Marine Biotechnology

, Volume 13, Issue 6, pp 1062–1073

Bioactivity, Chemical Profiling, and 16S rRNA-Based Phylogeny of Pseudoalteromonas Strains Collected on a Global Research Cruise

Authors

    • National Food InstituteTechnical University of Denmark
  • Maria Månsson
    • Department of Systems BiologyTechnical University of Denmark
  • Kristian F. Nielsen
    • Department of Systems BiologyTechnical University of Denmark
  • Lone Gram
    • National Food InstituteTechnical University of Denmark
Original Article

DOI: 10.1007/s10126-011-9369-4

Cite this article as:
Vynne, N.G., Månsson, M., Nielsen, K.F. et al. Mar Biotechnol (2011) 13: 1062. doi:10.1007/s10126-011-9369-4

Abstract

One hundred one antibacterial Pseudoalteromonas strains that inhibited growth of a Vibrio anguillarum test strain were collected on a global research cruise (Galathea 3), and 51 of the strains repeatedly demonstrated antibacterial activity. Here, we profile secondary metabolites of these strains to determine if particular compounds serve as strain or species markers and to determine if the secondary metabolite profile of one strain represents the bioactivity of the entire species. 16S rRNA gene similarity divided the strains into two primary groups: One group (51 strains) consisted of bacteria which retained antibacterial activity, 48 of which were pigmented, and another group (50 strains) of bacteria which lost antibacterial activity upon sub-culturing, two of which were pigmented. The group that retained antibacterial activity consisted of six clusters in which strains were identified as Pseudoalteromonas luteoviolacea, Pseudoalteromonas aurantia, Pseudoalteromonas phenolica, Pseudoalteromonas ruthenica, Pseudoalteromonas rubra, and Pseudoalteromonas piscicida. HPLC-UV/VIS analyses identified key peaks, such as violacein in P. luteoviolacea. Some compounds, such as a novel bromoalterochromide, were detected in several species. HPLC-UV/VIS detected systematic intra-species differences for some groups, and testing several strains of a species was required to determine these differences. The majority of non-antibacterial, non-pigmented strains were identified as Pseudoalteromonas agarivorans, and HPLC-UV/VIS did not further differentiate this group. Pseudoalteromonas retaining antibacterial were more likely to originate from biotic or abiotic surfaces in contrast to planktonic strains. Hence, the pigmented, antibacterial Pseudoalteromonas have a niche specificity, and sampling from marine biofilm environments is a strategy for isolating novel marine bacteria that produce antibacterial compounds.

Keywords

PseudoalteromonasAntibacterial activitySecondary metabolitesBioprospectingGalathea 3

Introduction

Compounds of relevance for the pharmaceutical and biotechnology industries are produced by marine microorganisms (Burgess et al. 1999), and it has been suggested that some compounds of pharmacological interest previously attributed to macroorganisms may in fact be of microbial origin (Bewley and Faulkner 1998; Simmons et al. 2008; Sudek et al. 2006). The emergence of multiresistant pathogenic bacterial strains and the failure of combinatorial and diversity-oriented chemistry to adequately supply the drug discovery pipeline (Newman 2008) have re-invigorated natural product chemistry as a path for discovery and development of new antibiotics. With this in mind, we isolated marine bacteria with antibacterial activity during the Danish Galathea 3 marine research expedition (Gram et al. 2010). The antibacterial strains were tentatively identified using 16S rRNA similarity, and one of the major groups of isolated bacteria was identified as Pseudoalteromonas (Gram et al. 2010).

The genus Pseudoalteromonas consists of Gram-negative marine bacteria belonging to the γ-proteobacteria and is present globally in marine waters where they constitute 0.5% to 6% of the total bacterioplankton (Wietz et al. 2010). They are heterotrophic aerobes and non-fermentative, and the cells are motile by one or more polar flagella. The genus divides into two groups: pigmented and non-pigmented species. The pigmented species are often producers of bioactive secondary metabolites (Bowman 2007) displaying cytotoxic (Zheng et al. 2006), antibacterial (Gauthier 1976b; Gauthier and Flatau 1976; Isnansetyo and Kamei 2003; Jiang et al. 2000; McCarthy et al. 1994), antifungal (Franks et al. 2006; Kalinovskaya et al. 2004), or antifouling (Egan et al. 2001; Holmström et al. 2002) effects. It has been hypothesized that bioactive Pseudoalteromonas are primarily associated with higher organisms (Holmström and Kjelleberg 1999), suggesting an ecological role in which some bioactive species might play an active part in host defense against pathogens and fouling organisms (Holmström et al. 1996; Armstrong et al. 2001; Egan et al. 2008). A link between surface colonization and antibacterial activity has not been experimentally verified, although several studies have successfully isolated epiphytic bacteria with antibacterial activity from algae and other marine organisms (Armstrong et al. 2001; Boyd et al. 1999; James et al. 1996; Penesyan et al. 2009). The group of non-pigmented species has highly similar 16S rRNA gene sequences (Ivanova et al. 2004) and is rarely inhibitory against other microorganisms, although, Pseudoalteromonas haloplanktis strain INH produces isovaleric acid and 2-methylbutyric acid showing a broad spectrum of bacterial inhibition (Hayashida-Soiza et al. 2008).

