Reversal of diastereoselectivity in palladium-arene interaction directed hydrogenative desymmetrization of 1,3-diketones

Article abstract of DOI:10.1007/s11426-019-9601-7

For the metal-catalyzed asymmetric hydrogenation of α-substituted ketones, cis reductive products are generally obtained due to steric hindrance of substituents. Herein, an unprecedented trans reductive products were observed in palladium-catalyzed hydrogenative desymmetrization of cyclic and acyclic 1,3-diketones, providing the chiral trans β-hydroxy ketones with two adjacent stereocenters including one α-tertiary or quaternary stereocenter with high enantioselectivity and diastereoselectivity. Mechanistic studies and DFT calculations suggested that the rarely observed diastereoselectivity reversal is ascribed to the charge-charge interaction between the palladium and aromatic ring of the substrate, which could not only result in the reversal of the diastereoselectivity, but also improve the reactivity.

Full text of DOI:10.1007/s11426-019-9601-7

SCIENCE CHINA
Chemistry
•ARTICLES•
https://doi.org/10.1007/s11426-019-9601-7
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reversal of diastereoselectivity in palladium-arene interaction
directed hydrogenative desymmetrization of 1,3-diketones
Chang-Bin Yu1†, Heng-Ding Wang1,2†, Bo Song1, Hong-Qiang Shen1, Hong-Jun Fan1* &
Yong-Gui Zhou1*
1State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China;
2University of Chinese Academy of Sciences, Beijing 100049, China
Received August 24, 2019; accepted August 28, 2019; published online October 22, 2019
For the metal-catalyzed asymmetric hydrogenation of α-substituted ketones, cis reductive products are generally obtained due to
steric hindrance of substituents. Herein, an unprecedented trans reductive products were observed in palladium-catalyzed
hydrogenative desymmetrization of cyclic and acyclic 1,3-diketones, providing the chiral trans β-hydroxy ketones with two
adjacent stereocenters including one α-tertiary or quaternary stereocenter with high enantioselectivity and diastereoselectivity.
Mechanistic studies and DFT calculations suggested that the rarely observed diastereoselectivity reversal is ascribed to the
charge-charge interaction between the palladium and aromatic ring of the substrate, which could not only result in the reversal of
the diastereoselectivity, but also improve the reactivity.
palladium-arene interaction, hydrogenative desymmetrization, 1, 3-diketones
Citation:
Yu CB, Wang HD, Song B, Shen HQ, Fan HJ, Zhou YG. Reversal of diastereoselectivity in palladium-arene interaction directed hydrogenative
desymmetrization of 1,3-diketones. Sci China Chem, 2019, 62, https://doi.org/10.1007/s11426-019-9601-7
1 Introduction
Neutral (hetero)arenes are capable of interaction to transi-
tion metal, such as, Rh [4], Ru [5], Pt [6] and other metals [7],
although the interaction is very weak. Among them, Pd-
arene interactions in complexes [8] have been observed. In
addition, the behavior of palladium-arene interactions was
also proposed as catalytic intermediates in palladium-cata-
lyzed reactions. For example, an intermediate was involved
in palladium-catalyzed Catellani reaction [9] or palladium-
catalyzed decarboxylative olefination of arene carboxylic
acids [10]. Although these interactions were discovered in
palladium complexes or intermediates, arenes are seldom
used as auxiliary ligands, or sub-coordination group in pal-
ladium-involved reaction because of the weak interaction
and a relatively limited range of substrates. Therefore, the
development of an efficient strategy to extend the process
scope to a wider range of substrates and realize its further
application in chemistry is still of great significance. We [11]
Asymmetric hydrogenation represents one of the most
practical methods to chiral molecules. Over the past decades,
significant advances have been made in development of
highly effective catalyst systems [1]. During the meantime,
asymmetric hydrogenation of ketones has been a long ex-
isting interest due to the abundance of chiral alcohols in
natural products and pharmaceuticals [2]. In general, cis
products would be obtained in the hydrogenation of cyclic α-
substituted ketones (Scheme 1) [3], which results from the
point that when the hydrogenation takes place, active M-H
species prefers to attack the carbonyl group from the less
steric face.
†These authors contributed equally to this work.
*Corresponding authors (email: fanhj@dicp.ac.cn; ygzhou@dicp.ac.cn)
© Science China Press and Springer-Verlag GmbH Germany, part of Springer Nature 2019
chem.scichina.com link.springer.com

2
Yu et al. Sci China Chem
the eluent to give the chiral reductive products.
