Kocurin
Kocurin
Kocurin
Abstract
The antibiotically bioactive thiopeptide compound kocurin was identified in extracts from a newly isolated Kocuria rosea
strain. The axenic strain was retrieved from a soil sample of the intertidal area at the Paracas National Park, Peru. The
genetic basis of this promising natural product with activity against methicillin-resistant Staphylococcus aureus (MRSA)
strains was revealed by comparative genome analysis of this new isolate and other reported thiopeptide producer strains.
The functionality of the predicted gene locus was experimentally proven by heterologous expression in Streptomyces
coelicolor M1146. Expression of the gene cluster under the control of a constitutive promoter enabled the transgenic strain to
produce kocurin in selected media. The kocurin biosynthetic gene cluster comprises nine open reading frames and spans
around 12 kbp of the genome.
The increasing antibiotic resistance of pathogenic bacteria MS/MS resulted in the identification of the known thiopep-
has become a serious threat towards human health [1]. tide compound kocurin as the molecule responsible for the
Hence, new treatment options, e.g. novel antibiotics, are antibiotic activity (Fig. S1, Table S1, available in the online
urgently needed [2]. To refill the antibiotic development Supplementary Material).
pipeline, bioprospecting projects have been revived, espe-
Kocurin is regarded as a compound with promising proper-
cially in the academic environment, all around the globe.
ties, which could potentially lead to the development of a
Natural products have always been the most rewarding
novel antibiotic [3]. The substance was first described as com-
resource for novel compounds with intriguing bioactivities.
pound PM181104, and a patent for the treatment of
Even though the low-hanging fruits seem to have been har-
Clostridium difficile-associated infections was claimed
vested, since previously known compounds have been re-
(WO2014102570 A1). In 2013 the group of Genilloud and
isolated in activity-based approaches, nature still has an
Reyes published an alternative structure to the one shown in
untapped reservoir of natural products harbouring interest-
the patent and named the compound kocurin. They investi-
ing activities from a pharmaceutical point of view [3].
gated marine actinomycetes of the family Micrococcaceae,
In the present report, the search for new antibiotically active whereby the strains had been isolated from sponges. Structure
compounds from marine bacteria, isolated from the Peru- elucidation was achieved using a combination of spectroscopic
vian coastline, led to the isolation of a strain whose ethyl and chemical methods, which included extensive NMR analy-
acetate crude extract showed strong antibiotic activity sis, HRMS and MS/MS fragmentation, as well as chemical
against Gram-positive test strains, i.e. Arthrobacter cristallo- degradation and Marfey’s analysis of the resulting amino acid
poietes, Bacillus megaterium and MRSA. 16S rDNA identifi- residues [4]. Thiazolyl peptide antibiotics are a class of natural
cation revealed the isolate to be 99 % identical to Kocuria products showing in vitro activity against Gram-positive bac-
rosea DSM 20447 (accession number: NR_044871.1). Ana- teria. Meanwhile, about 100 molecules of this type were
lysing the fractionated crude extract by subsequent HPLC- reported [5]. Based on the promising in vitro activity of
kocurin, in vivo tests were performed. Using a BALB/c mouse Sequencing revealed 100 % identity at the nucleotide level
septicaemia model, kocurin showed promising activity. It dis- with Kocuria sp. UCD-OTCP and 91 % with K. flava
played 100 % effective dose (ED100) values of 2.5 and 5.0 mg HO-9041. As a next step, K. rosea s17 was genome
kg 1 body weight against MRSA and of 10.0 mg kg 1 against sequenced using Illumina sequencing and assembled using
vancomycin-resistant enterococci [6]. Spades. Performing antiSMASH analysis on the assembled
sequence data, a contig was identified harbouring the com-
In this study, it was projected to identify the genetic basis of
plete gene locus linked to kocurin production. The overall
kocurin biosynthesis, i.e. to identify the corresponding bio-
identity at DNA level to sequences available in databases is
synthetic gene cluster (BGC). Genomic sequence data from
99.6 % and the organization of the cluster is identical.
several Kocuria strains were analysed using antiSMASH [7].
