A Toolbox for Nodule Development Studies in Chickpea: a Hairy-Root Transformation Protocol and an Efficient Laboratory Strain Of

bioRxiv preprint doi: https://doi.org/10.1101/362947; this version posted July 5, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 A toolbox for nodule development studies in chickpea: a hairy-root

2 transformation protocol and an efficient laboratory strain of Mesorhizobium sp.

3 Drishti Mandal1, and Senjuti Sinharoy1*

4 Affiliation: 1National Institute of Plant Genome Research, New Delhi – 110067. India

5 *Corresponding author: E-mail: [email protected]

6 Running title: Toolbox for Chickpea root nodule symbiosis study

7 Abstract: Mesorhizobium sp. produces root nodules in chickpea. Chickpea and model

8 legume Medicago truncatula are members of inverted repeat lacking clade (IRLC). The

9 rhizobia after internalization inside plant cell called ‘bacteroid’. Nodule Specific

10 Cysteine-rich (NCR) peptides in IRLC legumes guide bacteroids to a ‘terminally

11 differentiated swollen (TDS)’ form. Bacteroids in chickpea are less TDS than those in

12 Medicago. Nodule development in chickpea indicates recent evolutionary diversification

13 and merits further study. A hairy root transformation protocol and an efficient laboratory

14 strain are prerequisites for performing any genetic study on nodulation. We have

15 standardized a protocol for composite plant generation in chickpea with a transformation

16 frequency above 50%, as shown by fluorescent markers. This protocol also works well

17 in different ecotypes of chickpea. Localization of subcellular markers in these

18 transformed roots is similar to Medicago. When checked inside transformed nodules,

19 peroxisomes were concentrated along the periphery of the nodules, while ER and golgi

20 bodies surrounded the symbiosomes. Different Mesorhizobium strains were evaluated

21 for their ability to initiate nodule development, and efficiency of nitrogen fixation.

22 Inoculation with different strains resulted in different shapes of TDS bacteroids with

23 variable nitrogen fixation. Our study provides a toolbox to study nodule development in

24 the crop legume chickpea.

25 Introduction:

26 Root nodule symbiosis (RNS) is the most successful metabolism-dependent symbiosis

27 on earth. Leguminous plants get reduced nitrogen directly from RNS, at the expense of

28 photosynthate (Werner et al., 2015). The staple crop chickpea (Cicer arietinum) is

29 world’s second largest cultivated legume. Chickpea seeds are a valuable source of

30 dietary protein in lower socio-economic class. Additionally, due to the symbiotic

31 interaction with Mesorhizobium sp., chickpea can be cultivated in a sustainable way

32 (Jain et al., 2013; Varshney et al., 2013). Research on nodule development has been

33 centered upon model legumes Medicago truncatula and Lotus japonicus. Today, model

34 legumes are in the forefront of legume biology in terms of both available knowledge and

35 resources. The only disadvantage is that model legumes are not crop species. In spite

36 of being the most important grain legume in tropical and sub-tropical countries (Jukanti

37 et al., 2012), literature on chickpea nodule development is scarce. Genome sequencing

38 and the establishment of transcriptomic and proteomic resources have laid the pillars for

39 making chickpea a model amongst crop legumes (Jain et al., 2013; Varshney et al.,

40 2013; Ramalingam et al., 2015; Pandey et al., 2018).

41 Chickpea belongs to the inverted repeat lacking clade (IRLC), which diverged from

42 model legume Medicago ~10-20 million years ago (Jain et al., 2013; Varshney et al.,

43 2013). IRLC legumes develop indeterminate nodules, with a gradient of cells at different

44 stages of development can be seen from the distal to the proximal part of the nodule. A

45 persistent meristem (zone I) is present at the distal end of the nodule. Bacterial

46 endocytosis and colonization takes place in the postmeristematic cells of the infection

47 zone (zone II), where plant membrane-bound bacterial units are formed. These units

48 are considered as an ammonium-exporting organelle called symbiosome (Roth and

49 Stacey, 1989). The rhizobia inside the symbiosomes are called bacteroids. The

50 bacteroids divide inside the symbiosome and gradually colonize the whole cell. In the

51 nitrogen fixation zone (zone III) bacteroids differentiation is terminated, ammonium

52 assimilation genes are repressed, and nitrogen fixation genes are induced (Oldroyd,

53 2013; Udvardi and Poole, 2013).

54 Terminally differentiated, enlarged bacteroids with different morphotypes are a typical

55 feature of the IRLC legumes. The major determining factor behind these irreversibly

56 differentiated bacteroids is nodule specific cysteine-rich (NCR) peptides (Montiel et al.,

57 2016; Montiel et al., 2017). The molecular mechanism of NCR peptide regulated-endo-

58 reduplication of the symbiont genome has been worked out in model legume Medicago.

59 Medicago genome encodes more than 700 NCR genes. At least 138 NCR peptides get

60 processed in the endoplasmic reticulum (ER) and targeted towards symbiosomes

61 (Wang et al., 2010; Durgo et al., 2015). NCR peptides force endo-reduplication of the

62 symbiont genome. As a result, the bacteria (now bacteroid) lose their ability to divide

63 and re-grow on culture media, but their size increases up to ~10-fold (Mergaert et al.,

64 2006; Young et al., 2011; Sinharoy et al., 2013). NCR gene family evolved in Medicago

65 by a recent local gene duplication (Alunni et al., 2007). NCRs have been identified in

66 several IRLC legumes, including chickpea (Kant et al., 2016; Montiel et al., 2016;

67 Montiel et al., 2017). The number of NCR genes among different IRLC legumes varies

68 greatly. This results in different morphotypes of bacteroids in different legumes, such as

69 swollen/spherical, elongated, and elongated-branched (Montiel et al., 2017).

70 Mesorhizobium TAL620-induced nodules in chickpea express only 63 NCR genes,

71 while its’ bacteroids endo-reduplicated up to 4 fold (Kant et al., 2016; Montiel et al.,

72 2017). Chickpea and Medicago NCRs share less than 80% identity. Chickpea NCR

73 peptides have more identity with Glycyrrhiza uralensis, Onobrychis vicifolia, and

74 Astragalus canadensis, while phylogenetically chickpea is closer to Medicago (Montiel

75 et al., 2016; Montiel et al., 2017). Chickpea’s swollen/ spherical bacteroids are basal to

76 the evolution of NCR-guided morphogenesis (Montiel et al., 2017). In contrast to

77 chickpea, Medicago has more than 700 NCR genes produce elongated-branched

78 bacteroids depicting an advanced stage of this trait. Interestingly, this morphogenesis of

79 bacteroids and the genome endo-reduplication is thought to determine the efficiency of

80 nitrogen fixation in respective legumes (Oono and Denison, 2010). Thus, comparative

81 investigation of nodule development in chickpea and Medicago will enhance our

82 knowledge on the evolutionary link between variable nitrogen fixation efficiencies

83 amongst different legumes.

84 An efficient hairy-root transformation protocol is an indispensable tool to understand

85 nodule biology, enabling us to study the localization of any protein (fluorescent tag),

86 activities of promoters (GUS fusion), and the effects of over-expression or knock-down

87 of certain genes. We have undertaken an effort to establish hairy root transformation in

88 chickpea and study nodule development. Unlike A. tumefaciens, A. rhizogenes

89 generates transformed roots from the site of infection. A. rhizogenes contains root locus

90 (rol) genes in the Ri-plasmid, which promotes the formation of genetically transformed

91 adventitious hairy roots. A. rhizogenes carrying a recombinant Ri plasmid can generate

92 composite plants. These plants contain untransformed shoot and transform root

93 (Georgiev et al., 2012). When such an A. rhizogenes additionally carrying a gene of

94 interest in a binary vector is used for transformation, a certain percentage of co-

95 transformed roots are obtained. Both overexpression and downregulation of a specific

96 gene can be achieved by these co-transformed roots (Limpens et al., 2004; Sinharoy

97 and DasGupta, 2009; Sinharoy et al., 2015). Recently, it has been reported that

98 CRISPR/Cas9 mediated gene knock out is also possible in hairy roots (Ron et al., 2014;

99 Cai et al., 2015; Wang et al., 2016). Till date, all the successful protocol for hairy root

100 generation in other legumes have shown the transformed roots to be biologically similar

101 to untransformed roots, with no difference in normal nodule development (Stougaard et

102 al., 1987; Quandt et al., 1993; Stiller et al., 1997; Boisson-Dernier et al., 2001; Van-de-

103 Velde et al., 2003; Limpens et al., 2004; Estrada-Navarrete et al., 2006; Kereszt et al.,

104 2007; Sinharoy et al., 2009; Bonaldi et al., 2010; Imanishi et al., 2011; Brijwal and

105 Tamta, 2015; Thilip et al., 2015; Habibi et al., 2016; Thwe et al., 2016). A. rhizogenes‐

106 mediated hairy-root transformation is an efficient and less time-consuming alternative

107 method for the functional validation of genes. Here we are reporting an efficient protocol

108 for hairy root transformation in chickpea. We have developed both in vivo and ex vitro

109 protocols. We have used four different Mesorhizobium strains to follow nodule initiation

110 and the developmental program and evaluated the nitrogen fixation efficiency. We have

111 also checked the localization of several subcellular markers in transformed chickpea

112 roots and nodules. Additionally, we have confirmed that the shape of the bacteroids in

113 the mature nitrogen-fixing zone of chickpea nodules is directly correlated with the

114 efficiency of nitrogen fixation.

