Red-Brown Pigmentation of Acidipropionibacterium jensenii Is Tied to Haemolytic Activity and cyl-Like Gene Cluster

The novel Acidipropionibacterium genus encompasses species of industrial importance but also those associated with food spoilage. In particular, Acidipropionibacterium acidipropionici, Acidipropionibacterium thoenii, and Acidipropionibacterium jensenii play an important role in food fermentation, as biopreservatives, or as potential probiotics. Notably, A. jensenii and A. thoenii can cause brown spot defects in Swiss-type cheeses, which have been tied to the rhamnolipid pigment granadaene. In the pathogenic bacterium Streptococcus agalactiae, production of granadaene depends on the presence of a cyl gene cluster, an important virulence factor linked with haemolytic activity. Here, we show that the production of granadaene in pigmented Acidipropionibacterium, including A. jensenii, A. thoenii, and Acidipropionibacterium virtanenii, is tied to haemolytic activity and the presence of a cyl-like gene cluster. Furthermore, we propose a PCR-based test, which allows pinpointing acidipropionibacteria with the cyl-like gene cluster. Finally, we present the first two whole genome sequence analyses of the A. jensenii strains as well as testing phenotypic characteristics important for industrial applications. In conclusion, the present study sheds light on potential risks associated with the presence of pigmented Acidipropionibacterium strains in food fermentation. In addition, the results presented here provide ground for development of a quick and simple diagnostic test instrumental in avoiding potential negative effects of Acidipropionibacterium strains with haemolytic activity on food quality.

Thus, there is some controversy in the linking of pigmentation and haemolysis in propionibacteria species and it is not clear whether the phenomenon is restricted to members of the genus Acidipropionibacterium.
In the current work, PCR detection with cylG as a target shows diagnostic potential in detection of haemolytic and pigmented strains among several Acidipropionbacterium species. In addition, we report whole genome sequences of two novel strains of A. jensenii isolated in Finland from barley: JS279 and JS280, which display a distinct pigmented and non-pigmented phenotype, respectively. Here, the sequenced genomes were compared to those defined for the type strain A. jensenii DSM 20535 and other acidipropionibacteria. We also complement genomic findings related to B 12 vitamin synthesis, genome integrity, haemolytic activity and carbohydrate fermentation profiles with necessary bioinformatic and phenotypic analyses.

Growth Conditions
Bacterial strains used in this work are listed in Table 1. A. jensenii JS279 (VTT E-113203) and A. jensenii JS280 (VTT E-113204) were previously isolated from barley samples (data not shown) and deposited in the VTT Culture Collection. The strains were maintained in 15% glycerol stocks at −80 • C and routinely grown on yeast extract-lactate medium (YEL) [34], first for four days on solid medium supplemented with 1.5% agar (Difco, Becton, Dickinson and Company, Sparks, MD, USA) at 30 • C under an anaerobic atmosphere (Anaerocult, Merck, Darmstadt, Germany) and subsequently in 10 mL of liquid medium in 15-mL Falcon tubes for 3 days at 30 • C, under a microaerobic atmosphere.

Isolation of DNA, Genome Sequencing, and Assembly
Genomic DNA was isolated as described previously [36] using Illustra MiniSpin kit (GE Healthcare, Chicago, IL, USA) for JS279 and MagAttract HMW DNA Kit (Qiagen, Germantown, MD, USA) for JS280 samples. The sequencing was performed on PacBio RSII platform with the P5/C3 and P4/C2 chemistry for the strain JS279 and with the P5/C3 chemistry for the strain JS280. The genomes were assembled with Hierarchical Genome Assembly Process version 3 algorithm, with resulting 215X and 380X genome coverage for strains JS279 and JS280, respectively. Base modifications and motifs were detected using the RS Modification and Motif analysis protocol (SMRT Analysis package v.2.3.0). The whole genome sequences of the strains JS279 and JS280 together with the modification profiles were submitted to NCBI for annotation with the Prokaryotic Genome Annotation Pipeline (PGAP) and are available under accession numbers CP025571 and CP025570, respectively.

