LWT - Food Science and Technology 134 (2020) 110218

Contents lists available at ScienceDirect

LWT
journal homepage: www.elsevier.com/locate/lwt

The complex community structures and seasonal variations of

psychrotrophic bacteria in raw milk in Heilongjiang Province, China
Xinyan Yang 1, Xiaojie Guo 1, Weipeng Liu , Yazhen Tian , Pingping Gao , Yuwei Ren ,
Wei Zhang , Yujun Jiang *, Chaoxin Man **
Key Laboratory of Dairy Science, Ministry of Education, Department of Food Science, Northeast Agricultural University, Harbin, 150030, China

A R T I C L E I N F O A B S T R A C T

Keywords: Psychrotrophic bacteria are the dominant organisms in refrigerated raw milk, which may have potentially
Psychrotrophic bacteria adverse effects on the quality and shelf-life of dairy products. In this study, we investigated the concentration and
Raw milk microbial composition of psychrotrophic bacteria in raw milk from Heilongjiang Province (China) in winter and
Bacterial diversity
summer using traditional cultivation and single molecule real-time (SMRT) sequencing methods. The results
Single molecule real-time sequencing
showed that the mean counts of psychrotrophic bacteria were 3.73 log CFU/mL based on traditional cultivation
method. The isolated psychrotrophic bacteria (45 genera, 120 species) showed high diversity. Pseudomonas,
Lactococcus, Acinetobacter, Chryseobacterium and Staphylococcus were the most frequently isolated genera.
Moreover, 2 potential novel genera and 3 unknown species were discovered. The SMRT sequencing method
could accurately determine the relative abundance of psychrotrophic bacteria in raw milk. Brevundimonas
(10.1%), Janthinobacterium (10.0%), Acinetobacter (8.9%), Sphingomonas (8.5%) and Enterococcus (8.1%), as the
predominant genera, were present in all raw milk samples. The concentration and community structure of
psychrotrophic bacteria varied significantly relying on the season. This work contributes to attracting dairy
manufacturers’ attention to the contamination of psychrotrophic bacteria in raw milk and providing support for
the control of psychrotrophic bacteria.

1. Introduction Samaržija, Zamberlin, & Pogačić, 2012).

Several studies have reported that the most abundant psychrotrophic
The quality of raw milk is a key factor affecting the whole dairy bacteria are commonly Gram-negative genera, including Pseudomonas,
processing chain (Zucali et al., 2011). Raw milk, due to its high nutri­ Acinetobacter, Flavobacterium, Sphingobacterium and Serratia. Among
tional value and susceptibility to contamination, may promote the Gram-positive genera, Lactococcus, Aerococcus, Bacillus, Kurtha and
growth of microorganisms during handling, transportation and pro­ Staphylococcus are also detected in raw milk samples (Ercolini, Russo,
cessing (Champagne et al., 1994). Although the cold chain system is Ferrocino, & Villani, 2009; Júnior et al., 2018). The composition of the
widely used in the dairy industry, it is difficult to hinder the growth of psychrotrophic bacterial microbiota in raw milk is contingent on several
psychrotrophic bacteria in raw milk (Vithanage et al., 2017). Psychro­ factors, such as seasonality, storage conditions, geographical areas and
trophic bacteria are ubiquitous organisms capable of growing at below farm management practices (Giannino, Marzotto, Dellaglio, & Feligini,
7 ◦ C, in spite of their high optimum growth (Jonghe et al., 2011). Under 2009; Yamazi, Moreira, Cavicchioli, Burin, & Nero, 2013). Thus, it is
refrigerated temperatures, psychrotrophic bacteria continuously secrete vital to investigate the diversity of psychrotrophic bacteria in order to
thermostable extracellular proteases and lipases, which may cause se­ establish targeted control technology.
vere quality problems, such as the reduction of nutritional value, the In this work, raw milk samples were collected from 10 different
formation of bitter or rancid flavor and the gelation of UHT milk, commercial dairy farms in Heilongjiang Province (China) during the
thereby shortening the shelf-life of dairy products (Mateos et al., 2015; winter and summer periods. Based on traditional cultivation and single

* Corresponding author. .
** Corresponding author. .
E-mail addresses: [email protected] (Y. Jiang), [email protected] (C. Man).
1
These authors contributed equally to this work.

https://doi.org/10.1016/j.lwt.2020.110218
Received 19 June 2020; Received in revised form 8 September 2020; Accepted 10 September 2020
Available online 14 September 2020
0023-6438/© 2020 Elsevier Ltd. All rights reserved.
X. Yang et al. LWT 134 (2020) 110218

Fig. 1. Distribution of sampling locations in Heilongjiang Province, China. Farm’s average distance is about 135 km.

Fig. 2. Psychrotrophic bacteria counts of raw milk samples collected from 10 different dairy farms in winter and summer. Each bar represents mean ± SD, n = 3. *P
< 0.05, * *P < 0.01.

2
 Table 1

X. Yang et al.
Identification and distribution of psychrotrophic bacteria isolated from raw milk samples collected from Heilongjiang Province, China in winter and summer.
Genera (Total number) Species GramStaina Total No. of isolatesb Genera (Total Species GramStaina Total No. of isolatesb
number)
Winter Summer Winter Summer

