Accepted Manuscript

β-Glucosidase activity in almond seeds

Jorge Del Cueto, Birger Lindberg Møller, Federico Dicenta, Raquel Sánchez-Pérez

PII: S0981-9428(17)30429-1
DOI: 10.1016/j.plaphy.2017.12.028
Reference: PLAPHY 5092

To appear in: Plant Physiology and Biochemistry

Received Date: 29 September 2017

Revised Date: 12 December 2017
Accepted Date: 15 December 2017

Please cite this article as: J. Del Cueto, B.L. Møller, F. Dicenta, R. Sánchez-Pérez, β-Glucosidase
activity in almond seeds, Plant Physiology et Biochemistry (2018), doi: 10.1016/j.plaphy.2017.12.028.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
1 β-Glucosidase activity in almond seeds

2 Jorge Del Cueto1,2,3, Birger Lindberg Møller2,3 Federico Dicenta1 and Raquel Sánchez-

3 Pérez*1, 2, 3

PT
1
5 Department of Plant Breeding, CEBAS-CSIC, P.O. Box 164, 30100 Campus
6 Universitario de Espinardo, Murcia, SPAIN.

RI
2
7 University of Copenhagen, Faculty of Science, Plant Biochemistry Lab, DK-1871

SC
8 Copenhagen C, DENMARK
3
9 VILLUM Research Center for Plant Plasticity, DK-1871 Frederiksberg C, Denmark

U
10 *rasa@plen.ku.dk
11
AN
12
13
14
M

15
16
17
D

18
TE

19
20
21
22
EP

23
24
25
C

26
AC

27
28
29
30
31
32
33
34
35
36
 ACCEPTED MANUSCRIPT
37
38 Abstract

39 Almond bitterness is the most important trait for breeding programs since bitter-

40 kernelled seedlings are usually discarded. Amygdalin and its precursor prunasin are

41 degraded by specific enzymes called β-glucosidases. In order to better understand the

PT
42 genetic control of almond bitterness, some studies have shown differences in the

43 location of prunasin hydrolases (PH, the β-glucosidase that degrades prunasin) in sweet

RI
44 and bitter genotypes. The aim of this work was to isolate and characterize different PHs

SC
45 in sweet- and bitter-kernelled almonds to determine whether differences in their

46 genomic or protein sequences are responsible for the sweet or bitter taste of their seeds.

47
U
RNA was extracted from the tegument, nucellus and cotyledon of one sweet (Lauranne)
AN
48 and two bitter (D05-187 and S3067) almonds throughout fruit ripening. Sequences of

49 nine positive PH clones were then obtained from all of the genotypes by RT-PCR and
M

50 cloning. These clones, from mid ripening stage, were expressed in a heterologous
D

51 system in tobacco plants by agroinfiltration. The PH activity was detected using

TE

52 prunasin by the Feigl-Anger method and quantifying the cyanide released. Furthermore,

53 β-glucosidase activity was detected by Fast Blue BB salt and Umbelliferyl method.
EP

54 Differences at the sequence level (SNPs) and in the activity assays were found, although

55 no correlation with bitterness was concluded.

C
AC

56

57 Keywords: β-glucosidases, agroinfiltration, bitterness, cyanogenic glucosides, prunasin

58 hydrolase, Prunus dulcis.

59

60
 ACCEPTED MANUSCRIPT
61 1. Introduction

62 Bitterness in almonds (Prunus dulcis Miller D.A. Webb syn. Prunus amygdalus

63 Batsch) is one of the most important and widely studied traits in this species. Almond

64 bitterness is a monogenic trait, and the sweet taste allele is dominant over the bitter taste

PT
65 allele [1,2,3,4]. The gene responsible for bitterness in almonds is called Sweet kernel

66 (Sk), and it is located in linkage group five in all of the almond genetic linkage maps

RI
67 described so far [9,13,14]. Commercial sweet genotypes can be homozygous (SkSk) or

68 heterozygous (Sksk) for this trait, while slightly bitter genotypes are always

SC
69 heterozygous (Sksk), and bitter genotypes are homozygous recessive (sksk) [5,6]. The

U
70 bitter trait is controlled by the genotype of the seed mother [1,3,7,8,9]. As a result, all
AN
71 the fruits of an almond will be sweet, slightly bitter or bitter, and the influence of both

72 progenitors will be seen in the next generation [10,11,12].

M

73 Cyanogenic glucosides (CNGLcs) are β-glucosides of α-hydroxinitriles

D

74 [16,17,18]. These compounds are present in more than 3,000 plant species [19,20,21]
TE

75 like sorghum, cassava and Rosaceous stone fruits. Their main function is to defend

76 plants against pathogens and predators [22,23].

EP

77 There are two cyanogenic glucosides present in almonds: prunasin (mono-

78 glucoside of R-mandelonitrile) and amygdalin [di-glucoside of R-mandelonitrile with a

C

79 β-(1,6) bond, called gentiobioside] [16,20,25,26,27,28,29,30,31]. These two compounds

AC

80 are de novo synthesized in the almond seed [20], and the amygdalin content is

81 significantly higher in mature bitter almond seeds than in slightly bitter or sweet seeds

82 (between 200 and 1000-fold higher, respectively). On the other hand, there is no clear

83 relationship between the prunasin levels in the vegetative parts of the plant and the

84 bitterness of the fruit [30].

 ACCEPTED MANUSCRIPT
85 The almond fruit consists of maternal tissues (endocarp, mesocarp, tegument –

86 also called the seed coat – and nucellus) and parental tissues (endosperm (2n (mother) +

87 n (father)) and cotyledon (n (mother) + n (father)). The first difference between sweet

88 and bitter almonds can be found in the tegument, as prunasin only accumulates in the

89 bitter teguments throughout seed development [20].

PT
90 Figure 1 describes the metabolism of prunasin and amygdalin, which can be

RI
91 divided into the following three parts: biosynthesis, bioactivation and detoxification.

Biosynthesis: This route starts with the amino acid L-phenylalanine (Phe) [32], which is

SC
92

93 N-hydroxylated and converted to prunasin by two cytochromes P450 (CYP79 and

U
94 CYP71) and an UDP-glucosyltransferase (UGT1) [31]. Finally, prunasin is converted to
AN
95 amygdalin through another unknown UDP-glucosyltransferase (UGT2).

96 Bioactivation: Amygdalin can be degraded by amygdalin hydrolase (AH) to produce

M

97 prunasin and glucose. This prunasin is subsequently degraded by prunasin hydrolase

D

98 (PH), releasing mandelonitrile and glucose [17,20,33,34,35,36,37,38]. Both enzymes

TE

99 (AH and PH) are β-glucosidases, which belong to family 1 of the glycoside hydrolases.

100 This family of enzymes (β-glucosidases) catalyzes the hydrolysis of the β-glycosidic
EP

101 bonds between rests of carbohydrates or between a carbohydrate and a rest aglycone

102 [18]. Both cyanogenic glucosides and β-glucosidases are separated in different places of
C

103 the cell, in this moment they are inactive. Only when tissue is damaged by an attack of
AC

104 an herbivore for example, cyanogenic glucosides are bioactivated by β-glucosidases,

105 releasing, among other products, hydrogen cyanide [18,36,39,40,41], providing an

106 immediate chemical defence against herbivores. This compartmentalization of cyanogenic

107 glucosides and β-glucosidases prevents any premature HCN release prior to tissue

108 disruption [29,42,43]. This prevention also has been observed in another species like
 ACCEPTED MANUSCRIPT
109 Plantago lanceolate and Plantago major between iridoid glycosides and β-glucosidases

110 [44]. Finally, mandelonitrile lyase 1 (MDL1) hydrolyzes mandelonitrile [45,46]

111 producing benzaldehyde (which gives the bitter taste) and releasing cyanide (which is

112 toxic and mortal) [47]. Other plants and organisms like insects are able to release this

113 toxic compound in a process called cyanogenesis [42].

PT
114 Detoxification: There is a cyanide-detoxification pathway through the nitrilases

RI
115 producing a nitrogen source for the plant. This route consists of the following two steps:

116 a) β-cyanoalanine synthase catalyzes the reaction between HCN and L-cysteine to

SC
117 synthesize β-cyanoalanine Ala [48] and b) nitrilases catalyze the production of

ammonia, L-aspartic acid and L-asparagine from H2O and β-cyanoalanine

U
118 AN
119 [49,50,51,52].