115 Results:

116 Hairy root transformation of chickpea:

117 We used Agrobacterium rhizogenes R1000, ARqua1, and MSU440 strains for hairy root

118 transformation. All three strains were successful to generate hairy roots (data not

119 shown). For detailed characterization, we used ARqua1 strain. Dicotyledonous plants

120 can be transformed using both in vivo (tissue culture based), and ex vitro (without tissue

121 culture) protocols (Collier et al., 2005; Sinharoy et al., 2015). We attempted both the

122 methods. For in vivo transformation, chickpea seedlings were infected either by cutting

123 1-3 cm of the radicle and scraping it in an A. rhizogenes lawn culture or using a needle

124 containing A. rhizogenes to make a small wound at the hypocotyl region. Overview of

125 the methods is given in Figure 1. In both the cases, transformed roots began to emerge

126 from the infected region 14-15 days after infection (dai) (Figure 1 step-Ia). In case of

127 needle mediated rhizogenes infection, the original roots were removed from the plant

128 when transformed root started coming. These plants were transferred to kanamycin

129 containing selection medium to increase the efficiency of the transformation. We tried

130 two different concentration of kanamycin (50µg/ml, and 25µg/ml). The plants were

131 maintained in the antibiotic-containing medium for 10-12 days for antibiotic selection. All

132 the roots became black, and the plant died within two weeks when 50 µg/ml of

133 kanamycin was used (data not shown). But the transgenic roots grew normally in

134 presence of 25 µg/ml kanamycin. The shoots of composite plants had no difference in

135 morphology with their non-transgenic counterparts (Figure 1 step IIa). The transformed

136 roots exhibited characteristics of plagiotropic growth. The transformation frequency was

137 determined using either cauliflower mosaic virus (CaMV) 35S promoter, or Arabidopsis

138 ubiquitin 10 (AtUBQ10) promoter-driven expression of the DsRed fluorescent protein

139 (Figure 1 step IIIa). Both showed a similar frequency of transformed roots (Table 1).

140 Usually, we obtained a slightly greater frequency of transformed roots after antibiotic

141 selection (Table 1. 48.7% without antibiotic over 53.1% with an antibiotic in BDG256).

142 Plants generated by in vivo methods were transferred to the substrate (a mixer of 3:1

143 vermiculite and fireclay balls) 20-25 dai. Non-fluorescent roots were removed from the

144 plant at this point. Plants were maintained in this substrate for further 2-3 days under

145 nitrogen free condition. Nodules were observed in these transformed roots after 10 days

146 post-inoculation (dpi) with Mesorhizobium (Figure 1 Step IVa).

147 Ex-vitro composite plant generation is a rapid, efficient, simple, and low-cost method for

148 producing composite plants without tissue culture (Collier et al., 2005; Sinharoy et al.,

149 2015). The steps for ex-vivo transformation are shown on the right-hand side of Figure

150 1. For ex-vitro transformation, 25-30 days old apical stem tissue of young branches

151 were cut and dipped in the A. rhizogenes culture and transferred to the substrate

152 (Figure 1 Step1b). Further, the plants were maintained in a humid chamber for 30 days

153 to promote root production in presence of full nitrogen fertilization regime. Both

154 transformed and adventitious root production were observed by 20 dai. The transformed

155 roots were screened based on dsRed fluorescence (Figure 1 step IIb). At this point, the

156 substrate containing plants were washed thoroughly with distilled water, and transferred

157 to a new pot and maintained under nitrogen free condition. After 2-3 days the plants

158 were inoculated with Mesorhizobium culture for root nodule development, and nodules

159 were noticed within 10 dpi (Figure 1 step IIIb).

160 We have tested this protocol in different chickpea ecotypes. A list of different cultivars of

161 chickpea from different origin, where we successfully tested our transformation protocol,

162 has been given in Table 1. Both in vivo and ex vitro methods had comparable

163 transformation efficiency. While we obtained greater transformation efficiency in

164 BDG256, and ‘local’ variety (above 50%), ICC4958, ICC8261, and kabuli had a lower

165 efficiency. In summary, this method can be used in several different varieties of

166 chickpea cultivars.

167 Expression of different subcellular markers in transformed hairy roots and

168 nodules of chickpea:

169 To test the activity of transgenes transformed following our protocol we employed

170 already known markers for their subcellular localization in chickpea root and nodules.

171 These markers were already published for either Arabidopsis thaliana (At) (model

172 plant), or Medicago truncatula (Mt) (model legume) (Nelson et al., 2007; Ivanov and

173 Harrison, 2014). We expressed some of these markers in hairy roots and nodules of

174 chickpea. A construct created by fusing the signal peptide of AtWAK2 (Arabidopsis

175 thaliana wall-associated kinase 2) at the N-terminus of a mCherry sequence which also

176 contained endoplasmic reticulum (ER) retention signal HDEL at the C-terminus, was

177 used as ER marker. This mCherry was under AtUBQ10 promoter (He et al., 1999). An

178 extensive characteristic tubular and sheet-like ER network was seen throughout the

179 cytoplasm of chickpea root epidermal cells (Supplemental Figure 1A). As expected, ER

180 network around the nucleus was more uniform (Supplemental Figure 1A). The

181 mitochondrial marker was created by adding the ﬁrst 29 amino acids of Saccharomyces

182 cerevisiae cytochrome c oxidase IV with a mCherry sequence under pAtUBQ10 (Kohler

183 et al., 1997). Typical small and round mitochondria were present in high copy numbers

184 in every cell in chickpea hairy root epidermis. The mitochondria were generally

185 clustered in the periphery of the cells, and around the nucleus (Supplemental Figure

186 1B). A similar result was obtained earlier in Medicago (Ivanov and Harrison, 2014). We

187 have used pAtUBQ10 driven mCherry fused with microtubule binding domain of

188 mammalian microtubule-associated protein 4 (Marc et al., 1998) as the microtubule

189 marker. In root epidermal cells, we noticed extensive microtubular networks

190 (Supplemental Figure 1C). The marker for peroxisome was created by fusing SKL

191 signal peptide of AtPTS1 to the C-terminal of mCherry (Reumann, 2004). Innumerable

192 peroxisomes were seen in the cells of chickpea root epidermis (Supplemental Figure

193 1D). Unlike mitochondria, peroxisomes were scattered all around the cells (compare

194 Supplemental Figure 1B and D). To monitor the localization of apoplasts, we used the

195 pAtUBQ10-driven expression of Medicago blue copper protein’s (MtBCP1) 23 amino

196 acid signal peptide (BCPsp) fused to mCherry (Pumplin and Harrison, 2009). MtBCP1

197 signal peptide drives the routing of mCherry to apoplasts, as was also seen in Medicago

198 before (Supplemental Figure 1E) (Ivanov and Harrison, 2014). From the MtBCP1sp-

199 mCherry localization, it is clear that the epidermal cells are connected through

200 apoplastic connection (Supplemental Figure 1E). Additionally, we used p35S promoter

201 driven expression of a plasma membrane aquaporin (PIP2a) fused to cyan fluorescent

202 protein (CFP) (Cutler et al., 2000). We noticed uniform labeling of CFP along the

203 surface of the cell (Supplemental Figure 1F). A marker created by fusing

204 plasmodesmata-located protein 1 (PDLP1) with mCherry at N terminus under

205 pAtUBQ10 was used to visualize the cytoplasmic connections between the cells

206 (Thompson and Wolniak, 2008). A cross-section of chickpea root showed the sym-

207 plastic connections between the cortical cells (Supplemental Figure 1G). The golgi body

208 marker was created by fusing N-terminal 49 amino acids of GmMAN1 to the N terminus

209 of mCherry (Saint-Jore-Dupas et al., 2006). The golgi apparatus were scattered

210 throughout the cytoplasm (Supplemental Figure 1H) in the chickpea root epidermal cell.

211 To check the hormone responsiveness in these transformed roots, we used a construct

212 created by fusing a nuclear localization signal (NLS) at the C-terminus of GFP under

213 auxin responsive synthetic DR5 promoter (DR5-GFP-NLS) (Suzaki et al., 2013). We

214 noticed high GFP expression around the proliferating lateral root primordial cells

215 (Supplemental Figure 1I), but GFP fluorescence was not observed in the mature root

216 cortex (Supplemental Figure 1I). In summary, by using eight different subcellular

217 markers under both 35S and AtUBQ10 promoter, and the DR5-GFP construct, we

218 confirmed the normal physiological functioning of these transformed roots

219 (Supplemental Figure 1).

220 The transgenic hairy roots also showed efficient nodule development (Figure 1). After

221 the bacterial endocytosis in the post-mitotic cells, bacteroids multiply and rapidly fill up

222 the entire cell. The peribacteroid membrane (PBM) is a plant-derived membrane.

223 Hence, the division of bacteroids accompanies a huge expansion of the elements of

224 host cell membrane. Several parallel evidences suggest that membrane trafficking

225 towards symbiosomes increases significantly during this phase (Gavrin et al., 2017).

226 Symbiosomes in Medicago has a unique mosaic identity with plasma membrane

227 syntaxin SYP132, and endomembrane marker (regulatory small GTPases of the Rab

228 family) Rab7 (Limpens et al., 2009). This suggests that under certain circumstances,

229 established markers may behave differently inside the infected nodule cells. To

230 demonstrate the utility of this protocol, we used three subcellular markers (peroxisome,

231 ER, and golgi) to check their localization inside nodule cells. Peroxisomes are

232 organelles that play key roles in plant cell metabolism (Hu et al., 2012). The central

233 infected nodule tissue is surrounded by three layers of uninfected peripheral tissues.