Pseudomonas (153) fluorescens – 31 1 (1); 2 (1); 3 (3); 4 (3); 7 (2); 2 (2); 3 (2); 5 (6); 6 (1); 10 (4) Streptococcus (9) parauberis + 8 10 (2) 10 (6)
8 (6)
lurida – 18 7 (7); 8(3) 2 (8) equinus + 1 1 (1)
psychrophila – 11 3 (2); 6 (2) 5 (4); 8 (2); 9 (1) Chryseobacterium (53) carnipullorum – 10 5 (1); 10 (2) 1 (3); 4 (2);
5 (1); 8 (1)
gessardii – 10 3 (1); 6 (1); 7 (5); 8 (2) 2 (1) yeoncheonense – 7 9 (7)
azotoformans – 10 8 (1) 1 (1); 2 (1); 3 (2); 5 (1); 9 (3); halperniae – 4 3 (2); 6 (2)
10 (1)
fragi – 8 5 (2); 6 (2) 2 (4) scophthalmum – 4 4 (2); 6 (2)
jessenii – 7 5 (3); 8 (3) 5 (1) indologenes – 4 7 (2); 8 (2)
putida – 7 1 (2); 3 (2); 8 (1) 2 (1); 4 (1) piscium – 3 6 (1); 10 (2)
lundensis – 6 1 (1); 5 (1) 2 (3); 3 (1) aahli – 3 2 (2) 4 (1)
cedrina – 5 3 (1) 2 (2); 3 (2) joostei – 3 6 (2) 9 (1)
fulva – 4 4 (4) balustinum – 3 8 (3)
brenneri – 3 7 (2) 10 (1) bovis – 2 4 (2)
monteilii – 3 4 (3) jeonii – 2 9 (2)
silesiensis – 2 1 (2) chaponense – 2 1 (2)
protegens – 2 3 (2) xixisoli – 2 4 (2)
rhizosphaerae – 2 6 (2) haifense – 2 10 (2)
poae – 2 8 (2) anthropi – 2 10 (2)
helmanticensis – 2 5 (2) Stenotrophomonas (9) rhizophila – 7 3 (2); 4 (2) 4 (2); 6 (1)
japonica – 1 8 (1) maltophilia – 1 2 (1)
synxantha – 1 8 (1) pictorum – 1 7 (1)
umsongensis – 2 9 (2) Rothia (15) endophytica + 8 2 (2) 7 (6)
baetica – 2 4 (2) marina + 7 7 (6); 8 (1)
3

extremorientalis – 3 2 (2); 5 (1) Pseudoclavibacter (1) helvolus + 1 2 (1)

anguilliseptica – 3 1 (2); 3(1) Rhodococcus (6) erythropolis + 6 2 (1); 3 (1) 1 (1); 4 (1);
9 (2)
mandelii – 2 2 (2) Brevundimonas (4) vesicularis – 3 2 (1) 8 (1); 9 (1)
veronii – 3 5 (1) 5 (2) diminuta – 1 9 (1)
caeni – 1 1 (1) Janthinobacterium (24) lividum – 11 3 (1); 7 (1); 8 5 (1); 10 (5)
(3)
parafulva – 1 6 (1) svalbardensis – 7 4 (1); 7 (2); 8 1 (2)
(2)
migulae – 1 6 (1) agaricidamnosum – 6 8 (5) 10 (1)
Acinetobacter (59) albensis – 24 1 (2); 5 (2); 7 (1); 9 (2); 10 2 (2); 8 (2) Aerococcus (1) urinaeequi + 1 2 (1)
(13)
guillouiae – 18 1 (3); 3 (3); 4 (4); 5 (1) 4 (3); 5 (4) Empedobacter (2) brevis – 2 5 (1) 1 (1)
johnsonii – 12 1 (1); 3 (3); 9 (1) 3 (1); 5 (3); 6 (2); 7 (1) Yersinia (2) massiliensis – 1 3 (1)
lwoffii – 5 1 (2); 2 (1); 5 (1); 9 (1) intermedia – 1 6 (1)
Lactococcus (63) raffinolactis + 27 4 (4); 6 (1); 7 (1); 10 (4) 3 (4); 5 (6); 6 (3); 8 (4) Carnobacterium (9) maltaromaticum + 9 6 (5); 7 (4)
piscium + 24 4 (2); 5 (12); 9 (6); 10 (3) 1 (1) Massilia (4) aurea – 4 6 (1); 10 (1) 8 (2)
lactis + 9 4 (7) 3 (2) Enterococcus (14) faecalis + 14 3 (4); 9 (10)
garvieae + 1 1 (1) Clostridium (12) bifermentans + 12 3 (9); 9 (3)
laudensis + 2 6 (2) Pantoea (5) agglomerans – 2 7 (2)
Staphylococcus (40) equorum + 24 2 (10); 3 (3) 1 (5); 3 (1); 4 (5) septica – 2 9 (2)
succinus + 8 1 (4); 4 (2); 6 (2) ananatis – 1 1 (1)
xylosus 4 1 (1); 2 (1) 1 (2) Kurthia (3) gibsonii 3 6 (2); 8 (1)

LWT 134 (2020) 110218

+ +
fleurettii + 4 1 (2); 6 (1); 8 (1) Serratia (3) liquefaciens – 3 5 (3)
Rahnella (8) aquatilis – 8 1 (3); 6 (5) Leuconostoc (3) lactis + 3 6 (3)
Sphingobacterium (5) faecium – 3 10 (1) 6 (1); 8 (1) Cytophaga (1) aurantiaca – 1 1 (1)
kitahiroshimense – 2 1 (1); 10 (1) Microterricola (1) gilva – 1 1 (1)
Bacillus (6) amyloliquefaciens + 4 4 (2); 6 (1); 10 (1) Lysinibacillus (2) fusiformis + 2 1 (2)
velezensis + 2 1 (2) Kluyvera (1) intermedia – 1 5 (1)
(continued on next page)
 X. Yang et al. LWT 134 (2020) 110218

molecule real-time (SMRT) sequencing methods, the concentration and

Summer community structure of psychrotrophic bacteria and their variations

5 (2); 6 (2)

9 (1); 10 (1)

7 (6); 10 (2)
with seasonality in raw milk were investigated. The purpose of our
No. of isolatesb

8 (1)
8 (1)
3 (1)
6 (1)
3 (1)

9 (1)
6 (2)
research was to comprehensively analyze the microbial composition and

10 (2)
7 (1)
seasonal variations of psychrotrophic bacteria in raw milk of dairy
farms, which provides information on controlling dominant psychro­
Winter

trophic bacteria in raw milk.