3 Results and discussion
To validate the hypothesis, we initiated our exploration by
examining the palladium-catalyzed hydrogenative de-
symmetrization of 1,3-diketone 2-methyl-2-phenyl-1H-in-
dene-1,3(2H)-dione (1a). The reaction can proceed through
two possible pathways. According to previous reports, the
Pd-H species would attack the carbonyl group from the less
steric face leading to cis product. The other pathway is the
addition of Pd-H to the carbonyl group from the face of
aromatic ring because of the interaction between palladium
and arene in substrate generated in situ to give the trans
product. Delightedly, the reaction proceeded smoothly to
afford the trans β-hydroxy ketone 2a in 86% yield, 91% ee
and 10:1 diastereoselectivity (Scheme 2, Eq. 1), and no bis-
hydrogenated product was observed [13]. The absolute
configuration of the desymmetrization product (2R,3S)-2a
was determined by X-ray diffraction analysis (see the Sup-
porting Information online). In sharp contrast, cis β-hydroxy
ketone 2a′ was obtained with chiral ruthenium catalyst (Eq.
2). These results identified that a rare phenomenon of re-
versal of diastereoselectivity was involved in palladium-
catalyzed hydrogenative desymmetrization.
Encouraged by the above result, we next optimized the
reaction conditions to further improve the diastereoselec-
tivity. Delightedly, in the presence of additive trifluoro-
acetic acid (TFA), the enantio- and diastereoselectivity was
improved (Table 1, entry 2). Next, when the reaction tem-
perature was decreased, the diastereoselectivity was im-
proved to 17:1 (entry 3). The evaluation of acid additives
indicated that no desired product was obtained in the pre-
sence of strong Brønsted acids (entry 4). However, with
weak Brønsted acids, such as tartaric acid, benzoic acid and
salicylic acid, both excellent yield and enantioselectivity
were observed (entries 5–8). Upon further optimization from
the commercially available chiral bisphosphine ligands (en-
tries 9–11), (S)-SegPhos emerged as the best ligand. Thus,
the optimal condition was established: Pd(OCOCF3)2/(S)-
SegPhos/PhCO2H/TFE/H2 (300 psi)/0 oC/24 h.
Scheme 1 Diastereoselectivity control in the hydrogenation of α-sub-
stituted ketones: cis and trans reductive products.
and others [12] have been involved in the development of
palladium-catalyzed asymmetric hydrogenation of ketones.
Considering (hetero)arenes could in situ interact with pal-
ladium, we envisaged whether 2-aryl substituted 1,3-dike-
tones could be employed as substrates for the palladium-
catalyzed asymmetric hydrogenative desymmetrization
(Scheme 1), giving the trans chiral α-aryl substituted β-hy-
droxy ketones. Herein, an unprecedented trans reductive
products were observed in palladium-catalyzed asymmetric
hydrogenative desymmetrization of cyclic and acyclic 1,3-
diketones, providing the chiral trans β-hydroxy ketones with
two adjacent stereocenters including one α-tertiary or qua-
ternary stereocenter with high enantioselectivity and dia-
stereoselectivity. Mechanistic studies suggested that the
rarely observed diastereoselectivity reversal is ascribed to
the charge-charge interaction between the palladium and
aromatic ring of the substrates, which could not only result in
the reversal of the diastereoselectivity, but also improve the
reactivity.
2 Experimental
Ligand (S)-SegPhos (3.4 mg, 0.0055 mmol) and Pd
(OCOCF3)2 (1.7 mg, 0.005 mmol) were placed in a dried
Schlenk tube under nitrogen atmosphere, and degassed an-
hydrous acetone was added. The mixture was stirred at room
temperature for 1 h. The solvent was removed under vacuum
to give the catalyst. This catalyst was taken into a glove box
filled with nitrogen and dissolved in 2,2,2-trifluoroethanol
(1.0 mL). To the mixture of 1,3-diketones (0.25 mmol) and
benzoic acid (3.1 mg, 0.025 mmol) in 2,2,2-trifluoroethanol
(3.0 mL) was added this catalyst solution, and then the
mixture was transferred to an autoclave, which was charged
hydrogen gas (300 psi). The autoclave was stirred at 0 °C for
24 h. After release of the hydrogen gas, the autoclave was
opened and the reaction mixture was evaporated. Purification
was performed on silica gel using hexanes/ethyl acetate as
With the optimized condition in hand, the scope of this
hydrogenative desymmetrization was examined. This meth-
od is compatible with various functional groups (Scheme 3).