120 In almond the majority of the above mentioned enzymes have not yet been
M

121 characterised, except for UGT1 (UGT85A19 [31]); MDL1 [46]; and [38]. Finally, two

122 PHs have been identified in almond (Ph691 and Ph692) [38]. To date, however, no Ph
D

123 gene has been expressed and had its protein activity tested.
TE

124 As already mentioned, the bitter flavour in almond is produced by the presence
EP

125 of amygdalin in the almond kernel, so it would be logical to think that the difference

126 between sweet and bitter almonds would be related to differences in biosynthesis,
C

127 degradation or detoxification that would or not allow the accumulation of amygdalin in
AC

128 the seed. However, it was observed that all enzymes and substrates necessary to produce

129 or degrade prunasin and amygdalin were present in both bitter and sweet almonds [20].

130 Why then do only bitter almond kernels accumulate amygdalin?.

131 From the genetic point of view, given that bitter is the original trait in wild

132 almonds, the alteration of just one allele could prevent the accumulation of amygdalin.
 ACCEPTED MANUSCRIPT
133 The question is thus whether this change would prevent the synthesis of amygdalin or

134 favour its degradation. In the literature, there are two theories to explain why an almond

135 is sweet or bitter, the first based on biosynthesis and the second on degradation.

136 In the first theory, it was proposed that the anabolic enzyme glucosyltransferase

137 was responsible for the accumulation of amygdalin [26]. However, since we know that

PT
138 the bitter trait is recessive, an anabolic enzyme cannot be directly responsible for bitterness.

RI
139 In the second theory, it was focused on the catabolic enzymes, suggesting the existence

140 of a catabolic mechanism that is responsible for the presence of amygdalin in bitter

SC
141 almond kernels, based on the β-glucosidase activity and the different location of

142 prunasin hydrolases in the apoplast or symplast of the tegument [20, 37, 38]. For this

U
reason, we have focused in prunasin hydrolases in this study.
AN
143

144 The aim of this work was to isolate and characterize different tissue-localized
M

145 PHs in sweet- and bitter-kernelled almonds to determine whether differences in their

146 genomic or protein sequences are responsible for the sweet or bitter taste of their seeds.
D
TE

147

148 2. Materials and Methods

EP

149 2.1. Plant material

C

150 The almonds (Prunus dulcis (Miller) D.A. Webb) used in this study, from the
AC

151 genotypes Lauranne (SkSk, sweet), D05-187 (sksk, bitter) and S3067 (sksk, bitter), were

152 provided by the Almond Breeding Program of the “Centro de Edafología y Biología

153 Aplicada del Segura” (CEBAS-CSIC, Murcia, Spain).

154

155
 ACCEPTED MANUSCRIPT
156 2.2. The evolution of PHs during kernel development

157 In order to study the evolution of the PHs synthesis, fruits from the Lauranne,

158 D05-187 and S3067 genotypes were harvested during the beginning (JD89), middle

159 (JD130) and final (JD167) stages of kernel development. “JD” stands for Julian Days,

160 meaning days after the first of January. The teguments, nucellus+endosperms (difficult

PT
161 to separate) and cotyledons were isolated, ground with liquid nitrogen, stored at -80 ºC

RI
162 and analyzed separately. RNA extraction was done with 100 mg of each tissue using an

163 Ultra Clean Plant RNA Isolation kit (MO BIO) according to manufacturer’s

SC
164 instructions.

U
165 As previously shown, PH691 and PH692 are present in both sweet and bitter
AN
166 almonds [38]. PH691 was selected for the present study due to its high homology (73%-

167 94% protein similarity) compared to other PHs previously described and characterized
M

168 in P. serotina [36]. PH692, on the other hand, showed between 71% and 79% protein

169 similarity with other PHs.

D
TE

170 RT-PCR was performed to obtain cDNA using SuperScript III Taq DNA

171 Polymerase (Invitrogen, http//www.invitrogen.com/), where 1µg of total RNA was used
EP

172 with the primers 691ATG and 691UGA (Table 1).

173 The amplification conditions were as follows: 30 min at 55 ºC, 30 cycles of (2

C

174 min at 94 °C, 15 s at 94 ºC, 30 s at 59 ºC) and one cycle of 2:30 min at 68 ºC, and one
AC

175 cycle of 5 min at 68 ºC. PCR products were analysed by electrophoresis in 1% agarose

176 gel with a 1kb Plus Ladder (Invitrogen) or Lambda Hind III (Thermofisher Scientific).

177 The expected band size was 1638 bp, which was cut and purified following the protocol

178 of the Ultra Clean DNA Purification Kit (MO BIO).

179
 ACCEPTED MANUSCRIPT
180 2.3. Comparison of PH clone sequences

181 2.3.1. Cloning in Escherichia coli: The amplified cDNA was cloned into pGEM-T easy

182 Vector (Promega). After overnight ligation at 4 ºC, chemical transformation took place

183 by adding 2 µL of ligation to E. coli competent cells in a total volume of 50 µL. The

184 mixture was put in ice for 20 min and then subjected to heat shock for 45 s at 42 ºC,

PT
185 followed by an additional 2 min in ice. Immediately afterwards, 950 µL SOC medium

RI
186 was added, and then the mixture was incubated for 1.5 h at 37 ºC with shaking. Finally,

187 100 µL of each transformation culture was spread on LB agar plates with kanamycin

SC
188 (50 µg/mL) and then incubated overnight at 37 ºC. Nested PCR was performed with the

internal primer pairs 27F and 27R (Table 1) to check the positive clones using the

U
189 AN
190 following PCR conditions: 2 min at 95 ºC, 30 cycles of 30s at 95 ºC, 30 s at 57 ºC and 1

191 min at 72 ºC, and one cycle of 5 min at 72 ºC. The expected band size was 1050 bp.
M

192 2.3.2. Minipreps and sequence analyses: Once positive colonies were identified by

193 nested PCR, plasmid was purified using a Pure Link Quick Plasmid Miniprep Kit
D

194 (Invitrogen). DNA was quantified based on A280 (Nanodrop; Thermo Scientific,
TE

195 http://www.thermofisher.com) and sent to sequence Secugen (http://www.secugen.es/)

EP

196 for sequencing. Sequences were visualized with the software BioEdit and BlastX versus

197 National Center for Biotechnology Information (NCBI).

C

198 Phylogenetic tree, signal peptide and motif analyses: Neighbor-joining phylogenetic
AC

199 trees were constructed using MEGA 5.5 software [53];

200 http://www.megasoftware.net/index.html) with 1,000 bootstrap trials performed. For

201 signal peptide prediction, three aspects were analysed: the peptide size, cleavage site

202 and the place the protein would be led to. To do this, we used the following two web

203 servers: SignalP 4.1 (http://www.cbs.dtu.dk/services/SignalP/) and WoLF PSORT

 ACCEPTED MANUSCRIPT
204 (http:// wolfpsort.org/). For the N-glycosylation sites, the NetNGlyc 1.0 Server was used

205 (http://www.cbs.dtu.dk/services/NetNGlyc/).

206 2.4. Detection of β-glucosidase activity

207 The β-glucosidase activity was determined in Nicotiana benthamiana plants

PT
208 transformed with Agrobacterium tumefaciens. Clones used for this experiment came

209 from the mid ripening stage (JD 130) as in this stage the Ph691 was expressed in all the

RI
210 tissues studied.

SC
211 2.4.1. Cloning in Agrobacterium tumefaciens: Nine (three genotypes times three tissues)

212 cDNA sequences of the Ph691 gene were amplified with the primers 691B1F and

U
213 692B2R containing the attB1 and attB2 Gateway cloning sites (Table 1) using PCR Hot
AN
214 Master under the following conditions: 2 min at 94 ºC, 35 cycles of 20 s at 94 ºC, 20 s at

215 57 ºC and 2 min at 65 ºC, and one cycle of 5 min at 65 ºC. The PCR products were
M

216 cloned by Gateway recombination in two cloning steps: 1) using Gateway BP Clonase
D

217 II Enzyme Mix (Invitrogen), where the PCR products were introduced into the
TE

218 pDONOR207 entry clone, and 2) using LR Clonase II Enzyme Mix (Invitrogen) into

219 pJAM1502 as the destination vector [54]. The bacteria strains used in the transformation
EP

220 were E. coli DH5α chemically competent cells (Invitrogen) and AGL1 (A. tumefaciens)

221 electro-competent cells (2.2 V, 25 µF, 400Ω in a BIORAD machine). Finally, AGL1
C

222 cells from the different tissues containing Ph691 were selected on LB agar + rifampicin
AC

223 (25 µg/ml) + kanamycin (50 µg/ml).