1 (1); 2

1 (1); 2
1 (1)
2 (3)

(1)

(2)
2. Materials and methods
Total

2.1. Sample collection

11
1
1
1
1
1
4
1
2
1
5
3

2
GramStaina

Twenty raw milk samples from bulk milk tanks were collected from
10 different commercial dairy farms (location as shown in Fig. 1) in the
Heilongjiang Province of northeast China in winter (November 2018)
+

+
+

+
–

–
–
–
–
–
–
–

and summer (June 2019), respectively. All farms were in compliance

with Sanitary Specification for Dairy Farm (GB 16568–2006). Samples
per farm were transported to the laboratory under 4 ◦ C and analyzed
Species

ornithinolytica
michiganensis

immediately.
amnigenus
submarina

testaceum
amnigena
aquaticus

sakazakii

osloensis
foliorum
oxydans
aerolata
oxytoca

2.2. Psychrotrophic bacteria counts (PBC) and isolation

The samples were serially diluted five times with 0.1% aseptic
Microbacterium (10)

peptone salt solution (w/v), followed by plating on sterile Psychro­

Genera (Total

Sphingomonas (1)
Enterobacter (4)
number)

Deinococcus (1)

Cronobacter (1)

Moraxella (11)

trophic Bacterium Count Agar consisting of 0.5% tryptone, 0.25% yeast

Raoultella (1)
Klebsiella (2)

Lelliottia (2)
Devosia (1)

extract, 0.1% glucose, 0.1% skimmed milk powder and 1.5% agar (MPC,
Qingdao Hope Bio-Technology, China) in triplicate. The plates were
incubated at 7 ◦ C for 7–10 d to count aerobic psychrotrophic bacteria
(Yuan et al., 2017). A total of 25–35 colonies with different morphology
were randomly selected from countable plates, and then cultured on
MPC plates for three consecutive times to obtain pure cultures. More­
3 (1); 6 (2); 7 (1); 10 (2)

over, 600 isolates were subjected to Gram staining prior to 16S rRNA
Summer

gene sequencing and stored in 25% glycerol (v/v) at − 80 ◦ C for subse­

quent analysis.
1 (1); 5 (1)

2.3. DNA extraction and PCR amplification

No. of isolatesb

10 (1)
6 (1)

5 (2)

All bacterial were grown at 25 ◦ C for 24 h in Luria Bertani (LB,

Qingdao Hope Bio-Technology, China) broth. Genomic DNA was
extracted from 2 mL of bacteria cultures using Gen Elute Bacterial
Genomic DNA Kit (TianGen, Beijing, China) according to the manufac­
turer’s instructions. The pellet was suspended in TE buffer and then
Winter

stored at − 20 ◦ C until further use. All extracts were amplified with 16S
rRNA gene sequences using universal primers 27F (5′ -AGAGTTT­
1 (4); 6 (2)

2 (2); 4 (2)

5 (3); 9 (2)

GATCCTGGCTCAG-3′ ) and 1492R (5′ -TACGGCTACCTTGTTACGACTT-

10 (7)
2 (2)
6 (1)

1 (2)
5 (2)

5 (2)

3′ ). PCR amplification was performed as described by Lu et al. (2014).

The amplified products were verified by 1% agarose gel electrophoresis
Total

(AGE), and then sent to Sangon Biotech Co., Ltd. (Shanghai, China) for
595
10
6
2
1
1
1
2
2
2

7
5

2
2

sequencing analysis.
The total genomic DNA of each raw milk sample (approximately 50
GramStaina

mL) was extracted according to the manufacturer’s instructions of the

OMEGA DNA isolation kit (Omega, D5625-01, USA). The concentration
+
+
+
–
–
–
–
–

–
–
–

–
–

of DNA samples was determined using a NanoDrop ND-1000 spectro­

photometer (Thermo Fisher Scientific, Waltham, MA, USA). The nearly
psychrolactophilus

full-length bacterial 16S rRNA gene sequences (about 1.5 kb) were
Species

alimentarius

oncorhynchi

plurextorum
caseolyticus

frigidimaris
frigidarium
arilaitensis

amplified by PCR using 27F (5′ -AGAGTTTGATCMTGGCTCAG-3′ ) and

maritimus
fulvigenes
sanguinis
pulmonis

alpinus

1492R (5′ -ACCTTGTTACGACTT-3′ ) primers. Sample-specific 16 bp

Farm (Number of isolates).

barcodes were incorporated into the primers. PCR amplification was

– Negative; + Positive.

carried out as described by She, Cai, and Liu (2018). The products were
Genera (Total number)

purified with Agencourt AMPure Beads (Beckman Coulter, Indianapolis,

Table 1 (continued )

IN, USA) and quantified using the PicoGreen dsDNA Assay Kit (Invi­
Flavobacterium (16)
Psychrobacter (11)

Macrococcus (10)

trogen, Carlsbad, CA, USA). All PCR primers were synthesized by

Arthrobacter (6)

SinoGenoMax Co., Ltd. (Beijing, China).

Total

b
a

4
X. Yang et al. LWT 134 (2020) 110218

Fig. 3. Rarefaction curves of Shannon’s diversity index of different samples (A) and two season groups (B). S: summer; W: winter; S1–S10 and W1–W10: raw milk
samples collected in Farm 1–10 in summer and winter, respectively.

Fig. 4. Classification tree of psychrotrophic bacteria in raw milk samples based on GraPhlAn.

5
X. Yang et al. LWT 134 (2020) 110218

Fig. 5. The relative abundance of psychrotrophic bacteria at the genus-level (A) and species-level (B) in the winter and summer groups.

6
X. Yang et al. LWT 134 (2020) 110218

Fig. 6. Linear discriminant analysis of the community structure of psychrotrophic bacteria in raw milk samples. Green or red nodes represent the OTUs with sig­
nificant differences. The node size corresponds to the relative abundance of the OTUs. The letter identifies the name of the OTUs. (For interpretation of the references
to colour in this figure legend, the reader is referred to the Web version of this article.)