In general, various 1,3-cyclopentanediones 1 were converted
to chiral β-hydroxy ketones 2 bearing two adjacent stereo-
centers including α-quaternary stereo center with high en-
antioselectivity regardless of the electronic and steric
properties. Notably, the non-benzofused 1,3-diketone (1l)
was also compatible. The spiro diketone 1m could be hy-
drogenated with moderate 61% ee, and diketones 1n and 1o
bearing the ester group could also be smoothly hydro-

3
Yu et al. Sci China Chem
Scheme 2 Diastereoselectivity switch in the hydrogenative desymme-
trization of 1,3-diketones 1a with Pd and Ru catalyst.
Table 1 Optimization of the reaction conditionsa)
Entry
1e
2e
3
Acid
–
L
Yield (%)b)
drc)
10:1
11:1
17:1
–
ee (%)d)
91
Scheme 3 Substrate scope: α-quaternary stereocenters; For 1m and 1o:
Pd(OCOCF3)2 (5.0 mol%), (S)-SegPhos (6.0 mol%), trifluoroacetic acid
(100 mol%), 70 °C, 42–48 h; For 1n: Pd(OCOCF3)2 (5.0 mol%), (S)-Seg-
Phos (6.0 mol%), TFA (100 mol%), 50 °C, 48 h.
L1
L1
L1
L1
L1
L1
L1
L1
L2
L3
L4
86
86
89
9
TFA
93
TFA
95
4
TsOH·H2O
D-TA
–
5
93
92
92
91
80
93
90
20:1
20:1
>20:1
17:1
17:1
18:1
19:1
96
phenyl ring had negligible effect on the yield, diastereos-
electivities and enantioselectivities (3b, 3e, 3f and 3i). The
absolute configuration of the desymmetrization product
(2R,3R)-4f was determined by X-ray diffraction analysis (see
Supporting Information online).
To further determine the substrate generality of this pro-
tocol, the naphthalene-derived 2,2-disubstituted 1,3-cyclo-
hexanediones 5 were also investigated, and the results were
summarized in Scheme 5. To our delight, the hydrogenative
desymmetrization afforded the chiral β-hydroxy ketones in
high enantioselectivity. Besides, acyclic simple 1,3-dike-
tones 7 were also suitable partner, giving the linear chiral β-
hydroxyketones 8 with 93%–94% ee and 85%–90% yields
(Scheme 5).
To demonstrate the synthetic utility of this attractive pro-
tocol, the asymmetric hydrogenation of 1a (Scheme 6, Eq. 3)
was performed at gram scale to give the desired product 2a in
93% yield and 95% ee without loss of activity, diastereos-
electivity and enantioselectivity. Meanwhile, the remaining
carbonyl group is an attractive functional group for further
transformations. For example, nucleophilic addition with
methyl lithium gave the chiral diol 9 with 89% yield and high
6
L-TA
96
7
PhCO2H
SA
96
8
95
9
PhCO2H
PhCO2H
PhCO2H
95
10
11
96
95
a) Conditions: 1a (0.25 mmol), Pd(OCOCF3)2 (2.0 mol%), ligand (2.2
mol%), acid (10 mol%), H2 (300 psi), TFE (4.0 mL), 0 °C, 24 h. b) Isolated
yield. c) Determined by 1H NMR. d) Determined by HPLC. e) 5 h and 25 °
C. TFA=trifluoroacetic acid. TsOH•H2O=p-toluenesulfonic acid mono-
hydrate. SA=salicylic acid. D-TA =D-tartaric acid.
genated, followed by transes terification to give chiral tri-
cyclic compounds with 91% ee and 84% ee, respectively.
The above method is also applicable to 2-aryl substituted
1,3-cyclopentanediones 3 through a simple reoptimization
(see Supporting Information online). As depicted in Scheme
4, both electron-donating and withdrawing substituents on
the phenyl ring were tolerable, giving a series of chiral β-
hydroxy ketones bearing two continuous stereocenters in
excellent enantioselectivities and diastereoselectivities (3a–
3l). Furthermore, the sterically hindered substituents on the

4
Yu et al. Sci China Chem
Scheme 6 Experiment at gram scale and product transformations.