Bioinformatics Analyses and Classification of the Strains
The average nucleotide identity (ANIb) values were estimated by BLAST algorithm with the aid of JSpecies web software [37]. For the genome-wide descriptions, the genomes of the strains JS279 and JS280 together with the whole genome of A. jensenii NCTC13652 (accession number LR134473.1) (equivalent of the type strain A. jensenii DSM 20535) downloaded from NCBI were aligned with Progressive Mauve [38], re-annotated with PROKKA [39] and subjected to comparative genomics analysis with Roary [40].

Detection of Haemolysis and Pigmentation
The haemolytic capability of the strains was assessed by growth on tryptic-soy agar (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) supplemented with 5% sheep blood (Labema, Helsinki, Finland). A fresh colony from a YEL agar plate was streaked on blood agar and incubated for one week at 30 • C under an anaerobic atmosphere (Anaerocult, Merck, Darmstadt, Germany) after which the plates were examined for haemolysis. The presence of clearing around the bacterial growth was interpreted as β-haemolysis. For detection of pigmentation, fresh colonies were streaked on YEL agar plates and incubated as above followed by visual inspection. Staphylococcus aureus 8325-4 [41] was used as a positive control of haemolytic activity. The assay was repeated at least three times for each tested strain.

PCR Amplification of the cylG Gene Fragment
Oligonucleotide primers fw 5 CACCGCGATCGAGATGGC 3 and rev 5 GGAACATGTCGGTGTC GACGA 3 were used for amplification of an internal fragment of the cylG gene. The primer annealing to a conserved region in the cylG gene, were designed on the basis of BLASTn alignment between the strains A. jensenii LMGT 2818 (FJ617193.1), A. jensenii JS279 (CP025571), A. thoenii DSM 20276 T (NZ_KE384018.1), and A. virtanenii DSM 106790 T (CP025198). For crude DNA template preparation, YEL-grown cells were harvested by centrifugation at 3220× g for 10 min and lysed by bead beating with glass beads in TE-buffer (10 mM Tris-HCl, 0.1mM EDTA, pH 8.0) with a FastPrep ® -24 Homogenizer (MP Biomedicals, Irvine, CA, USA) in three cycles of 30 s. Cell debris and glass beads were removed by centrifugation at 12,000× g for 5 min and an aliquot of the supernatant was used as a template in PCR. The amplification was carried out using an MJ Mini ® Personal Thermal Cycler (Bio-Rad, Hercules, CA, USA) in a reaction volume of 25 µL in Phusion Master Mix HF buffer (Thermo Fisher Scientific, Waltham, MA, USA) including 12.5 pM of fw and rev primers and 1 µL of a crude template DNA preparation. The following thermocycling conditions were used: 98 • C for 2 min, in 30 cycles of 98 • C for 10 s, 70 • C for 15 s, and 72 • C for 15 s. A 5-µL aliquot of the PCR reaction was mixed with 2 µL of loading buffer and run on a 1.0 % agarose gel at 60 V in 1× TBE buffer (89 mM Tris, 89 mM boric acid and 2 mM EDTA) and visualized in an AlphaImager (ProteinSimple, San Jose, CA, USA) device under UV light setting. The sizes of PCR products were determined by comparing against a phage Lambda DNA digested with PstI. The expected size of the cylG specific amplicon was 518 bp. Confirmation of the identity of amplified products was performed by Sanger sequencing the amplicons obtained from A. thoenii DSM 20276 using BigDye Terminator v3.1 cycle sequencing with the primers fw and rev. Sequences were generated on an Applied Biosystems 3730XL high-throughput capillary sequencer (Thermo Fisher Scientific, Waltham, MA, USA). The resulting sequences were subjected to a BLASTn search of the NCBI database with default parameters [42]. The genomes of the strains JS279 and JS280 were further subjected to the search of clustered regularly interspaced short palindromic repeats (CRISPR) with the aid of CRISPRFinder [43]. For the search of genomic islands IslandViewer IV [44] and putative prophages Prophinder [45] and Phaster [46] were used. Methylation profiles and Restriction-Modification systems were analysed at REBASE [47].

Phenotypic Characterisation
The carbohydrate fermentation profiles of the strains JS279 and JS280 were determined with the aid of API CHL50 test (Biomerieux, Marcy I'Etoile, France) as described previously [48]. The ability to grow in modified YEL with pH values adjusted to 4.0, 4.5, 5.0 and 9.0, in the presence of salt concentrations adjusted to 256.7 mM (1.5%), 513.3 mM (3%), and 1.1 M (6.5%) NaCl as well as growth in unmodified YEL at 12 • C and 42 • C. These experiments were performed on 96-well microtiter plates (353077, Falcon, Corning incorporated, Durham, NC, USA) as previously described [48].