7
X. Yang et al. LWT 134 (2020) 110218

Fig. 7. The abundance distribution of the top 20 OTUs with significant variance at species-level between seasons.

Fig. 8. Nonmetric Multidimensional Scaling (NMDS) for the community structure of psychrotrophic bacteria in raw milk samples collected from winter and summer.

8
X. Yang et al. LWT 134 (2020) 110218

Fig. 9. The phylogenetic tree of potential novel genera and species with reference strains based on 16S rDNA genes.

generated by averaging 100 evenly resampled OTU subsets under the

Table 2
90% of the minimum sequencing depth to minimize the difference of
Potential novel genera and species isolated from raw milk samples.
sequencing depth for further analysis.
Genera/Species No. of Related species in Identity Distribution
isolates GenBank database (%)
2.5. Diversity and statistical analysis
Gen.nov.sp.nov 1 Moraxella osloensis 93.93 W9
(1) CCUG 350a
The data were expressed as mean value ± standard deviation (SD).
Gen.nov.sp.nov 1 Moraxella osloensis 93.93 W9
(2) CCUG 350a The results of psychrotrophic bacteria counts between winter and
Pedobacter sp.nov 1 Pedobacter agri 96.98 S8 summer were calculated using a one-sample t-test by SPSS statistic 20.0
(1) PB92a software (SPSS, Chicago, IL, USA). Values of P < 0.05 were considered as
Pedobacter sp.nov 1 Pedobacter agri 96.90 S8 significant.
(2) PB92a
Chryseobacterium 1 Chryseobacterium 96.90 S4
The BLAST algorithm in the GenBank database was conducted on the
sp.nov (1) hungaricum DSM identification of the 16S rRNA sequences.
19684a QIIME and R software packages (v3.2.0) were mainly used for
a
Type strain.
sequence data analysis. For alpha diversity analysis, rarefaction curves
were generated to evaluate the adequacy of the sequencing depth and
compared the diversity of each sample to a certain extent. The compo­
2.4. SMRT sequencing analysis
sition of psychrotrophic bacterial microbiota was investigated by sub­
mitting the psychrotrophic bacteria database to the raw milk microbiota
The purified PCR products were sequenced by Pacbio Sequel plat­
database. The taxonomy compositions and abundances were visualized
form at Personal Biotechnology Co., Ltd. (Shanghai, China). The PacBio
using GraPhlAn. Taxa abundances at the species level were statistically
circular consensus sequencing (CCS) was used to ensure that the pre­
compared among groups by Metastats (White, Nagarajan, & Pop, 2009).
diction accuracy was not less than 90%, while the filtration was at least 3
Linear discriminant analysis effect size (LEfSe) was performed to detect
full passes. The sequencing data were processed by the Quantitative
differentially abundant taxa across groups using the 3 default parame­
Insights Into Microbial Ecology (QIIME, v1.8.0) as previously described
ters (Segata et al., 2011). Beta diversity analysis was carried out to
(Caporaso et al., 2010). Briefly, the original sequences were assigned to
determine the diversity variation of psychrotrophic bacteria between
each sample according to the primer and the barcode information and
groups via nonmetric multidimensional scaling (NMDS).
identified to obtain valid sequences. After trimming off the interrogative
sequence, the remaining high-quality sequences of each sample were
2.6. Nucleotide sequence accession numbers
clustered into the operational taxonomic unit (OTU) with a similarity
cut-off value of 97% using the UCLUST algorithm (Edgar, 2010). The
Representative nucleotide sequences of closely related species iso­
most abundant sequence in each OTU was selected as the representative
lated from this study were deposited in GenBank under accession
sequence. OTU taxonomic classification of the representative sequence
numbers MN758761 to MN758883.
was compared with the NCBI 16S ribosomal RNA Database. OTUs with
abundance values lower than 0.001% of the total sequences in all
samples were removed. An averaged, rounded rarefied OTU table was

9
X. Yang et al. LWT 134 (2020) 110218

3. Results 3.4. Variance of psychrotrophic bacterial diversity between seasons

3.1. Analysis of psychrotrophic bacteria Based on LEfSe analysis, a total of 103 OTUs with significant dif­
ferences between the summer and winter groups were observed (Fig. 6).
PBC was conducted to investigate the level of psychrotrophic bac­ Among them, 53 OTUs had high abundance in the summer group and 50
teria in raw milk. As shown in Fig. 2, the concentration of psychro­ OTUs had high abundance in the winter group. Additionally, Metastats
trophic bacteria ranged from 2.80 to 5.58 log CFU/mL, and the mean analysis was also used to determine the seasonal variation of OTUs. In
value was 3.73 log CFU/mL (winter 3.54, summer 3.92). The concen­ the mass, 3 phylum-level, 22 genus-level and 44 species-level OTUs were
tration of psychrotrophic bacteria in summer was significantly higher significantly different between summer and winter groups. Fig. 7 illus­
than that in winter (P < 0.05). trates the top 20 OTUs with significant variance at species-level between
groups. For instance, E. faecalis exhibited the highest abundance
3.2. Identification of culturable psychrotrophic bacteria (15.6%) in the raw milk microbiota sampled from summer. However, it
contributed only a minor part (0.6%) of the winter group.
A total of 600 psychrotrophic bacteria colonies with different The differences in the community structure of psychrotrophic bac­
morphology were isolated from 20 raw milk samples via cultivation- teria between seasons were evaluated by NMDS. As shown in Fig. 8, the
based methods. The results of Gram staining indicated that 395 iso­ summer group was clearly separated from the winter group, indicating
lates were Gram-negative bacteria, accounting for a relatively high that the community structure of psychrotrophic bacteria was signifi­
proportion of 65.8%. Meanwhile, all isolates were further identified cantly different between the two seasons.
using 16S rRNA gene sequencing. As listed in Table 1, all psychrotrophic
bacteria isolates were classified into 45 genera and 120 species. The 3.5. Potential novel genera and species
summer group represented 40 genera (83 species), while the winter
group included 27 genera (75 species), indicating that the psychro­ In virtue of the low sequence similarity percentage (<98%) of the
trophic bacterial communities in the summer group were more complex 16S rRNA gene sequences of 5 psychrotrophic bacteria in the GenBank
than that in the winter group. database, they were not able to be classified into species level. The
The genera with the highest isolation frequency were Pseudomonas, evolutionary relationships of these strains were revealed by the
Lactococcus, Acinetobacter, Chryseobacterium and Staphylococcus. Mean­ neighbour-joining method (Fig. 9). The results of sequence similarity
while, P. fluorescens, L. raffinolactis, A. albensis, S. equorum and L. piscium analysis by comparing with the sequence of reference strains of related
were the predominant isolated species. Pseudomonas was detected in 18 species are shown in Table 2. Two isolates showed sequence identity
of 20 samples, which included 29 species, demonstrating the highest percentages below 95%, which might be the potential novel genera. The
biodiversity. Among this genus, P. fluorescens was isolated from 11 other three potential novel species might belong to Chryseobacterium sp.
samples. Detailed information on strain distribution is shown in Table 1. and Pedobacter sp. (the identity < 98%). The characteristics of these
strains may require further exploration in phenotype, physiology and
3.3. Composition of psychrotrophic bacterial microbiota biochemistry.