Scheme 4 Substrate scope: α-tertiary stereocenters.
details in Supporting Information online) calculations were
carried out based on the optimized reaction conditions. We
proposed four possible reaction mechanisms [14], four-
membered ring inner-sphere model, TFA promoted four-
membered ring model, outer sphere six-membered ring
model, outer sphere eight-membered ring model, and found
that four-membered ring inner-sphere model is the most
plausible (see details in Supporting Information online). In
this mechanism (Scheme 7), CF3COO− ligand is directly
detached from the metal hydride A to generate an un-
occupied site, INT B. The 1,3-diketone 1a coordinates with
this active center (INT B) to generate INT C. The generation
of INT E by hydride transfer from Pd-H to 1a is the rate-
determining step (TS D). INT E reacts with TFA to generate
INT F, following by releasing the product 2a. TFA can be
generated during the hydrogen activation step; therefore, the
hydrogenation can be able to proceed without additional acid
additive (Table 1, entry 1). Some acid additives were found
to be able to slightly increase the reactivity of the hydro-
genation, we propose that this is because the concentration of
TFA generated by H2 activation is pretty low, and additional
acid could enhance the tranformation from INT E to 2a. The
key features of the reactions are similar with or without acid
additives. We calculated four reaction pathways of 1a to 2a
with four-membered inner-sphere model (Scheme 7), which
showed a trend of enantioselectivity and diastereoselectivity
of trans-(2R,3S)-2a>trans-(2S,3R)-2a>cis-(2S,3S)-2a>cis-
(2R,3R)-2a when (S)-SegPhos was used as chiral ligand.
Lower barriers for formation of trans-(2R,3S)-2a is in good
consistent with the high diastereoselectivity observed in
Scheme 5 Substrate scopes: the naphthalene-derived 1,3-diketones and
acyclic 1,3-diketones.
diastereoselectivity (Eq. 4). Cyclic homoallylic alcohol 12
was obtained through hydroxyl protection with TBS in the
presence of lutidine, Wittig olefination and deprotection with
TBAF in THF (Eq. 5). Monoprotected diol 14 with three
continuous stereogenic centers could be synthesized through
TBS protection and reduction with sodium borohydride with
high yield (Eq. 6). Notably, for the above all transformations,
no loss of optical purity was observed.
To understand the origin of unprecedented diastereos-
electivity, density functional theory (M06-2X/def2-TZVP//
M06-2X/def2-TZVP for Pd and 6-31g(d,p) for other atoms,

5
Yu et al. Sci China Chem
ring points to the cationic Pd(II) center. Late transition metal-
arene interaction is known in many transition metal com-
pounds, and very recently it is found that such interactions
could be very important to catalytic C–H/C–H cross cou-
pling reaction [4a]. In our reaction system, the phenyl ring
cannot coordinate strongly to the Pd(II) due to steric hin-
drance, however, the distance between Pd-H fragment and
phenyl ring is in the range of weak interaction, and the in-
teraction between them could be the driven force for the
trans selectivity. To validate this idea, we have calculated the
interaction energy between Pd-H and phenyl ring/methyl
group of substrate 1a with trans-(2R,3S)-TS D and cis-
(2S,3S)-TS D using ADF 2018 and Interacting Quantum
Atoms (IQA) method [15]. Indeed, we found that the Pd-H
and phenyl interaction energy in trans-(2R,3S)-TS D is −7.0
kcal/mol, much larger than that for Pd-H and Me (−3.5 kcal/
mol) in cis-(2S,3S)-TS D. The energy difference of these two
interactions is −3.5 kcal/mol, which is very close to the total
electronic energy difference between trans-(2R,3S)-TS D
and cis-(2S,3S)-TS D (−3.9 kcal/mol), suggesting that the
weak interaction of the Pd-H and phenyl ring is the dom-
inating factor of the trans selectivity. Since phenyl ring does
not coordinate with Pd(II) we suppose this weak interaction
between Pd-H and phenyl ring mostly arise from charge-
charge interaction.
Scheme 7 DFT calculations of four types of reaction pathways for dia-
stereoselective reversal (color online).
experiment.
To verify this assumption, we conducted the computational
experiments to study the trans/cis selectivity by replacing
phenyl to other groups (such as –NO2, t-butyl, –CF3, –OTf
and halogens) in the trans-(2R,3S)-TS D and cis-(2R,3S)-TS
D (details in Supporting Information online). Indeed, the
strong trans selectivity was observed for the negative
charged groups, and trans selectivity dramatically decreased
when we changed phenyl group into H atom. The bulky t-
Butyl group, surrounded by the positively charged H atoms,
is the only group which is predicted to have cis selectivity.