Phylogeny and differentiation of bacterial species rely heavily on 16S rRNA gene similarity (Stackebrandt et al. 2002); however, 16S rRNA gene similarity does not provide sufficient differentiation below species level (Fox et al. 1992). Comparison of secondary metabolite production has supported species delineation within the actinomycete genus Salinispora, where strains with high 16S rRNA similarity were shown to belong to distinct species each with different specific metabolite profiles (Jensen et al. 2007). In mycology, comparison of chemical profiles (e.g., TLC, direct-infusion mass spectrometry, and HPLC with various detectors such as UV/VIS and/or mass spectrometry) of secondary metabolites has been widely used to identify and differentiate filamentous fungi (chemotaxonomy) also at sub-species level (Frisvad et al. 2008), and the chemophylogeny correlates well with phylogenetic analysis of sequences of specific housekeeping genes (e.g., chitin synthase, β-tubulin, and calmodulin; Geiser et al. 2007). Since several Pseudoalteromonas species produce a range of secondary metabolites, we hypothesized that chemical profiling and specific marker compounds could be indicative of bioactive potential and at the same time be useful in species identification or differentiation (Jensen et al. 2007).

The aim of the present study was to profile the secondary metabolites of these strains to determine if particular compounds serve as markers of strains or species with antibacterial activity and to determine if several strains of each species must be tested to assess the full bioactivity potential. As part of this, we provide accurate identification and phylogeny of these organisms by detailed 16S rRNA gene sequence comparative analysis. Since the bacteria were isolated from different sample types, our collection also allows us to address aspects of Pseudoalteromonas ecology such as the possible link between surface or planktonic lifestyle and antibacterial activity.

Materials and Methods

Strain Isolation

Approximately 500 marine bacterial strains with antagonist activity against Vibrio anguillarum strain 90-11-287 (Skov et al. 1995) were isolated during the Danish Galathea 3 research expedition (Gram et al. 2010). One hundred one of these strains tentatively identified as Pseudoalteromonas species were included in the present study. V. anguillarum 90-11-287 was used as target strain since the expedition ship was not equipped to handle, e.g., potential human pathogens, and this Vibrio strain is in our experience very sensitive to antibacterial compounds from other marine bacteria (Hjelm et al. 2004).

Growth Media and Culture Conditions

Pseudoalteromonas strains were grown in marine broth (MB) 2216 (Difco, Detroit, MI, USA) and on marine agar (MA) 2216 (Difco, Detroit, MI, USA) prepared in accordance with the manufacturer’s instructions. Broth cultures were incubated under stagnant conditions at 25°C. Pigment production was determined by visual inspection of 48-h-old culture broths (MB) and colonies grown on MA for 24 to 48 h.

Antibacterial Activity

Instant Ocean (IO) bioassay agar plates were prepared as described by Hjelm et al. (2004). Ten grams per liter agar, 3.3 g/l casamino acids (Difco 223050, Detroit, MI, USA), and 30 g/l Instant Ocean aquatic salts (Instant Ocean® Aquarium systems Inc., Sarrebourg, France) were added to distilled water and autoclaved. Glucose (0.4%) and 10 μl/ml of V. anguillarum overnight culture were added to the cooled (44°C) IO and plates poured. The plates were allowed to dry for 15 min, and if used for well diffusion agar assays (WDAA), wells (diameter 6 mm) were punched. The inhibitory activity of live Pseudoalteromonas bacterial cells was tested by spotting 48-h-old MA grown colonies on freshly prepared IO agar plates containing V. anguillarum. Plates were incubated at 25°C and inspected for clearing zones in the growth of V. anguillarum after 24 h.

Cell-free supernatants were prepared to test for the presence of water-soluble antibacterial compounds secreted to the broth, and ethyl acetate extracts were prepared to test for production of non-polar antibacterial compounds. Each strain was grown in 20 ml of MB for 48 h. A 1.5-ml sample was withdrawn for 0.2 μm filtering, and subsequently the remainder of the culture was extracted with an equal volume of ethyl acetate. The ethyl acetate fraction was transferred to a new vessel, evaporated to dryness, and redissolved in 2 × 0.5 ml ethyl acetate. The 1.5-ml cell-free sterile-filtered supernatant and the ethyl acetate extracts were stored at −20°C until tested in the WDAA (50 μl sample per well) based on IO agar plates containing V. anguillarum. Controls (sterile MB and pure ethyl acetate) did not cause any inhibition zones.