The bacterial diversity of 20 raw milk samples was assessed by 4. Discussion

Pacbio SMRT sequencing technology. The rarefaction curves are shown
in Fig. 3, suggesting that the sequencing depth was enough to predict the The contaminating psychrotrophic bacteria in raw milk can produce
diversity within the sample. Moreover, the species diversity of the heat-resistant proteolytic and lipolytic enzymes, which can remain
summer group was higher than that of the winter group. active after heat treatment, thereby potentially affecting the quality and
A classification tree was constructed based on GraPhlAn. Fig. 4 re­ shelf-life of dairy products (Chen, Daniel, & Coolbear, 2003). Hence, it is
veals that Proteobacteria, Firmicutes and Bacteroidetes were the dominant necessary to investigate the psychrotrophic bacteria in raw milk in order
phyla, accounting for 96.6% of the obtained bacterial communities. to control the contamination of psychrotrophic bacteria from the source.
Brevundimonas, Janthinobacterium and Acinetobacter, as the commonly Studies have reported that season is one of the important factors
predominant genera, were present in all raw milk samples. B. vesicularis affecting the community structure of psychrotrophic bacteria in raw
was the dominant species. milk (Yuan, Sadiq, Burmolle, Wang, & He, 2019; Zhang, Palmer, Teh,
The composition of psychrotrophic bacterial microbiota was inves­ Biggs, & Flint, 2019). However, so far, there are few studies on the
tigated by submitting the psychrotrophic bacteria database to the raw differences between the seasons of psychrophilic bacterial communities
milk microbiota database. The genus-level with the highest relative in China. The objective of this current study provides key information
abundance is shown in Fig. 5A. Brevundimonas (10.1%), Janthinobacte­ about the community structures and seasonal variations of psychro­
rium (10.0%), Acinetobacter (8.9%), Sphingomonas (8.5%) and Entero­ trophic bacteria in raw milk in China.
coccus (8.1%) were the predominant genera. The summer group was In 20 raw milk samples, the values of PBC ranged from 2.80 to 5.58
mainly comprised of Enterococcus (summer (S): 15.6%; winter (W): log CFU/mL, with a mean of 3.73 log CFU/mL. It has been reported that
0.6%), Enterobacter (S: 13.7%; W: 1.3%) and Janthinobacterium (S: 9.1%; the difference in PBC may be related to the milking environment, animal
W: 11.0%). The winter group was mainly comprised of Brevundimonas health, packaging and handling (Burke, O’Dwyer, Southern, & Adley,
(W: 14.6%; S: 5.7%), Acinetobacter (W: 13.2%; S: 4.6%) and Sphingo­ 2017). Thus, PBC is considered to be an important indicator that de­
monas (W: 13.2%; S: 3.7%). At the species level, in the summer group, termines the quality of raw milk and final dairy products (Yuan et al.,
the highest proportion strain was E. faecalis (S: 15.6%; W: 0.6%), fol­ 2017). According to Matta, Punj, and Kanwar (1997) and Birkeland,
lowed by E. amnigenus (S: 13.7%; W: 1.3%) and S. kitahiroshimense (S: Stepaniak, and Sorhaug (1985), the PBC generally required to initiate
7.0%; W: 1.6%). The winter group was dominated by B. vesicularis (W: spoilage in milk is about 106 CFU/mL. These results indicated that the
13.9%; S: 5.1%), followed by S. aerolata (W: 13.2%; S: 3.7%) and raw milk produced in the sampled farms in this study had high micro­
A. guillouiae (W: 10.4%; S: 0.8%) (Fig. 5B). It was observed that the biological quality because only the PBC of the sample from Farm 5 was
dominant strains and community structures of psychrotrophic bacteria higher than 105 CFU/mL. In this study, it also can be observed that PBC
were apparently varied relying on the season. in summer was significantly higher than that in winter, which is
consistent with the previous reports (Vithanage et al., 2016; Zucali et al.,
2011). The possible reason is that the environment humidity and high
temperature in summer promoted the growth of microorganisms.