These results support that the weak charge-charge interaction
is the dominating factor for trans selectivity.
Then, some control experiments were also conducted to
provide the further insight into the mechanism. Firstly, 1,3-
diketone 1a was reduced using the aliphatic ligand 1,2-bis
(dicyclohexylphosphanyl)-ethane (DCPE) as to identify
whether the chirality of the ligand or π-π stacking effect
between the aryl in ligand and the substrate induced trans
configuration of the product (Scheme 8, Eq. 7). Trans pro-
duct was obtained, which proved that the ligand has no effect
on diastereoselectivity, and there was also no π-π stacking
effect in the transition state. When cis-4a′ was conducted
under the standard condition (Eq. 8), no reaction occurred,
which suggested that chiral palladium complex cannot pro-
mote the isomerisation of cis product to thermodynamically
stable trans product.
Computational experiments have been carried out to give
a further understanding of the observed trans selectivity.
We first considered that possibly the selectivity arises from
the interaction between the ligand and the substrate.
Therefore, we removed Pd-H fragment from the optimized
structure of trans-(2R,3S)-TS D and cis-(2S,3S)-TS D to
direct compare total substrate/ligand interaction in the two
transition states. We found the energy difference (ΔEtotal
)
between trans-(2R,3S)-TS D and cis-(2S,3S)-TS D de-
crease dramatically from −3.9 to −1.2 kcal/mol. This result
showed that the total substrate/ligand interaction is not the
dominant factor for the observed trans selectivity. To gain
more insight, the substrate/ligand interaction was analyzed
with the distortion/interaction model (details in SI). We
found the advantageous interaction-energy terms (ΔEint) in
trans-(2R,3S)-TS D is remarkably canceled by distortion
energy terms (ΔEdist).
On the other hand, if the ligand was removed from the
optimized structure of trans-(2R,3S)-TS D and cis-(2S,3S)-
TS D the energy difference between the two transition states
only changed slightly from −3.9 to −3.3 kcal/mol, which
suggested that the trans selectivity is maintained even if the
ligand is not there.
Next, the cause of Pd-H and substrate interaction leading
to the trans product was inspected. We noticed that in the
trans transition state the Pd(II) cation is closer to the phenyl
ring of the substrate, and the negative charge of the phenyl
Next, the influence of substituents of phenyl ring on re-
activity and diastereoselectivity was investigated (Scheme

6
Yu et al. Sci China Chem
Acknowledgements This work was supported by the National Natural
Science Foundation of China (21871255, 21532006, 21873096) and Chi-
nese Academy of Sciences (XDB17020300, XDB17010200).
Conflict of interest The authors declare that they have no conflict of
interest.
Supporting information The supporting information is available online at
http://chem.scichina.com and http://link.springer.com/journal/11426. The
supporting materials are published as submitted, without typesetting or
editing. The responsibility for scientific accuracy and content remains en-
tirely with the authors.
Scheme 8 The control experiments.
1
(a) Jäkel C, Paciello R. Chem Rev, 2006, 106: 2912–2942; (b) Tang
W, Zhang X. Chem Rev, 2003, 103: 3029–3070; (c) Verendel JJ,
Pàmies O, Diéguez M, Andersson PG. Chem Rev, 2014, 114: 2130–
2169; (d) Johnson NB, Lennon IC, Moran PH, Ramsden JA. Acc
Chem Res, 2007, 40: 1291–1299; (e) Wang DS, Chen QA, Lu SM,
Zhou YG. Chem Rev, 2012, 112: 2557–2590; (f) Zhang Z, Butt NA,
Zhang W. Chem Rev, 2016, 116: 14769–14827; (g) Xie JH, Zhu SF,
Zhou QL. Chem Rev, 2011, 111: 1713–1760
2
3
(a) Noyori R, Ohkuma T. Angew Chem Int Ed, 2001, 40: 40–73;
(b) Morris RH. Chem Soc Rev, 2009, 38: 2282; (c) Malacea R, Poli R,
Manoury E. Coord Chem Rev, 2010, 254: 729–752; (d) Baiker A.