Classification of Strains JS279 and JS280 as A. jensenii
The results of the ANIb analysis ( Figure 1) clearly show that the newly sequenced strains JS279 and JS280 belong to the species A. jensenii with the ANIb % identity values ≥97% [37]. The summary details of their genomes are presented in Table 2, together with the details of the previously sequenced type strains of other species belonging to the genus Acidipropionibacterium: A. jensenii DSM 20535 (NCTC13652), A. acidipropionici DSM 4900 (CGMCC1.2230), A. thoenii DSM 20276, and A. virtanenii DSM 106790 (JS278). In addition, the type strain of the closely related Propionibacterium freudenreichii DSM 20271 (JS16) and the most studied strain of Cutibacterium acnes DSM 16379 (KPA171202) were included. Genomes of the Acidipropionibacterium species are generally larger than those of Cutibacterium and Propionibacterium representatives, however, A. jensenii genomes are smaller than those of other acidipropionibacteria. Interestingly, the type strain A. jensenii DSM 20535 has an approximately 5% larger genome than the other two strains and it possesses an additional ribosomal RNA operon.

Detection of Defence Mechanisms and Mobile Genetic Elements
The genomes of the strains JS279 and JS280 were further subjected to the search of clustered regularly interspaced short palindromic repeats (CRISPR) with the aid of CRISPRFinder [43]. For the search of genomic islands IslandViewer IV [44] and putative prophages Prophinder [45] and Phaster [46] were used. Methylation profiles and Restriction-Modification systems were analysed at REBASE [47].

Phenotypic Characterisation
The carbohydrate fermentation profiles of the strains JS279 and JS280 were determined with the aid of API CHL50 test (Biomerieux, Marcy I'Etoile, France) as described previously [48]. The ability to grow in modified YEL with pH values adjusted to 4.0, 4.5, 5.0 and 9.0, in the presence of salt concentrations adjusted to 256.7 mM (1.5%), 513.3 mM (3%), and 1.1 M (6.5%) NaCl as well as growth in unmodified YEL at 12 °C and 42 °C . These experiments were performed on 96-well microtiter plates (353077, Falcon, Corning incorporated, Durham, NC, USA) as previously described [48].

Classification of Strains JS279 and JS280 as A. jensenii
The results of the ANIb analysis ( Figure 1) clearly show that the newly sequenced strains JS279 and JS280 belong to the species A. jensenii with the ANIb % identity values ≥97% [37]. The summary details of their genomes are presented in Table 2, together with the details of the previously sequenced type strains of other species belonging to the genus Acidipropionibacterium: A. jensenii DSM 20535 (NCTC13652), A. acidipropionici DSM 4900 (CGMCC1.2230), A. thoenii DSM 20276, and A. virtanenii DSM 106790 (JS278). In addition, the type strain of the closely related Propionibacterium freudenreichii DSM 20271 (JS16) and the most studied strain of Cutibacterium acnes DSM 16379 (KPA171202) were included. Genomes of the Acidipropionibacterium species are generally larger than those of Cutibacterium and Propionibacterium representatives, however, A. jensenii genomes are smaller than those of other acidipropionibacteria. Interestingly, the type strain A. jensenii DSM 20535 has an approximately 5% larger genome than the other two strains and it possesses an additional ribosomal RNA operon.    [50] * N/A-not available; ** The rRNA numbers are not final as this genome is at the draft level; T indicates type strain of the species.