234 Namely from inside to outside, the nodule parenchyma, endodermis, and cortex (Xiao et

235 al., 2014). We noticed the presence of a significantly higher number of peroxisomes in

236 the nodule cortex around the vascular tissue (Figure 2A). The number of peroxisomes

237 were significantly low in nodule parenchyma, and endodermis (Figure 2A). But the

238 concentration of peroxisomes were more or less similar in all infected zones (Figure 2B-

239 C). The golgi apparatus (cis-golgi) were noticed everywhere inside the infected cells

240 surrounding the symbiosomes (Figure 2D). We failed to detect any golgi apparatus

241 surrounding the infection thread (Figure 2D). ERs also show similar localization pattern

242 to the golgi, close to the symbiosomes (Figure 2E). Additionally, we were able to detect

243 ERs around the infection thread (Figure 2F).

244 Evaluating the symbiotic potential of different Mesorhizobium strains:

245 An efficient laboratory strain is a pre-requisite to study nodule development. We

246 examined several different M. ciceri strains originating from different sources to evaluate

247 their potential for healthy nodule development, and N2 fixation efficiency. M. ciceri

248 CC1192 is an Australian isolate used as a commercial inoculant there, and recently the

249 complete genome sequence has been published (Haskett et al., 2016). M. ciceri IC-59

250 was isolated by ICRISAT, India (Rupela and Sudarshana, 1990; Esfahani et al., 2014),

251 and being used for different field trials (Esfahani et al., 2014). M. ciceri TAL620 was

252 used for studies in past (Chandra and Pareek, 1987), and recently has been used for

253 the transcriptome sequencing of chickpea root nodules (Kant et al., 2016). Our own

254 isolation effort from the experimental garden of National Institute of Plant Genome

255 Research (NIPGR), India, has resulted to the culturing of NIPGR Isolate 7 or NI-7.

256 Chickpea plants with healthy pink nodules were collected from the experimental garden.

257 The pink nodules were dissected, surface sterilized, and rhizobia were isolated from the

258 nodules (detail in material and method). The most efficient strain (NI-7) was used for the

259 comparative study. The 16S rDNA sequence of the NI-7 strain has been submitted to

260 NCBI (MH517450). Interestingly, NI-7 shows the highest similarity with Mesorhizobium

261 tianshanense. M. tianshanense nodulating Cicer arietinum and Cicer canariense have

262 already been reported (Rivas et al., 2007; Armas-Capote et al., 2014). We generated

263 nodules by inoculation of all four above mentioned Mesorhizobium strains

264 (Supplemental Figure 2). All four strains were able to initiate nodules by 6 dpi (bump

265 formation) (Supplemental Figure 2 A-D). Numerous small white round shaped nodules

266 were observed by 10 dpi (Supplemental Figure 2 E-H). At 15 dpi, nodules were pink

267 (due to the expression of leghemoglobin) in case of CC1192, IC59, and NI-7

268 (Supplemental Figure 2 I-K) but remained white in case of TAL620 (Supplemental

269 Figure 2L). When followed further, all four nodules become pink by 21 dpi

270 (Supplemental Figure 2 M-P). By 28 and 35dpi CC1192, IC59, and NI-7 nodules

271 showed similar developmental fate by becoming elongated and cylindrical

272 (Supplemental Figure 2 Q-S and U-W), while TAL620 nodules were comparably smaller

273 and pale pink (Supplemental Figure 2 T and X). Senescence zones, a typical feature of

274 indeterminate nodules, started appearing by 35 dpi in most of the nodules

275 (Supplemental Figure 2 U-X). In case of TAL620 nodules, the senescence zone was

276 visible even from 28 dpi (Supplemental Figure 2 T). This suggests that TAL620 is a

277 comparably inefficient strain for chickpea nodule development. We performed acetylene

278 reduction assay (ARA) at 21 and 35 dpi to determine the nitrogen fixation efficiency of

279 these strains. CC1192 showed highest ARA, followed by NI-7, IC59, and TAL620 at

280 both the time points (Figure 3A). TAL620 showed significantly low ARA compared to

281 CC1192 (the most efficient strain), at both the time points. IC59 shows significantly low

282 ARA only at 35 dpi. Taken together, our data show that CC1192 is the most efficient

283 strain for chickpea followed by NI-7. Further, we have used SYTO dyes to stain the

284 nitrogen fixation zone of these chickpea nodules at 21 dpi. Interestingly, we found small

285 spherical symbiosomes inside CC1192, IC59, and NI-7 generated nodules (Figure 3 B-

286 D), whereas in case of TAL620 the symbiosomes were small and elliptical (Figure 3E).

287 Our result shows a direct correlation of the shape of bacteroids with ARA. Acidification

288 of symbiosome space is crucial for functional nitrogen fixation. LysoTracker DND189

289 has been successfully used in Medicago for in planta detection of peribacteriod space

290 (PBS, also called symbiosome space) acidification (Pierre et al., 2013). We then asked

291 the question whether there is any correlation between the PBS acidity and ARA by the

292 above mentioned strains. We used pH sensitive dyes DND189 to stain the most efficient

293 (CC1192), and the least efficient (TAL620) Mesorhizobium-induced nodules. DND189 is

294 a weak base linked with a fluorophore. It gets protonated in acidic organelles and is

295 retained in the membranes of those organelles. The biggest advantage of DND-189

296 over most of the lysotrackers is, it is almost non-fluorescent except when inside acidic

297 compartments (Lopez et al., 2005). DND189 starts fluorescing at a pH of ≤ 5.2 and

298 reaches an optimal emission between pH 4 and 5. We used DND189 in combination

299 with SYTO82/PI to check the acidity of chickpea symbiosome space. In case of

300 CC1192, DND-189 was accumulated in the PBS and we observed the clear spherical

301 shape of the acidic zone (PBS) (Figure 4A-B). The bacterial nucleic acid showed clear

302 red staining when it was counter-stained with SYTO82 (Figure 4A). There was some

303 acidity in the PBS of the symbiosomes made by strain TAL620, but we rarely obtained a

304 clear demarcation of PBS boundary when stained with DND189 (Figure 4C). To test the

305 viability of bacteria inside TAL620-induced nodules, we used propidium iodide (PI) and

306 SYTO staining (Haag et al., 2011). Live bacteria with intact cytoplasmic membranes

307 would be stained by membrane-permeable SYTO dye, while dead bacteria with

308 degraded cytoplasmic membranes would be stained by membrane-impermeable PI.

309 When TAL620 induced nodules were stained with PI, the nucleic acid in most of the

310 samples did not take up the stain. Interestingly, the few instances where sporadic

311 symbiosomes took up the stain, they were spherical (Figure 4D). But the frequency of

312 occurrence of such zone inside the nodules is very scare. Whereas SYTO dye shows

313 live bacteroids inside TAL620 induced nodules (Figure 3E) Hence, from our

314 observation it appears that DND-189 is accumulated less specifically in the PBS of

315 TAL620 induced symbiosomes, indicative of less acidity.

316

317 Discussion:

318 Unlike model legumes, nodule development studies in chickpea are limited. Nodule

319 development in chickpea and Medicago follow a similar course. Hence, the knowledge

320 on Medicago nodule development holds true for many aspects of nodule development

321 in chickpea. Nonetheless, differentiation of bacteroids in both the systems show major

322 dissimilarity. Bacteroid differentiation takes place in the invasion zone of both Medicago

323 and chickpea nodules. Bacteroids inside Medicago nodules endoreduplicate up to 64

324 fold, whereas a maximum of 4 fold endo-reduplication is observed in chickpea (Montiel

325 et al., 2017). Certain NCR genes (dnf4-NCR211 and dnf7-NCR169) are indispensable

326 for nitrogen fixation in Medicago (Horvath et al., 2015; Kim et al., 2015). Recently, it has

327 been shown that two NCR, nitrogen fixation specificity (NFS1 and 2) genes also

328 function as negative regulator of symbiont persistence (Wang et al., 2017; Yang et al.,

329 2017). Such a specific role of few NCR peptides amidst overall redundancy among the

330 members of this family suggests complex regulation behind the evolution of terminal

331 differentiation of bacteroids. Also, the presence of a much lower number of NCR genes

332 and absence of the symbiotically essential NCRs in chickpea genome suggest major

333 differences in the molecular regulation of bacteroid differentiation (Montiel et al., 2017).

334 The recalcitrance of chickpea towards transformation and it’s low regeneration

335 efficiency is a bottleneck at this moment (Indurker et al., 2007; Varshney et al., 2010);

336 Hairy root transformation is an easy alternative to stable lines, for root biologists. The

337 co-transformed roots can be used for over-expression, knockdown or CRISPR/Cas9-

338 mediated deletion of a gene of interest (Limpens et al., 2009; Ron et al., 2014). Till date,

339 the majority of literature on RNS in model legume involves hairy root transformation.

