At the crossroads of botanical collections and molecular genetics laboratory: testing methods to obtain amplifiable DNA from moss herbarium material

Background. Museum collections, including herbarium specimens, are considered an invaluable source of DNA. They constitute a source of a precious commodity, particularly when it is difficult to obtain living material of rare species, or extant populations occurred only in hard to access geographical territories. It is apparent that herbaria should be directly linked with molecular genetics laboratories making them a quick, open-source for molecular projects. However, herbarium DNA is inherently characterised by high degradation and chemical modifications such as the presence of various secondary compounds. A wide range of DNA molecular techniques dedicated to the preserved plant material has been published so far. However, contrasting with a general interest in the application of molecular analyses in moss biology, no comprehensive assessment of DNA isolation and amplification methods from moss herbarium material, is available. Methods. To assess the feasibility of using DNA from moss herbarium specimens, we have tested and compared the silica column-based method and three variants of CTAB-based DNA extraction protocol. We used herbarium collections of twenty-five moss species collected between 1979 and 2013 and specifically focused on austral polar regions to assess the potential of herbarium as a source of biological material from geographical regions of difficult and restricted access. Results. Here, we present an optimized CTAB-based approach which effectively suppresses inhibitors in the herbarium DNA as was measured by amplification success. In this report, DNA purity and the length of the target genetic region are the fundamental agents which drive the successful PCR reaction. Conversely, the specimen age seems to be less relevant. Moreover, the size distribution of the DNA fragments extracted using Qiagen protocol is shown to be comparable to our original CTAB-based approach. Our modified CTAB-based method provides a high-purity genomic DNA allowing efficient downstream amplification. It is not outcompeted by the column-based method and appears as a method of choice in molecular studies of moss herbarium material.

117 Specimens were in the age range from 12 to 39 years with a median of 19. As far as we could 118 reconstruct, all specimens used in this study were air-dried immediately after collection. 119 Preparation steps 120 Before sample preparation and DNA extractions, the bench top was cleaned with Fugaten Spray 121 (Medilab, Poland) with one-minute incubation. Forceps were sterilised with ethanol and flamed 122 before each specimen handling. All disposable consumables were DNA-free. Sterile filter tips 123 were used for all experimental procedures. During the preparatory step, whenever possible, green 124 gametophyte vegetative shoots were taken. Considering that large amounts of herbarium voucher 125 material are usually not available, we applied to the presented DNA extraction protocols less 126 than 10 mg of dried tissue, typically around 8 mg. Selected fragments of the dried tissue from 127 herbarium voucher specimens were weighted and disrupted in a mixer mill (MM400 -Qiagen 128 TissueLyser II, Retsch, Germany), using one tungsten bead per sample. Samples were ground 129 two times for 30 seconds at 20 Hz and subsequently used for DNA extraction.  Table 1. It should be noted that the number of samples and taxa used in Qiagen 139 test and CTAB extraction tests are different. This is primarily related to the fact, that originally 140 both data were obtained as independent tests performed within different framework. However, 141 the two tests presented provide together a well complementary view on the possible 142 methodological approaches.

143
DNA quality after extractions was evaluated using two criteria: (1) DNA yield, and (2) 144 PCR amplification success (for a detailed description see below). Here, PCR success was 145 selected as a proxy for evaluating DNA content and purity. The presence of primary and 146 secondary chemicals in plant cells are expected to have inhibiting properties on PCR reaction. 147 Hence, inadequate purification of genomic DNA, especially from polyphenols and 148 polysaccharides, could result in a lack of amplification. In this test, we analysed 21 moss species of different age (12-39 years old). Isolation 157 output was tested using PCR amplification of 10 genomic loci of variable length: nuclear 158 ribosomal DNA (5.8SR-ITS2, 18S, adk and phy2), and plastid marker regions (psbAF-trnHR2, 159 atpI-atpH, trnL-trnF, rps4, atpB1-rbcL1, psbB-clpP). Genetic studies using herbarium specimens 160 often highlighted the degraded nature of ancient DNA. Hence, when the above PCR tests were 161 negative, we additionally analysed selected short fragments of the plastid trnS-trnF region.