In general, members of the family Micrococcaceae possess a To prove the predicted link between the BGC and the
small genome (2.7–4 Mbp) compared with other families of metabolite experimentally, heterologous expression was
actinobacteria such as Streptomycetaceae [8]. The overall envisaged. Therefore, the boundaries of the cluster were first
number of BGCs observed after analysis of the eight avail- defined. A closer examination of the predicted function of
able genomes of Kocuria species ranged between 2 and 9. the proteins encoded by the open reading frames neigh-
Most of the predicted BGCs were type III polyketide syn- bouring the propeptide sequence enabled the identification
thases (PKSs), bacteriocins, siderophores and thiopeptides, of eight genes downstream in the same orientation that
with only a very small number of modular PKSs and non- have a predicted role in kocurin biosynthesis (Fig. 1). This
ribosomal peptide synthetases (NRPSs). region spans 12.1 kbp and the genes were named kocA-I
(from kocurin). Upstream of the propeptide encoding kocA,
Formerly it was speculated that the biosynthesis of kocurin
a gene coding for a HEAT repeat containing protein was
is encoded by a NRPS/PKS system [3]. However, this was
identified in the opposite orientation, but its function was
proven wrong following investigation of the biosynthesis of
considered not to be involved in kocurin biosynthesis.
thiomuracins in Nonomuraea species [9], and a closer look
Downstream of the ABC transporter gene kocI, a group of
at the predicted BGCs in our strain revealed that no hybrid
genes involved in general stress response was detected,
NRPS/PKS BGC seems to be associated with kocurin pro-
forming the other border of the BGC (Fig. 1).
duction. Instead, a predicted ribosomally biosynthesized
and post-translational modified peptide (RiPP) gene cluster Two different Streptomyces strains were chosen for heterolo-
harboured a small gene that might encode the precursor gous expression. One strain was the well-characterized
peptide of kocurin. Hence, this BGC was attributed to Streptomyces coelicolor M1146 that is optimized for the
kocurin production. expression of specialized metabolites due to the deletion of
host-specific BGCs [12]. The thiopeptide GE2270 was
In silico analysis revealed two putative kocurin BGCs, in Kocu-
expressed successfully in this strain [10]. The second strain
ria UCD-OTCP (accession number: NZ_AOSQ01000025)
was Streptomyces sp. s120, a strain that was isolated in the
and in Kocuria flava HO-9041 (accession number:
same bioprospecting project as K. rosea s17. This strain was
NZ_CP013254), that could be related to thiopeptide biosyn-
chosen due to its fast growth rate and genetic tractability.
thesis. The biosynthesis of the thiopeptide antibiotics GE2270
(from Planobispora rosea) and GE37468 (from Streptomyces To clone the kocurin BGC to express it in a heterologous
ATCC55365) was recently characterized [10, 11]. Both mole- host, it was amplified by PCR in three parts using the
cules show structural similarity, as well as a related mode of reverse and forward primer pairs koc1-3 (Table S2). The
action to kocurin. Both groups reported the ribosomal internal primers created a 35 bp homologous region
origin of the investigated peptide. The RiPPs are initially syn- between the DNA fragments generated by PCR and the
thesized from a larger propeptide, which is post-translationally external primers added 35 bp with homology to both flanks
modified by heterocyclization, serine dehydratation, amida- of the HindIII-linearized, integrative expression vector pSE-
tion and macrocyclization to the final product. The resulting T152ermE* [13]. The three PCR products were joined with
molecules normally have a mass in the range of 1100– the linearized vector using Gibson assembly (1 h at 50 C).