340 The transformation efficiency of model legume ranges around 60-70%. On the other

341 hand, the transformation efficiency of the large-seeded grain legumes soybean, Vigna,

342 and pea varies greatly between 20-80% (Sinharoy et al., 2015). According to our

343 protocol, the transformation efficiency of chickpea varies greatly between ecotypes,

344 ranging between 26-53% (Table 1). This is comparable to other large-seeded grain

345 legumes. These transformed roots develop healthy and active nodules (Figure 1) which

346 can fix nitrogen efficiently (data not shown). We have tested our protocol using nine

347 different constructs carrying different transgenes, and all of them showing an expected

348 pattern and subcellular localization (Figure 2 and Supplemental Figure 1). It should be

349 noted that the transformation frequency using pCMU-ACTLr (actin cytoskeleton)

350 construct was very low (data not shown).

351 Our data show that peroxisomes are enriched around the vascular bundle in the nodule

352 cortex (Figure 2). Peroxisomes play a major role in ureide transport inside the

353 determinant nodules of soybean (Hanks et al., 1981). The role of peroxisomes in amide

354 transporting nodules of temperate legumes is not clear. Scavenging of toxic H2O2

355 around the nodule cortex could be one possibility. Presence of active peroxisomes has

356 been reported also in the meristem and the invasion zones of pea nodules (Borucki,

357 2007). On contrary, we have noticed a higher number of peroxisomes in the nodule

358 cortex (Figure 2A). Discrimination between active and inactive peroxisomes was beyond

359 the limits of the present study. Our observation demands further studies on the

360 significance of peroxisome accumulation around the nodule cortex of temperate

361 legumes like chickpea. In Medicago, the ER apparatus is assembled within the

362 epidermal cells around an area where the future infection thread would penetrate, even

363 before the actual penetrance (Genre et al., 2005). The localization of ER around the

364 infection threads in chickpea nodules indicates similar organization (Figure 2F). ER-

365 medicated trafficking is important during endocytosis of bacteroids. Normally, proteins of

366 exocytosis pathway go from ER to the membrane, through the golgi bodies (Ivanov et

367 al., 2012; Gavrin et al., 2017). Surprisingly, we did not find golgi bodies around the

368 infection threads. The significance of selective concentration of ERs in combination with

369 the absence of golgi bodies around the infection threads is not clear.

370 The efficiency of nitrogen fixation differs significantly depending on bacterial genotypes.

371 A systematic effort to evaluate the behavior of different Mesorhizobium strains during

372 chickpea nodule development has not been given till date. Our study highlights a

373 striking difference of symbiotic performances between different Mesorhizobium strains

374 (Figure 3 and Supplemental Figure 2). Presence of strain-specific negatively regulating

375 NCR gene in Medicago genome (Wang et al., 2017; Yang et al., 2017) opens up the

376 question that is there any NCR gene in chickpea genome that causes toxic effect in

377 TAL620 induces nodule? We detected sporadic spherical-shaped dead bacteroids

378 inside TAL620 induced nodules (Figure 4D). This, together with the late onset of pink

379 coloration and early senescence of the above-mentioned nodules indicate a delay in

380 bacteriod differentiation and less survival time. 63 NCRs express in chickpea BDG256

381 nodules induced by TAL620, as shown by transcriptomic experiments (Kant et al., 2016;

382 Montiel et al., 2017). This suggests that combinations of at least 63 NCRs are required

383 for the development of this uniquely-shaped symbiosome. Our study highlights the

384 importance of transcriptomic study needs to be performed on chickpea nodules

385 containing normal spherical-shaped symbiosomes. A comparative transcriptomic study

386 between nodules containing spherical and elliptical symbiosomes will highlight the

387 NCRs behind the transition between these two unique shapes. It is also very important

388 to screen diverse chickpea ecotypes inoculated by both TAL620 and CC1192/NI-7 to

389 see whether different hosts can generate different morphotypes of bacteroids.

390 Overall, our studies with different bacterial strains along with the efficient hairy root

391 transformation protocol will render the community of nodule biologists a new platform on

392 which genetic experiments of chickpea could easily be performed. Moreover, we have

393 presented enough intriguing features of nodule development in chickpea, which merits

394 further attention in future.

395 MATERIALS AND METHODS

396 Plant material and growth condition:

397 Seeds of chickpea (Cicer arietinum) varieties ICC4958, BDG256, ICC17258, ICC1885,

398 ICC8261 and local varieties (Desi) were surface sterilized by 30% commercial bleach

399 solution (RIN, Hindustan Unilever Limited) containing active sodium hypochlorite 1.2%

400 for 3 minutes. The seeds were thoroughly washed with sterile water. They were placed

401 on plates containing sterile wet filter paper for germination. The plates were kept in 20-

402 22ºC temperature in dark. After 2-3 days, the germinated seedlings were placed in pots

403 with growing substrate (A mixture of 3:1 vermiculite and fire clay balls). The pots were

404 placed in a growth chamber under 200 μE m–2 s–1 for 16 hours light and 8 hours dark

405 cycle, at 22ºC and 40-60% relative humidity.

406 Bacterial material and growth condition:

407 The Agrobacterium rhizogenes R1000, ARqua1 & MSU440 strains were grown at 28ºC

408 in solidified Luria Bertani medium with appropriate antibiotics. The Agrobacterium

409 strains were transformed with the vectors (pCMU-ERr, pCMU-APr, pCMU-ACTLr,

410 pCMU-MITr, pCMU-PERr, pCMU-GAr, pCMU-PDESr, pm-ck CFP CD3 1001, pDR5-

411 GFP-NLS and pKGW Red root (Karimi et al., 2002; Nelson et al., 2007; Suzaki et al.,

412 2013; Ivanov and Harrison, 2014) by the freeze-thaw method (Hofgen and Willmitzer,

413 1988). For hairy root transformation, Agrobacterium was grown for 48 hours at 30ºC in

414 LB agar plate with suitable antibiotics to produce a lawn in the plate. M. ciceri IC59, M.

415 ciceri TAL 620, M. ciceri CC1192 and Mesorhizobium sp. NI-7 were cultivated at 28ºC in

416 Yeast Mannitol broth.

417 Hairy root transformation:

418 To conduct the in-vivo method of transformation, seeds were surface sterilized and

419 germinated. The growing tender roots were cut (1-3cm) below epicotyls close to the

420 cotyledonary region, 2-4 days after germination, with a sterile blade. The wounded

421 region of the cotyledon was rubbed in Agrobacterium lawn with the appropriate vectors.

422 Alternatively, the germinated seedlings were pierced by a needle containing

423 Agrobacterium in the tip. The needle was pricked in epicotyls close to the cotyledonary

424 region of the root. The infected seeds were transferred to a plate containing solidified

425 Fahraeus medium (Boisson-Dernier et al., 2001) without antibiotics. The lower part of

426 the plate was covered with Aluminum foil. The plates were placed longitudinally allowing

427 the root to grow gravitropically, at 22ºC under 16h/8h light and dark cycle. After 7 days

428 hairy root production was observed. When the roots were coming from the injection site,

429 the main roots were cut. Transformed roots were screened under either the Leica stereo

430 fluorescence microscope M205FA equipped with a Leica DFC310FX digital camera

431 (Leica Microsystems) or stereo fluorescence microscope Nikon AZ100 equipped with

432 Nikon digital camera (Nikon Digital Sight DS-Ri1) (Nikon). Non-transformed roots were

433 removed by cutting the roots below the injection region. After 14-15 days the composite

434 plants were transferred to a solidified Fahraeus medium with suitable antibiotics in

435 culture tubes (25x150mm). After 12-15 days the seeds were transferred to growing

436 substrate. Alternatively, the plants with transformed roots were directly transferred in

437 growing substrate without any prior antibiotic selection. The growing substrate was

438 pretreated with full nitrogen B and D medium (Broughton and Dilworth, 1971). The

439 plants were bottom watered four times per week with sterile water. After 10 days in both

440 cases, the plants were treated with full nitrogen B and D medium. After 10-15 days the

441 growing substrate containing plants were washed thoroughly with distilled water and

442 transferred to a new pot for removal of nitrogen. For 2-3 days the plants were

443 maintained in this nitrogen free condition and after that, the plants were inoculated with

444 Mesorhizobium for nodulation.

445 For ex-vitro transformation, plants were grown in pots in 22ºC. A slanted cut (5-6cm)

446 was made in the young branches of 25-30 days grown plants. The wounded region was

447 dipped in a bacterial culture plate (Containing 10mM MES pH 5.6 and 20µM

448 acetosyringone) of Agrobacterium strains with the appropriate vector. Infected

449 segments were transferred into 4 cm long pots with growing substrate. The pots were

450 placed in a tray containing sufficient full nitrogen B and D medium to overcome water

451 stress. A humid chamber was created around the tray with clear wrap and placed in a

452 growth chamber under 16h/8h light and dark cycle at 22ºC. Plants were regularly

453 checked for maintaining water and humidity. After 15 days roots were seen and after 30

454 days the growing substrate containing plants were washed thoroughly with distilled

455 water and transferred to a new pot for removal of nitrogen. After maintaining this

456 nitrogen free condition for 2-3 days the plants were inoculated with Mesorhizobium for

457 nodulation.

458 Nodulation Assay:

459 The growing substrate was vigorously washed for removal of nitrogen prior to

460 nodulation. It was tested for ammoniacal nitrogen and nitrate nitrogen with the soil

461 testing kit (Himedia). The inoculum was prepared from different Mesorhizobium culture

462 of OD600 (0.7-0.9) in Yeast Mannitol medium. The bacterial culture was harvested,

463 media was discarded and diluted in nitrogen-free B and D media. Each pot was

464 inoculated with 50 ml of diluted 0.03-0.05 OD600 bacterial culture in B and D medium.