224 2.4.2. Agroinfiltration

225 Overnight cultures of A. tumefaciens containing Ph691 from the different tissues

226 (tegument, nucellus and cotyledon) from the three genotypes in pJAM1502 and the

227 gene-silencing inhibitor protein p19 [55] were grown in LB-containing kanamycin (50
 ACCEPTED MANUSCRIPT
228 µg/ml) and rifampicin (25 µg/ml), while p19 was grown alone with kanamycin,

229 rifampicin and tetracycline (10 µg/ml). Cultures were harvested by centrifugation at

230 3,000 rpm for 10 min and resuspended to an OD600 of 2.0 in 1 mL of acetosyringone

231 solution: 50 mL water, 0.5 mL MES 10 mM, 0.5 mL MgCl2 10mM and 5 µL

232 acetosyringone 100µM. After 2 hours of incubation at room temperature, a combined

PT
233 solution was made in a total volume of 2 mL: 0.5 mL of p19, 0.5 mL of Agro solution

RI
234 (OD600 of 2.0) containing Ph691 from each genotype and 1 mL of acetosyringone

235 solution. N. benthamiana leaves were agroinfiltrated using a 1 mL syringe. After 4 to 5

SC
236 days, leaf discs (1 cm diameter) were cut from infiltrated leaves. To check if the nine

237 transformed clones expressing Ph691 had β-glucosidase activity, the following two

238 methods were used.

U
AN
239 2.4.3. Fast Blue BB salt assay: This assay for measuring β-glucosidase activity was
M

240 based on of Sánchez-Pérez et al., 2009 [37], consisting in the incubation of BNG

241 (glucopyranoside) (Sigma Aldrich) as a general substrate for β-glucosidases in the presence
D

242 of Fast Blue BB salt (Sigma-Aldrich), which gives a reddish colour. Firstly, agroinfiltrated
TE

243 N. benthamiana leaf samples were ground in liquid nitrogen. Then, 20 mM MES buffer,

244 pH 6, was added in a 1:3 ratio. Protein extracts were centrifuged at 20,000 g for 30 min
EP

245 at 4 ºC. The supernatant was collected and quantified based on A280 (Nanodrop;
C

246 Thermo Scientific, http://www.thermofisher.com). The entire process was carried out in
AC

247 cold conditions. The protein composition was analyzed by SDS–PAGE (12% gel)

248 following the application of protein (18 µl, concentration around 10 mg/ml), combined

249 with 10% bromo-phenol blue (2 µl), 100% glycerol (10 µl) and 0.5% SDS (10 µl) in a

250 final volume of 40 µl. At the end of electrophoresis (2 h, 175 V, 4 ºC), the gels were

251 washed (2 x 10 min) in Fast Blue BB buffer [50 mM sodium citrate, 100 mM phosphate

252 (pH 5.8)]. Solution A (15 mg Fast Blue BB salt + 20 ml Fast BB Buffer) and solution B
 ACCEPTED MANUSCRIPT
253 [20 mg 6 bromo-2-naphthyl β-D-glucopyranoside + 200 µL DMF

254 (dimethylformamide)] were mixed and put with the gel by incubation at 37ºC for 2

255 hours with gentle shaking in foil paper.

256 2.4.4. Umbellyferil substrate: This second method uses umbelliferyl 4-methyl-β-D-

257 glucoside (Sigma Aldrich), which is an excellent fluorogenic substrate for detecting β-

PT
258 glucosidases. Leaf discs were put in a tube with 150 µL MES buffer (Sigma Aldrich) 20

RI
259 mM pH 6 and 6 µL of general substrate 4-methyl-umbelliferyl-β-D-glucoside (Sigma

260 Aldrich) 25 mM. Control was made with a leaf disc of p19. The tubes were incubated

SC
261 for 15 min at 37 ºC and observed under UV light. The β-glucosidase activity was

detected by fluorescence.

U
262 AN
263 2.5. Verification of PH activity

264 Clones used for this experiment came from the mid ripening stage (JD 130), as in this
M

265 stage the Ph691 was expressed in all the tissues studied.
D

266 2.5.1. Feigl-Anger: The Feigl-Anger method is a qualitative assay used to detect
TE

267 hydrogen cyanide released. If PH was active versus prunasin, then cyanide could be

268 detected using Feigl-Anger paper [56]. To prepare the paper, we separately dissolved 5
EP

269 g of copper ethylacetoacetate (Alfa Aesar, http://www.alfa.com/) and 4.4-methylenebis

270 (N, N-dimethylaniline) (Sigma-Aldrich) in 0.5 liters of water each and then combined
C

271 both solutions. We then cut Whatman 3MM paper (Sigma-Aldrich) to 8 x 12 cm, wet it
AC

272 with the solution and then dried it. The paper was stored at 4 ºC until use. A total of 200

273 µL of 20 mM MES buffer pH 6.5 plus 15 µL substrate 4 mM were added to a 96-well

274 microtiter plate. The following substrates were used: prunasin, dhurrin, linamarin,

275 lotaustralin (cyanogenic monoglucosides), amygdalin and linustatin (cyanogenic

276 diglucosides). An agroinfiltrated leaf disc from an N. benthamiana plant was added and
 ACCEPTED MANUSCRIPT
277 triturated in each well; three ELISA well-discs were made per Ph691construction.

278 Feigl-Anger paper was put between the plate and the lid. Cyanogenesis was detected

279 after two hours of exposure, when the Feigl-Anger paper turned blue, showing the

280 oxidation of a tetrabase in the presence of a copper salt due to the cyanide released.

281 2.5.2. Cyanide release: The cyanide released from the reaction of crude protein extracts

PT
282 and substrate was determined using the colorimetric method described by Lambert [57]

RI
283 and modified by Halkier and Møller [58]. The reaction mix (total volume: 100 µL)

284 contained 10 µL of substrate 1 mM, 10 µL of protein (1 mg/ml) and 80 µL of MES 20

SC
285 mM buffer pH 6. Samples without substrate were used as control. The following step

was an incubation (10 min, 30 ºC, 300 rpm shaking) and the samples were frozen in

U
286 AN
287 closed Eppendorf tubes with the presence of liquid nitrogen to avoid loss of volatile

288 HCN formed. While samples thawed at room temperature, 40 mL of 6 M NaOH was
M

289 added. A 60-mL aliquot from each sample was transferred to a 96-well microtiter plate.

290 Immediately, the following reagents were added to each well: 12.5 µL of 100% glacial
D

291 acetic acid (Sigma-Aldrich), 50 µL of reagent A (made by dissolving 50mg of N-

TE

292 chlorosuccinimide (Sigma-Aldrich) and 125 mg of succinimide (Sigma-Aldrich) in 50

293 mL of water) and 50 µL of reagent B (made by mixing 15 mL of pyridine (Sigma-

EP

294 Aldrich) and 3 g of barbituric acid (Fluka) in 50 ml of water). After 5 min of incubation
C

295 at room temperature, the wells were scanned between 450 and 700 nm with a peak
AC

296 reading made at 584 nm. The cyanide released was calculated against KCN standards

297 made in 1 M NaOH. For the standard deviation, at least, two replicates per genotype

298 were carried out.

299

300
 ACCEPTED MANUSCRIPT
301 3. Results

302 3.1. PH evolution during kernel development

303 Our findings regarding the presence or absence of PHs in each tissue are

304 described below.

PT
305 3.1.1. Stage 1: Beginning of development (JD 89) (Figure 2.a): During this period,

306 the immature kernel was composed of a nucellus and tegument. Cotyledons had not yet

RI
307 developed. In this stage, the DNA band characteristic of the Phs (1,638 bp) was not

SC
308 detected in the sweet genotype. In the bitter genotype (S3067), it was more intense in

309 the tegument than in the nucellus. At the beginning we could only amplified S3067

310

U
since it was the earliest genotype to ripen.
AN
311 3.1.2. Stage 2: Halfway through development (JD 130) (Figure 2.b): In this period,
M

312 endosperm and cotyledons began to grow inside of the tegument, and the nucellus was

313 still visible. The amplified DNA bands of Ph691 from the bitter genotype S3067 were
D

314 weaker in the tegument than in Stage 1, but they were similar in the nucellus. For the
TE

315 first time, we can detect it in the cotyledon, being so strong that the band migrated a bit

316 further when compared to the bands amplified in the other tissues. In the sweet
EP

317 Lauranne, Ph691 was also amplified in the three tissues.

C

318 3.1.3. Stage 3: Final development (JD 167) (Figure 2.c): In this last stage, the kernel
AC

319 was only made up of the cotyledons and the dried tegument. The Ph691 gene was only

320 amplified from the cotyledons and not in the tegument, in both the sweet and bitter

321 genotypes. This makes sense, as at this stage the tegument is a very dry tissue

322 surrounding the cotyledon.

