may project to MPOA (not
illustrated) or to a downstream target to which MPOA also
projects.
MPOA:
GOF
: Optogenetic activation of MPOA
ESR1
neurons evokes mounting very inefficiently
8
(
Fig. 3c,
d
).
This may be because the
ESR1
+
population in MPOA is heterogeneous, and
contains multiple
subpopulations with different functions
15
; if these subpopulations interfere with each other, then their
simultaneous activation may cancel each other out. Intersectional activation of GABAergic
ESR1
+
3
neurons stimulates a more restricted
population and therefore these competing effects are eliminated,
yielding more efficient promotion of mounting
⑥
(
Fig. 3c, d, ED Fig. 7
). Activation of
MPOA
ESR1
∩VGAT
neurons promotes both mounting towards intruders, and USVs in solitary males.
Whether the
se are the same or different subsets of MPOA
ESR1
∩VGAT
neurons cannot be distinguished at
present. Gao et al. recently reported that optogenetic stimulation of MPOA
VGAT
neurons promotes
USVs
23
, but whether these neurons were
ESR1
+
was not determined. Stimul
ation of MPOA
ESR1
∩VGAT
terminals in VMHvl suppresses aggression
⑦
(
Fig. 4s, t
). Whether this occurs via collateral projections
of the same MPOA neurons as promote mounting, or via a different subpopulation of MPOA neurons, is
not yet clear.
One might predict that optogenetic activation of the male
-
selective neurons in MPOA would promote
aggressive
behaviour
⑧
. However, male sexual
behaviour
dominates when MPOA
ESR1
∩VGAT
neurons
are optogenetically stimulated, because the female
-
selective neuro
ns outnumber the male
-
selective
neurons in MPOA by ~2:1
(
Fig. 2f, g
)
.
LOF
: Silencing MPOA
ESR1
neurons reduces both mounting
7
and USVs
9
⑥
(
Fig. 2m
-
o,
ED Fig. 9c
-
e
).
There is no effect on aggressive
behaviour
(
ED Fig. 9g
-
i
), perhaps because such silencing also
inactivates MPOA neurons that normally inhibit aggression via projections to VMHvl
⑦
, thereby
increasing VMHvl activity.
In rare cases, we have observed that silencing MPOA
releases aggression
towards females.
Single
-
cell RNA sequencing has identified 6
-
7 transcriptomically distinct subsets of
ESR1
+
neurons in
VMHvl
16
. A similar diversity of
ESR1
+
neurons has been identified in the POA
15
. Given that binary
intersectional techniques were required to selectively activa
te
ESR1
+
, VGAT
+
neurons in MPOA that
promote sexual
behaviour
(this study), it is likely that triple or even higher
-
order genetic intersectional
techniques will be required to functionally isolate smaller subsets of these cells. Such techniques are
currently challenging
in mice, but may become more
feasible
in the future.
45. Leypold, B. G. et al. Altered sexual and social
behaviors
in trp2 mutant mice.
Proc. Natl Acad. Sci.
USA
99
, 6376
–
6381 (2002).