465 Acetylene Reduction Assay:

466 Acetylene reduction assay was performed as described previously (Oke and Long, 1999

467 ). Briefly, plants were grown on growing substrate and inoculated separately with the

468 four respective bacteria. At 21 dpi and 35 dpi plants were harvested (5 biological

469 replicates of plants inoculated with each bacterium in each time point). Entire root

470 systems were placed in a test tube and seal with suba-seal septa (Sigma-Aldrich) (1

471 plant per tube). The tubes were kept in dark at room temperature for 16 hours. Ethylene

472 production was measured by gas chromatography (Shimadzu GC-2010 equipped with

473 HP-PLOT ‘S’ Al2O3 50m, 0.53 mm column (Agilent technologies). The mean and

474 standard error of mean (SEM) were calculated for five biological replicates.

475 Confocal microscopy:

476 The plants were harvested and the sample was hand sectioned and mounted with

477 water. The staining was done by SYTO13 (DNA, Green) (Invitrogen), SYTO82 (DNA,

478 red) (Invitrogen), Propidium iodide (PI) (DNA of dead cell, red) or DND189 (Acidic

479 compartment, Green) (Invitrogen). The confocal microscopy was performed either in

480 Olympus Model IX81 or in Leica TCS SP5 confocal microscope using an excitation

481 wavelength and emission bandpass of 488 nm and 500-530 nm for SYTO13, 541 nm

482 and 550-580 nm for SYTO82, 448 and 500-530 nm for DND189, 587 nm 600-630 nm

483 for mCherry, 535 nm and 610-640 nm for PI.

484 Isolation of Mesorhizobium from chickpea nodules:

485 The plants were collected from NIPGR experimental field. Healthy looking pink nodule

486 containing plants were selected. Roots were washed thoroughly to remove soil. About

487 10 pink nodules were collected by cutting the nodules from the root about 0.5 cm on

488 each side. The nodules were placed in sterile water with 0.12g glass beads and

489 vortexed for 30-60 seconds in repetitive cycles to remove adhering soil particles. It was

490 treated with 30% commercial bleach solution (RIN, Hindustan Unilever Limited)

491 containing active sodium hypochlorite 1.2% for 3 minutes and washed with sterile water

492 for 5-6 times. After surface sterilization, the nodules were rolled over solidified Yeast

493 Mannitol medium and incubated for 2 to 3 days at 28ºC (control). Immediately, the same

494 nodules were squashed in sterile water. It was serially diluted with sterile water and

495 plated in solidified Yeast Mannitol medium. If the control plate does not shows growth,

496 then only the corresponding experimental plates were processed for further

497 characterization. The experimental plates were placed in 28ºC for 5 to 6 days. The

498 growth was checked regularly. The bacterial colony in the experimental plates were

499 subculture to purify the bacterial culture and further used for the nodule developmental

500 experiment.

501 Cloning of 16S rDNA:

502 NI-7 16S rDNA was amplified by forward 5′ TAACACATGCAAGTCGAACG 3′ and

503 reverse 5′ ACGGGCGGTGTGTAC 3′ primers using PrimeSTAR Max DNA polymerase

504 (Takara, Clontec). It was then cloned in pJET 1.2/blunt vector (Clone JET PCR cloning

505 kit, Thermo Fisher). Sanger sequencing was done with sequencing primers provided by

506 Clone JET PCR cloning kit (forward 5′ CGACTCACTATAGGGAGAGCGGC 3′ and

507 reverse 5′ AAGAACATCGATTTTCCATGGCAG 3′). The sequence was assembled and

508 processed with Geneious v9.1.8. The 16S rDNA sequence of the NI-7 strain has been

509 submitted to NCBI (MH517450)

510 Table 1: Agrobacterium rhizogenes mediated transformation of different chickpea

511 ecotypes with pKGW-Red-Root:

Genotype Region Method of transformation Efficiency of transformation

In vivo Ex vitro In vivo Ex vitro

Cut Inject Cut Inject

ICC4958 India 20(16) ND ND 26.027% ND ND

(DESI)

BDG256 India 20(19) 20(17) 20(14) 48.717% 39.614% 38.333%

(DESI)

BDG256 India 25(21) ND ND 53.135%* ND ND

(DESI)

ICC17258 India 20(20) ND ND 40.331% ND ND

(DESI)

ICC1885 India 20(18) ND ND 37.313% ND ND

(DESI)

ICC8261 Turkey 20(9) ND ND 26.984% ND ND

(KABULI)

Local Variety India 35(35) 35(31) 35(34) 50.259%* 51.437%* 40.107%

(DESI)

512 * % after antibiotic selection

513

514 Figure 1: Agrobacterium rhizogenes induced hairy roots in chickpea. In vivo

515 transformation: Germinated seedlings were infected with Agrobacterium rhizogenes and

516 plated on Fahraeus medium. 14-15 days after infection hairy roots were seen and the

517 composite plants were transferred in selective antibiotics (depicted in step Ia). After 10-

518 12 days, the plants were transferred to growing substrate (depicted in step IIa).

519 Transformed roots were screen under the red channel of the stereo-microscope

520 (depicted in step IIIa). Composite plants were inoculated with M. ciceri CC1192 after 10

521 days (depicted in step IVa). Ex vitro transformation: The wounded portion was

522 inoculated with Agrobacterium rhizogenes. 30 days after infection the hairy roots were

523 seen (depicted in step Ib) and transformed roots were screened under the red channel

524 of stereo-microscope (depicted in step IIb). The composite plants were inoculated with

525 M. ciceri CC1192 for nodulation (depicted in step IIIb). The transformed root pictures

526 were taken under stereo-microscope either under bright field or under red channel.

527 Figure 2: Localization of different subcellular protein markers in transgenic chickpea

528 nodules: A-F, Transformed nodules were hand sectioned and stained with SYTO13

529 (DNA, green). A-C, mCherry-SKL showing localization A, whole nodule B, invasion

530 zone C, nitrogen fixation zone D, MAN49-mCherry localization in the nitrogen fixation

531 zone E-F, mCherry-HDEL localization in endoplasmic reticulum in the nitrogen fixation

532 zone of transgenic chickpea nodules induced by M. ciceri CC1192. Nucleus marked by

533 N and infection thread marked by IT. Scale bars in A, = 50µm and B-F, = 5µm.

534 Figure 3: Characterization of different chickpea nodules: A, Acetylene reduction activity

535 (ARA) with Mesorhizobium ciceri CC1192, Mesorhizobium ciceri IC59, Mesorhizobium

536 sp. NI-7 and Mesorhizobium ciceri TAL620 induced nodules at 21 dpi and 35 dpi. ARA

537 was measured per plant. Mean and SEM of five biological replicates are presented in

538 each case. B-D, 21 dpi chickpea nodules were hand sliced and stained with SYTO82

539 and E, SYTO13. B, Mesorhizobium ciceri CC1192 C, Mesorhizobium ciceri IC59 D,

540 Mesorhizobium sp. NI-7. E, Mesorhizobium ciceri TAL 620 induced nodules in BDG256

541 chickpea ecotype. Scale bars in B-E, = 5µm. Asterisks indicate significantly less

542 acetylene reduction by IC59 and TAL620 induced nodules: *P ≤ 0.05, **P ≤ 0.01

543 Figure 4: Characterization of the acidity of the peribacteroid space of chickpea nodules:

544 A-D, 21 dpi chickpea nodule were hand sliced and stained with DND189 (Acidic

545 compartment, green). A, Counterstain with SYTO82 (DNA, a live cell). D,

546 Counterstained with PI (DNA of a dead cell). A-B, Nodules were induced by

547 Mesorhizobium ciceri CC1192. C-D, Nodules were induced by Mesorhizobium ciceri

548 TAL 620. Scale bars in A-C, = 5µm and D, = 10 µm.

549 Supplemental Figure 1: Localization of different subcellular protein markers in

550 chickpea hairy roots: A, mCherry-HDEL localization to the ER. B, COX4-mCherry

551 localization to mitochondria C, LifeAct-mCherry localization to actin microfilaments, D,

552 mCherry-SKL localization to peroxisome E, BCPsp-mCherry localization to apoplastic

553 space F, PIP2A-CFP localization to the plasma membrane in the chickpea epidermal

554 cells. G, PDLP1-mCherry localization to plasmodesmata in the chickpea cortical cells H,

555 MAN49-mCherry localization to golgi apparatus (cis-golgi) in the chickpea epidermal

556 cells I, Localization of GFP-NLS to the nucleus of chickpea root primordia driven by the

557 DR5 promoter. Nucleus marked by N and vascular bundle marked by V. Scale bars in

558 B-D, G, = 10µm, A, E-F, H, = 5µm and I, = 50µm.

559 Supplemental Figure 2: Nodule developmental time course of chickpea nodules: A-D,

560 6 dpi nodules. E-H, 10 dpi nodules I-L, 15 dpi nodules M-P, 21 dpi nodules Q-T, 28 dpi

561 nodules U-X, 35 dpi nodules A, E, I, M, Q, U, Mesorhizobium ciceri CC1192 induced

562 nodules B, F, J, N, R, V, Mesorhizobium ciceri IC59 induced nodules C, G, K, O, S, W,

563 Mesorhizobium sp. NI-7 induced nodules D, H, L, P, T, X, Mesorhizobium ciceri TAL

564 620 induced nodules. Scale bars A-X, = 1mm.