10
X. Yang et al. LWT 134 (2020) 110218

Overall, the number of psychrotrophic bacteria in the evaluated raw reason may be that it is suitable for growth at a higher temperature. The
milk samples was low, so there was no risk of milk spoilage caused by genera Brevundimonas, Sphingomonas and Acinetobacter possessed rela­
psychrotrophic bacteria in the sampled farms. tively high abundance in winter because of the lower optimal temper­
The traditional culture-based method, as a gold standard, is widely atures. In addition, during summer and winter, the humidity of the
used to isolate psychrotrophic bacteria. 16S rRNA gene sequencing the environment and feeding modes of cows may lead to the variation of the
most commonly used identification system, which determines species or psychrotrophic bacterial community (Zhang, Huo, Zhu, & Mao, 2015).
subspecies level identification (Vithanage, Yeager, Jadhav, Palombo, & Since seasonal changes are inevitable, it is essential to improve farm
Datta, 2014). In this study, through 16S rRNA gene sequencing, all hygiene and implement the means to monitor microbial contamination
isolates of psychrotrophic bacteria obtained by the culture-based in terms of different seasons.
method were classified into 45 genera and 120 species, which Two potential novel genera and three unknown species were found
revealed the complex communities of psychrotrophic bacteria in raw in raw milk samples in this study. This phenomenon indicated that the
milk in Heilongjiang. Three genera with high separation frequency were structure of the psychrotrophic bacterial community still requires to be
identified, namely Pseudomonas, Lactococcus and Acinetobacter, isolated further studied. In virtue of low relative abundance of some psychro­
from 18, 12 and 15 raw milk samples, respectively. Similar results have trophic bacteria or underestimation of certain novel species in raw milk,
been reported by previous studies (Hahne, Isele, Berning, & Lipski, it is challenging to obtain the complete profile and spoilage potential of
2019; Li et al., 2018; Neubeck et al., 2015). Pseudomonas becomes the psychrotrophic bacteria. We need to dig into the composition of psy­
most dominant psychrotrophic bacterial community, which might be chrotrophic bacterial communities in raw milk, because they are the key
due to its short generation time at 0–7 ◦ C (Leser, Boye, & Hendriksen, factors that determine the quality of raw milk, and further affect the
1995). Species in this genus, especially P. fluorescens play an important quality and shelf-life of dairy products. Therefore, the high diversity of
role in causing the spoilage of milk and dairy product (Ercolini et al., psychrotrophic bacteria in this study indicated that it is still necessary to
2009). In this study, P. fluorescens was the dominant species, isolated focus on the contamination of psychrotrophic bacteria in raw milk. This
from 11 out of 20 samples. The low isolation frequency of P. fluorescens can provide important information as a reference for dairy manufac­
may be due to the good microbiological quality of the milk produced in turers to control psychrotrophic bacteria.
the farms sampled, as confirmed by the low PBC values. Lactococcus and
Acinetobacter are known psychrotrophic genus with high isolation rates 5. Conclusion
in raw milk (Gurung et al., 2013; Júnior et al., 2018). In addition, the
psychrotrophic bacterial communities in summer were more complex This study demonstrated the community structure of psychrotrophic
than that in winter, suggesting that psychrotrophic bacterial commu­ bacteria in raw milk collected from 10 dairy farms through cultivation-
nities appeared to be season dependent. The main advantage of the based and SMRT sequencing methods. A total of 45 genera and 120
traditional culture-based method is that it can provide strains for further species were isolated, indicating that the relative diversity of psychro­
research on psychrotrophic bacteria. However, this method is not suf­ trophic bacteria was high. Two potential novel genera and three un­
ficient to determine the relative abundance and diversity of psychro­ known species were detected. SMRT sequencing could accurately
trophic bacteria in raw milk. determine the abundance of psychrotrophic bacteria in raw milk.
In this study, in order to determine the abundance of psychrotrophic Furthermore, there were significant differences in the concentration and
bacteria more accurately, the analysis method of submitting the psy­ community structure of psychotrophic bacteria between seasons. In
chrotrophic bacteria database obtained by traditional cultivation to the summary, our research can provide information on controlling domi­
raw milk microbiota database was used for the first time. The results nant psychrotrophic bacteria in raw milk. The characteristics of the
suggested that in addition to the above-mentioned genera Pseudomonas, novel bacteria and the spoilage potential of the isolated strains require
Lactococcus and Acinetobacter, the risk of contamination of Brevundi­ to be further explored.
monas, Janthinobacterium, Sphingomonas and Enterococcus in raw milk
was also high, which was reported in previous investigations (Vithanage CRediT authorship contribution statement
et al., 2016; Yuan et al., 2017). However, based on the traditional
cultivation method, only 4 Brevundimonas, 24 Janthinobacterium, 1 Xinyan Yang: Methodology, Formal analysis, Investigation, Writing
Sphingomonas and 14 Enterococcus were isolated from 600 strains of - original draft. Xiaojie Guo: Methodology, Investigation, Writing -
psychrotrophic bacteria. In addition, the SMRT sequencing result original draft. Weipeng Liu: Software, Writing - review & editing.
showed that B. vesicularis was the dominant species, which was different Yazhen Tian: Software, Writing - review & editing. Pingping Gao:
from P. fluorescens as the predominant isolated species in the cultivation Visualization. Yuwei Ren: Validation. Wei Zhang: Data curation.
method. The genus Brevundimonas has growth requirements for specific Yujun Jiang: Conceptualization, Resources, Project administration,
vitamins, including pantothenate, biotin, and cyanocobalamin, resulting Funding acquisition. Chaoxin Man: Conceptualization, Supervision,
in growth slowly on ordinary nutrient media (Lipuma, Currie, Peacock, Resources, Funding acquisition.
& VanDamme, 2011; Ryan & Pembroke, 2018). Janthinobacterium is
proposed to grow at an optimum temperature of 25–30 ◦ C (Friedrich Declaration of competing interest
et al., 2020). These may be the main reasons for the differences in the
diversity of the two genera obtained by cultivation and SMRT The authors declare no conflict of interest.
sequencing methods. Therefore, it is reasonable to speculate that the
differences between the results of these two methods may be related to Acknowledgements
the bacteria characteristics. The results suggested that combining the
traditional cultivation method with the SMRT sequencing method is This work was supported by the National Natural Science Foundation
helpful for studying the complex community structure and diversity of of China, China (No. 31871828) and the National Key Research and
psychrotrophic bacteria in raw milk. Development Project of China, China (No. 2018YFE0120500).
In this work, psychrotrophic bacteria counts and microbial compo­
sition were significantly diverse among the seasons. These results are References
consistent with previous studies (Marchand et al., 2009; Porcellato et al.,
2018), which may be associated with optimum growth temperature. Birkeland, S. E., Stepaniak, L., & Sorhaug, T. (1985). Quantitative studies of heat-stable
proteinase from Pseudomonas fluorescens P1 by the enzyme-linked immunosorbent
There was a clear trend that the abundance of Enterococcus relied on the assay. Applied and Environmental Microbiology, 49, 382–387. https://doi.org/
seasons, most of them existed in summer and fewest in winter. The 10.1128/AEM.49.2.382-387.1985.