323 3.2. Comparison of PH clone sequences

 ACCEPTED MANUSCRIPT
324 The amplified sequences of the Ph691 gene contained 13 exons and 12 introns (Figure

325 3). The full-length cDNA in our 9 clones had 1,635 bp, codifying a 544 amino acid

326 protein, which contained 5 putative N-glycosylation sites (Table 2). The PH691s studied

327 here had ITENG and NEP motifs characteristic of the active site of the glycoside

328 hydrolase family 1 and INKKGIEYY motif specific for PH (Figure 3). No differences

PT
329 were detected between the active sites in sweet and bitter genotypes.

RI
330 Regarding the signal peptide (Table 2), the size was 27 amino acids for most of

331 the genotypes and tissues and 23 amino acids for the Lauranne tegument, the Lauranne

SC
332 cotyledon and the D05-187 cotyledon. The cleavage site, between amino acids 22 and

23, was therefore different in these last three genotypes. In the rest of the genotypes, the

U
333 AN
334 site was between amino acids 26 and 27. According to signal peptide sequences found

335 with the SignalP 4.1 Server program, the protein was always targeted to the vacuole
M

336 except in the tegument of S3067, in which case the protein was extracellular.

337 Table 3 shows amino acid differences between the 9 PH sequences studied here.
D

338 Although some nucleotide differences were detected in the exons, they were only
TE

339 translated into an amino acid change in 17 cases (in exons 1, 6, 7, 9, 11, 12 and 13).
EP

340 Three of these changes were found inside the signal peptide, located in the amino acids

341 12, 20 and 22. Differences were not consistent between the sweet and bitter genotypes
C

342 or between tissues. However, that small differences of each genotype could have a
AC

343 consequence in the activity of the protein expressed.

344 Also we compared sequences of tissues between genotypes (tegument, nucellus

345 and cotyledon) (Supplemental Figure 1), we found four SNPs in the amino acids 12,

346 340, 410 and 446. In the four cases the difference was detected in the nucellus in

347 comparison with tegument and cotyledon.

 ACCEPTED MANUSCRIPT
348 When we compared the PHs studied here, the percentage of nucleotide similarity

349 ranged between 98 and 99 in our 9 sequences (Table 4). No differences were found

350 between sweet and bitter genotypes. As expected, there was high homology (99%) with

351 the two sequences of Ph691 described in almond [38]. Compared to black cherry

352 sequences, PsPh1 and PsPh5 were the most similar (94% and 96%, respectively), and

PT
353 PsPh3 showed a greater difference than the others (Table 4).

RI
354 Phylogenetic analysis clearly separated one group of sequences from the rest based on

355 the positions of the clones in the tree and the percentage of bootstrap support (Figure 4).

SC
356 This group contained the nine PHs characterized in this study and, as expected, the two

PdPH691s from a sweet (Ramillete, R) and a bitter (S3067, S) [38]. According to the

U
357 AN
358 PHs from black cherry, PsPH5 had the closest amino acid sequence to the PHs from

359 almond, which agrees with the nucleotide similarity described before. In this analysis,
M

360 AH from P. serotina appears as a distinct branch, so we clearly observed the differences

361 between all the PHs and the AH.

D

362 3.3. Detection of β-glucosidase activity

TE

363 3.3.1. Fast Blue BB method: With this method, the nine PHs showed β-glucosidase
EP

364 activity, except in D05-187 tegument and the control p19 (Figure 5a). Different bands

365 of different sizes (between 50 and 75 kD) were observed. Moreover, some clones
C

366 showed more than one band (Lauranne nucellus and cotyledon), which could indicate
AC

367 the presence of different PH isoforms. These differences were not related to the sweet or

368 bitter phenotype.

369 3.3.2. Umbelliferyl method: With this method, β-glucosidase activity was detected in

370 all the clones except in the D05-187 tegument, which showed low fluorescence, similar
 ACCEPTED MANUSCRIPT
371 to the p19 control (background) (Figure 5b). The results were thus similar to the Fast

372 BB results described above.

373 3.4. Verification of PH activity

374 3.4.1. Feigl-Anger assay: The addition of prunasin to the extract from agroinfiltrated N.

PT
375 benthamiana leaves showed the PH activity of our nine PHs, both in the sweet and

376 bitter genotypes. As expected, the negative control p19 did not show any activity

RI
377 (Figure 6). Hydrolase activity was also detected in Ln, Lc, Dn and Dc with the

monoglucoside dhurrin, but not with the diglucosides amygdalin and linustatin or with

SC
378

379 the other monoglucosides assayed (linamarin and lotaustralin) (Figure 6), showing that

U
380 the PH has a high substrate specificity for prunasin.
AN
381 3.4.2. CN assay: The assay to quantify the PH activity based on cyanide release was

382 performed with the substrates that had positive hydrolase activity (prunasin and dhurrin)
M

383 and with amygdalin. As expected, the results showed high PH activity, low hydrolase
D

384 activity when incubated with dhurrin, and very low activity with amygdalin, as the
TE

385 levels detected were similar to those obtained with the negative control p19 (Figure 7).

386 With the substrate prunasin, the PH from the Lauranne nucellus liberated the
EP

387 highest amount of cyanide, with more than 500 nmol CN-/g per agroinfiltrated leaf.

388 After the PH of the Lauranne nucellus, the Lauranne cotyledon PH liberated the next
C

389 highest amount of cyanide, followed by the PHs of the D05-187 cotyledon and the D05-
AC

390 187 nucellus, which liberated between 300 and 400 nmol CN-/g. The rest of the PHs

391 liberated between 100 and 200 nmol CN-/g. Only the D05-187 tegument PH showed a

392 very low level of cyanide release, which agrees with the results of the Fast BB and

393 umbelliferyl assays described above. With dhurrin, only the Lauranne tegument PH had
 ACCEPTED MANUSCRIPT
394 significant cyanide release (almost 200 nmol CN-/g), whereas the rest of the PHs

395 showed very low activity.

396 4. Discussion

397 4.1. The evolution of PHs during kernel development

PT
398 The kernel development stages described in this work were similar to those

399 previously observed in almond [59, 60] and in black cherry [28]. Ph691 appeared in the

RI
400 tegument and nucellus of the bitter S3067 at the beginning of the development period;

SC
401 in the tegument, nucellus and cotyledon in both the bitter and sweet genotypes halfway

402 through the development period; and only in both cotyledons in the final development

403 stage.
U
AN
404 4.2. Comparison of PH clone sequences
M

405 The amplified sequences of the Ph691 contained 13 exons and 12 introns (Figure

406 3) as the same number of exons previously described in almond [38] and in black cherry
D

407 [36]. On the other hand, the 5 PHs identified in black cherry had between 537 and 549
TE

408 amino acids, compared with the 544 amino acids of our PHs, and between 2 and 8 N-

409 glycosylation sites [36] (5 in our work).

EP

410 Regarding the signal peptide (Table 2), according to Sánchez-Pérez [38], the
C

411 predicted cleavage site for PH691 was between amino acids 26 and 27 (TNA-AR),
AC

412 whereas for PH692 it was between amino acids 22 and 23 (ALA-DT). Therefore, the

413 signal peptide size of 27 amino acids for most of the genotypes and tissues in this study

414 was more similar to PH691 as described [38]. On the other hand, the Lauranne

415 tegument, Lauranne cotyledon and D05-187 cotyledon had 23 amino acids, making

416 them more similar to PH692.

 ACCEPTED MANUSCRIPT
417 In comparison, it was observed small differences in the N-terminals of different

418 isoforms of PH from black cherry [35]. Moreover, it was also shown [18] that β-

419 glucosidases contained an N-terminal signal peptide that placed them in the symplast or

420 apoplast. Previously, it was observed a different localization (symplastic versus

421 apoplastic) for these catabolic enzymes in sweet and bitter genotypes [38]. Our results

PT
422 disagree with this former work since we were not able to distinguish between sweet and

RI
423 bitter genotypes based on differences in the signal peptide. One reason could be that

424 [38] inmunolocalized the PHs with an antibody common to more than one PH. It would

SC
425 be interesting to clone Ph692 to see how it behaves in the bitter teguments. The PH691s

426 studied here had ITENG and NEP motifs characteristic of the active site of the

427
U
glycoside hydrolase family 1 and INKKGIEYY motif specific for PH (Figure 3). These
AN
428 active-site motifs have also been described in PHs [34] and AH [17] from black cherry.
M

429 No differences were detected between the active sites in sweet and bitter genotypes.

430 Another difference was observed between amino acid sequences (position 414)
D

431 in sweet and bitter genotypes [36], but we did not detect this difference in our clones.
TE

432 Respect to the Supplemental Figure 1 in which we observed four SNPs, these
EP

433 differences found in the nucellus respect to the rest of the tissues were not translated in

434 any difference in the activity.