565 Acknowledgement:

566 We thank R. Varshney and S. Gopalakrishnan, ICRISAT, India; S. Bhatia, NIPGR-New

567 Delhi, J. Terpolilli, Murdoch University, Australia for providing M. ciceri IC59, M. ciceri

568 TAL 620, and M. ciceri CC1192 respectively; H.D. Upadhyaya and D. Sastry, ICRISAT,

569 India for providing chickpea seeds (ICC4958, ICC17258, ICC1882, ICC8261); S. Bhatia,

570 NIPGR for providing BDG256. D. J. Chattopadhyay, Amity University; M. DasGupta of

571 Department of Biochemistry, University of Calcutta, and A. Seal, Department of

572 Biotechnology, University of Calcutta, for their enormous support and allowing us to use

573 their facility. A. Seal for providing pm-ck CFP CD3 1001 construct, Maria J. Harrison,

574 Boyce Thompson Institute for Plant Research Ithaca, USA for providing pAtUb driven

575 subcellular marker constructs and Suzaki Takuya, University of Tsukuba, Japan for

576 providing DR5-GFP-NLS construct. DBT-CU-IPLS and NIPGR for their confocal

577 facilities; CIF-NIPGR; NIPGR-DELCON for their support. T Khanna, Jamia Millia

578 Islamia, New Delhi for technical support. This work was supported by core research

579 grant from National Institute of Plant Genome Research and Ramalingwaswami Re-

580 entry grant, DBT (BT/RLF/Re-entry/41/2013).

581 Reference:

582 Alunni, B., Kevei, Z., Redondo-Nieto, M., Kondorosi, A., Mergaert, P., and Kondorosi, E. 583 2007. Genomic organization and evolutionary insights on GRP and NCR genes, 584 two large nodule-specific gene families in Medicago truncatula. Mol Plant 585 Microbe Interact 20:1138-1148.

586 Armas-Capote, N., Perez-Yepez, J., Martinez-Hidalgo, P., Garzon-Machado, V., Del 587 Arco-Aguilar, M., Velazquez, E., and Leon-Barrios, M. 2014. Core and symbiotic 588 genes reveal nine Mesorhizobium genospecies and three symbiotic lineages 589 among the rhizobia nodulating Cicer canariense in its natural habitat (La Palma, 590 Canary Islands). Syst Appl Microbiol 37:140-148. 591 Boisson-Dernier, A., Chabaud, M., Garcia, F., Becard, G., Rosenberg, C., and Barker, 592 D.G. 2001. Agrobacterium rhizogenes-transformed roots of Medicago truncatula 593 for the study of nitrogen-fixing and endomycorrhizal symbiotic associations. Mol 594 Plant Microbe Interact 14:695-700. 595 Bonaldi, K., Gherbi, H., Franche, C., Bastien, G., Fardoux, J., Barker, D., Giraud, E., 596 and Cartieaux, F. 2010. The Nod factor-independent symbiotic signaling 597 pathway: development of Agrobacterium rhizogenes-mediated transformation for 598 the legume Aeschynomene indica. Mol Plant Microbe Interact 23:1537-1544. 599 Borucki, W. 2007. Proliferation of peroxisomes in pea root nodules - an influence of 600 NaCI- or Hg2+- stress conditions. ACTA Societatis Botanicorum Poloniae 601 76:287-298. 602 Brijwal, L., and Tamta, S. 2015. Agrobacterium rhizogenes mediated hairy root 603 induction in endangered Berberis aristata DC. Springerplus 4:443. 604 Broughton, W.J., and Dilworth, M.J. 1971. Control of leghaemoglobin synthesis in snake 605 beans. Biochem J. 125:1075–1080. 606 Cai, Y., Chen, L., Liu, X., Sun, S., Wu, C., Jiang, B., Han, T., and Hou, W. 2015. 607 CRISPR/Cas9-Mediated Genome Editing in Soybean Hairy Roots. PLoS One 608 10:e0136064. 609 Chandra, R., and Pareek, R.P. 1987. Effect of inoculum rate on the performance of 610 chickpea (Cicer arietinum L.) Rhizobium strains in the field. Biology and Fertility 611 of Soils 5:83–87. 612 Collier, R., Fuchs, B., Walter, N., Kevin Lutke, W., and Taylor, C.G. 2005. Ex vitro 613 composite plants: an inexpensive, rapid method for root biology. Plant J 43:449- 614 457. 615 Cutler, S.R., Ehrhardt, D.W., Griffitts, J.S., and Somerville, C.R. 2000. Random 616 GFP::cDNA fusions enable visualization of subcellular structures in cells of 617 Arabidopsis at a high frequency. Proc Natl Acad Sci U S A 97:3718-3723. 618 Durgo, H., Klement, E., Hunyadi-Gulyas, E., Szucs, A., Kereszt, A., Medzihradszky, 619 K.F., and Kondorosi, E. 2015. Identification of nodule-specific cysteine-rich plant 620 peptides in endosymbiotic bacteria. Proteomics 15:2291-2295. 621 Esfahani, M.N., Sulieman, S., Schulze, J., Yamaguchi-Shinozaki, K., Shinozaki, K., and 622 Tran, L.S. 2014. Approaches for enhancement of N(2) fixation efficiency of 623 chickpea (Cicer arietinum L.) under limiting nitrogen conditions. Plant Biotechnol 624 J 12:387-397. 625 Estrada-Navarrete, G., Alvarado-Affantranger, X., Olivares, J.E., Diaz-Camino, C., 626 Santana, O., Murillo, E., Guillen, G., Sanchez-Guevara, N., Acosta, J., Quinto, 627 C., Li, D., Gresshoff, P.M., and Sanchez, F. 2006. Agrobacterium rhizogenes 628 transformation of the Phaseolus spp.: a tool for functional genomics. Mol Plant 629 Microbe Interact 19:1385-1393.

630 Gavrin, A., Kulikova, O., Bisseling, T., and Fedorova, E.E. 2017. Interface Symbiotic 631 Membrane Formation in Root Nodules of Medicago truncatula: the Role of 632 Synaptotagmins MtSyt1, MtSyt2 and MtSyt3. Front Plant Sci 8:201. 633 Genre, A., Chabaud, M., Timmers, T., Bonfante, P., and Barker, D.G. 2005. Arbuscular 634 mycorrhizal fungi elicit a novel intracellular apparatus in Medicago truncatula root 635 epidermal cells before infection. Plant Cell 17:3489-3499. 636 Georgiev, M.I., Agostini, E., Ludwig-Muller, J., and Xu, J. 2012. Genetically transformed 637 roots: from plant disease to biotechnological resource. Trends Biotechnol 30:528- 638 537. 639 Haag, A.F., Baloban, M., Sani, M., Kerscher, B., Pierre, O., Farkas, A., Longhi, R., 640 Boncompagni, E., Herouart, D., Dall'angelo, S., Kondorosi, E., Zanda, M., 641 Mergaert, P., and Ferguson, G.P. 2011. Protection of Sinorhizobium against host 642 cysteine-rich antimicrobial peptides is critical for symbiosis. PLoS Biol 643 9:e1001169. 644 Habibi, P., de Sa, M.F., da Silva, A.L., Makhzoum, A., da Luz Costa, J., Borghetti, I.A., 645 and Soccol, C.R. 2016. Efficient genetic transformation and regeneration system 646 from hairy root of Origanum vulgare. Physiol Mol Biol Plants 22:271-277. 647 Hanks, J.F., Tolbert, N.E., and Schubert, K.R. 1981. Localization of Enzymes of Ureide 648 Biosynthesis in Peroxisomes and Microsomes of Nodules. Plant Physiol. 68:65- 649 69. 650 Haskett, T., Wang, P., Ramsay, J., O'Hara, G., Reeve, W., Howieson, J., and Terpolilli, 651 J. 2016. Complete Genome Sequence of Mesorhizobium ciceri Strain CC1192, 652 an Efficient Nitrogen-Fixing Microsymbiont of Cicer arietinum. Genome Announc 653 4. 654 He, Z.H., Cheeseman, I., He, D., and Kohorn, B.D. 1999. A cluster of five cell wall- 655 associated receptor kinase genes, Wak1-5, are expressed in specific organs of 656 Arabidopsis. Plant Mol Biol 39:1189-1196. 657 Hofgen, R., and Willmitzer, L. 1988. Storage of competent cells for Agrobacterium 658 transformation. Nucleic Acids Res 16:9877. 659 Horvath, B., Domonkos, A., Kereszt, A., Szucs, A., Abraham, E., Ayaydin, F., Boka, K., 660 Chen, Y., Chen, R., Murray, J.D., Udvardi, M.K., Kondorosi, E., and Kalo, P. 661 2015. Loss of the nodule-specific cysteine rich peptide, NCR169, abolishes 662 symbiotic nitrogen fixation in the Medicago truncatula dnf7 mutant. Proc Natl 663 Acad Sci U S A 112:15232-15237. 664 Hu, J., Baker, A., Bartel, B., Linka, N., Mullen, R.T., Reumann, S., and Zolman, B.K. 665 2012. Plant peroxisomes: biogenesis and function. Plant Cell 24:2279-2303. 666 Imanishi, L., Vayssieres, A., Franche, C., Bogusz, D., Wall, L., and Svistoonoff, S. 2011. 667 Transformed hairy roots of Discaria trinervis: a valuable tool for studying 668 actinorhizal symbiosis in the context of intercellular infection. Mol Plant Microbe 669 Interact 24:1317-1324. 670 Indurker, S., Misra, H.S., and Eapen, S. 2007. Genetic transformation of chickpea 671 (Cicer arietinum L.) with insecticidal crystal protein gene using particle gun 672 bombardment. Plant Cell Rep 26:755-763. 673 Ivanov, S., and Harrison, M.J. 2014. A set of fluorescent protein-based markers 674 expressed from constitutive and arbuscular mycorrhiza-inducible promoters to