11
X. Yang et al. LWT 134 (2020) 110218

Burke, N., O’Dwyer, J., Southern, M., & Adley, C. C. (2017). An analysis of the microbial the stability of UHT milk. International Dairy Journal, 49, 78–88. https://doi.org/
quality of water in a milk production plant. Lebensmittel-Wissenschaft und 10.1016/j.idairyj.2015.04.008.
-Technologie- Food Science and Technology, 78, 41–46. https://doi.org/10.1016/j. Matta, H., Punj, V., & Kanwar, S. S. (1997). An immuno-dot blot assay for detection of
lwt.2016.11.058. thermostable protease from Pseudomonas sp. AFT-36 of dairy origin. Letters in
Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D., Applied Microbiology, 25, 300–302. https://doi.org/10.1046/j.1472-
Costello, E. K., et al. (2010). QIIME allows analysis of high-throughput community 765X.1997.00228.x.
sequencing data. Nature Methods, 7, 335–336. https://doi.org/10.1038/nmeth.f.303. Neubeck, V. M., Baur, C., Krewinkel, M., Stoeckel, M., Kranz, B., Stressler, T., et al.
Champagne, C. P., Laing, R. R., Roy, D., Mafu, A. A., Griffiths, M. W., & White, C. (1994). (2015). Biodiversity of refrigerated raw milk microbiota and their enzymatic
Psychrotrophs in dairy products: Their effects and their control. Critical Reviews in spoilage potential. International Journal of Food Microbiology, 211, 57–65. https://
Food Science and Nutrition, 34, 1–30. https://doi.org/10.1080/10408399409527648. doi.org/10.1016/j.ijfoodmicro.2015.07.001.
Chen, L., Daniel, R. M., & Coolbear, T. (2003). Detection and impact of protease and Porcellato, D., Aspholm, M., Skeie, S. B., Monshaugen, M., Brendehaug, J., &
lipase activities in milk and milk powders. International Dairy Journal, 13, 255–275. Mellegard, H. (2018). Microbial diversity of consumption milk during processing
https://doi.org/10.1016/s0958-6946(02)00171-1. and storage. International Journal of Food Microbiology, 266, 21–30. https://doi.org/
Edgar, R. C. (2010). Search and clustering orders of magnitude faster than BLAST. 10.1016/j.ijfoodmicro.2017.11.004.
Bioinformatics, 26, 2460–2461. https://doi.org/10.1093/bioinformatics/btq461. Ryan, M. P., & Pembroke, J. T. (2018). Brevundimonas spp: Emerging global
Ercolini, D., Russo, F., Ferrocino, I., & Villani, F. (2009). Molecular identification of opportunistic pathogens. Virulence, 9, 480–493. https://doi.org/10.1080/
mesophilic and psychrotrophic bacteria from raw cow’s milk. Food Microbiology, 26, 21505594.2017.1419116.
228–231. https://doi.org/10.1016/j.fm.2008.09.005. Samaržija, D., Zamberlin, Š., & Pogačić, T. (2012). Psychrotrophic bacteria and their
Friedrich, I., Hollensteiner, J., Schneider, D., Poehlein, A., Hertel, R., & Daniel, R. (2020). negative effects on milk and dairy products quality. Mljekarstvo, 62, 77–95. https://
First complete genome sequences of Janthinobacterium lividum EIF1 and EIF2 and doi.org/10.1186/1297-9686-44-10.
their comparative genome analysis. Genome Biology and Evolution. , Article evaa148. Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W. S., et al. (2011).
https://doi.org/10.1093/gbe/evaa148. Metagenomic biomarker discovery and explanation. Genome Biology, 12, 1–18.
Giannino, M. L., Marzotto, M., Dellaglio, F., & Feligini, M. (2009). Study of microbial https://doi.org/10.1186/gb-2011-12-6-r60.
diversity in raw milk and fresh curd used for Fontina cheese production by culture- She, Y., Cai, H. Y., & Liu, G. H. (2018). Effects of antibiotic on microflora in ileum and
independent methods. International Journal of Food Microbiology, 130, 188–195. cecum for broilers by 16S rRNA sequence analysis. Animal Science Journal, 89,
https://doi.org/10.1016/j.ijfoodmicro.2009.01.022. 1680–1691. https://doi.org/10.1111/asj.13113.
Gurung, M., Nam, H. M., Tamang, M. D., Chae, M. H., Jang, G. C., Jung, S. C., et al. Vithanage, N. R., Dissanayake, M., Bolge, G., Palombo, E. A., Yeager, T. R., & Datta, N.
(2013). Prevalence and antimicrobial susceptibility of Acinetobacter from raw bulk (2016). Biodiversity of culturable psychrotrophic microbiota in raw milk
tank milk in Korea. Journal of Dairy Science, 96, 1997–2002. https://doi.org/ attributable to refrigeration conditions, seasonality and their spoilage potential.
10.3168/jds.2012-5965. International Dairy Journal, 57, 80–90. https://doi.org/10.1016/j.
Hahne, J., Isele, D., Berning, J., & Lipski, A. (2019). The contribution of fast growing, idairyj.2016.02.042.
psychrotrophic microorganisms on biodiversity of refrigerated raw cow’s milk with Vithanage, N. R., Dissanayake, M., Bolge, G., Palombo, E. A., Yeager, T. R., & Datta, N.