The number of antibacterial Pseudoalteromonas strains in surface samples (e.g., algae, driftwood, fish, and sediment samples) was compared to their numbers in water samples by the Fisher’s exact test (Fisher 1958). A 2 × 2 contingency table was used to test the hypothesis that presumed antibacterial strains with stable antibacterial activity were equally likely to be isolated from water samples and surface samples.

16S rRNA Gene Sequence Analyses

A detailed phylogenetic analysis was performed on 16S rRNA sequences obtained in a previous study (Gram et al. 2010). For the analysis in this study, we conducted a BLAST (http://blast.ncbi.nlm.nih.gov) search against a compilation of Pseudoalteromonas type strain sequences retrieved from GenBank (list of type strains obtained from http://www.bacterio.cict.fr), and sequences of the type strains with a BLAST match in our strain collection were included in 16S rRNA sequence analysis to obtain a robust phylogenetic tree. Sequences from two additional Pseudoalteromonas strains were included: The genus type strain P. haloplanktis and the bioactive Pseudoalteromonas tunicata. Salinispora arenicola CNS-205 was used as outgroup. The sequences were aligned by the MAFFT online software (http://www.ebi.ac.uk/Tools/mafft/; Katoh et al. 2002) and curated with the Gblocks software on its least stringent settings (Castresana 2000; Talavera and Castresana 2007). The resulting alignment was processed using the MEGA4 software (Tamura et al. 2007) to create neighbor-joining and minimum evolution trees. PhyML 3.0 was used to generate a maximum likelihood tree (Guindon and Gascuel 2003). Phylogenetic trees were generated under default parameters with 1,000 bootstrap replications for neighbor-joining and minimum evolution trees and 100 bootstrap replications for the maximum likelihood tree. GenBank accession numbers for the Pseudoalteromonas strains used in this study are included in Supplementary Table 1.

HPLC-UV/VIS Analysis of Secondary Metabolites

The strains were grown in static cultures in 10 ml MB for 3 days at 25°C, and for each species, one strain was cultured in triplicate to establish extraction and growth variation. Cultures were extracted with equal volumes of ethyl acetate, centrifuged, and the ethyl acetate was evaporated under N2 to dryness. Samples were redissolved in 300 μl acetonitrile–water (1:1 v/v) and filtered through a 13-mm ID PFTE syringe filter. A subsample of 2 μl was then analyzed by reversed phase HPLC on an Agilent 1100 System equipped with a UV/VIS photo diode array detector (scanning 200–600 nm). Separation was done on a 100 mm × 2 mm i.d., 3 μm Gemini C6-phenyl column (Phenomenex, Torrance, CA, USA), running at 40°C using a binary linear solvent system of water (A) and acetonitrile (B; both buffered with 50 μl/l trifluoroacetic acid) at a flow of 300 μl/min. The gradient profile was t = 0 min, 5% B; t = 22 min, 70% B; t = 24.5, 100% B; t = 27 min, 100% B; and t = 29 min, B = 5%, holding this for 8 min prior to the next injection. The chromatographic profiles were compared, subtracting peaks present in media blank extracts. Samples were analyzed in random order, and six of the first extracts were analyzed several times during the sequence to determine any retention time shifts. Cluster analysis was done on a matrix of detected / non-detected peaks (1/0) using NTSYSpc 2.20q (Exeter Software, Setauket, NY, USA). SAHN clustering was used by unweighted pair-group method (UPGMA) and simple distance measurement. Representative extracts were also analyzed by HPLC-UV/VIS-TOFMS in both positive and negative electrospray (Nielsen and Smedsgaard 2003). Peaks were tentatively identified by UV spectra and accurate mass data by matching in Antibase 2009 (35 930 microbial secondary metabolites; Wiley & Sons, Hoboken, NJ, USA; Nielsen et al. 2006).