1600 Da. Subsequently, 5 µl of the reaction mixture were added to
200 µl CaCl-competent E. coli XL1blue cells. Transforma-
Screening the sequence data for a putative kocurin propep-
tion was achieved by 20 min on ice, 90 s heat shock at 42 C,
tide (amino acid sequence: STNCFCYPCCSCSAPSS)
followed by addition of 1 ml Luria Bertani (LB) medium
revealed suitable genes in both predicted BGCs. Based on
and 90 min incubation at 37 C for regeneration. Subse-
this strong indication for the in silico identification of the
quently, aliquots were plated onto apramycin-containing
kocurin BGC, the presence of a homologous BGC in our
(50 µg ml 1) LB agar plates. After overnight incubation few
producer strain was attempted. The primer pair Koc_scre_f
colonies were observed. These were tested by colony PCR
and Koc_scre_r was thus designed, which targets the pro-
for the presence of the BGC (Supporting information).
peptide gene region and a fragment of the adjacent dehydra-
tase gene (Supplementary information). PCR using the From positive clones, the plasmid pSET152-Koc was isolated
genomic DNA of K. rosea s17 as template resulted in the (Pureyield Promega plasmid isolation Kit, following the man-
amplification of a fragment with an expected size of 1.1 kbp. ufacture’s protocol) and restricted with SacI. Eight out of ten
Downloaded from www.microbiologyresearch.org by
IP: 194.99.105.233
1410
On: Mon, 08 Jul 2019 22:11:43
Linares-Otoya et al., Microbiology 2017;163:1409–1414
Fig. 1. Biosynthetic gene cluster (BGC) of kocurin and homologues. (a) Gene cluster organization of the related thiopeptide molecules
kocurin GE2270 and GE37468. The colour coding indicates the predicted function of the encoded proteins. The names of the respective
genes are indicated [10, 11]. Genes are drawn to scale; size bar is on the top line. (b) Amino acid alignment of selected thiopeptide pro-
peptides. Three kocurin producers (K. rosea s17, K. flava HO9041 and K. UCD-OTCP), as well as the producers of GE37468 (S. ATCC
55365) and GE2270 (P. rosea), are shown. Black lines indicate the three parts of the propeptide chain; both the leader peptide and the
C-terminal residues are cleaved off to form the mature peptide.
plasmids showed the expected restriction pattern and were new apramycin/nalidixic acid containing SFM agar plates.
sequenced. The validated plasmid pSET152-Koc was trans- After 72 h single colonies were picked and tested by colony
ferred into E. coli ET12567, a methylation-deficient strain PCR for kocurin BGC integration. Subsequent sequencing
[14]. Then, triparental conjugation was performed using addi- verified the integration of the kocurin BGC for both strains.
tionally E. coli ET12567 cells carrying the pUB307 helper plas-
mid and the targeted Streptomyces host strain. In brief, the Heterologous expression of thiopeptide gene clusters is most
E. coli strains were cultured overnight in chloramphenicol challenging. Hence, the projected expression of the thiopep-
(25 µg ml 1) containing LB broth plus kanamycin (50 µg tide TP1161 in Streptomyces coelicolor M512 remained unsuc-
ml 1) in the case of E. coli ET12567-pUB307, and apramycin cessful [15], and the compound GE2270 was only expressed
(50 µg ml 1) for E. coli ET12567-pSET152-Koc. A volume of in 1 out of 35 different media tested [10]. Using the two trans-
200 µl of this preculture was used to inoculate 20 ml LB broth genic strains generated here, five different production media
containing the same antibiotics, and incubation was per- were used, i.e. SM13 and medium C used previously for
formed at 37 C. When the cultures reached an OD600 of 0.5, GE2270 expression in Streptomyces and Planobispora [10, 16],
cells were recovered by centrifugation (3500 g, 5 min) and as well as AF, SFM and ISP2 medium [17]. A preculture was
washed twice with 20 ml LB medium. The resulting pellet was inoculated by using spores of the transgenic strains in ISP2
resuspended in 500–1000 µl LB medium. A volume of 100 µl medium and incubated 48 h at 30 C. Then, 2 ml were used to
of each was mixed together with 100 µl of a Streptomyces spore inoculate 100 ml of the various media; fermentation was per-
solution, which was previously activated (10 min, 50 C in formed for 4 d at 30 C in 300 ml baffled flasks. Subsequently,
2YT medium). The mixture was directly plated on SFM agar the cultures were extracted using 2300 ml of ethyl acetate
plates supplemented with 10 mM MgCl2. After 16 h at 30 C, for liquid–liquid separation. The organic layer was collected,
the plates were covered with 1 ml H2O containing 400 µg nali- combined and dried by evaporation. The resulting extract was
dixic acid and 1 mg apramycin. Conjugants were observed dissolved in methanol (2 mg ml 1). Samples were analysed by
after 24 and 48 h for Streptomyces sp. S120 and S. coelicolor HPLC/MS-MS [gradient: 10 to 100 % acetonitrile over 20 min,
M1146, respectively. Colonies were picked and streaked onto EC10/2 Nucleoshell C18 2.7 µm column (Macherey-Nagel)].