C

435 Given the similarity between the amino acid sequences of our PHs and other
AC

436 sequences of different species, we can deduce that almond PHs belong to the BGA

437 family of β-glucosidases for their: a) similar nucleotide and amino acid lengths, b)

438 signal peptide size, c) ITENG, NEP and INKKGIEYY motifs, d) number of exons and

439 introns, and e) N-glycosylation sites. BGAs (glycoside hydrolase family 1) are enzymes

440 that catalyse glucoside hydrolysis. They have an amino acid residue (Ile/Val-Thr-Glu-
 ACCEPTED MANUSCRIPT
441 Asn-Gly) motif. This family includes β-glucosidases, phosphor-β-glucosidases, thio-β-

442 glucosidases and β-galactosidases from archaebacteria, bacteria, plants and mammals.

443 We observed similarities between our PH clones and other PHs described before

444 in almond and black cherry. We could not, however, see any differences that would

445 allow us to separate sweet and bitter genotypes at the sequence level or in terms of their

PT
446 location. There were small differences but none exclusive to one specific genotype. The

RI
447 next step was therefore to determine whether these clones had β-glucosidase and PH

448 activity and whether the small differences in the sequences had any influence over the

SC
449 prunasin hydrolase activity.

U
450 4.3. Detection of β-glucosidase activity
AN
451 We observed that all PHs, except in the D05-187 tegument, had β-glucosidase

452 activity. We could not, however, differentiate between sweet and bitter genotypes in
M

453 terms of the activity of this enzyme. It looks like there was a higher activity in the sweet
D

454 genotype, but we cannot conclude this since we are talking about semi-quantitative
TE

455 experiments. Furthermore, no relationship was found between the differences in the

456 sequences and the activity of the PHs in either the sweet or bitter genotypes. These
EP

457 results disagree with [60], who detected a high level of β-glucosidase activity in the

458 cotyledons of sweet and bitter almond genotypes, very low activity in the nucellus and
C

459 endosperm, and no activity in the tegument. One reason for these results could be the
AC

460 stage of ripening when the samples were collected. As shown in [38], the localization of

461 the PH differs from the beginning to the end of the ripening season.

462 With respect to the β-glucosidase bands, we observed different sizes between 50

463 and 75 kD. Similarly, the presence of a common structure for most β-glucosidases with

464 a molecular mass of 55 to 65 kD was detected before [42]. Four AH isoenzymes of 52

 ACCEPTED MANUSCRIPT
465 kD from black cherry have been isolated (AH I, I´, II and II´) [35]. Furthermore, PHs of

466 60 kD for P. domestica and of 68 kD for black cherry [30] and five PHs in black cherry

467 with sizes of between 52 and 68 kD were described [36].

468 4.4. Verification of PH activity

PT
469 None of the techniques used in this study showed differences related to the

470 bitterness phenotype. We still need to determine, however, why the PH from the D05-

RI
471 187 tegument showed a lower β-glucosidase activity.

SC
472 In accordance with our results, [35] also described the specificity of PH towards

473 prunasin in black cherry. Furthermore, PH and AH specificity towards prunasin and

U
474 amygdalin in black cherry and almond was shown [36, 37], respectively. In fact, in
AN
475 2009, this activity was found to come from protein extracts from almond but not from

476 the heterologous expression of a prunasin hydrolase. This current study is therefore the
M

477 first time a prunasin hydrolase from almond has been expressed and its activity with
D

478 respect to prunasin has been tested.

TE

479 As expected, PH does not show specific activity with amygdalin, linamarin or

480 linustatin [33].

EP

481 It has been observed that only a few amino acid residues are responsible for
C

482 enzyme specificity towards a substrate through the change the active site. It was
AC

483 observed that a change of two amino acids (position 220 and 394) in AH was able to

484 confer PH activity [37]. Moreover, only a single amino acid polymorphism in the

485 aglycone binding region of the active site (G211 in BGD2 and V211 in BGD4)

486 explained the difference in substrate specificity for the cyanogenic glucosides linamarin

487 and lotaustralin [61].

488
 ACCEPTED MANUSCRIPT
489 5. Conclusions

490 In this study, we have characterized Ph691 from the following three tissues

491 during kernel development in sweet and bitter almonds: the tegument, nucellus and

492 cotyledon. We found differences in the nucleotide sequences (SNPs) and amino acids

PT
493 between the nine PH sequences studied, but none of these differences were related to

494 the sweet or bitter flavour. Only PH691 from the D05-187 tegument showed no β-

RI
495 glucosidase or lower PH activity. In order to determine whether bitterness in almond is

related to PH activity, other PHs will have to be studied. We found no significant

SC
496

497 differences between the expression profiles of the PHs of sweet and bitter genotypes.

U
498 Acknowledgements
AN
499 This work was financed by the projects “Mejora Genética del Almendro”
M

500 (MINECO-Spain), “The molecular mechanisms to break flower bud dormancy in fruit

501 trees” to RS-P within Villum Young Investigator Program at the VILLUM Research
D

502 Center for Plant Plasticity-Denmark and “Breeding stone fruit species assisted by
TE

503 molecular tools” (Fundación Séneca-Spain).

504
EP

505
C
AC
 ACCEPTED MANUSCRIPT
506 References

507 [1] J. Heppner, The factor for bitterness in the sweet almond, Genetics 8 (1923) 390-

508 392.

509 [2] J. Heppner, Further evidence on the factor for bitterness in the sweet almond,

PT
510 Genetics. 11 (1926) 605-606.

511 [3] F. Dicenta, J.E. García, Inheritance of the kernel flavor in almond, Heredity. 70

RI
512 (1993) 308-312.

SC
513 [4] F. Dicenta, E. Ortega, P. Martinez-Gomez, Use of recessive homozygous genotypes

514 to assess the genetic control of kernel bitterness in almond, Euphytica 153 (2007) 221-

515 225.
U
AN
516 [5] C. Grasselly, P. Crossa-Raynaud, L'amandier, GP Maison et Larose, París, 1980.
M

517 [6] F.J. Vargas, M.A. Romero, I. Batlle, Kernel taste inheritance in almond, In: Ak B.E.

(ed.) 11 GREMPA Seminar on pistachios and almonds. Cahiers Options

D

518

519 Méditerranéennes. 56 (2001) 129-134.

TE

520 [7] D.E. Kester, T.M.Gradziel, Almonds (Prunus) En: Fruit Breeding. Editores: Moore,
EP

521 J.N. y Janick, J. New York: Wiley, 1996, pp. 1-97.

522 [8] R. Socias i Company, Fruit tree genetics at a turning point: the almond example,
C

523 Theor. Appl. Genet. 96 (1998) 588-601.

AC

524 [9] R. Sánchez-Pérez, W. Howad, J. Garcia-Mas, P. Arus, P. Martinez-Gomez, F.

525 Dicenta, Molecular markers for kernel bitterness in almond, Tree Genetics &

526 Genomes. 6 (2010) 237-245.

527 [10] D.E. Kester, R.N. Asay, Almonds. En: Advances in fruit breeding, Ed: Janick J y

528 Moore JN. Purdue University Press, West Lafayette, Indiana, 1975, pp. 387-419.
 ACCEPTED MANUSCRIPT
529 [11] F. Dicenta, P. Martínez-Gómez, E. Ortega, H. Duval, Cultivar pollinizer does not

530 affect almond flavor, HortScience 35 (2000) 1153-1154.

531 [12] G. Arrázola, R. Sánchez-Pérez, F. Dicenta, N. Grané, Content of the cyanogenic

532 glucoside amygdalin in almond seeds related to the bitterness genotype, Agron.

533 Colomb. 30 (2012) 260-265.

PT
534 [13] T. Joobeur, M.A. Viruel, M.C. de Vicente, B. Jáuregui, J. Ballester, M.T. Dettori, I.

RI
535 Verde, M.J. Truco, R. Messeguer, I. Battle, R. Quarta, E. Dirlewanger, P. Arús,

Construction of a saturated linkage map for Prunus using an almond × peach F2

SC
536

537 progeny, Theor. Appl. Genet. 97 (1998) 1034-1041.

U
538 [14] F.A. Bliss, S. Arulsekar, M.R. Foolad, A.M. Becerra, A. Gillen, M.L. Warburton,
AN
539 A.M. Dandekar, G.M. Kocsisne, K.K. Mydin, An expanded genetic linkage map of

540 Prunus based on an interspecific cross between almond and peach, Genome. 45
M

541 (2002) 520-529.