high bacterial counts and their food spoilage potential. Food Microbiology, 79, 11–19. (2017). Microbiological quality of raw milk attributable to prolonged refrigeration
https://doi.org/10.1016/j.fm.2018.10.019. conditions. Journal of Dairy Research, 84, 92–101. https://doi.org/10.1017/
Jonghe, V. D., Coorevits, A., Hoorde, K. V., Messens, W., Landschoot, A. V., Vos, P. D., s0022029916000728.
et al. (2011). Influence of storage conditions on the growth of Pseudomonas species Vithanage, N. R., Yeager, T. R., Jadhav, S. R., Palombo, E. A., & Datta, N. (2014).
in refrigerated raw milk. Applied and Environmental Microbiology, 77, 460–470. Comparison of identification systems for psychrotrophic bacteria isolated from raw
https://doi.org/10.1128/aem.00521-10. bovine milk. International Journal of Food Microbiology, 189, 26–38. https://doi.org/
Júnior, J. C. R., Oliveira, A. M. D., Silva, F. D. G., Tamanini, R., Oliveira, A. L. M. D., & 10.1016/j.ijfoodmicro.2014.07.023.
Beloti, V. (2018). The main spoilage-related psychrotrophic bacteria in refrigerated White, J. R., Nagarajan, N., & Pop, M. (2009). Statistical methods for detecting
raw milk. Journal of Dairy Science, 101, 75–83. https://doi.org/10.3168/jds.2017- differentially abundant features in clinical metagenomic samples. PLoS
13069. Computational Biology, 5, Article e1000352. https://doi.org/10.1371/journal.
Leser, T. D., Boye, M., & Hendriksen, N. B. (1995). Survival and activity of Pseudomonas pcbi.1000352.
sp. strain B13(FR1) in a marine microcosm determined by quantitative PCR and an Yamazi, A. K., Moreira, T. S., Cavicchioli, V. Q., Burin, K. R. C., & Nero, L. A. (2013).
rRNA-targeting probe and its effect on the indigenous bacterioplankton. Applied and Long cold storage influences the microbiological quality of raw goat milk. Small
Environmental Microbiology, 61, 1201–1207. https://doi.org/10.1128/ Ruminant Research, 113, 205–210. https://doi.org/10.1016/j.
AEM.61.4.1201-1207.1995. smallrumres.2013.02.004.
Lipuma, J. J., Currie, B. J., Peacock, S. J., & VanDamme, P. A. R. (2011). Burkholderia, Yuan, L., Sadiq, F. A., Burmolle, M., Wang, N. I., & He, G. (2019). Insights into
stenotrophomonas, ralstonia, cupriavidus, pandoraea, Brevundimonas, comamonas, psychrotrophic bacteria in raw milk: A review. Journal of Food Protection, 82,
delftia, and acidovorax. Manual of Clinical Microbiology, 692–713. https://doi.org/ 1148–1159. https://doi.org/10.4315/0362-028X.JFP-19-032.
10.1128/9781555816728.ch41. Yuan, L., Sadiq, F. A., Liu, T., Flint, S., Chen, J., Yang, H., et al. (2017). Psychrotrophic
Li, N., Wang, Y., You, C., Ren, J., Chen, W., Zheng, H., et al. (2018). Variation in raw milk bacterial populations in Chinese raw dairy milk. Lwt-Food Science and Technology, 84,
microbiota throughout 12 Months and the impact of weather conditions. Scientific 409–418. https://doi.org/10.1016/j.lwt.2017.05.023.
Reports, 8, 2371. https://doi.org/10.1038/s41598-018-20862-8. Zhang, R., Huo, W., Zhu, W., & Mao, S. (2015). Characterization of bacterial community
Lu, Y., Chen, Y., Lu, X. A., Lv, J., Man, C. X., Chai, Y. L., et al. (2014). Comparison of of raw milk from dairy cows during subacute ruminal acidosis challenge by high-
methods for the microbiological identification and typing of Cronobacter species in throughput sequencing. Journal of the Science of Food and Agriculture, 95, 1072–1079.
infant formula. Journal of Dairy Science, 97, 632–641. https://doi.org/10.3168/ https://doi.org/10.1002/jsfa.6800.
jds.2013-7147. Zhang, D., Palmer, J., Teh, K. H., Biggs, P., & Flint, S. (2019). 16S rDNA high-throughput
Marchand, S., Heylen, K., Messens, W., Coudijzer, K., De Vos, P., Dewettinck, K., et al. sequencing and MALDI-TOF MS are complementary when studying psychrotrophic
(2009). Seasonal influence on heat-resistant proteolytic capacity of Pseudomonas bacterial diversity of raw cows’ milk. International Dairy Journal, 97, 86–91. https://
lundensis and Pseudomonas fragi, predominant milk spoilers isolated from Belgian doi.org/10.1016/j.idairyj.2019.06.001.
raw milk samples. Environmental Microbiology, 11, 467–482. https://doi.org/ Zucali, M., Bava, L., Tamburini, A., Brasca, M., Vanoni, L., & Sandrucci, A. (2011). Effects
10.1111/j.1462-2920.2008.01785.x. of season, milking routine and cow cleanliness on bacterial and somatic cell counts
Mateos, A., Guyard-Nicodeme, M., Bagliniere, F., Jardin, J., Gaucheron, F., Dary, A., of bulk tank milk. Journal of Dairy Research, 78, 436–441. https://doi.org/10.1017/
et al. (2015). Proteolysis of milk proteins by AprX, an extracellular protease s0022029911000598.
identified in Pseudomonas LBSA1 isolated from bulk raw milk, and implications for

12