675 label organelles, membranes and cytoskeletal elements in Medicago truncatula. 676 Plant J 80:1151-1163. 677 Ivanov, S., Fedorova, E.E., Limpens, E., De Mita, S., Genre, A., Bonfante, P., and 678 Bisseling, T. 2012. Rhizobium-legume symbiosis shares an exocytotic pathway 679 required for arbuscule formation. Proc Natl Acad Sci U S A 109:8316-8321. 680 Jain, M., Misra, G., Patel, R.K., Priya, P., Jhanwar, S., Khan, A.W., Shah, N., Singh, 681 V.K., Garg, R., Jeena, G., Yadav, M., Kant, C., Sharma, P., Yadav, G., Bhatia, 682 S., Tyagi, A.K., and Chattopadhyay, D. 2013. A draft genome sequence of the 683 pulse crop chickpea (Cicer arietinum L.). Plant J 74:715-729. 684 Jukanti, A.K., Gaur, P.M., Gowda, C.L., and Chibbar, R.N. 2012. Nutritional quality and 685 health benefits of chickpea (Cicer arietinum L.): a review. Br J Nutr 108 Suppl 686 1:S11-26. 687 Kant, C., Pradhan, S., and Bhatia, S. 2016. Dissecting the Root Nodule Transcriptome 688 of Chickpea (Cicer arietinum L.). PLoS One 11:e0157908. 689 Karimi, M., Inze, D., and Depicker, A. 2002. GATEWAY vectors for Agrobacterium- 690 mediated plant transformation. Trends Plant Sci 7:193-195. 691 Kereszt, A., Li, D., Indrasumunar, A., Nguyen, C.D., Nontachaiyapoom, S., Kinkema, 692 M., and Gresshoff, P.M. 2007. Agrobacterium rhizogenes-mediated 693 transformation of soybean to study root biology. Nature protocols 2:948-952. 694 Kim, M., Chen, Y., Xi, J., Waters, C., Chen, R., and Wang, D. 2015. An antimicrobial 695 peptide essential for bacterial survival in the nitrogen-fixing symbiosis. Proc Natl 696 Acad Sci U S A 112:15238-15243. 697 Kohler, R.H., Cao, J., Zipfel, W.R., Webb, W.W., and Hanson, M.R. 1997. Exchange of 698 protein molecules through connections between higher plant plastids. Science 699 276:2039-2042. 700 Limpens, E., Ivanov, S., van Esse, W., Voets, G., Fedorova, E., and Bisseling, T. 2009. 701 Medicago N2-fixing symbiosomes acquire the endocytic identity marker Rab7 but 702 delay the acquisition of vacuolar identity. Plant Cell 21:2811-2828. 703 Limpens, E., Ramos, J., Franken, C., Raz, V., Compaan, B., Franssen, H., Bisseling, T., 704 and Geurts, R. 2004. RNA interference in Agrobacterium rhizogenes-transformed 705 roots of Arabidopsis and Medicago truncatula. J Exp Bot 55:983-992. 706 Lopez, J.J., Camello-Almaraz, C., Pariente, J.A., Salido, G.M., and Rosado, J.A. 2005. 707 Ca2+ accumulation into acidic organelles mediated by Ca2+- and vacuolar H+- 708 ATPases in human platelets. Biochem J 390:243-252. 709 Marc, J., Granger, C.L., Brincat, J., Fisher, D.D., Kao, T., McCubbin, A.G., and Cyr, R.J. 710 1998. A GFP-MAP4 reporter gene for visualizing cortical microtubule 711 rearrangements in living epidermal cells. Plant Cell 10:1927-1940. 712 Mergaert, P., Uchiumi, T., Alunni, B., Evanno, G., Cheron, A., Catrice, O., Mausset, 713 A.E., Barloy-Hubler, F., Galibert, F., Kondorosi, A., and Kondorosi, E. 2006. 714 Eukaryotic control on bacterial cell cycle and differentiation in the Rhizobium- 715 legume symbiosis. Proc Natl Acad Sci U S A 103:5230-5235. 716 Montiel, J., Szucs, A., Boboescu, I.Z., Gherman, V.D., Kondorosi, E., and Kereszt, A. 717 2016. Terminal Bacteroid Differentiation Is Associated With Variable 718 Morphological Changes in Legume Species Belonging to the Inverted Repeat- 719 Lacking Clade. Mol Plant Microbe Interact 29:210-219.

720 Montiel, J., Downie, J.A., Farkas, A., Bihari, P., Herczeg, R., Balint, B., Mergaert, P., 721 Kereszt, A., and Kondorosi, E. 2017. Morphotype of bacteroids in different 722 legumes correlates with the number and type of symbiotic NCR peptides. Proc 723 Natl Acad Sci U S A 114:5041-5046. 724 Nelson, B.K., Cai, X., and Nebenfuhr, A. 2007. A multicolored set of in vivo organelle 725 markers for co-localization studies in Arabidopsis and other plants. Plant J 726 51:1126-1136. 727 Oke, V., and Long, S.R. 1999 Bacterial genes induced within the nodule during the 728 Rhizobium-legume symbiosis. Mol Microbiol. 32:837-849. 729 Oldroyd, G.E. 2013. Speak, friend, and enter: signalling systems that promote beneficial 730 symbiotic associations in plants. Nat Rev Microbiol 11:252-263. 731 Oono, R., and Denison, R.F. 2010. Comparing symbiotic efficiency between swollen 732 versus nonswollen rhizobial bacteroids. Plant Physiol 154:1541-1548. 733 Pandey, A., Chakraborty, S., and Chakraborty, N. 2018. Nuclear Proteome: Isolation of 734 Intact Nuclei, Extraction of Nuclear Proteins, and 2-DE Analysis. Methods Mol 735 Biol 1696:41-55. 736 Pierre, O., Engler, G., Hopkins, J., Brau, F., Boncompagni, E., and Herouart, D. 2013. 737 Peribacteroid space acidification: a marker of mature bacteroid functioning in 738 Medicago truncatula nodules. Plant Cell Environ 36:2059-2070. 739 Pumplin, N., and Harrison, M.J. 2009. Live-cell imaging reveals periarbuscular 740 membrane domains and organelle location in Medicago truncatula roots during 741 arbuscular mycorrhizal symbiosis. Plant Physiol 151:809-819. 742 Quandt, H.J., Puhler, A., and Broer, I. 1993. Transgenic root nodules of Vicia hirsuta. A 743 fast and efficient system for the study of gene expression in indeterminate-type 744 nodules. Mol Plant Microbe Interact 6:699-703. 745 Ramalingam, A., Kudapa, H., Pazhamala, L.T., Weckwerth, W., and Varshney, R.K. 746 2015. Proteomics and Metabolomics: Two Emerging Areas for Legume 747 Improvement. Front Plant Sci 6:1116. 748 Reumann, S. 2004. Specification of the peroxisome targeting signals type 1 and type 2 749 of plant peroxisomes by bioinformatics analyses. Plant Physiol 135:783-800. 750 Rivas, R., Laranjo, M., Mateos, P.F., Oliveira, S., Martinez-Molina, E., and Velazquez, 751 E. 2007. Strains of Mesorhizobium amorphae and Mesorhizobium tianshanense, 752 carrying symbiotic genes of common chickpea endosymbiotic species, constitute 753 a novel biovar (ciceri) capable of nodulating Cicer arietinum. Lett Appl Microbiol 754 44:412-418. 755 Ron, M., Kajala, K., Pauluzzi, G., Wang, D., Reynoso, M.A., Zumstein, K., Garcha, J., 756 Winte, S., Masson, H., Inagaki, S., Federici, F., Sinha, N., Deal, R.B., Bailey- 757 Serres, J., and Brady, S.M. 2014. Hairy root transformation using Agrobacterium 758 rhizogenes as a tool for exploring cell type-specific gene expression and function 759 using tomato as a model. Plant Physiol 166:455-469. 760 Roth, L.E., and Stacey, G. 1989. Bacterium release into host cells of nitrogen-fixing 761 soybean nodules: The symbiosome membrane comes from three sources. Eur. 762 J. Cell Biol. 49:13-23. 763 Rupela, O.P., and Sudarshana, M.R. 1990. Displacement of native rhizobia nodulating 764 chickpea (Cicer arietinum L.) by an inoculant strain through soil solarization. 765 Biology and Fertility of Soils 3:207-212.