D

542 [15] R. Sánchez-Pérez, H. Howad, F. Dicenta, P. Arús, P. Martínez-Gómez, Mapping

TE

543 major genes and quantitative trait loci controlling agronomic traits in almond, Plant

544 Breeding. 126 (2007) 310-318.

EP

545 [16] E. Conn E, Cyanogenic compounds, Annu. Rev. Plant Physiol. 31 (2007)433-51.
C

546 [17] L. Zheng, J.E. Poulton, Temporal and spatial expression of amygdalin hydrolase
AC

547 and (R)-(+) - mandelonitrile lyase in black cherry seeds, Plant Physiol. 109 (1995)

548 31-39.

549
 ACCEPTED MANUSCRIPT
550 [18] A.V. Morant A.V, K. Jørgensen K, C. Jørgensen, S.M. Paquette, R. Sánchez-Pérez,

551 B.L. Møller, S. Bak, β-Glucosidases as detonators of plant chemical defense,

552 Phytochemistry 69 (2008) 1795-1813.

553 [19] S. Bak, S.M. Paquette, M. Moran, A.V. Morant, S. Saito, N. Bjarnholt, M.

554 Zagrobelny, K. Jørgensen, S. Osmani, T. Hamann, H.T. Simonsen, R. Sánchez-Pérez

PT
555 , T.B. van Heeswijck, K. Jørgensen, B.L. Møller BL, Cyanogenic glycosides: a case

RI
556 study for evolution and application of cytochromes P450, Phytochem Reviews. 5

557 (2006) 309-329.

SC
558 [20] R. Sánchez-Pérez, K. Jørgensen, C. Erik Olsen, F. Dicenta, B.L. Møller, Bitterness

U
559 in almonds, Pant Physiol. 146 (2008) 1040-1052.
AN
560 [21] R.M. Gleadow, B.L. Møller, Cyanogenic Glycosides: Synthesis, Physiology, and

561 Phenotypic Plasticity, Annual Review of Plant Biology. 65 (2014) 155-185.

M

562 [22] S. Bak, C.E. Olsen, B.A. Halkier, B.L. Møller, Transgenic tobacco and Arabidopsis
D

563 plants expressing the two multifunctional sorghum cytochrome P450 enzymes,
TE

564 CYP79A1 and CYP71E1, are cyanogenic and accumulate metabolites derived from

565 intermediates in Dhurrin biosynthesis, Plant Physiology 123 (2006) 1347-1448.

EP

566 [23] M. Zagrobelny, S. Bak, B.L. Møller, Cyanogenesis in plants arthropods.

C

567 Phytochemistry, 69 (2008) 1457-1468.

AC

568 [24] J. R. Ketudat Cairnsa, B. Mahongc, S. Baiyaa, J-S. Jeonc, β-Glucosidases:

569 Multitasking, moonlighting or simply misunderstood? Plant Sciences 241 (2015)

570 246-259.

571 [25] C.D. McCarty, J.W. Leslie, H.B. Frost, Bitterness of kernels of almond 3 peach

572 hybrids and their parents, Proc. Am. Soc. Hortic. Sci. 59 (1952) 254-258.
 ACCEPTED MANUSCRIPT
573 [26] M. Frehner, M. Scalet, E.E. Conn, Pattern of the cyanide-potential in developing

574 fruits, Plant Physiol. 94 (1990) 28-34.

575 [27] B.L. Møller, D.S. Seigler, Biosynthesis of cyanogenic glycosides, cyanolipids and

576 related compounds. In BK Singh, ed, Plant Amino Acids, Biochemistry and

577 Biotechnology, Marcel Dekker, New York, 1991, pp563-609.

PT
578 [28] E. Swain, C.P. Li, J.E. Poulton, Development of the potential for cyanogenesis in

RI
579 maturing black cherry (Prunus serotina Ehrh.) fruits, Plant Physiol. 98 (1992 a)

1423-1428.

SC
580

581 [29] J.E. Poulton, C. Ping Li, Tissue level compartmentation of (R)-amygdalin and

U
582 amygdalin hidrolase prevents large-scale cyanogenesis in undamaged Prunus seeds,
AN
583 Plant Physiol. 104 (1994) 29-35.

584 [30] F. Dicenta, P. Martínez-Gómez, N. Grané, M.L. Martín, A. León, J.A. Cánovas, V.
M

585 Berenguer, Relationship between cyanogenic Compounds in kernels, leaves, and

D

586 roots of sweet and bitter kernelled almonds, J Agric. Food Chem. 50 (2002) 2149-
TE

587 2152.

588 [31] T.K. Franks, A. Yadollahi, M.G. Wirthensohn, J. Guerin, B.N. Kaiser, M. Sedgley,
EP

589 C.M. Ford, A seed coat cyanohydrin glucosyltransferase is associated with bitterness
C

590 in almond (Prunus dulcis) kernels, Functional Plant Biology. 35 (2008) 236-246.
AC

591 [32] C. Mentzer, J. Favre-Bonvin, Sur la biogenèse du glucoside cyanogénétique des

592 feuilles de laurier-cerise (Prunus lauro-cerasus), C. R. Acad. Sci. Ser. III Sci. Vie.

593 253 (1961) 1072-1074.

594 [33] J. Del Cueto, I. A. Ionescu, M. Piˇcmanová, O. Gericke, M. S. Motawia, C. E.

595 Olsen, J. A. Campoy, F. Dicenta, B. L. Møller, R. Sánchez-Pérez, Cyanogenic

 ACCEPTED MANUSCRIPT
596 Glucosides and Derivatives in Almond and Sweet Cherry Flower Buds from

597 Dormancy to Flowering, Frontiers in Plant Sciences. 800 (2017) 1-16.

598 [34] G.W. Kuroki, J.E. Poulton, Isolation and characterization of multiple forms of

599 prunasin hydrolase from black cherry (Prunus serotina Ehrh.) seeds, Biochemistry

600 and Biophysis. 255 (1987) 19-26.

PT
601 [35] C.P. Li, E. Swain, J.E. Poulton, Prunus serotina amygdalin hydrolase and prunasin

RI
602 hydrolase. Plant Physiol. 100 (1992) 282-290.

SC
603 [36] J. Zhou, S. Hartmann, B.K. Shepherd, J.E. Poulton, Investigation of the

604 microheterogeneity and aglycone specificity-conferring residues of black cherry

U
605 prunasin hydrolases, Plant Physiol. 129 (2002) 1252-1264.
AN
606 [37] R. Sánchez-Pérez, K. Jørgensen, M.S. Motawia, F. Dicenta, B.L. Møller, Tissue

607 and cellular localization of individual β-glycosidases using a substrate-specific sugar

M

608 reducing assay, The Plant Journal 60 (2009) 894-906.

D

609 [38] R. Sánchez-Pérez, F.S. Belmonte, J. Borch, F. Dicenta, B.L. Møller, K. Jørgensen,
TE

610 Prunasin Hydrolases during Fruit Development in Sweet and Bitter Almonds. Plant

611 Physiol. 158 (2012) 1916-1932

EP

612 [39] E. Conn E, Cyanogenic glycosides. J Agric. Food Cham. 17 (1969) 519-526.
C

613 [40] M. Morant, S. Bak, B.L. Møller, D. Werck-Reichhart, Plant cytochromes P450:
AC

614 tools for pharmacology, plant protection and phytoremediation. Curr. Opin.

615 Biotechnol. 14 (2003) 151-162.

616 [41] M. Zagrobelny, S. Bak, A.V. Rasmussen, B. Jørgensen, C.M. Naumann, B.L.

617 Møller, Cyanogenic glucosides and plant-insect interactions, Phytochemistry. 65

618 (2004) 293-306.

 ACCEPTED MANUSCRIPT
619 [42] J.E. Poulton, Cyanogenesis in plants, Plant Physiol. 94 (1990) 401-405.

620 [43] E. Swain, C.P. Li, J.E. Poulton, Tissue and subcellular localization of enzymes

621 catabolizing (R)-amygdalin in mature Prunus serotina seeds, Plant Physiol. 100

622 (1992 b)291-300.

PT
623 [44] H. Pankoke, T. Buschmann, C. Müller, Role of plant β-glucosidases in the dual

624 defense system of iridoid glycosides and their hydrolyzing enzymes in Plantago

RI
625 lanceolata and Plantago major, Phytochemistry. 94 (2013) 99-107.