Chem Soc Rev, 2015, 44: 7449–7464; (e) Li YY, Yu SL, Shen WY,
Gao JX. Acc Chem Res, 2015, 48: 2587–2598
For selected examples on asymmetric hydrogenation of α-alkyl or aryl
substituted ketones: (a) Ohkuma T, Li J, Noyori R. SynLett, 2004, 8:
1383; (b) Lin H, Xiao LJ, Zhou MJ, Yu HM, Xie JH, Zhou QL. Org
Lett, 2016, 18: 1434–1437; (c) Peach P, Cross DJ, Kenny JA, Mann I,
Houson I, Campbell L, Walsgrove T, Wills M. Tetrahedron, 2006, 62:
1864–1876; (d) Nishiyama H, Park SB, Itoh K. Tetrahedron-Asym-
metry, 1992, 3: 1029–1034; (e) Baratta W, Barbato C, Magnolia S,
Siega K, Rigo P. Chem Eur J, 2010, 16: 3201–3206; (f) Fernández R,
Ros A, Magriz A, Dietrich H, Lassaletta JM. Tetrahedron, 2007, 63:
6755–6763; (g) Liu C, Xie JH, Li YL, Chen JQ, Zhou QL. Angew
Chem Int Ed, 2013, 52: 593–596; (h) Cheng LJ, Xie JH, Chen Y,
Wang LX, Zhou QL. Org Lett, 2013, 15: 764–767
(a) She Z, Wang Y, Wang D, Zhao Y, Wang T, Zheng X, Yu ZX, Gao
G, You J. J Am Chem Soc, 2018, 140: 12566–12573; (b) Mikhel IS,
Ruegger H, Butti P, Camponovo F, Huber D, Mezzetti A. Organo-
metallics, 2008, 27: 2937–2948
(a) Huber D, Kumar PGA, Pregosin PS, Mezzetti A. Organometallics,
2005, 24: 5221–5223; (b) Nagashima H, Fukahori T, Aoki K, Itoh K.
J Am Chem Soc, 1993, 115: 10430–10431
(a) Berenguer JR, Lalinde E, Moreno MT, Sánchez S, Torroba J. Inorg
Chem, 2012, 51: 11665–11679; (b) Falceto A, Carmona E, Alvarez S.
Organometallics, 2014, 33: 6660–6668; (c) Fornies J, Menjon B,
Gomez N, Tomas M. Organometallics, 1992, 11: 1187–1193; (d)
McGuire R Jr., McGuire MC, McMillin DR. Coord Chem Rev, 2010,
254: 2574–2583
(a) Liu S, Rogachev AY. ChemPhysChem, 2018, 19: 2579–2588; (b)
Elschenbroich C, Salzer A. in Organometallics. A Concise Introduc-
tion. 2nd ed. Weinheim: VCH, 1992. 346; (c) Belt ST, Helliwell M,
Jones WD, Partridge MG, Perutz RN. J Am Chem Soc, 1993, 115:
1429–1440
(a) Maestri G, Motti E, Della Ca' N, Malacria M, Derat E, Catellani
M. J Am Chem Soc, 2011, 133: 8574–8585; (b) Yamamoto K, Kimura
S, Murahashi T. Angew Chem Int Ed, 2016, 55: 5322–5326; (c) Li Y,
Wang WH, He KH, Shi ZJ. Organometallics, 2012, 31: 4397–4400;
(d) Canty AJ, van Koten G. Acc Chem Res, 1995, 28: 406–413; (e)
Falvello LR, Forniés J, Navarro R, Sicilia V, Tomás M. Angew Chem
Int Ed Engl, 1990, 29: 891–893; (f) Cámpora J, López JA, Palma P,
Valerga P, Spillner E, Carmona E. Angew Chem Int Ed, 1999, 38:
147–151; (g) Vignolle J, Gornitzka H, Donnadieu B, Bourissou D,
Bertrand G. Angew Chem Int Ed, 2008, 47: 2271–2274; (h) Murahashi
Scheme 9 Hydrogenation of 1,3-diketones bearing different substituents.
9). When electron-donating group methoxy was introduced
to phenyl (1p), trans/cis (>20:1) and full conversion was
obtained. Electron-withdrawing group –CF3 was introduced
to phenyl (1q), lower trans/cis (14:1) was obtained, and
when two –CF3 were introduced (1r), the yield and trans/cis
decreased to 21% and 11:1, respectively. The above results
showed that the diastereoselectivity and reactivity gradually
weakens with the decrease of electron cloud density on the
aromatic ring, which leads to the decrease of interaction
between palladium and aromatic ring. Notably, when aryl on
2-position was replaced by isobutyl, no product was ob-
tained, which indicated that the aryl is very important to not
only keep high diastereoselectivity, but also improve the
reactivity.