An extract of the native producer K. rosea s17 served as EF-Tu of the original producer can be resistant towards the
positive control, and S. coelicolor M1146 carrying the empty thiopeptide as is the case for the GE2270 producer Planobis-
pSET152 vector and the respective cell-free production pora rosea [10]. This might explain why growth of the het-
medium as negative control. A peak matching the expected erologous host used in this study was retarded, since no
kocurin retention time (12.2–12.4 min) was only detected in additional resistant version of an EF-Tu gene was cloned
S. coelicolor M1146-Koc grown in ISP2 medium. The observed with the BGC. Another option is to use Nonomuraea species
mass was 1515.3745 m z 1 [M+H+] (calculated: 1515.3733 as the heterologous host, since GE2270 production levels
m z 1). In AF medium, only traces were observed, while no could be increased up to 250 mg l 1 [18].
kocurin production was observed for Streptomyces sp s120 in
any of the culture media used (Fig. 2). A putative kocurin biosynthesis is proposed based on compar-
ative analysis with the BGCs of GE2270 and GE37468 and
The observed production level is very low under the condi- considering the recent work performed by the van der Donk
tions used. This is in accordance with previous heterologous and Mitchell groups [19] (Fig. 3 and Table S3). The propep-
expression attempts for thiopeptides in Streptomyces, tide-coding gene kocA is located at the upstream end of the
whereby the media used seem to have a strong influence on BGC and codes for 58 amino acids, divided into a leader pep-
expression [10]. In future, further optimization will be nec- tide (39 amino acids), a core peptide (17 amino acids) and a
essary to improve the kocurin titre. It might be speculated C-terminal Ser-Ala residue that is cleaved off during matura-
that the toxicity of the compound can affect the growth and tion. Also in GE2270 and GE37468, either terminal Ser-Ala or
metabolism of the heterologous host at early stages of the Asp is excised during maturation (Fig. 1). The genes coding
life cycle, since visible biomass production started 3 days for the post-translational modifying enzymes are situated
post-inoculation. Hence, using inducible promoters could downstream of kocA. First, the oxazole and thiazol(ine) rings
be a feasible option to overcome this. Another strategy are formed, catalysed by KocE-G (Fig. 3). These proteins
would be the overexpression of the resistance gene. It was resemble McbC-type dehydrogenases (KocE, G) and a YcaO-
shown before that this resulted in improved GE2270 pro- like cyclodehydratase (KocF) [20].