SC
626 [45] E. Swain, J.E. Poulton, Immunocytochemical localization of prunasin hydrolase

627 and mandelonitrile lyase in stems and leaves of Prunus serotina, Plant Physiol. 106

U
628 (1994) 1285-1291.
AN
629 [46] M. Suelves, P. Puigdoménech, Molecular cloning of the cDNA coding for the (R)-

630 (+)-mandelonitrile lyase of Prunus amygdalus: temporal and spatial expression

M

631 patterns in flowers and mature seeds, Planta 206 (1998) 388-393.
D

632 [47] V.A. Evreinoff, Quelques observations biologiques sur l’amandier, Revue
TE

633 internationale de botanique appliquèe et d’agriculture tropicale. 32 (1952) 442-459.

[46] HG. Floss, L. Hadwiger, E.E. Conn, Enzymatic formation of β-cyanoalanine from
EP

634

635 cyanide, Nature. 208 (1965) 1207-1208.

C

636 [49] M. Piotrowski, S. Schönfelder, E.W. Weiler, The Arabidopsis thaliana isogene
AC

637 NIT4 and its orthologs in tobacco encode beta-cyano-L-alanine hydratase/nitrilase, J

638 Biol Chem. 276 (2001) 2616-2621.

639 [50] M. Piotrowski, J.J. Volmer, Cyanide metabolism in higher plants: cyanoalanine

640 hydratase is a NIT4 homolog, Plant Mol Biol. 61 (2006) 111-122.

 ACCEPTED MANUSCRIPT
641 [51] R. Jenrich, I. Trompetter, S. Bak, C.E. Olsen, B.L. Møller, M. Piotrowski,

642 Evolution of heteromeric nitrilase complexes in Poaceae with new functions in nitrile

643 metabolism. Proc Natl Acad Sci USA. 104 (2007) 18848-18853.

644 [52] V. Kriechbaumer, W.J. Park, M. Piotrowski, R.B. Meeley, A. Gierl, E.

645 Glawischnig, Maize nitrilases have a dual role in auxin homeostasis and β-

PT
646 cyanoalanine hydrolysis, J Exp Bot. 58 (2007) 4225-4233.

RI
647 [53] K. Tamura, D. Peterson, N. Peterson, G. Stecher, M. Nei, S. Kumar, MEGA5:

molecular evolutionary genetics analysis using maximum likelihood, evolutionary

SC
648

649 distance, and maximum parsimony methods, Mol. Biol. Evol. 28 (2011) 2731-273.

U
650 [54] J. Luo, Y. Nishiyama, C. Fuell, G. Taguchi, K. Elliott, L. Hill, Y. Tanaka, M.
AN
651 Kitayama, M. Yamazaki, P. Bailey, A. Parr, A.J. Michael, K. Saito, C. Martin,

652 Convergent evolution in the BAHD family of acyltransferases: identification and

M

653 characterization of anthocyanin acyltransferases from Arabidopsis thaliana, The

D

654 Plant Journal. 50 (2007) 678-695.

TE

655 [55] O. Voinnet, S. Rivas, P. Mestre, D. Baulcombe, An enhanced transient expression

656 system in plants based on suppression of gene silencing by the p19 protein of tomato
EP

657 bushy stunt virus, Plant J. 33 (2003) 949-956.

C

658 [56] F. Feigl, V. Anger, Replacement of benzidine by copper ethylacetoacetate and tetra
AC

659 basis as spot-test reagent for hydrogen cyanide and cyanogen, Analyst (Lond.). 91

660 (1966) 282-284.

661 [57] J.L. Lambert, J. Ramasamy, J.V. Paukstelis, Stable reagents for the colorimetric

662 determination of cyanide by modified König reactions, Anal. Chem. 47 (1975) 916-

663 918.
 ACCEPTED MANUSCRIPT
664 [58] B.A. Halkier, B.L. Møller, Biosynthesis of the cyanogenic glucoside dhurrin in

665 seedlings of Sorghum bicolor (L.) Moench and partial purification of the enzyme

666 system involved, Plant Physiol. 90 (1989) 1552-1559.

667 [59] P. Martínez-Gómez, R. Sánchez Pérez, F. Dicenta, Fruit development in almond

668 for fresh consumption, Journal of the American Pomological Society. 62 (2008) 82-

PT
669 86.

RI
670 [60] L. Abarrategui, Bitterness in almonds, Master thesis. University of Copenhagen.

2010.

SC
671

672 [61] D. Lai, M. Picmanová, M. Abou Hachem, M.S. Motawia, C.E. Olsen, B.L, F.

U
673 Rook , A.M. Takos, Lotus japonicus flowers are defended by a cyanogenic â-
AN
674 glucosidase with highly restricted expression to essential reproductive organs, Plant

675 Molecular Biology. 89 (2015) 21-34.

M

676
D

677
TE

678
EP

679

680
C
AC

681

682

683
 ACCEPTED MANUSCRIPT

Table 1. Designed primers for almond cDNA, annealing temperatures, and expected product
size.

Primer Sequence (5’-3’) Tm (°C) Expected size (bp)

PT
691ATG ATGGCATTGCAATTCCGCTCTTTGCTCTTGTG 57 1638

691UGA TCAAATTTGATACACAAATTTGGTAGCCCTA 57 1638

RI
27F AATTATGAAGGATATGGGGTTGG 57 1050

27R CACCTCTGTTACCTCCAAGCA 57 1050

SC
691B1F GGGGACAAGTTTGTACAAAAAAGCAGGCTATGGCATTGCAA 1638

TTCCGCTCTTTGCTCTTGTG
57

U
691B2R TCAAATTTGATACACAAATTTGGTAGCCCTATCAAATTTGAT 1638

ACACAAATTTGGTAGCCCTA
AN
57
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 2. Comparison of major features of P. dulcis and P. serotina nucleotides and amino
acid sequences. Sequence length, polypeptide length, N-glycosylation sites, signal peptide and
location of the enzyme are indicated. L: Lauranne (sweet); D: D05-187 (bitter); S: S3067
(bitter); t: tegument; n: nucellus; and c: cotyledon. PsPh1, PsPh2, PsPh3, PsPh4 and PsPh5
(Zhou et al., 2002). PdPhR691, PdPhS691, PdPhR692 and PdPhS692 (Sánchez-Pérez et al.,
2012).

PT
Sequence Polypeptide N-glycosylation Signal Location
length length sites peptide

RI
PdPh691Lt 1635 544 5 23 Vacuolar
PdPh691Ln 1635 544 5 27 Vacuolar

SC
PdPh691Lc 1635 544 5 23 Vacuolar
PdPh691Dt 1635 544 5 27 Vacuolar
PdPh691Dn 1635 544 5 27 Vacuolar

U
PdPh691Dc 1635 544 5 23 Vacuolar
PdPh691St 1635 544 5 27 Extracellular
AN
PdPh691Sn 1635 544 5 27 Vacuolar
PdPh691Sc 1635 544 5 27 Vacuolar
M

PsPh1 2056 549 8 23 Vacuolar

PsPh2 2059 544 2 23 Vacuolar
PsPh3 1960 537 6 27 Vacuolar
D

PsPh4 1911 545 7 23 Vacuolar

PsPh5 1819 542 5 23 Vacuolar
TE

PdPhR691 1635 544 5 27 Vacuolar

PdPhS691 1635 544 5 27 Vacuolar
EP

PdPhR692 1629 542 5 23 Vacuolar

PdPhS692 1629 542 5 23 Vacuolar
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

Table 3. Amino acid differences between sequences of our nine PHs. Exon, gDNA, cDNA
and aa positions of the differences are indicated. These nine PHs are compared with five PHs
described in P. serotina by Zhou et al. (2002) and with two PHs described in sweet
(PdPHR691 and PdPHR692) and bitter (PdPHS91 and PdPHS92) almonds (Sánchez-Pérez et
al., 2012) (L: Lauranne, D: D05-187, S: S3067).