Results

Pigmentation and Antibacterial Activity

The one hundred one Pseudoalteromonas strains were originally isolated for their ability to inhibit V. anguillarum (Gram et al. 2010). However, on re-cultivation and re-testing for the ability to inhibit V. anguillarum after storage at −80°C for several months, only 51 strains retained inhibitory activity (Table 1). These 51 strains were all inhibitory when tested as live cultures in the “spot test” assay. Twenty of the strains produced water-soluble, diffusible, antibacterial substances as indicated by the ability of cell-free sterile-filtered supernatant to inhibit growth of V. anguillarum in the WDAA. These 20 strains were identified as Pseudoalteromonas phenolica (one strain), Pseudoalteromonas luteoviolacea (five strains), Pseudoalteromonas rubra (nine strains), Pseudoalteromonas citrea (one strain), and Pseudoalteromonas aurantia (four strains). Ethyl acetate extraction of culture broths resulted in 19 crude extracts which inhibited growth of V. anguillarum in the WDAA. Four of these were identified as P. luteoviolacea and were the only strains where both cell-free supernatant and crude ethyl acetate extracts inhibited Vibrio growth. The remaining 15 inhibitory crude extracts all originated from strains identified as Pseudoalteromonas ruthenica. The crude ethyl acetate extracts of 26 strains were intensely colored; however, only some of these extracts inhibited growth of Vibrio indicating that the pigments were not universally antibacterial. Cell-free culture supernatants and ethyl acetate crude extracts of strains with no growth inhibition of V. anguillarum in the “spot test” assay were also tested but showed no growth inhibition.
Table 1

Identity, antibacterial activity, and pigmentation of Pseudoalteromonas strains from a global collection

16S rRNA cluster

No. of strains

Related type straina

No. of strains from

No. of strains inhibiting Vibrio

Inhibition of Vibrio by EtAc extracts

Pigmented

Non-pigmented

Surface samples

Water samples

Pigmented

Non-pigmented

I

1

37

P. agarivorans

16

22

0

0

0

II

5

0

P. aurantia

2

3

5

0

0

III

0

9

P. prydzensis

8

1

0

0

0

IV

3

3

P. phenolica

3

3

3

2

0

V

4

0

P. luteoviolacea

4

0

4

0

4

VI

9

0

P. rubra

9

0

9

0

0

VII

13

0

P. flavipulchra

13

0

13

0

0

VIII

15

0

P. ruthenica

14

1

15

0

15

S1727

0

1

P. mariniglutinosa

0

1

0

0

0

S3655

1

0

P. spongiae

1

0

0

0

0

Total

51

50

 

70

31

49

2

19

aThe type strain which the majority of the strains in the cluster were most closely related to

Forty-eight of the 51 antibacterial strains were pigmented, while two (S3431 and S3655) of the 50 non-active strains were pigmented (Table 1). Antibacterial activity was significantly more likely to be produced by pigmented strains as determined by Fisher’s exact test (two-tailed p value of 0.0000). In total, 70 strains were isolated from surface swabs and 31 from water samples. Of the surface-associated strains, 45 remained active in the spot-assay in comparison to six of the water sample strains. The Fisher’s exact test demonstrated a significant relation between surface association and stable antibacterial activity (two-tailed p value of 0.0000).

Pseudoalteromonas strains were isolated on all parts of the global cruise in both tropical and temperate waters (Fig. 1). Our strain collection is not large enough to allow for complete biogeographic analysis, but pigmented strains appeared to be more frequent in coastal areas whereas the non-pigmented strains appeared associated with open waters (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10126-011-9369-4/MediaObjects/10126_2011_9369_Fig1_HTML.gif
Fig. 1

Isolation sites of 101 Pseudoalteromonas strains with each color code representing one of the eight clusters as indicated in Table 1. Each circle represents the isolation of one or more strains belonging to the given cluster. White = cluster I, yellow = cluster II, gray = cluster III, green = cluster IV, violet = cluster V, red = cluster VI, orange = cluster VII, pink = cluster VII, and blue = non-clustered strain

16S rRNA Gene Sequence Analyses

We initially performed a BLAST search querying the 16S rRNA gene sequence of each strain against the GenBank database. The results of this analysis were ambiguous, as some sequences returned more than 40 hits all with identical scores in the BLAST results (data not shown), frequently including the sequences of several different Pseudoalteromonas species. Therefore, a BLAST analysis was carried out querying the sequences against a complete set of Pseudoalteromonas type strains, and hence, each strain is matched with the best type strain BLAST match (Suppl Table 1).

The 16S rRNA gene sequences were used to cluster the strains by constructing a neighbor-joining tree, and branch support was verified by comparison to minimum evolution and maximum likelihood trees (Fig. 2). Nodes supported by an 80% bootstrap cutoff were collapsed when three or more strains were included in the cluster (Fig. 2). An exception was made for clusters VI and VII, which are shown as separate clusters due to obvious differences in phenotype (pigment, bioactivity, secondary metabolite profile). Ninety-nine of the strains fell into one of eight primary clusters.
https://static-content.springer.com/image/art%3A10.1007%2Fs10126-011-9369-4/MediaObjects/10126_2011_9369_Fig2_HTML.gif
Fig. 2