duction yields [10]. However, a reported self-resistance gene
for an alternative copy of the elongation factor EF-Tu was The serine residues present in the core peptide should be
not found in the entire genome. Therefore, it can be dehydrated through the action of KocB and KocC, both lan-
assumed that the ABC transporter KocI fulfils the immunity tibiotic-type dehydratases. However, the exact timing of the
function, as suggested for GE37468 [11]. Furthermore, the reactions is still unknown. The C-terminus is shortened by
Fig. 2. HPLC/MS-MS analysis of crude extracts of selected S. coelicolor M1146-Koc fermentations and controls. MS spectrum of the
kocurin peak observed in the native producer K. rosea s17 and (b) of S. coelicolor M1146-Koc. (c) Extracted-ion chromatogram (m z 1:
1515.37±0.01) of K. rosea s17 extract [fermented in Marine broth (Difco) medium], ISP2 medium and S. coelicolor M1146-pSET152
(both negative controls), and the heterologous producer strain S. coelicolor M1146-Koc (fermented in ISP2 medium). The peak corre-
sponding to kocurin at 12.2 min is highlighted; the scale is the same in all chromatograms.
the excision of the last two amino acid residues and ami- Hence, in addition to biosynthetic derivatization it might
dated by the action of KocH. For the thiopeptides bernina- become feasible to produce sufficient quantities in a sustain-
mycin and microccocin C, it was shown that this step is able approach to perform chemical modifications on the
essential for the final cyclization reaction [21, 22]. Finally, mother molecule. On the other hand, the generation of
the last step is the +1 to +11 cyclization, yielding a 29-mem- derivatives, e.g. molecules with exchanged amino acids, is
bered macrocycle, by forming the central pyridine ring facilitated by this system.
through an aza–Diels–Alder reaction catalysed by KocD. In
koc-BGC, kocI is additionally found at the downstream end
encoding for an ABC transporter. Funding information
The authors would like to thank the German Academic Exchange Ser-
In conclusion, the results presented here provide the experi- vice (DAAD) for funding (PROPERU, project 57236327). Substantial
mental proof for kocurin BGC and enable future studies on funding of this project by Peruvian Ministry of Education through
Fondo Nacional de Desarrollo Cientifico, Tecnologico y de Innovacion
this antibiotic molecule. On the one hand, a heterologous Tecnologica (Convenio de subvencion 092–2014-FONDECYT, CONCY-
expression system using well-investigated actinobacterial TEC) is kindly acknowledged. L. L-O. thanks Fondo para la Innovación,
strains is the method of choice for up-scaling experiments. la Ciencia y la Tecnología (FINCyT-Innovate Peru) for his fellowship.
Conflicts of interest 12. Gomez-Escribano JP, Bibb MJ. Engineering Streptomyces coeli-
The authors declare that there are no conflicts of interest. color for heterologous expression of secondary metabolite gene
clusters. Microb Biotechnol 2011;4:207–215.
13. Stegmann E, Pelzer S, Bischoff D, Puk O, Stockert S et al. Genetic
References analysis of the balhimycin (vancomycin-type) oxygenase genes.
1. Willyard C. The drug-resistant bacteria that pose the greatest J Biotechnol 2006;124:640–653.
health threats. Nature 2017;543:15.
14. Du L, Liu RH, Ying L, Zhao GR. An efficient intergeneric conju-
2. Sch€aberle TF, Hack IM. Overcoming the current deadlock in anti- gation of DNA from Escherichia coli to mycelia of the lincomy-
biotic research. Trends Microbiol 2014;22:165–167. cin-producer Streptomyces lincolnensis. Int J Mol Sci 2012;13:
3. Schinke C, Martins T, Queiroz SCN, Melo IS, Reyes FGR. Antibac- 4797–4806.
terial compounds from marine bacteria, 2010–2015. J Nat Prod 15. Engelhardt K, Degnes KF, Zotchev SB. Isolation and characteriza-
2017;80:1215–1228. tion of the gene cluster for biosynthesis of the thiopeptide antibi-
4. Martín J, da S Sousa T, Crespo G, Palomo S, Gonzalez I et al. otic TP-1161. Appl Environ Microbiol 2010;76:7093–7101.
Kocurin, the true structure of PM181104, an anti-methicillin- 16. Gastaldo L, Marinelli F. Changes in GE2270 antibiotic production
resistant Staphylococcus aureus (MRSA) thiazolyl peptide from the in Planobispora rosea through modulation of methylation metabo-
marine-derived bacterium Kocuria palustris. Mar Drugs 2013;11: lism. Microbiology 2003;149:1523–1532.