162
The total DNA concentration was measured in all samples tested using Invitrogen Qubit 163 2.0 Fluorometer (Life Technologies, USA) with the Qubit dsDNA High Sensitivity Assay Kit.
164 CTAB extraction test 165 In the second step of the tests, we compared the DNeasy Plant Mini Kit based isolation with 166 modified CTAB extraction protocols, less costly and potentially yielding a higher amount of 167 isolated DNA. Here, we used 7 different moss species which were collected over a period of 6-168 20 years.

169
To check the quality of the extracted genomic DNA, PCR amplification was performed 170 for genetic regions of the nuclear ribosomal (ITS5bryo-ITSCbryo, ITSDbryo-ITS4bryo), and 171 plastid (trnT-trnF, rps4) DNA regions. Within CTAB extraction protocols, the type of 172 precipitation solutions, i.e. ethanol (C 2 H 5 OH) combined with the sodium chloride (NaCl), and 173 isopropanol (C 3 H 8 O), as well as the proportions of the ethanol/sodium chloride used in relation 174 to total sample volume, were the key determinants to test the effects on downstream molecular 175 applications. We proposed a modified proportion of ethanol/sodium chloride component (here, 176 protocol CTAB-ethanol/NaCl b ), differing from the method used by Healey et al. (2014) (here, 177 protocol CTAB-ethanol/NaCl a ). The modification applied is supposed to increase DNA purity 178 although possibly decreasing DNA concentration. Thus, we have checked whether DNA purity 179 or concentration is more relevant for obtaining PCR-amplifiable DNA from herbarium moss 180 tissue.

181
CTAB-based methods often provide a weakly purified DNA with contaminants having 182 inhibitory effects on downstream enzymatic treatments, thus we attempted to additionally purify 183 our CTAB extracted samples. To this end, we used the Genomic DNA Clean & Concentrator-10 184 kit (Zymo Research, USA) according to the manufacturer's recommendation. With this protocol, 185 we used 10 μL of input genomic DNA. Following purification, DNA was eluted from the matrix 186 with 15 μL of the DNA Elution Buffer preheated to 65°C. Usually, the Zymo-Spin matrix 187 absorbs approximately 5 μL volume of the DNA Elution Buffer and the final output was around 188 10 μL of the purified genomic DNA. In the PCR reactions, we utilized two types of genomic 189 DNA samples, before and after cleaning on the Zymo-Spin matrix, to compare PCR success rate 190 between samples with and without purification. Although silica binding based protocols provide 191 extractions of highly purified DNA samples, Qiagen DNA isolates were additionally purified 192 using Zymo-Spin matrix, in order to allow for a comparison of the final results within this assay.