4
5
6
4 Conclusions
In summary, a successful diastereoselective reversal was
achieved in palladium-catalyzed asymmetric hydrogenation
of 1,3-diketones through desymmetrization, providing the
chiral β-hydroxy ketones bearing two adjacent stereo centers
including one α-tertiary or quaternary stereocenter with high
yields, enantioselectivities and diastereoselectivities [16].
The control experiments and theoretical studies based on
DFT calculations revealed that the observed diastereoselec-
tivity reversal was resulted from the charge-charge interac-
tion between the palladium and the aromatic ring of the
substrate, which also promoted the reactivity. We envision
that the in situ interaction strategy would open a new window
for the development of asymmetric synthesis.
7
8

7
Yu et al. Sci China Chem
T, Takase K, Oka M, Ogoshi S. J Am Chem Soc, 2011, 133: 14908–
14911; (i) Kocovský P, Vyskocil Š, Císařová I, Sejbal J, Tišlerová I,
Smrcina M, Lloyd-Jones GC, Stephen SC, Butts CP, Murray M,
Langer V. J Am Chem Soc, 1999, 121: 7714–7715; (j) Yin J, Rainka
MP, Zhang XX, Buchwald SL. J Am Chem Soc, 2002, 124: 71162;
(k) Sköld C, Kleimark J, Trejos A, Odell LR, Nilsson Lill SO, Norrby
PO, Larhed M. Chem Eur J, 2012, 18: 4714–4722
(a) Catellani M, Mealli C, Motti E, Paoli P, Perez-Carreño E, Pregosin
PS. J Am Chem Soc, 2002, 124: 4336–4346; (b) Della Ca’ N, Fontana
M, Motti E, Catellani M. Acc Chem Res, 2016, 49: 1389
Chem, 2009, 2009(4): 510–515; (c) Chen J, Liu D, Butt N, Li C, Fan
D, Liu Y, Zhang W. Angew Chem Int Ed, 2013, 52: 11632–11636;
(d) Jiang W, Zhao Q, Tang W. Chin J Chem, 2018, 36: 153–156
13 For coordinative group (acetamido, toluenesulfonamido and benza-
mido) substituted ketones, trans products could be also obtained with
high enantioselectivity owing to the coordination-directed interaction,
for some recent examples, see: (a) Lu B, Wu X, Li C, Ding G, Li W,
Xie X, Zhang Z. J Org Chem, 2019, 84: 3201–3213; (b) Cotman AE,
Lozinšek M, Wang B, Stephan M, Mohar B. Org Lett, 2019, 21: 3644–
3648
9
10 Tanaka D, Romeril SP, Myers AG. J Am Chem Soc, 2005, 127:
10323–10333
14 Duan Y, Li L, Chen MW, Yu CB, Fan HJ, Zhou YG. J Am Chem Soc,
2014, 136: 7688–7700
11 (a) Chen QA, Ye ZS, Duan Y, Zhou YG. Chem Soc Rev, 2013, 42:
497–511; (b) Xie J, Zhou Q. Acta Chim Sin, 2012, 70: 1427; (c) Wang
YQ, Lu SM, Zhou YG. Org Lett, 2005, 7: 3235–3238; (d) Zhou XY,
Wang DS, Bao M, Zhou YG. Tetrahedron Lett, 2011, 52: 2826–2829
12 (a) Raja R, Thomas JM, Jones MD, Johnson BFG, Vaughan DEW. J
Am Chem Soc, 2003, 125: 14982–14983; (b) Goulioukina NS, Bon-
darenko GN, Bogdanov AV, Gavrilov KN, Beletskaya IP. Eur J Org
15 The summed interaction between the active Pd-H and phenyl/methyl
also evaluated. See details in Supporting Information.
16 During our preparation of this manuscript, a related irdium-catalyzed
hydrogenative desymmetrization of 1,3-diketones was reported by
Zhang’s group, please see: The chiral Ir/f-ampha complex catalyst was
used, and cis product was obtained. In our work, trans products were
observed.. Gong Q, Wen J, Zhang X. Chem Sci, 2019, 10: 6350–6353