387–398. 17. Donadio S, Monciardini P, Sosio M. Chapter 1. Approaches to dis-
5. Bagley MC, Dale JW, Merritt EA, Xiong X. Thiopeptide antibiotics. covering novel antibacterial and antifungal agents. Meth Enzymol
Chem Rev 2005;105:685–714. 2009;458:3–28.
6. Mahajan G, Thomas B, Parab R, Patel ZE, Kuldharan S et al. In 18. Tocchetti A, Maffioli S, Iorio M, Alt S, Mazzei E et al. Capturing lin-
vitro and in vivo activities of antibiotic PM181104. Antimicrob ear intermediates and C-terminal variants during maturation of
Agents Chemother 2013;57:5315–5319. the thiopeptide GE2270. Chem Biol 2013;20:1067–1077.
7. Weber T, Blin K, Duddela S, Krug D, Kim HU et al. antiSMASH 3.0- 19. Zhang Z, Hudson GA, Mahanta N, Tietz JI, van der Donk WA et al.
a comprehensive resource for the genome mining of biosynthetic Biosynthetic timing and substrate specificity for the thiopeptide
gene clusters. Nucleic Acids Res 2015;43:W237–W243. thiomuracin. J Am Chem Soc 2016;138:15511–15514.
8. Young M, Artsatbanov V, Beller HR, Chandra G, Chater KF et al. 20. Melby JO, Li X, Mitchell DA. Orchestration of enzymatic process-
Genome sequence of the Fleming strain of Micrococcus luteus, a ing by thiazole/oxazole-modified microcin dehydrogenases.
simple free-living actinobacterium. J Bacteriol 2010;192:841–860. Biochemistry 2014;53:413–422.
9. Morris RP, Leeds JA, Naegeli HU, Oberer L, Memmert K et al. 21. Malcolmson SJ, Young TS, Ruby JG, Skewes-Cox P, Walsh CT.
Ribosomally synthesized thiopeptide antibiotics targeting elonga- The posttranslational modification cascade to the thiopeptide ber-
tion factor Tu. J Am Chem Soc 2009;131:5946–5955. ninamycin generates linear forms and altered macrocyclic scaf-
folds. Proc Natl Acad Sci USA 2013;110:8483–8488.
10. Flinspach K, Kapitzke C, Tocchetti A, Sosio M, Apel AK. Heterolo-
gous expression of the thiopeptide antibiotic GE2270 from Plano- 22. Bewley KD, Bennallack PR, Burlingame MA, Robison RA, Griffitts
bispora rosea ATCC 53733 in Streptomyces coelicolor requires JS et al. Capture of micrococcin biosynthetic intermediates
deletion of ribosomal genes from the expression construct. PLoS reveals C-terminal processing as an obligatory step for in vivo
One 2014;9:e90499. maturation. Proc Natl Acad Sci USA 2016;113:12450–12455.
11. Young TS, Walsh CT. Identification of the thiazolyl peptide
GE37468 gene cluster from Streptomyces ATCC 55365 and heter-
ologous expression in Streptomyces lividans. Proc Natl Acad Sci
USA 2011;108:13053–13058. Edited by: G. H. Thomas
Five reasons to publish your next article with a Microbiology Society journal
1. The Microbiology Society is a not-for-profit organization.
2. We offer fast and rigorous peer review – average time to first decision is 4–6 weeks.
3. Our journals have a global readership with subscriptions held in research institutions around
the world.
4. 80% of our authors rate our submission process as ‘excellent’ or ‘very good’.
5. Your article will be published on an interactive journal platform with advanced metrics.