766 Saint-Jore-Dupas, C., Nebenfuhr, A., Boulaflous, A., Follet-Gueye, M.L., Plasson, C., 767 Hawes, C., Driouich, A., Faye, L., and Gomord, V. 2006. Plant N-glycan 768 processing enzymes employ different targeting mechanisms for their spatial 769 arrangement along the secretory pathway. Plant Cell 18:3182-3200. 770 Sinharoy, S., and DasGupta, M. 2009. RNA interference highlights the role of CCaMK in 771 dissemination of endosymbionts in the Aeschynomeneae legume Arachis. Mol 772 Plant Microbe Interact 22:1466-1475. 773 Sinharoy, S., Pislariu, C.I., and Udvardi, M.K. 2015. A high-throughput RNA interference 774 (RNAi)-based approach using hairy roots for the study of plant-rhizobia 775 interactions. Methods Mol Biol 1287:159-178. 776 Sinharoy, S., Saha, S., Chaudhury, S.R., and Dasgupta, M. 2009. Transformed hairy 777 roots of Arachis hypogea: a tool for studying root nodule symbiosis in a non- 778 infection thread legume of the Aeschynomeneae tribe. Mol Plant Microbe Interact 779 22:132-142. 780 Sinharoy, S., Torres-Jerez, I., Bandyopadhyay, K., Kereszt, A., Pislariu, C.I., 781 Nakashima, J., Benedito, V.A., Kondorosi, E., and Udvardi, M.K. 2013. The 782 C2H2 transcription factor regulator of symbiosome differentiation represses 783 transcription of the secretory pathway gene VAMP721a and promotes 784 symbiosome development in Medicago truncatula. Plant Cell 25:3584-3601. 785 Starker, C.G., Parra-Colmenares, A.L., Smith, L., Mitra, R.M., and Long, S.R. 2006. 786 Nitrogen fixation mutants of Medicago truncatula fail to support plant and 787 bacterial symbiotic gene expression. Plant Physiol 140:671-680. 788 Stiller, J., Martirani, L., Tuppale, S., Chian, R.-J., Chiurazzi, M., and Gresshoff, P.M. 789 1997. High frequency transformation and regeneration of transgenic plants in the 790 model legume Lotus japonicus. Journal of Experimental Botany 48:1357-1365. 791 Stougaard, J., Petersen, T.E., and Marcker, K.A. 1987. Expression of a complete 792 soybean leghemoglobin gene in root nodules of transgenic Lotus corniculatus. 793 Proc Natl Acad Sci USA 84:5754-5757. 794 Suzaki, T., Ito, M., and Kawaguchi, M. 2013. Induction of localized auxin response 795 during spontaneous nodule development in Lotus japonicus. Plant Signal Behav 796 8:e23359. 797 Thilip, C., Soundar Raju, C., Varutharaju, K., Aslam, A., and Shajahan, A. 2015. 798 Improved Agrobacterium rhizogenes-mediated hairy root culture system of 799 Withania somnifera (L.) Dunal using sonication and heat treatment. 3 Biotech 800 5:949-956. 801 Thompson, M.V., and Wolniak, S.M. 2008. A plasma membrane-anchored fluorescent 802 protein fusion illuminates sieve element plasma membranes in Arabidopsis and 803 tobacco. Plant Physiol 146:1599-1610. 804 Thwe, A., Valan Arasu, M., Li, X., Park, C.H., Kim, S.J., Al-Dhabi, N.A., and Park, S.U. 805 2016. Effect of Different Agrobacterium rhizogenes Strains on Hairy Root 806 Induction and Phenylpropanoid Biosynthesis in Tartary Buckwheat (Fagopyrum 807 tataricum Gaertn). Front Microbiol 7:318. 808 Udvardi, M., and Poole, P.S. 2013. Transport and metabolism in legume-rhizobia 809 symbioses. Annu Rev Plant Biol 64:781-805.

810 Van-de-Velde, W., Mergeay, J., Holsters, M., and Goormachtig, S. 2003. Agrobacterium 811 rhizogenes-mediated transformation of Sesbania rostrata. Plant Science 812 165:1281–1288. 813 Varshney, R.K., Thudi, M., May, G.D., and Jackson, S.A. 2010. Legume genomics and 814 breeding. Plant Breed Rev. 33:257. 815 Varshney, R.K., Song, C., Saxena, R.K., Azam, S., Yu, S., Sharpe, A.G., Cannon, S., 816 Baek, J., Rosen, B.D., Tar'an, B., Millan, T., Zhang, X., Ramsay, L.D., Iwata, A., 817 Wang, Y., Nelson, W., Farmer, A.D., Gaur, P.M., Soderlund, C., Penmetsa, R.V., 818 Xu, C., Bharti, A.K., He, W., Winter, P., Zhao, S., Hane, J.K., Carrasquilla-Garcia, 819 N., Condie, J.A., Upadhyaya, H.D., Luo, M.C., Thudi, M., Gowda, C.L., Singh, 820 N.P., Lichtenzveig, J., Gali, K.K., Rubio, J., Nadarajan, N., Dolezel, J., Bansal, 821 K.C., Xu, X., Edwards, D., Zhang, G., Kahl, G., Gil, J., Singh, K.B., Datta, S.K., 822 Jackson, S.A., Wang, J., and Cook, D.R. 2013. Draft genome sequence of 823 chickpea (Cicer arietinum) provides a resource for trait improvement. Nat 824 Biotechnol 31:240-246. 825 Wang, D., Griffitts, J., Starker, C., Fedorova, E., Limpens, E., Ivanov, S., Bisseling, T., 826 and Long, S. 2010. A nodule-specific protein secretory pathway required for 827 nitrogen-fixing symbiosis. Science 327:1126-1129. 828 Wang, L., Wang, L., Tan, Q., Fan, Q., Zhu, H., Hong, Z., Zhang, Z., and Duanmu, D. 829 2016. Efficient Inactivation of Symbiotic Nitrogen Fixation Related Genes in Lotus 830 japonicus Using CRISPR-Cas9. Front Plant Sci 7:1333. 831 Wang, Q., Yang, S., Liu, J., Terecskei, K., Abraham, E., Gombar, A., Domonkos, A., 832 Szucs, A., Kormoczi, P., Wang, T., Fodor, L., Mao, L., Fei, Z., Kondorosi, E., 833 Kalo, P., Kereszt, A., and Zhu, H. 2017. Host-secreted antimicrobial peptide 834 enforces symbiotic selectivity in Medicago truncatula. Proc Natl Acad Sci U S A. 835 Werner, G.D., Cornwell, W.K., Cornelissen, J.H., and Kiers, E.T. 2015. Evolutionary 836 signals of symbiotic persistence in the legume-rhizobia mutualism. Proc Natl 837 Acad Sci U S A 112:10262-10269. 838 Xiao, T.T., Schilderink, S., Moling, S., Deinum, E.E., Kondorosi, E., Franssen, H., 839 Kulikova, O., Niebel, A., and Bisseling, T. 2014. Fate map of Medicago truncatula 840 root nodules. Development 141:3517-3528. 841 Yang, S., Wang, Q., Fedorova, E., Liu, J., Qin, Q., Zheng, Q., Price, P.A., Pan, H., 842 Wang, D., Griffitts, J.S., Bisseling, T., and Zhu, H. 2017. Microsymbiont 843 discrimination mediated by a host-secreted peptide in Medicago truncatula. Proc 844 Natl Acad Sci U S A. 845 Young, N.D., Debelle, F., Oldroyd, G.E., Geurts, R., Cannon, S.B., Udvardi, M.K., 846 Benedito, V.A., Mayer, K.F., Gouzy, J., Schoof, H., Van de Peer, Y., Proost, S., 847 Cook, D.R., Meyers, B.C., Spannagl, M., Cheung, F., De Mita, S., Krishnakumar, 848 V., Gundlach, H., Zhou, S., Mudge, J., Bharti, A.K., Murray, J.D., Naoumkina, 849 M.A., Rosen, B., Silverstein, K.A., Tang, H., Rombauts, S., Zhao, P.X., Zhou, P., 850 Barbe, V., Bardou, P., Bechner, M., Bellec, A., Berger, A., Berges, H., Bidwell, 851 S., Bisseling, T., Choisne, N., Couloux, A., Denny, R., Deshpande, S., Dai, X., 852 Doyle, J.J., Dudez, A.M., Farmer, A.D., Fouteau, S., Franken, C., Gibelin, C., 853 Gish, J., Goldstein, S., Gonzalez, A.J., Green, P.J., Hallab, A., Hartog, M., Hua, 854 A., Humphray, S.J., Jeong, D.H., Jing, Y., Jocker, A., Kenton, S.M., Kim, D.J., 855 Klee, K., Lai, H., Lang, C., Lin, S., Macmil, S.L., Magdelenat, G., Matthews, L.,

856 McCorrison, J., Monaghan, E.L., Mun, J.H., Najar, F.Z., Nicholson, C., Noirot, C., 857 O'Bleness, M., Paule, C.R., Poulain, J., Prion, F., Qin, B., Qu, C., Retzel, E.F., 858 Riddle, C., Sallet, E., Samain, S., Samson, N., Sanders, I., Saurat, O., Scarpelli, 859 C., Schiex, T., Segurens, B., Severin, A.J., Sherrier, D.J., Shi, R., Sims, S., 860 Singer, S.R., Sinharoy, S., Sterck, L., Viollet, A., Wang, B.B., Wang, K., Wang, 861 M., Wang, X., Warfsmann, J., Weissenbach, J., White, D.D., White, J.D., Wiley, 862 G.B., Wincker, P., Xing, Y., Yang, L., Yao, Z., Ying, F., Zhai, J., Zhou, L., Zuber, 863 A., Denarie, J., Dixon, R.A., May, G.D., Schwartz, D.C., Rogers, J., Quetier, F., 864 Town, C.D., and Roe, B.A. 2011. The Medicago genome provides insight into the 865 evolution of rhizobial symbioses. Nature 480:520-524.

866 ;

34 bioRxiv preprint doi: https://doi.org/10.1101/362947; this version posted July 5, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/362947; this version posted July 5, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/362947; this version posted July 5, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/362947; this version posted July 5, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.