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Position Tegument Nucellus Cotyledon Prunus serotina Prunus dulcis

PT
Difference EXON aa L D S L D S L D S Ps Ps Ps Ps Ps PdPH PdPH PdPH PdPH
PH1 PH2 PH3 PH4 PH5 R691 S691 R692 692

RI
V/M/C EXON 1 12 V M V V M M V V M V V C V C M M V V

A/T EXON 1 20 A A A T A A A A A A A A A A A A A A

SC
A/E/R EXON 1 22 A A E A A A A A A A A R A A A A A A

U
I/V/S/G EXON 1 34 I I I V I I I I I V V S G I V I V V

AN
Y/C EXON 6 176 C Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

F/L EXON 7 199 F F F F L F F F F F F F F F F F F F

M
T/A EXON 7 208 T T T T T A T T T T T T T T T T T T

E/G/K EXON 9 337 E E E E E E E E G E E E E E K E K K

D
S/L EXON 9 341 S S S L S L S S S S S L S S S S S S

TE
S/I/N/T EXON 9 343 S S S S S I S S S N I N T S S S S S

V/L EXON 11 412 V V V L V L V V V V V L V V L V L L

EP
F/L EXON 12 434 L F F F R F F F F F F F F F F F F F
C

N/D/Q EXON 12 436 N N N N N N N D N N N N V N Q N Q Q

AC

G/A/D/S/T/V EXON 12 450 G G G A G A D G G A S T A V S G S S

E/D EXON 13 469 E E E E E E E E D E E D E E E E E E

R/P EXON 13 494 R R R R R R R P R R R R R R R R R R

V/L/G EXON 13 545 V V V V V L V V V V V G V V V V V V

 ACCEPTED MANUSCRIPT

Table 4. Percentage of nucleotide similarity between our nine Phs and other sequences of P.
serotina and P. dulcis. L: Lauranne (sweet); D: D05-187 (bitter); S: S3067 (bitter); t:
tegument; n: nucellus; and c: cotyledon. PsPh1, PsPh2, PsPh3, PsPh4 and PsPh5 (Zhou et
al., 2002). PdPhR691, PdPhS691, PdPhR692 and PdPhS692 (Sánchez-Pérez et al., 2012). Lt
stands for PdPh691Lt, Ln means PdPh691Ln, and so on.

Sweet Bitter

PT
Lt Ln Lc Dt Dn Dc St Sn Sc
PdPh691Lt 100.00 98.90 99.69 99.45 99.33 99.79 99.27 98.96 99.57
PdPh691Ln - 100.00 98.90 98.96 98.71 98.90 98.78 99.45 98.84

RI
PdPh691Lc - - 100.00 99.39 99.27 99.72 99.20 98.96 99.51
PdPh691Dt - - - 100.00 99.39 99.59 99.27 99.14 99.51
PdPh691Dn - - - - 100.00 99.45 99.14 98.99 99.39

SC
PdPh691Dc - - - - - 100.00 99.24 99.04 99.60
PdPh691St - - - - - - 100.00 98.83 99.20
PdPh691Sn - - - - - - - 100.00 99.02

U
PdPh691Sc - - - - - - - - 100.00
PdPhR691 99.00 99.00 99.00 99.00 99.00 99.00 99.00 99.00 99.00
AN
PdPhS691 99.00 99.00 99.00 99.00 99.00 99.00 99.00 99.00 99.00
PdPhR692 79.00 79.00 79.00 79.00 79.00 79.00 79.00 79.00 79.00
PdPhS692 79.00 79.00 79.00 79.00 79.00 79.00 79.00 79.00 79.00
M

PsPh1 94.00 94.00 94.00 94.00 93.00 94.00 94.00 94.00 94.00
PsPh2 91.00 91.00 91.00 91.00 91.00 91.00 91.00 91.00 91.00
PsPh3 82.00 83.00 82.00 82.00 82.00 83.00 82.00 83.00 82.00
D

PsPh4 92.00 92.00 92.00 92.00 92.00 92.00 92.00 92.00 92.00
PsPh5 96.00 96.00 96.00 96.00 96.00 96.00 96.00 96.00 96.00
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 1. The metabolic pathways for the synthesis and catabolism of the cyanogenic glucosides prunasin and amygdalin in almonds. Biosynthetic enzymes (solid lines):
CYP79 and CYP71, Cytochrome P450 monooxygenases, UGT1, UDPG-mandelonitrile glucosyltransferase; UGT2, UDPG-prunasin glucosyltransferase. Catabolic enzymes
(dotted lines): AH, amygdalin hydrolase; PH, prunasin hydrolase; MDL1, mandelonitrile lyase; ADGH*, amygdalin diglucosidase (putative). Detoxification enzymes (dashed
lines) CAS: β-cyanoalanine synthase; NIT4: nitrilases 4. Source: adapted from Sánchez-Pérez et al. (2008) and Del Cueto and Ionescu et al. (2017). Blue indicates
biosynthesis, red indicates bioactivation and yellow indicates a detoxification pathway.
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
M1 EP M1 M2

1638 pb
C
AC

Figure 2. Kernel development and detection of prunasin hydrolases by RT-PCR in the three development stages: a) beginning (JD 89); b) halfway
(JD 130); c) final (JD 167). T: tegument; N: nucellus; E: endosperm; C: cotyledon. The arrow indicates the band size of 1638 pb expected for
amplification of prunasin hydrolase 691 (Ph691). M1, Ladder 1 kb+; M2: Ladder Lambda HindIII. L: Lauranne (sweet); D: D05-187 (bitter); S:
S3067 (bitter).
 ACCEPTED MANUSCRIPT

DCCJLA1

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 3. Amino acid sequences of prunasin hydrolases (PHs) from P. serotina (Ps) and P. dulcis (Pd). Differences in amino acids are marked in color. Red frames: signal peptide and NEP and
ITENG motif of the active site and INKKGIEYY motif of PH. Black arrows: introns. Red arrows: SNPs. L: Lauranne (sweet); D: D05-187 (bitter); S: S3067 (bitter); t: tegument; n: nucellus; and c:
cotyledon. PsPH1, PsPH2, PsPH3, PsPH4 and PsPH5 (Zhou et al., 2002). PdPHR691, PdPHS691, PdPHR692 and PdPHS692 (Sánchez-Pérez et al., 2012).
 ACCEPTED MANUSCRIPT

Slide 3

DCCJLA1 Modificado INKKGIEYY Motif

Del Cueto Chocano Jorge Luis Agroscope, 06-12-2017

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 4. Phylogenetic tree among prunasin hydrolases (PHs) characterized in this study and other PHs from P. serotina (Ps) and P. dulcis
(Pd). The numbers represent percentage bootstrap support (1,000 replicates). L: Lauranne (sweet); D: D05-187 (bitter); S: S3067 (bitter); t:
tegument; n: nucellus; and c: cotyledon. PsPH1, PsPH2, PsPH3, PsPH4 and PsPH5 (Zhou et al., 2002). PdPHR691, PdPHS691, PdPHR692 and
PdPHS692 (Sánchez-Pérez et al., 2012), and PsAH (Li et al., 1992).
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 5. β-glucosidase activity, from agroinfiltrated N. benthamiana leaves expressing prunasin hydrolase 691 belonging to
sweet and bitter cultivars from mid ripening stage (JD 130), detected by the following two methods: a) Fast Blue BB salt. L: The band
size ranged between 50 and 75 kD; and b) Umbellifery method. Fluorescence under UV light using the substrate 4-methyl-umbelliferyl-
β-D-glucoside. Lauranne (sweet); D: D05-187 (bitter); S: S3067 (bitter); p19: negative control; t: tegument; n: nucellus; and c:
cotyledon.
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 6. Substrate specificity detected by Feigl-Anger method from agroinfiltrated N. benthamiana leaves expressing
prunasin hydrolase 691 from mid ripening stage (JD 130), incubated with dhurrin, linamarin, lotaustralin, prunasin, amygdalin
and linustatin. p19: negative control. Blue indicates the positive reaction liberating CN from the PH691 activity. L: Lauranne
(sweet); D: D05-187 (bitter); S: S3067 (bitter); t: tegument; n: nucellus; c: cotyledon.
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP

Figure 7. Cyanide (nmol CN/g fresh tissue) released from agroinfiltrated N. benthamiana leaves from mid ripening stage (JD 130), with the nine Ph691s against the
AC

substrates prunasin, amygdalin and dhurrin. L: Lauranne (sweet); D: D05-187 (bitter); S: S3067 (bitter); t: tegument; n: nucellus; and c: cotyledon.
 ACCEPTED MANUSCRIPT
β-Glucosidase activity in almond seeds

Jorge Del Cueto1,2, Federico Dicenta1, Birger Lindberg Møller2 and Raquel Sánchez-

Pérez*1, 2

PT
Highlights:

RI
We focused on prunasin hydrolases, one of the two β-glucosidases responsible
for the bitter taste in almond, by liberating glucose, benzaldehyde and hydrogen

SC
cyanide.

U
β-glucosidase and prunasin hydrolase activity was detected in different parts of
AN
the seeds: tegument, nucella and endosperm and cotyledon.
M

Sequences alignments showed several SNPs that could explain differences

observed in the substrate specificity and in the CN released assay from
D

agroinfiltrated N. benthamiana leaves.

TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC