ORIGINAL ARTICLE

Distribution of glucomannans and xylans in poplar xylemand their changes under tension stress

Jong Sik Kim • Geoffrey Daniel

Received: 23 December 2011 / Accepted: 4 January 2012 / Published online: 19 January 2012

� Springer-Verlag 2012

Abstract Present work investigated glucomannan (GM)

and xylan distribution in poplar xylem cells of normal-

(NW), opposite- (OW) and tension wood (TW) with

immunolocalization methods. GM labeling was mostly

detected in the middle- and inner S2 (?S3) layer of NW and

OW fibers, while xylan labeling was observed in the whole

secondary cell wall. GM labeling in vessels of NW and

OW was much weaker than in fibers and mostly detected in

the S2 layer, whereas slightly stronger xylan labeling than

fibers was detected in the whole secondary cell wall of

vessels. Ray cells in NW and OW showed no GM labeling,

but strong xylan labeling. These results indicate that GMs

and xylans are spatially distributed in poplar xylem cells

with different concentrations present in different cell types.

Surprisingly, TW showed significant decrease of GM

labeling in the normal secondary cell wall of gelatinous

(G) fibers compared to NW and OW, while xylan labeling

was almost identical indicating that the GM and xylan

synthetic pathways in fibers have different reaction mech-

anisms against tension stress. Unlike fibers, no notable

changes in GM labeling were detected in vessels of TW,

suggesting that GM synthesis in vessels may not be

affected by tension stress. GM and xylan was also detected

in the G-layer with slightly stronger and much weaker

labeling than the normal secondary cell wall of G-fibers.

Differences in GM and xylan distribution are also dis-

cussed for the same functional cells found in hardwoods

and softwoods.

Keywords Gelatinous layer � Glucomannan �Opposite wood � Poplar � Tension wood � Xylan

Abbreviations

CML Compound middle lamella

G-layer Gelatinous layer

GGM Galactoglucomannan

GM Glucomannan

NW Normal wood

OW Opposite wood

TW Tension wood

Introduction

Wood is composed of different cells and its formation is

highly regulated by environmental conditions. In response to

gravitropic changes, trees develop reaction (i.e. abnormal)

wood and a unique type of cell is formed during reaction wood

formation (Timell 1986). In hardwoods (angiosperms), reac-

tion wood known as tension wood (TW) occurs in the upper

side of leaning stems or branches to maintain its original

growing position (Timell 1969). Although several phytohor-

mones, such as auxin (Moyle et al. 2002), ethylene (Andersson-

Gunneras et al. 2003) and gibberellins (Funada et al. 2008)

have been suggested as key internal factors involved in TW

formation in combination with external environmental stress,

the physiological process of TW formation is not fully under-

stood (Andersson-Gunneras et al. 2006).

Chemically, TW contains less lignin but more cellulose

than normal wood (NW, reviewed by Gorshkova et al.

2010 and Pilate et al. 2004). Several earlier chemical

studies have also suggested that TW contains less gluco-

mannan (GM) and xylan and more galactans than NW or

J. S. Kim � G. Daniel (&)

Wood Science, Department of Forest Products,

Swedish University of Agricultural Sciences,

P.O. Box 7008, SE-750 07 Uppsala, Sweden

e-mail: geoffrey.daniel@sprod.slu.se

123

Planta (2012) 236:35–50

DOI 10.1007/s00425-012-1588-z

opposite wood (OW, Fujii et al. 1982; Timell 1967, 1969).

Recent transcript profiling and gene expression data have

also shown that the biosynthetic activities of mannans and

xylans are decreased during TW formation (Andersson-

Gunneras et al. 2006; Decou et al. 2009). In contrast, some

reports have also indicated that TW contains more xylans

(Aguayo et al. 2010) and GMs (Moon et al. 2011) than

OW.

36 Planta (2012) 236:35–50

123

Most prominent feature of most but not all tension wood

formed by hardwoods is the formation of gelatinous

(G) fibers that are distinct from normal wood fibers by their

development of a characteristic G-layer (Clair et al. 2006;

Fisher and Stevenson 1981). The G-layer is rich in cellu-

lose with an almost parallel microfibril angle to the fiber

axis (reviewed by Gorshkova et al. 2010). Recent immu-

nocytochemical studies have shown that G-layer also

contains galactan, rhamnogalacturonan I (RGI), arabino-

galactan, arabinogalactan protein (AGP) and xyloglucan

(Arend 2008; Bowling and Vaughn 2008; Nishikubo et al.

2007). In contrast, xylan labeling was not detected in the

G-layer of sweetgum, hackberry and poplar of TW

(Bowling and Vaughn 2008; Decou et al. 2009) which is

generally recognized as a key feature of G-layers in all

plant (i.e. non- and woody plants) gelatinous fibers

(reviewed by Gorshkova et al. 2010).

Regarding GM, although Nishikubo et al. (2007)

showed the presence of b-1,4-mannan in the G-layer by

neutral sugar analysis of isolated G-layer from poplar TW,

the GM distribution in the G-layer has not so far been

clearly characterized. In particular, no microscopy studies

have been reported on GM distribution in G-layer. Even in

NW, although some microscopy observations have repor-

ted on GM distribution in hardwood cell walls (Baba et al.

1994; Kaneda et al. 2010), the information is rather limited.

In this study, we describe the distribution of GMs and

xylans not only in the G-layer, but also normal secondary

cell wall of G-fibers, vessels and ray cells in comparison

with NW and OW using immunolocalization methods. To

increase recognition of different epitopes of GMs and

xylans, several different mannan (LM21, LM22, BGM

C6) and xylan (LM10, LM11)-specific monoclonal anti-

bodies were used in combination with immunofluores-

cence and gold labeling methods. Different localization

properties of GMs and xylans in poplar xylem cells were

also discussed.

Materials and methods

Plant materials

Wood discs were taken from a 10-year-old leaning poplar

tree (Populus tremula L.) grown in a forest site near the

campus of the Swedish University of Agricultural Sciences

on August 16, 2011. Small sectors were cut from the TW

and OW sides of the stem and fixed with 2% v/v parafor-

maldehyde ?2.5% glutaraldehyde in 0.05 M sodium cac-

odylate buffer (pH 7.2) for 4 h at room temperature. After

washing in three changes of cacodylate buffer for 20 min

each, the sectors were dehydrated through a graded ethanol

series and infiltrated in a mixture of LR White resin

(London Resin Co., UK) and ethanol (v/v = 1:3, 1:1, 3:1

each overnight, and 100% resin for 2 days). Sectors were

then embedded in pure LR White resin and polymerized at

60�C for 2 days. Some sectors of NW were also collected

from a mature poplar tree and embedded in LR White resin

according to the procedures described above.

Immunofluorescence labeling

Labeling was conducted according to procedures descri-

bed previously (Kim et al. 2010a, b). Semi-thin sections

(ca 1 lm) prepared from resin embedded blocks with a

diamond knife were mounted on slides coated and treated

with 50 mM glycine/phosphate-based saline (PBS) solu-

tion for 15 min. After washing with PBS buffer for 5 min,

sections were suspended in blocking buffer (pH 7.2, PBS

buffer containing 3% w/v bovine serum albumin, BSA)

for 30 min at room temperature. After washing with PBS

buffer for 5 min, sections were incubated in mannan

(LM21, LM22, Marcus et al. 2010, PlantProbes, UK;

BGM C6, Pettolino et al. 2001, Biosupplies, Australia;

1:20, 1:20, and 1:100 dilution in PBS buffer, respectively)

or xylan (LM10, LM11, McCartney et al. 2005, PlantP-

robes, UK; 1:20 dilution in PBS buffer)-specific mono-

clonal antibodies for 2 days at 4�C. After washing in

three changes of PBS buffer for 10 min each, sections

were incubated with anti-rat (LM10, LM11, LM21,

LM22) or mouse (BGM C6) IgG Alexa Fluor 488

(Invitrogen, USA; 1:100 dilution in PBS buffer) for 2 h at

35�C. For control, some sections were incubated with

Alexa Fluor 488 only. After three washes with PBS buffer

for 10 min, sections were examined under a fluorescence

microscope (Leica DMRE, Germany) with I3 filter cube

(excitation 450–490 nm, emission [515 nm). Lignin

autofluorescence was completely eliminated through a

short exposure time (not shown). Some serial sections

were also stained with 1% w/v toluidine blue solution in

0.1% borax buffer and observed using a Leica DMBL

light microscope.

Fig. 1 Immunofluorescence localization of GMs in NW. a Differen-

tiating (Di) and mature (Ma) xylem stained with toluidine blue (TB).

Note thick secondary cell wall of fiber (F) and vessel (V) at the early

stage of differentiating xylem (inset). b–d Serial sections of a. Each

number indicates the same cell as in a. Sections were labeled by

LM21, LM22 and BGM C6, respectively. Strong labeling was only

detected in xylem. No labeling was detected in the cambial zone (Ca).

The strongest and weakest labeling by LM21 (b) and BGM C6 (d),

respectively. e–f Differentiating xylem. Enlargement of rectangles in

b–d, respectively. Vessels (V) and ray cells (R) showed much weaker

labeling than fibers with the antibodies. No (g) or very weak (e,

f) labeling of phloem fibers (Pf) and some specific labeling in the

cytoplasm of parenchyma cells by LM22 (arrows in f). h–j Mature

xylem. Enlargement of squares in b–d, respectively. Fibers showed

much stronger labeling than vessels (V) and ray cells (R) with the

antibodies. Stronger labeling in the inner- than outer secondary cell

wall layer (OuSW) and compound middle lamella (CML) (inset in j).Bar 100 lm (a–d), 25 lm (e–j)

b

Planta (2012) 236:35–50 37

123

Immunogold labeling

Immunogold labeling was conducted according to proce-

dures described by Kim et al. (2010a, b) with minor

modification. Transverse ultrathin sections (ca 90 nm)

prepared from LR white embedded blocks were mounted

on nickel grids and incubated in blocking buffer (pH 8.2,

Tris-buffered saline (TBS) containing 1% w/v BSA and

38 Planta (2012) 236:35–50

123

0.1% w/v NaN3) for 30 min at room temperature. Grids

were then incubated with mannan and xylan specific anti-

bodies (see below for incubation conditions). After three

washes with blocking buffer for 15 min each, grids were

incubated with secondary antibodies (see below for incu-

bation conditions). For control, some grids were incubated

with secondary antibodies only. After washing in five

changes of blocking buffer for 15 min, grids were post-

stained with 4% w/v uranyl acetate for 10 min and exam-

ined with a Philips CM12 transmission electron microscope

(TEM, USA) operated at 80 kV. Negative TEM films were

scanned using an Epson Perfection Pro 750 film scanner.

Antibodies were incubated under the following conditions:

BGM C6: A 2-day incubation at 4�C (1:500 dilution in

blocking buffer), followed by incubation with anti-mouse

secondary antibody labeled with 10-nm colloidal gold

particles (BB International, UK; 1:50 dilution in blocking

buffer) for 4 h at 35�C.

LM10, LM21 and LM22: A 2-day incubation at 4�C

(1:20 dilution in blocking buffer), followed by incubation

with anti-rat secondary antibody labeled with 10-nm col-

loidal gold particles (BB International, UK; 1:50 dilution in

blocking buffer) for 4 h at 35�C.

LM11: A 2-day incubation at 4�C (1:20 dilution in blocking

buffer), followed by incubation with anti-rat secondary anti-

body labeled with 10-nm colloidal gold particles (1:100

dilution in blocking buffer) for 4 h at room temperature.

Results

General anatomy of NW, OW and TW

NW, TW and OW showed almost identical developing

stages of secondary xylem (Figs. 1a, 2a, 3a). The normal

stages of cambium activity and fibers, vessels and ray

secondary cell wall formation (i.e. S2/S1 formation) was

not observed in the developing xylem (inset in Fig. 1a).

OW (Fig. 2a) showed a much shorter developing xylem

compared to NW (Fig. 1a) and TW (Fig. 3a). Fibers having

shorter radial diameter were also detected in mature OW

xylem (inset in Fig. 2a). G-layer formation in TW was

initially detected in the third or fourth fibers from the

cambium cells (inset in Fig. 3a). A part of S2 including

the entire S3 layer was replaced by the G-layer (S1 ? S2 ?

G-layer, inset in Fig. 3h).

Immunofluorescence localization of GMs in NW

and OW

NW and OW showed almost identical labeling patterns in

developing and mature xylem regardless of antibody type,

although labeling was stronger in NW than OW (Figs. 1, 2).

All antibodies showed similar labeling patterns in develop-

ing and mature xylem but with differences in labeling

intensity (Figs. 1b–d, 2b–d). LM21 and BGM C6 showed

strongest and weakest labeling in the xylem, respectively.

Labeling was not detected in the cambial zone, whereas

developing and mature xylem showed strong labeling

(Figs. 1b–d, 2b–d). Fibers showed much stronger labeling

than vessels and ray cells in developing and mature xylem

(Figs. 1e–j, 2e–g). In mature fibers, stronger labeling was

detected in the inner- than outer secondary cell wall layer

including the compound middle lamella (CML) (inset in

Figs. 1j, 2g). Latewood (LW) in OW showed much weaker

labeling than earlywood (EW) regardless of antibody type

(Fig. 2b–d, h). Unlike LM21 and BGM C6, specific labeling

was detected in the cytoplasm of parenchyma cells by LM22

(arrows in Figs. 1f, 2f, h). In addition, phloem fibers showed

no or very weak labeling in the cell wall (Figs. 1a–g, 2a–f).

Immunofluorescence localization of GMs in TW

GM labeling was detected in developing and mature TW

xylem (Fig. 3b–d), but was much weaker than NW

(Fig. 1b–d) and OW (Fig. 2b–d). Unlike NW and OW, TW

showed clear differences in labeling patterns between

antibodies. LM21 and BGM C6 showed stronger labeling

in vessels than fibers (Fig. 3i, k), whereas LM22 revealed

almost identical weak labeling of vessels and fibers

(Fig. 3j). Some strong labeling was also detected in

G-layers at the late stage of TW formation compared to

early- and mature stages by LM21 and BGM C6 (Fig. 3b,

d, g), but it was not detected by LM22 (Fig. 3c). Like NW,

specific labeling was detected in the cytoplasm of paren-

chyma cells by LM22 (Fig. 3c, f). Phloem fibers showed

very weak labeling as shown in NW regardless of antibody

type (Fig. 3e, f).

Fig. 2 Immunofluorescence localization of GMs in OW. a Differen-

tiating (Di) and mature (Ma) xylem stained with toluidine blue (TB).

Shorter differentiating zone compared to NW (Fig. 1a) and TW

(Fig. 3a) and variations of fiber shape in mature xylem (inset).b–d Serial sections of a. Each number indicates the same cells as in

a. Sections were labeled by LM21, LM22 and BGM C6, respectively.

Labeling was weaker and more uneven than that in normal wood

(Fig. 1b–d), but much stronger than tension wood (Fig. 3b–d).

e–f Differentiating xylem. Enlargement of squares in b and c. Vessels

(V) and ray cells (R) showed much weaker labeling than fibers during

early stages of xylem formation. Very weak labeling in phloem fibers

(Pf) and some specific labeling in the cytoplasm of parenchyma cells

by LM22 (arrows in f) as shown in normal (f) and tension wood

(Fig. 3f). g Mature xylem. Enlargement of square in b. Fibers showed

much stronger labeling than vessels (V) and ray cells (R). Stronger

labeling in the inner secondary cell wall layer than the outer (OuSW)

and compound middle lamella (CML) (inset in g) as shown in normal

wood (Fig. 1j). h Growth ring area. Enlargement of square in

c. Weaker labeling was observed in late- (LW) and transition wood

(TW) compared to earlywood (EW). Some specific labeling in the

cytoplasm of terminal axial parenchyma (asterisks in inset) and ray

cells (R, arrows) by LM22. Bar 100 lm (a–d), 25 lm (e–h)

b

Planta (2012) 236:35–50 39

123

40 Planta (2012) 236:35–50

123

Immunogold localization of GMs in NW and OW

Immunofluorescence observations were further advanced

by immunogold labeling. Labeling patterns of OW (not

shown) were almost identical to those in NW, except for

weaker labeling in the S2 layer of developing fibers of TW

(Fig. 6a–c) than NW (Fig. 4a–c). No GM labeling was

detected in cambium cells regardless of antibody type

(Fig. 4a–c). Strong labeling was detected in the S2 layer of

developing NW fibers with more intense labeling by LM21

and LM22 than BGM C6 (Fig. 4a–c). However, most

labeling was detected in the middle S2 layer and only very

few gold particles were detected in the outer secondary

cell wall of developing fibers regardless of antibody

(Fig. 4a–c). Some strong labeling was also observed in

the primary cell wall by LM21 (arrows in Fig. 4a). In

mature xylem, strong labeling was detected in the S2 layer

of fibers, but it was mostly limited to the middle- and

inner S2 (?S3) layer (Fig. 4d–f). Almost no labeling was

detected in the S1- and outer S2 layer of fibers, no matter

of antibody type (Fig. 4d–f). Vessels also showed some

strong labeling, but it was much weaker than fibers

(Fig. 5a–c). The S1- and S3 layers of vessels showed

weaker labeling than the S2 layer (Fig. 5a–c). Like

developing fibers, LM21 and LM22 (Fig. 5a, b) showed

much stronger labeling in vessels than BGM C6 (Fig. 5c)

even though the labeling patterns were almost identical.

Almost no labeling was detected in ray cell walls with the

three GM antibodies (Fig. 5d–f).

Immunogold localization of GMs in TW

Labeling was detected in developing TW fibers (Fig. 6a–

c), but it was much weaker than in NW (Fig. 4a–c). The

S1- and outer S2 layer of fibers (Fig. 6a–c) showed almost

no labeling as in NW (Fig. 4a–c), although some labeling

was detected in the primary cell wall (arrows in Fig. 6a–c).

In mature xylem, GM labeling was detected in the S2 and

G-layer with stronger labeling by LM21 and LM22 than

BGM C6 (Fig. 6d–f). Almost no labeling was detected in

the S1 layer (Fig. 6d–f). The G-layer showed stronger

labeling than the S2 layer by LM21 and LM22 (Fig. 6d, e).

Some labeling was also detected in the CML between

G-fibers by all antibodies (Fig. 6d–f). In vessels, labeling

was mostly limited to the S2 layer with different patterns

recognized between the antibodies (Fig. 7a–c). LM21 and

BGM C6 showed stronger labeling in the S2 layer of ves-

sels than fibers (Fig. 7a, c), while LM22 showed almost

similar labeling density between vessels and fibers

(Fig. 7b). Some labeling was also detected in the CML

between fibers and vessels by the three antibodies (Fig. 7a–

c). No specific labeling was detected in ray cell walls

(Fig. 7d–f).

Immunofluorescence localization of xylans in NW,

TW and OW

Labeling was detected in developing and mature xylem

regardless of sample and antibody type (Fig. 8a–f). No

labeling was observed in cambium and phloem regions

except for phloem fibers (Fig. 8a–f). Strong labeling was

detected in the whole secondary cell walls of fibers, ves-

sels, and ray cells of NW and OW with all antibodies

(Fig. 8g, h). TW also showed strong labeling in the outer

cell wall of G-fibers, vessels and ray cells (Fig. 8i). How-

ever, the presence of xylans in the G-layer was not clearly

recognized under immunofluorescence microscopy even

though some weak labeling was detected in the inner part

of cell walls (inset in Fig. 8i). Unlike GMs (Fig. 2b–d),

LW in OW showed strong labeling like other xylem areas

(Fig. 8e, f).

Immunogold localization of xylans in NW,

TW and OW

Almost identical labeling patterns were detected in NW and

OW (not shown) fibers, vessels and ray cells regardless of

antibody type (Figs. 9a–c, 10a–c). LM10 showed slightly

stronger labeling in the outer- than inner secondary cell

wall of NW fibers (Fig. 9b, c), while LM11 showed uni-

form distribution in the whole secondary cell wall

(Fig. 10b, c). LM10 also showed stronger labeling in ves-

sels than fibers in NW, especially in the inner S2 layer

(Fig. 9b). The normal secondary cell wall of G-fibers

showed almost uniform labeling no matter antibody type

(Figs. 9d–h, 10d–h). Some specific labeling was also

detected in the G-layer from early stages of G-layer

Fig. 3 Immunofluorescence localization of GMs in TW. a Differen-

tiating (Di) and mature (Ma) xylem stained with toluidine blue (TB).

G-layer formation (asterisks in inset) was initially detected in the

third or fourth fibers (marked 3 and 4 in inset) by cambium cells. b–dSerial sections of a. Each number indicates the same cells as in

a. Sections were labeled by LM21, LM22 and BGM C6, respectively.

Labeling in the xylem area was significantly decreased compared to

NW (Fig. 1b–d) and OW (Fig. 2b–d) with all antibodies. Note the

similar labeling patterns between LM21 and BGM C6 with stronger

labeling in vessels than fibers. e, f Differentiating xylem. Enlargement

of squares in b and c, respectively. Note no labeling in the cambial

zone (Ca) and very weak labeling of phloem fibers (Pf). Specific

labeling was also detected in cytoplasm of parenchyma cells by LM22

(arrows in f). g Late stages of G-fiber (Gf) formation. Enlargement of

square in d. Note some strong labeling in the inner parts of the fiber

cell walls. h–k Mature xylem. The formation of thick G-layer (inset in

h) stained pink color with toluidine blue. Stronger labeling was

detected in vessels (V) than G-fibers by LM21 (i) and BGM C6 (k),

while LM22 (j) showed similar labeling between vessels and G-fibers.

Some strong labeling in the inner secondary cell wall of G-fibers by

LM22 (j). Bar 100 lm (a–d), 25 lm (e–k)

b

Planta (2012) 236:35–50 41

123

formation by LM11 (Fig. 10e, f). In contrast, no labeling was

detected in G-layer during stages of G-layer formation by

LM 10 (Fig. 9e, f). Vessels and ray cells in TW (Figs. 9g, h,

10g, h) showed almost identical labeling patterns as those in

NW (Figs. 9b, c, 10b, c).

Discussion

Although the chemical structure of hemicelluloses in

hardwoods is well understood, information on the micro-

distribution of hemicelluloses in their cell walls is limited.

Fig. 4 Immunogold localization of GMs in NW fibers labeled by

LM21 (a, d), LM22 (b, e) and BGM C6 (c, f). a–c Developing fibers.

Most labeling was detected in the middle part of the S2 layer. The

outer secondary cell wall (S1, outer S2) and inner S2 layer showed no

or very weaker labeling with all antibodies. Some labeling in primary

cell wall by LM21 and BGM C6 (arrows in a, c). No labeling was

detected in cambium cells (Ca). d–f Mature fibers. Like in the

developing stage, most labeling was limited to the inner secondary

cell wall regardless of antibody type. Note much weaker labeling by

BGM C6 than LM21 and LM22. Bar 500 nm

42 Planta (2012) 236:35–50

123

In particular, only a few studies on GM localization in

hardwood xylem cells are available presumably because

GM represents only a minor hemicellulose component in

hardwoods. This work describes GM and xylan distribution

in different poplar cell walls with emphasis on GM dis-

tribution. The work also provides new insights into GM

and xylan reorganization in wood cell walls under tension

stress.

Fig. 5 Immunogold localization of GMs in NW vessels and ray cells

labeled by LM21 (a, d), LM22 (b, e) and BGM C6 (c, f). a–c Mature

vessels and fibers. Vessels showed much weaker labeling than fibers

with the antibodies. The S2 of vessels showed stronger labeling than

the S1- and S3 layers. Note almost no labeling in the S1- and S3 layer

of vessels by LM22 and some labeling in the compound middle

lamella (CML) by LM21. d–f Ray cells and mature fibers. Note

almost no labeling in ray secondary cell wall (SW) regardless of

antibody type. Bar 500 nm

Planta (2012) 236:35–50 43

123

Comparison of GM localization patterns

between antibodies

LM10 and LM11 antibodies are frequently used for xylan

localization in plant sciences (Herve et al. 2009;

