Unexpected Diversity during Community Succession in the Apple Flower Microbiome

I recently had the pleasure of reading about the succession of microbes in flowers as they emerge, open, and eventually fall from the tree. This interesting paper (published by the Handelsman group and found here:http://mbio.asm.org/content/4/2/e00602-12) used 454 pyrosequencing to identify bacterial community members in flowers across 5 different trees throughout time. I especially liked figure 4 (below):

They used the appearance of different groups at different parts of the season to identify "poineer", early, mid, late, climax and "generalist" microbes.  Although the list of sequences or microbes in each of these groups isn't provided, they go on to state: “Late” group members peaked when flowers had been open for 3 days and included a high abundance of Lactobacillusand Acetobacter taxa, whose occurrences aligned with previously reported conditions of flower decomposition by yeast"  I was surprised that beyond a casual mention in the paper's discussion, there was no attempt to probe whether or not the communities found on flowers resembled those found associated with pollinators.

As it turns out, the genera Lactobacillus and Acetobacter are also associated with sweat bees, honey bees and bumble bees (see http://aem.asm.org/content/early/2013/01/03/AEM.03681-12.abstract and http://www.biomedcentral.com/1471-2180/12/221).  I downloaded the entire apple flower dataset from MG-RAST and checked to see if any would affiliate with the bee specific sequences (using a 60% threshold).  Here is a quick table below, sorted alphabetically by "family" level. Out of a total of ~35,000 sequences, a subset (~5,600) have been previously found in bees. 

Family	Count
Acetobacetraceae	9
Acetobacteraceae	663
Actinomycetales	713
Actinomycetales_incertae_sedis	1271
alpha-1	53
alpha-2.2	11
alpha-2.2	47
Anaplasmataceae	1
beta	6
Bifidobacteriaceae	54
Brucellaceae	288
Burkholderiaceae	66
Burkholderiales_bacterium_YT0099	92
Comamonadaceae	38
Comamonadaceae_incertae_sedis	19
Comamonodaceae_incertae_sedis	194
Enterobacteriaceae	636
firm-4	3
gamma-1	3
Lactobacillaceae	293
Moraxellaceae	55
Neisseriaceae	19
Oxalobacteraceae	520
Staphylococcaceae	560

This picture changes when you classify the dataset using a more "complete" training set - the SILVA+Bees training set we developed (http://www.biomedcentral.com/1471-2180/12/221).  Sequences that found homes in the honey bee specific training set alone (within the honey bee specific clades), now find homes elsewhere and very few remain (highlighted below in red).  It's also worth mentioning that one major conclusion of the paper -- the "novel" sequences uncovered by the experiment -- was true in my hands as well. A very large fraction of the sequences  (~16,000) were "unclassified" at the family level, even at a 60% threshold using the NBC-RDPII and the SILVA + honey bee specific training set (see classifications below).  

Acetobacteraceae	821
Acholeplasmataceae	1
Acidimicrobiaceae	1
Acidothermaceae	5
Actinomycetaceae	16
Actinomycetales_incertae_sedis	1
Actinospicaceae	1
Actinosynnemataceae	1
Aerococcaceae	18
Aeromonadaceae	5
Alcaligenaceae	55
Anaplasmataceae	3
Aurantimonadaceae	15
Bacillaceae	344
Bacillales_incertae_sedis	7
Bacteriovoracaceae	29
Bdellovibrionaceae	19
Beijerinckiaceae	28
beta	4
Bifidobacteriaceae	1
Bradyrhizobiaceae	118
Brevibacteriaceae	6
Brucellaceae	8
Burkholderiaceae	89
Burkholderiales_bacterium_YT0099	3
Burkholderiales_incertae_sedis	91
Caldilineaceae	2
Campylobacteraceae	5
Carnobacteriaceae	13
Caryophanaceae	1
Caulobacteraceae	322
Cellulomonadaceae	90
Chitinophagaceae	1357
Clostridiaceae	113
Comamonadaceae	264
Conexibacteraceae	112
Coriobacteriaceae	10
Corynebacteriaceae	79
Cryomorphaceae	8
Cyclobacteriaceae	127
Cystobacteraceae	17
Cytophagaceae	1357
Deinococcaceae	696
Dermabacteraceae	3
Dermatophilaceae	32
Desulfovibrionaceae	1
Dietziaceae	21
Enterobacteriaceae	570
Enterococcaceae	24
Erysipelotrichaceae	69
Erythrobacteraceae	288
Eubacteriaceae	6
Flammeovirgaceae	12
Flavobacteriaceae	795
Fusobacteriaceae	2
gamma-1	1
Gemmatimonadaceae	236
Geobacteraceae	9
Geodermatophilaceae	75
Glycomycetaceae	1
Gracilibacteraceae	6
Haliangiaceae	1
Holophagaceae	4
Hydrogenophilaceae	3
Hyphomicrobiaceae	138
Hyphomonadaceae	13
Iamiaceae	161
Ilumatobacter	19
Incertae_Sedis_XI	33
Incertae_Sedis_XII	3
Incertae_Sedis_XIII	3
Incertae_Sedis_XIV	7
Incertae_Sedis_XVIII	2
Intrasporangiaceae	386
Kineosporiaceae	44
Lachnospiraceae	60
Lactobacillaceae	422
Leptotrichiaceae	5
Leuconostocaceae	21
Methylobacteriaceae	188
Methylocystaceae	3
Methylophilaceae	12
Microbacteriaceae	340
Micrococcaceae	178
Micrococcineae_incertae_sedis	5
Micromonosporaceae	23
Moraxellaceae	190
Mycobacteriaceae	31
Myxococcaceae	9
Nakamurellaceae	23
Nannocystaceae	18
Neisseriaceae	15
Nitriliruptoraceae	2
Nitrosomonadaceae	27
Nitrospiraceae	15
Nocardiaceae	45
Nocardioidaceae	406
Nocardiopsaceae	11
Oxalobacteraceae	396
Paenibacillaceae	90
Parachlamydiaceae	9
Pasteurellaceae	4
Peptococcaceae	11
Peptostreptococcaceae	117
Phyllobacteriaceae	110
Planococcaceae	162
Polyangiaceae	38
Promicromonosporaceae	1
Propionibacteriaceae	39
Pseudomonadaceae	21
Pseudonocardiaceae	92
Rhizobiaceae	42
Rhodobacteraceae	350
Rhodobiaceae	3
Rhodocyclaceae	38
Rhodospirillaceae	57
Rhodothermaceae	30
Rickettsiaceae	30
Rubrobacteraceae	4
Ruminococcaceae	35
Sanguibacteraceae	57
Saprospiraceae	17
Sinobacteraceae	86
Sneathiellaceae	1
Solirubrobacteraceae	44
Sphaerobacteraceae	10
Sphingobacteriaceae	401
Sphingomonadaceae	1673
Spirochaetaceae	2
Sporichthyaceae	4
Staphylococcaceae	19
Streptococcaceae	42
Streptomycetaceae	26
Streptophyta	28
Streptosporangiaceae	3
Syntrophobacteraceae	1
Syntrophomonadaceae	2
Thermoactinomycetaceae	77
Thermoanaerobacteraceae	7
Thermomonosporaceae	5
Trueperaceae	2107

Is this a surprising result? If pollinator-affiliated sequences were found in abundance on the flowers you could explain it by the simple fact that flowers are visited by these insects and therefore bacteria from bees might be shed onto the flower parts.  The relative paucity suggests instead that these bacteria, although inoculated into the environment by the bees, are specific to their hosts and perhaps do not thrive on the flower.