Phylogenetic tree based on 16S rRNA gene sequences of the Galathea 3 Pseudoalteromonas strains. Sequences were aligned by MAFFT (default options), and the resulting alignment was used to generate a neighbor-joining tree in the MEGA4 software package (default settings, 1,000 bootstraps). S. arenicola CNS-205, GenBank accession number CP000850, GeneID: 5705939 was used as outgroup. Clusters containing two or more non-type strain sequences were collapsed. The scale bar represents 0.02 amino acid substitutions per nucleotide position. Circle nodes also occurred in minimum evolution and maximum likelihood trees

Clusters I and III consisted of non-pigmented non-inhibitory strains (Table 1). Cluster I included 38 strains and the type strains of P. haloplanktis, Pseudoalteromonas agarivorans, Pseudoalteromonas tetraodonis, Pseudoalteromonas paragorgicola, Pseudoalteromonas distincta, Pseudoalteromonas arctica, Pseudoalteromonas nigrifaciens, Pseudoalteromonas elyakovii, Pseudoalteromonas carrageenovora, Pseudoalteromonas marina, and Pseudoalteromonas aliena. Strain S3431—a black-pigmented strain in cluster I—did not show more than 97% similarity even when compared to the full GenBank database which suggests that S3431 could represent a novel Pseudoalteromonas species. Despite the low BLAST similarity score, phylogenetic analysis and tree construction placed strain S3431 in the diverse cluster I (the non-collapsed cluster I is shown in Supplementary Figure 1). Cluster III contained no type strains which supported the BLAST analysis where the strains in cluster III were 98% similar to the best type strain match (Supplementary Table 1).

The remaining six of the eight clusters contained pigmented strains. Pale yellow strains clustered with the type strains of P. citrea and P. aurantia (cluster II) and four intensely purple strains grouped in cluster V with the type strain of P. luteoviolacea. Cluster VI contained nine red-pigmented strains and the type strain of P. rubra, and cluster VII consisted of 12 intensely yellow strains, one pale yellow strain, and the Pseudoalteromonas flavipulchra, Pseudoalteromonas maricaloris, and Pseudoalteromonas piscicida type strains. Fifteen strains and their nearest BLAST match, P. ruthenica, formed cluster VII. These strains all produced a pale brown pigment. Cluster IV contained six strains and the type strain of P. phenolica. Four of the strains in this cluster had P. phenolica as their best type strain BLAST match; however, strain S1093 had P. luteoviolacea as its best match at 98% identity, while P. rubra and P. luteoviolacea type strains scored identically (97%) as the best matches for S2724. The strains in cluster IV were heterogeneous with respect to pigmentation, some were non-pigmented and others appeared brown.

Profiling of Secondary Metabolites

The 101 strains and select type strains were separable in discrete groups by HPLC-UV/VIS (Fig. 3), and the triplicate profiles from an isolate of each species were very reproducible and could be superimposed on each other (data not shown). All of the 38 strains of the 16S rRNA cluster I fell into group A, in which no UV/VIS peaks were unique compared to the media blanks indicating that no secondary metabolites were produced. This large group also included all nine strains from 16S rRNA cluster III and strains of P. phenolica and P. ruthenica less proficient in secondary metabolite production. A summary of the detected compounds is shown in Table 2, and the producer organisms are shown by 16S rRNA cluster in Table 3.
https://static-content.springer.com/image/art%3A10.1007%2Fs10126-011-9369-4/MediaObjects/10126_2011_9369_Fig3_HTML.gif
Fig. 3

Dendrogram from cluster analysis of detected peaks from HPLC-UV/VIS detection of compounds in the ethyl acetate extracted broth and biomass. Data were processed in NTSYSpc 2.20q, with SAHN clustering by UPGMA and simple distance measurement

Table 2

Ethyl acetate extractable secondary metabolites produced by pigmented Pseudoalteromonas strains

Compound

MI (Da)c

UV-max data

RT (min)

Code

Indolmycina

257

 

11.21

A

2-Pentyl-4-quinolinolb

215

 

12.31

B

Novel mono-brominated indole

280

 

13.78

C

Violaceina

343

 

14.29

D

Unidentified

244

212 nm (100%), 250 nm (48%)

15.21

E

2-n-Heptyl-(1H)-quinolin-4-oneb

243

 

15.47

F

Unidentified

NI

228 nm (45%), 308 nm (100%)

15.70

G

Unidentified

386

<200 nm

15.78

H

Unidentified

316

286 nm (100%)

15.80

I

Unidentified

NI

<200 nm

15.81

J

Unidentified

NI

228 nm (45%), 308 nm (100%)

15.98

K

Unidentified

676

<200 nm

15.99

L

Prodigiosina

323

 

16.00

M

Unidentified

NI

<200 nm

16.26

N

Unidentified

NI

228 nm (45%), 308 nm (100%)

16.30

O

Bromoalterochromide Ab

843

 