193
In this test, Agilent 2100 Bioanalyzer (Agilent Technologies, Germany) an automated 194 on-chip electrophoresis system, was used to evaluate the size distribution of the DNA fragments. 195 For this purpose, the Agilent High Sensitivity DNA kit was selected to provide the optimal 277 CTAB extraction test 278 The efficiency of DNA extraction using CTAB-ethanol/NaCl a , CTAB-ethanol/NaCl b , CTAB-279 isopropanol, and Qiagen procedures are compared according to PCR success (Fig. 4). Four DNA 280 fragments (nuclear ITS5bryo-ITSCbryo, ITSDbryo-ITS4bryo, and plastid trnT-trnF, rps4), were 281 amplified before and after additional purification using Zymo-Spin matrix. The comparison of 282 four extraction methods showed differences in the number of successfully amplified target 283 regions. The extraction methods of CTAB-ethanol/NaCl b with a modified proportion of 284 ethanol/NaCl components and Qiagen had the best overall PCR success. However, when we 285 compared CTAB-ethanol/NaCl b , and Qiagen extraction, the method that yielded the most 286 amplifiable DNA before purification was CTAB-ethanol/NaCl b , whereas, after purification 287 treatment, the best performance showed Qiagen extraction. Remaining tested protocols, CTAB-288 ethanol/NaCl a , and CTAB-isopropanol had significantly worse performance, in particular before 289 additional DNA cleaning. 290 We found that additional purification and concentration using Zymo-spin matrix 291 significantly improved the PCR output in all DNA extractions methods tested. The highest PCR 292 success increase was reported in the CTAB-isopropanol method (an increase of 40,6%). For the 293 remaining DNA extractions, the success of PCR amplification after Zymo-spin matrix 294 purification was increased by 28,2% for the Qiagen, 21,9% for the CTAB-ethanol/NaCl a , and 295 9,4% for the CTAB-ethanol/NaCl b procedure. Furthermore, an improvement of the genomic 296 DNA concentration was observed after using Zymo-Spin kit through all extraction methods 297 tested (Table 4). 298 The length of the target region appeared to strongly influence the amplification success. 299 Accordingly, the trnT-trnF locus was the longest (ca.1500 bp), and the most difficult to amplify. 300 Nevertheless, the Qiagen extraction method in combination with Genomic DNA Clean & 301 Concetrator-10 kit has proved to be the most effective for the successful amplification of the 302 abovementioned genetic region. Remaining loci (rps4 ca. 600 bp; ITSDbryo-ITS4bryo ca. 450 303 bp; ITS5bryo-ITSCbryo ca. 380 bp) were comparable with respect to PCR success with a small 304 advantage for nuclear ITS regions. The electropherograms obtained by automatic fragment sizing within all extraction 306 methods showed a broad distribution of bands which has indicated that genomic DNA was 307 highly fragmented (Fig. 5, 6, 7). The average sizing of DNA isolate did not vary significantly 308 across methods and is ranged from ca. 400 to 500 bp. Despite the fragment size distribution in all 309 electropherograms remains comparable, the most similar shape of the genomic DNA profiles can 310 be observed among CTAB-ethanol/NaCl b , and Qiagen extraction method, which could be also 311 reflected in comparable PCR success rate within these two assays. 312 Discussion 313 DNA purity rather than concentration as a key factor 314 Our comparisons highlight the key importance of DNA purity after isolation from herbarium 315 sample, rather than DNA quantity, for successful PCR. Thus, particular attention should be paid 316 to separating DNA from naturally occurring plant cell contaminants, rather than strenuous efforts 317 to obtain high DNA quantity. Among protocols tested in our study, the CTAB-based DNA 318 extraction method provides such a solution, making it a superior choice relative to silica gel 319 column-based commercial kits for DNA extraction.

320
In the CTAB extraction protocols applied, the performance of the ethanol/NaCl mixture 321 has proven to be crucial for obtaining pure DNA. More precisely, decreasing the volume of the 322 ethanol/NaCl solution to the total volume of extracted sample, i.e. CTAB-ethanol/NaCl b 323 protocol, caused a significant increase of PCR success as opposed to the original proportions 324 applied by Healey et al. (2014). In turn, the proportion of the ethanol/NaCl ingredients has 325 remained unchanged in both variants of the CTAB-ethanol/NaCl based protocols. Likely, 326 reducing the volume of ethanol in our original CTAB protocol may have resulted in a reduced 327 amount of precipitated genomic DNA but in the same time in a significantly lowered 328 concentration of the co-precipitated PCR inhibitors, such as polysaccharides, phenols, and other 329 organic compounds. In general, the addition of a high salt buffer (here, NaCl) could increase 330 genomic DNA purity by a boost of polysaccharides solubility in ethanol, allowing their removal 331 when DNA is pelleted under centrifugation step.

332
In our CTAB-ethanol/NaCl b extraction protocol, the measured concentration values 333 were the lowest across the tested protocols. It is assumed that DNA yield is good enough to 334 obtain acceptable PCR products if ranged between 6.0-100 ng/μl (Do & Drábková, 2017). In 335 our study, we obtained successful PCR reactions from samples with concentration values lower 336 than 1 ng/μl. Nevertheless, the best performing DNA protocols should be aimed to obtain high 337 purity combined with high DNA yield, which is particularly important in respect of high-338 throughput sequencing methods. In CTAB extraction tests, the Genomic DNA Clean & 339 Concentrator-10 kit was additionally applied to all prepared DNA extracts. This kit is expected 340 to provide ultra-pure, high-yield genomic DNA. Accordingly, DNA concentration and 341 percentage of the successfully amplified samples has risen significantly after Zymo-Spin 342 cleaning. The main increase in PCR success was observed in the case of the most potentially 343 contaminated extracts, which derived in this study from CTAB-isopropanol, and CTAB-344 ethanol/NaCl a protocols.