Haemolytic Activity, Pigmentation and cylG Gene Detection by PCR
In S. agalactiae, the pigment with haemolytic activity is encoded by the cyl operon consisting of 10 genes: cylI, cylF [51], cylE, cylB, cylA, cylZ, acpC, cylG, cylD, and cylX [52]. Later, two additional genes cylJ and cylK were also identified [53]. While the S. agalactiae mutants deficient in either the acpC, cylZ, cylA, cylB, or cylE gene lost both the pigmentation and the haemolytic activity completely [52], the mutations in cylF, cylI, cylJ, and cylK ( [52,53] led to milder activities. BLASTp search of the JS279 theoretical proteome against the proteins coded by the S. agalactiae cyl cluster gave very low scores, with an average sequence identity of~30%. However, some of the hits were organised in a cluster and subsequent search against Conserved Domains Database [54] revealed that despite low sequence conservation the functional protein domains belonged to the same families (Figure 2a). BLASTn analyses of available propionibacteria genomes revealed the presence of a putative cyl cluster in at least one strain of A. thoenii (DSM 20276) and A. virtanenii (JS278) (Figure 2b) but not in any of the A. acidipropionici strains. The presence of a cyl gene cluster was previously reported for a genome assembled from a metagenomic sample, Propionibacterium sp. 5 U 42AFAA [26]. The sample is most likely a strain of Cutibacterium acnes subsp. acnes based on an ANIb value of 99.7% when compared to Cutibacterium acnes type strain ATCC 6919. However, as it is impossible to determine the phenotype of this isolate, it was not pursued further. The subsequent alignment of the putative cyl cluster from all acidipropionibacteria showed a high degree of conservation within the cylG gene region, which was selected for primer design and subsequent detection of the cyl cluster by PCR. Seven available A. jensenii strains, two A. thoenii strains, and the type strains of the recently described species A. virtanenii and that of P. freudenreichii as a negative control (see Table 1) were used. In addition, the strains were tested for haemolytic activity and pigment formation on sheep blood plates and YEL plates, respectively.
Clearance around bacterial growth on blood plates indicative of β-haemolytic activity was observed for five strains including two out of five A. jensenii strains (JS279 and DSM29275), the A. virtanenii strain, and both A. thoenii strains (Figure 3a). Four A. jensenii strains and the P. freudenreichii The subsequent alignment of the putative cyl cluster from all acidipropionibacteria showed a high degree of conservation within the cylG gene region, which was selected for primer design and subsequent detection of the cyl cluster by PCR. Seven available A. jensenii strains, two A. thoenii strains, and the type strains of the recently described species A. virtanenii and that of P. freudenreichii as a negative control (see Table 1) were used. In addition, the strains were tested for haemolytic activity and pigment formation on sheep blood plates and YEL plates, respectively.
Clearance around bacterial growth on blood plates indicative of β-haemolytic activity was observed for five strains including two out of five A. jensenii strains (JS279 and DSM29275), the A. virtanenii strain, and both A. thoenii strains (Figure 3a). Four A. jensenii strains and the P. freudenreichii strain were judged non-haemolytic on sheep blood plates under the conditions used (Figure 3a). The strains showing β-haemolytic activity also produced red/orange pigment on YEL-plates whereas no pigmentation was observed in non-haemolytic strains (Figure 3b), which is in accordance with previous reports linking haemolytic activity and red pigmentation in P. jensenii and P. thoenii [20]. Our primers produced amplicons of the expected size (518 bp) from each of the five β-haemolytic and pigmented Acidipropionibacterium strains representing three species but not from any of the non-haemolytic and non-pigmented strains (Figure 3c). Sanger sequencing of the PCR product from A. thoenii HAMBI 247 and subsequent BLASTn analysis revealed a best hit (97% identity over 499 bp sequence stretch) with cylG in the putative granadaene gene cluster of Propionibacterium jensenii LMGT2818 (nucleotides 6177-6675 in GenBank entry FJ617193) confirming the specificity of the primers. Thus, PCR detection with cylG as a target shows diagnostic potential in detection of haemolytic and pigmented strains among several Acidipropionibacterium species. In the present study, we used all the A. jensenii, A. thoenii, and A. virtanenii strains accessible in DSMZ and HAMBI culture collections. A more thorough assessment of molecular detection will become feasible when more strains are available for research.  About 1-5% of human S. agalactiae isolates are non-haemolytic, which is often associated with the loss of activity due to the presence of an insertion sequence in one of the cyl genes [25]. Conversely, in A. jensenii strains for which genomic sequence is available and no haemolytic activity can be observed, the granadaene gene cluster is absent altogether. This could indicate that in this species haemolysis is a result of acquisition of genes through horizontal gene transfer, which affected only a few of the strains, rather than being a core feature of the species. On the other hand, the widespread occurrence of pigmentation among A. thoenii strains could suggest that production of granadaene is About 1-5% of human S. agalactiae isolates are non-haemolytic, which is often associated with the loss of activity due to the presence of an insertion sequence in one of the cyl genes [25]. Conversely, in A. jensenii strains for which genomic sequence is available and no haemolytic activity can be observed, the granadaene gene cluster is absent altogether. This could indicate that in this species haemolysis is a result of acquisition of genes through horizontal gene transfer, which affected only a few of the strains, rather than being a core feature of the species. On the other hand, the widespread occurrence of pigmentation among A. thoenii strains could suggest that production of granadaene is intrinsic to this species. However, a larger number of strains is necessary to confirm this notion.

Biosynthesis of Vitamin B 12
Propionibacteria, particularly P. freudenreichii, are known for their ability to produce active vitamin B 12 . However, we previously reported that the members of the genus Acidipropionibacterium lack the bluB part of the fusion gene bluB/cobT2 necessary for the production of 5,6-dimethylbenzimidazole (DMBI), the lower ligand of the active form of vitamin B 12 [35]. Also, the presence of aspartate in the active site of the product of cobT gene is decisive for its function. CobT enzyme is responsible for activation of the lower ligand prior to its attachment. The charged amino acid in the active site of acidipropionibacteria indicates preference for activation of lower ligand bases less hydrophobic than DMBI, such as adenine, which results in a preferential production of pseudovitamin B 12 [55]. A similar organisation of the B 12 biosynthetic gene clusters can be observed in A. jensenii (Figure 4a) as in A. virtanenii [35] and A. acidipropionici [56], with the exception of the gene clusters I and III being co-located. Accordingly, in our previous report we determined the preference for production of pseudovitamin B 12 by the strains JS279 and JS280 and that only a small amount of active vitamin is produced if DMBI is provided [57]. in A. jensenii (Figure 4a) as in A. virtanenii [35] and A. acidipropionici [56], with the exception of the gene clusters I and III being co-located. Accordingly, in our previous report we determined the preference for production of pseudovitamin B12 by the strains JS279 and JS280 and that only a small amount of active vitamin is produced if DMBI is provided [57].

Mobile Elements and Defence Mechanisms
Analysis with CRISPRfinder identified 14 potential CRISPR arrays, two of which are associated with a complete CRISPR-Cas system (Figure 4b) in each JS279 and JS280 strains. The arrays PossibleCRISPR1 and CRISPR2 of the strain JS279 share a direct repeat (DR) consensus sequence CGGATTACCCCCGCTCGCGCGGGGACGAC. CRISPR2 consists of 76 spacers located between 280463-285125 and is associated with a complete CRISPR-Cas type IE system (Figure 4a). CRISPR1 consists of two spacers and is located directly upstream of the CRISPR2 array (279029-279180). CRISPR1 and CRISPR2 are separated by three genes coding for hypothetical proteins, each with similarity to transposases. Both the organisation of the Cas genes and the DR consensus sequence are highly similar to the CRISPR-Cas system type IE identified in P. freudenreichii [48], while DR also is nearly identical to the one associated with the incomplete CRISPR-Cas system type IU identified in A. virtanenii (GGGCTCACCCCCGCATACGCGGGGCGGAT) [35]. The first two arrays of the strain JS280, CRISPR1 and CRISPR2 consist of 55 and 15 spacers, respectively, and share the DR sequence GTCGCTCCTCTTTCCAGAGGAGCCCTTCGTTGAGGC. The arrays are separated by a gene coding for ISL3 family transposase (C0Z10_01425) and are located directly upstream of a complete CRISPR-Cas system type IU, which shares high similarity with the corresponding system identified in P. freudenreichii [48].
The remaining 12 arrays in each strain consist of one to two spacers each and DR sequences 23-43 nt long. While they do not appear to be associated with CRISPR-Cas systems, the presence of DR and their mostly intergenic locations may imply some other, yet unknown role. BLAST search against the spacers did not return hits to any known bacteriophages or mobile elements. Both CRISPR-Cas systems type IE and IU were previously detected in other propionibacteria [35,48], with a similar organisation, but with different DR sequences. It is worth noting that the novel IU system in A. jensenii and A. virtanenii differs from the one reported in P. freudenreichii by the presence of a hypothetical protein, which contains a predicted DNA-binding domain.
Four complete Restriction Modification (RM) systems were identified in both the strain JS279 ( Figure S1) and JS280 ( Figure S2). The PacBio-aided detection of methylation allowed identification of four 6mA methylation motifs: GAAYNNNNNTGG, CGCATC, GCAGCC and GGAABCG in the strain JS280, the first of which was successfully tied to the activity of the Type I RM system (COZ10_10535-COZ10_10545). No definitive motifs were identified in strain JS279. However, it is worth mentioning that the predicted system (C0Z11_09855, C0Z11_09860, C0Z11_09875) constitutes a part of a complete BREX [58] system (C0Z11_09845-C0Z11_09890).
Nineteen genomic islands were identified for each strain JS279 and JS280, while no prophages were identified in either. Further details of the bioinformatics analyses can be seen in Table S1 (strain JS279) and Table S2 (strain JS280).
Comparison of the genome organisation between the three complete genomes of A. jensenii revealed the presence of multiple inversions ( Figure S3). A similar phenomenon was reported in P. freudenreichii [48], but not in C. acnes [59]. Results of comparative genomics analyses between the A. jensenii type strain and the two newly sequenced strains can be seen in Tables S1 and S2 as well as Figures S3 and S4.