16.49

P

Novel bromoalterochromide, 2 bromine

921

 

16.60

Q

Novel bromoalterochromide, 1 bromine

857

 

17.10

R

Unidentified

333

310 nm (100%)

17.20

S

2-n-Nonyl-(1H)-quinolin-4-oneb

271

 

17.96

T

Nonyl-quinolinone analogb

271

 

18.28

U

Unidentified

315

362 nm (100%)

18.90

V

Unidentified

244

218 nm (100%), 280 nm (82%)

19.42

W

Unidentified

NI

218 nm (100%), 288 nm (82%)

19.78

X

Unidentified

NI

250 nm (100), 280 nm (86)

19.92

Y

Pentabromopseudilinb

549

 

22.65

Z

NI no ionization or MI could not be assigned using ESI+ and ESI

aValidated reference standard used for verification

bAccurate mass and UV data fit the data from Antibase2009

cMono-isotopic mass

Table 3

Secondary metabolites produced by Pseudoalteromonas species clustered by 16S rRNA gene similarity

16S cluster

# strains

Peak at retention time present in Pseudoalteromonas strain/organism

None

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z

I

38

x

                          

P. tetraodonis

DSM9166

x

                          

P. prydzensis

LMG21428

x

                          

II

1

  

x

   

x

             

x

      
 

2

  

x

   

x

             

x

x

     
 

1

  

x

   

x

          

x

   

x

     
 

1

  

x

   

x

          

x

   

x

     

P. aurantia

DSM6057

  

x

                

x

       

P. citrea

DSM8771

  

x

   

x

          

x

         

III

9

x

                          

IV

3

x

                          
 

1

                

x

x

x

        
 

2

                          

x

P. phenolica

DSM21460

x

                          

V

2

    

x

                     

x

 

2

 

x

  

x

                      

P. luteoviolacea

ATCC33492

    

x

                     

x

VI

1

            

x

              
 

1

             

x

  

x

x

x

      

x

 
 

1

   

x

 

x

   

x

             

x

x

  
 

1

          

x

  

x

     

x

  

x

    
 

1

       

x

    

x

x

x

            
 

2

            

x

              
 

1

             

x

             
 

1

            

x

x

             

P. rubra

DSM6842

            

x

              

VII

9

                

x

x

x

        
 

3

                

x

x

x

      

x

 
 

1

  

x

   

x

                    

P. flavipulchra

LMG20361

                

x

x

x

        

VIII

2

x

                          
 

13

        

x

  

x

   

x

           

P. ruthenica

LMG19699

x

                          

S1727

1

x

                          

S3655

1

x

                          

Identification of peak by capital letter in Table 2

Based on their production of specific metabolites, the majority of pigmented bacteria formed six main groups not including four P. rubra strains (Fig. 3). In the pigmented bacteria, a total of 26 distinct peaks were detected and included in the cluster analysis. We identified indolmycin, violacein, and prodigiosin among the significant peaks based on reference standards. Furthermore, nine peaks could be tentatively identified based on HPLC-UV/VIS-TOFMS results and data in Antibase2009. These nine included two likely novel bromoalterochromides and a brominated indole (Table 2).

Comparing the 16S rRNA gene sequence clusters with the chemical profiling revealed several patterns. Some compounds were exclusively produced by strains belonging to the same cluster whereas other compounds were produced across several strains from different clusters. All strains in cluster II (P. aurantia/P. citrea) shared production of compound B (retention time, RT 12.31 min) whereas the production of four other compounds F, Q, T, and U (RT 15.47, 16.60, 17.96, and 18.28 min) were scattered in the group. This included a novel bromoalterochromide (compound Q, RT 16.60 min) that was also found in cluster VII (P. flavipulchra/P. piscicida). Three of the other compounds were identified as quinolines based on distinct UV spectra and accurate mass.

P. luteoviolacea strains (cluster V) shared production of compound D (violacein, RT 14.29) in all four strains and the type strain but were sub-divided by compound A (indolmycin, RT 11.21) produced by two strains and compound Z (pentabromopseudilin, RT 22.65) produced by the two other strains and the type strain. This division is visible in Fig. 4, which shows chromatograms of the four strains in cluster V. Interestingly, also two P. phenolica strains produced compound Z.
https://static-content.springer.com/image/art%3A10.1007%2Fs10126-011-9369-4/MediaObjects/10126_2011_9369_Fig4_HTML.gif
Fig. 4

HPLC-UV/VIS profiles of the ethyl acetate extracted broth and biomass of four P. luteoviolacea with the tailing violacein peak at 14–15 min and in two traces pentabromopseudilin at 22-23 min in contrast to indolmycin at 11 min in the two other traces

Cluster VII (P. flavipulchra and P. piscicida) was chemically very homogeneous. All strains except one produced three bromoalterochromides P, Q, and R (RT 16.49, 16.60, and 17.10) of which Q and R were novel compounds. In contrast to cluster V and compound D, the production of P, Q, and R was not a unique marker for strains of this cluster as P, Q, and R were also detected in one strain from cluster II and one strain from cluster VI.