345
Our results are congruent with several studies which concluded that DNA purity is more 346 important for amplification success than DNA yield (e.g. Höss (Klavina et al., 2012, Klavina, 2015. Interestingly, Sabovljević, Bijelović &Dragoljub (2001) 351 described bryophytes as "remarkable reservoir" of natural products and/or secondary compounds 352 such as terpenoids, phenols, glycosides, fatty acids and rare aromatic ingredients. This is also 353 confirmed by Soni & Kumar (2009) who underlined that extraction of DNA from bryophytes 354 could be very difficult due to the presence of secondary compounds inhibiting downstream 355 applications.
356 Effects of target amplicon size and specimen age on successful PCR 357 In our tests of extraction protocols, the length of the selected target regions was correlated with 358 the PCR amplification success. This appears an obvious tendency for highly degraded genomic 359 material and our results are in agreement with Särkinen et al. (2012) and Do & Drábková (2017) 360 who indicated that the most easily amplifiable DNA fragments from herbarium material were 361 those below 500 bp. In our report, based on Qiagen extraction test, the best-performing locus are 362 psbA-trnH (ca. 250 bp), trnL-trnF (ca. 450 bp), and 5.8SR-ITS2 (ca. 450 bp). The most 363 pronounced decrease in PCR success was observed in amplicons around 1000 bp (18S, adk, 364 phy2, psbB-clpP, including trnT-trnF region from CTAB extraction test), and was more evident 365 in nuclear regions. However, it was possible in some cases to amplify target genomic regions of 366 up to 1500 bp. Thus, even though amplification success declines with target amplicon size for 367 herbarium-based isolates, some collections may provide DNA quality high enough to provide 368 adequate data for molecular analyses. Since short fragments prevail in herbarium DNA, it is 369 expected that PCR of smaller regions has a higher success rate. It is, admittedly, under 370 abovementioned observations, but on the other hand, an attempt to amplify short, barcode 371 regions using samples which failed previously in PCR reaction (within Qiagen extraction test) 372 was still unsuccessful. In a case like this, DNA un-purity may play a more significant role in 373 inhibiting PCR reaction than DNA fragmentation. Possibly, in this particular case, PCR 374 optimisation, using both fresh and herbarium material may result in improvement of successful 375 amplification.

376
Although in the Qiagen extraction test we did not have an equal share of specimens for 377 a given age, we found no correlation for the age of specimen and PCR success in the age range 378 examined. Successful amplification rate was comparable for the oldest (39 years old) and 379 youngest specimens (12) and rather other factors affecting the collection history seem decisive. 380 Previous studies have also shown that age of herbarium samples had no significant effect on 381 PCR success, pointing out the importance of locus types to be amplified rather than the age 382 (Särkinen et al., 2012; Do & Drábková, 2017). Summarizing, the age of moss specimen should 383 not deter bryologists from their usage in molecular research although certainly at the sample age 384 much exceeding those tested here the impact of DNA fragmentation may gradually appear 385 preponderant. 386 387 The rate of DNA degradation in moss herbarium material -CTAB test 388 Our extraction tests also took into consideration the level of DNA fragmentation in moss 389 herbarium samples. The overall strand breaks of DNA retrieved from the selected moss 390 herbarium specimens was high and only slightly varied between applied extraction methods, and

Add extraction buffer
Add to the each sample 1mL of preheated to 65 0 C 2x CTAB buffer containing β-mercapthoethanol.
Centrifuge at 5,000 rcf for 5 min to pellet and remove un-lysed leaf tissue. Transfer the extract to a new 2 mL tubes.

Protein extraction and RNAse treatment
Add an equal volume of chloroform : isoamyl alcohol (24:1) to the extract and mix gently. Extract for 30 min by rocking on orbital shaker.
Centrifuge at 13,000 rpm for 10 min.
Transfer the upper phase (containing DNA) to a new 2 mL tubes. Take care to avoid the aqueous/organic layer interface.
Add 1 μL of RNase A solution (10 mg/mL) per 100 μL DNA solution and incubate at 37 0 C for 15 min with periodic, gentle mixing.
Repeat the chloroform : isoamyl alcohol extraction to clear the aqueous phase.

Precipitation
Add X volume of 5M NaCl to the transferred aqueous phase and mix gently by inversion. Then add Y volume(s) of prechilled (-20 0 C) 95% ethanol and mix gently by inversion. Incubate at -20 0 C for 60 min. Note: do not leave the sample at -20 0 C for more than 60 min as both the CTAB and NaCl can precipitate from solution, preventing DNA isolation.