Phenotypes
Aside from the obvious difference in pigmentation associated with production of granadaene by strain JS279, strains JS279 and JS280 display similar phenotypes ( Table 3). The exceptions include a weak growth of strain JS280 already at pH 4.5, while both JS279 and JS280 grew well at pH 5.0. In addition, the strain JS280 was able to grow at 42 • C, while the maximum temperature for growth recorded for strain JS279 was 37 • C. The only difference concerning fermentation profiles was the ability of strain JS280 to utilise salicin as carbon source. More pronounced differences in fermentation profiles can be observed when both strains of cereal origin are compared against the type strain A. jensenii DSM 20535 isolated from buttermilk. Both cereal strains are able to grow on methyl-α-D-glucopyranoside, starch, and xylitol, while the dairy strain was able to use D-cellobiose and gentiobiose. The ability to ferment raffinose by the strains of A. jensenii and A. virtanenii analysed here could be used in the fermentation of pulses. Pulses are considered valuable source of protein where the presence of raffinose is considered one of the antinutritional factors, and currently efforts are made to eliminate it through the use of microbial fermentation [60].   It is worth noting that, unlike the type strain A. virtanenii JS278, which produced granadaene only when grown on lactate or turanose, strain JS279 invariably produced granadaene also on galactose, inositol, maltose, saccharose, trehalose, melezitose, xylitol, and L-arabitol (data not shown). However, the implications of this difference are currently unknown.

Conclusions
Due to their beneficial properties, the dairy propionibacteria including strains of A. jensenii and A. thoenii have gained interest as secondary and/or protective starter cultures in various food fermentations [7,[63][64][65] and even as probiotic candidates [66]. Linking the presence of the cyl operon to haemolytic activity and the production of red-pigment allows identification of haemolytic/pigmented strains by whole genome sequencing. Furthermore, it makes it possible to develop specific PCR methodology based on cylG gene specific primers. Lack of haemolysis/pigmentation-encoding genes can be used as one of the selection criteria for new propionibacteria strains to be used in food production and/or in search of new probiotics. This will allow to avoid potential negative effects on the food quality or on the health of consumers. While for the moment there are no reports indicating negative health effects of the consumption of cheeses containing pigmented propionibacteria, the presence of haemolytic pigment associated with virulence of the pathogenic S. agalactiae raises safety concerns. Until the safety of pigmented propionibacteria is confirmed, the risk of potential adverse effects on consumer health could be avoided by detecting haemolytic strains in raw milk. The strains can then be eliminated by heat treatment from batches intended for production of cheese varieties allowing growth of propionibacteria.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-2607/7/11/512/s1, Figure S1: REBASE schematics of the RM systems detected in strain JS279, Figure S2: REBASE schematics of the RM systems detected in strain JS280, Figure S3: Progressive mauve genome alignments. Figure S4: Venn diagram of orthologous groups assigned by ROARY analysis. Table S1: Cumulative results of bioinformatics analyses performed on strain JS279, Table S2: Cumulative results of bioinformatics analyses performed on strain JS280.
Funding: This research was funded by the Academy of Finland, grant numbers 257333,307856 and 325784 and the Finnish Cultural Foundation grant number 00190039.