Thirteen of the 15 strains in cluster VIII, identified as P. ruthenica, shared a unique chemical profile and produced the compounds H, K, and O (RT 15.78, 15.98, 16.30) with characteristic UV spectra. None of these matched compounds in Antibase2009 and potentially constitute novel antibacterials. These compounds were not detected among strains from other clusters, yet they were not suitable as a distinct chemical marker for cluster VIII since no secondary metabolites were detected in the remaining two strains in this cluster or the type strain.

The strains in cluster VI were identified as P. rubra and were chemically very heterogeneous. Five of nine strains produced the red pigment prodigiosin (compound M, RT 16.00) which was not detected in strains of other clusters. Additionally, a multitude of known and non-identified compounds were detected in one or more strains in the cluster. In total, 16 compounds were detected within the cluster, and 12 of these were unique for this cluster. Only two strains shared an identical production of secondary metabolites, further stressing the chemical diversity among the strains in this cluster.

Discussion

We demonstrate in this study, in agreement with earlier findings (Bowman 2007), that species within the Pseudoalteromonas genus produce a range of secondary metabolites, some with antibacterial activity. Several species of the genus are intensely pigmented, and it is hypothesized that pigmentation co-occur with antibacterial activity (Egan et al. 2002). In our global collection of Pseudoalteromonas strains that demonstrated antibacterial activity on initial isolation, strains that were pigmented were significantly more likely to retain antibacterial activity on re-growth than non-pigmented strains. Several intensely colored organic extracts were not inhibitory against V. anguillarum, and hence, we do not believe that the pigments, in general, are the cause of the antibacterial activity although, e.g., the purple pigment violacein is a known antibiotic compound (Lichstein and Vandesand 1945).

Nearly half of the isolated strains did not retain any antibacterial effect after frozen storage and sub-culturing despite being isolated on the original plates due to antibacterial activity. The observed loss of antibacterial activity may be due to a requirement for factors specific to local seawater, as initial tests for antibacterial activity were carried out using 50% local seawater (Gram et al. 2010). Furthermore, loss of antibacterial activity may be due to repression or inhibition of gene clusters encoding products that are required for secondary metabolite synthesis (e.g., by catabolite repression). A reduction in antibiotic production when the producer organism is grown in excess of a carbon source is a known phenomenon (Sanchez et al. 2010), and suppression of secondary metabolite production by excess concentrations of other substrate components is demonstrated in Streptomyces (Doull and Vining 1990). This could suggest that culturing the strains under nutrient limited conditions may reestablish production of antibacterial compounds. Also, during the original sampling and screening procedure, the agar plates may have harbored co-cultured microorganisms which potentially induce antibacterial activity as has been demonstrated by Mearns-Spragg et al. (1998). Hence, it may be possible to re-induce the antibacterial activity if the right conditions can be created.

Several bioactive Pseudoalteromonas have been isolated from higher organisms, and it has been hypothesized that antibacterial compounds may play a role in bacterial competition or as protective agents beneficial for the host organism (Holmström and Kjelleberg 1999). We provide statistical evidence that surface-associated presumed antibacterial pseudoalteromonads are significantly more likely to show stable production of antibacterial compounds than Pseudoalteromonas species isolated as planktonic cells. This suggests that production of antibacterial compounds may play an important role in the ability of Pseudoalteromonas strains to colonize and persist on surfaces submerged in the marine environment, as previously suggested for P. tunicata strain D2 (Rao et al. 2005).

The analysis of 16S rRNA gene sequences from our global collection of Pseudoalteromonas confirms that phylogenetic analysis results in a number of clusters encompassing predominantly pigmented species or non-pigmented species (Ivanova et al. 2004). Strain S3431 was the single pigmented strain in the so-called non-pigmented clusters. Novel diversity might be represented in cluster III which consisted of strains with 98% or less 16S rRNA gene sequence similarity to Pseudoalteromonas type strains and formed a separate cluster in the phylogenetic analysis. However, these strains showed no antibacterial activity and no small molecule metabolites were detected. Such novel diversity could still represent untapped biotechnological potential, producing, e.g., enzymes or peptides with biological activity, as known for other non-pigmented Pseudoalteromonas (Hoyoux et al. 2001; Violot et al. 2005).

Chemical profiling of the strains detected an array of secondary metabolites. In addition to complementing our analysis of 16S rRNA gene sequences, it also demonstrated that some compounds (e.g., violacein, prodigiosin) were characteristic of a species and other compounds were produced by several species, and we also detected intra-species clusters of different secondary metabolite profiles. In a broad sense, the clustering based on 16S rRNA gene similarity agreed with the groups resulting from the chemophylogenetic analysis. However, some compounds were produced by organisms of different species that then clustered together using the secondary metabolites as basis. The chemical analysis separated the four isolated P. luteoviolacea strains into two distinct sub-groups, showing intra-species chemical diversity. The P. luteoviolacea strains produced violacein and pentabromopseudilin which are active against gram-positive and gram-negative bacteria and the anti-staphylococcal agent indolmycin (Hornemann et al. 1971; Hurdle et al. 2004). Violacein and pentabromopseudilin have previously been detected in P. luteoviolacea (Gauthier 1976a; Laatsch and Pudleiner 1989), but to our knowledge, this is the first report of Pseudoalteromonas strains producing indolmycin (Månsson et al. 2010).

Within some species, all strains were consistently antibacterial. However, in others, such activity did not appear to be a consistent trait of the species. For instance, strains of the 16S cluster VI (P. phenolica) were heterogeneous in their ability to inhibit Vibrio in our assays, while all but one strain in the homogeneous cluster VII had identical metabolite profiles and all were inhibitory. Even more obvious was the heterogeneous chemical profiles within the P. rubra strains. All except one strain shared a chemical marker prodigiosin and/or RT 15.99 min but had major variations in 14 other compounds. This may in part be due to loss of ability to produce a compound. For instance, strain S2471 over time lost ability to produce the brominated indole (RT 13.78 min). Also, we note that the type strain DSM 6842 (ATCC 29570) did not in our culture produce prodigiosin which has been observed previously (Gauthier 1976b; Gauthier and Flatau 1976). The consistent bromoalterochromide production in the two species P. piscicida and P. flavipulchra/maricaloris (cluster VII) was expected (Speitling et al. 2007) and supported the high DNA sequence similarity between the two. This emphasizes the need to isolate and screen multiple strains from each species when bioprospecting within the genus Pseudoalteromonas, as even the homogeneous cluster VII contains one strain with a metabolite profile that does not share a single compound with the other strains in this cluster.

Several of the 26 detected peaks were known substances, with a majority known as antibacterials. These included violacein (Lichstein and Vandesand 1945), two bromopseudilins (Lovell 1966), two indolmycins (Werner 1980), four quinolines (Wratten et al. 1977), and prodigiosin (Kalesperis et al. 1975). Due to its very low aqueous solubility, violacein probably protects against predation rather than acts as a true antibiotic, and it has been shown to induce cell death in grazing organisms (Matz et al. 2008). Such compounds would be very beneficial for protection of a biofilm, which is likely how surface-associated Pseudoalteromonas would grow. The 14 compounds that could not be identified were mainly not identified due to poor ionization in ESI+ and ESI and/or several plausible candidates in Antibase2009. However, for chemotaxonomic studies, identity of the compounds is not necessary as long as they can be unambiguously identified between samples (Frisvad et al. 2008).

Within cluster VIII (P. ruthenica) and cluster II (aurantia/citrea), we found examples where strains with highly similar 16S rRNA gene sequences (>99%) and with identical chemotaxonomy originated from geographically distinct locations. This latter observation is in agreement with studies on Salinispora biogeography and secondary metabolite production in which the authors show how strains of the marine bacterium S. arenicola isolated from worldwide locations are highly related and produce identical patterns of secondary metabolites (Jensen and Mafnas 2006; Jensen et al. 2007). In contrast, the P. luteoviolacea and P. rubra strains showed both local and global variations in their secondary metabolite profile, which one might speculate is due to adaptation to local specific niches.

In conclusion, we believe sampling from specific niches, e.g., biofilms on surfaces, to be of importance in discovery of novel secondary metabolites from the genus Pseudoalteromonas. While differences in metabolite patterns among species encourage isolation and screening of novel diversity, bioprospecting known Pseudoalteromonas species should not be ruled out. Investigation of multiple strains of one Pseudoalteromonas species can yield novel compounds due to intra-species variations within secondary metabolite profiles.

Acknowledgments

We acknowledge Dr. Jesper B. Bruhn for valuable input during the early phase of this study. This study was supported by the Programme Commission on Health, Food and Welfare under the Danish Council for Strategic Research. The present work was carried out as part of the Galathea 3 expedition under the auspices of the Danish Expedition Foundation. This is Galathea 3 contribution no. p73.

Supplementary material

10126_2011_9369_MOESM1_ESM.doc (244 kb)
ESM 1(DOC 244 kb)

Copyright information

© Springer Science+Business Media, LLC 2011