McCartney et al. 2005) and their binding characteristics in

wood cell walls are well described (Kim et al. 2010b,

2011b). Here, we focus on differences in GM localization

using mannan specific antibodies (LM21, LM22, BGM C6)

whose binding characteristics particularly LM21 and LM22

Fig. 6 Immunogold localization of GMs in TW fibers labeled by

LM21 (a, d), LM22 (b, e) and BGM C6 (c, f). a–c Labeling was

detected in the S2 layer of developing fibers as in NW, but it was

much weaker than that in NW (Fig. 4a–c). No labeling was detected

in cambium cells (Ca). Labeling in primary cell wall by all antibodies

(arrows in a–c). d–f Mature G-fibers (Gf). Labeling was detected in

both normal secondary cell wall and G-layer, but it was much weaker

than NW (Fig. 4d–f). Stronger labeling was detected in G-layers than

normal secondary cell walls. Note also labeling in compound middle

lamella (CML) by the antibodies. Asterisks in e show a split between

the S2- and G-layer. Bar 500 nm

44 Planta (2012) 236:35–50

123

in wood cell walls are not fully understood. LM21 and

LM22 showed stronger labeling than BGM C6 regardless

of wood (i.e. NW, OW, TW) sample and cell type. If only

labeling intensity is considered, results suggest that LM21

and LM22 are better for GM localization in hardwoods

than BGM C6. BGM C6 is recommended for galactoglu-

comannan (GGM) labeling in softwoods because similar

strong labeling to those by LM21 and LM22 in this work

Fig. 7 Immunogold localization of GMs in TW vessels and ray cells

labeled by LM21 (a, d), LM22 (b, e) and BGM C6 (c, f). a–c Mature

vessels (V) and G-fibers (F). Strongest labeling in vessels was

detected in the S2 layer with the antibodies. The S1- and S3 layers

showed very weak labeling. Note stronger labeling in vessels than

G-fibers (Gf) by LM21 and BGM C6, and similar labeling by LM22.

Note also some labeling in compound middle lamella (CML) by all

antibodies. d–f Ray cells (R) and mature G-fibers (Gf). Almost no

labeling was observed in ray secondary cell walls (SW) with the

antibodies. Bar 500 nm

Planta (2012) 236:35–50 45

123

was detected earlier in tracheids by BGM C6 performed

using a similar approach (Kim et al. 2010a). LM21 and

BGM C6 showed similar localization characteristics

regardless of sample, although labeling was much stronger

in LM21 than BGM C6. In contrast, LM22 showed dif-

ferent labeling patterns compared to LM21 and BGM C6 in

46 Planta (2012) 236:35–50

123

TW and parenchyma cells, indicating that LM22 has some

different binding properties than LM21 (or BGM C6) as

suggested by Marcus et al. (2010).

GM and xylan distribution in NW and OW

Apart from slight differences in labeling intensity, the

general labeling patterns of GMs and xylans were almost

identical between NW and OW regardless of antibody. GMs

and xylans were not detected in cambium cells composed of

primary cell walls and intercellular layers even though some

strong GM labeling was detected in the primary cell wall of

differentiating fibers, indicating that GM deposition in the

primary cell wall may be a specific characteristic of sec-

ondary cell wall formation. During fiber maturation, GM

labeling was almost absent in the outer parts of the sec-

ondary cell wall, whereas xylan labeling was detected in the

whole secondary cell wall, indicating that there are differ-

ences not only in concentration, but also spatial distribution

between GMs and xylans in hardwood fibers. Interestingly,

LW in OW showed weaker GM labeling, but almost iden-

tical xylan labeling compared to EW. This result suggests

that hemicellulose composition may differ between LW and

EW in poplar xylem.

In vessels and ray cells, Kaneda et al. (2010) reported

that GM labeling was much weaker than that in fibers of

poplar using BGM C6. Present results showed also almost

identical patterns of GM labeling in vessels and ray cells by

all antibodies. However, as described above, LM21 and

LM22 showed much stronger GM labeling in vessels than

BGM C6 and provided further details of labeling patterns

in vessels even though ray cells showed almost no labeling

with the antibodies. Like fibers, the S1 layer of vessels

showed almost no GM labeling by LM21 and LM22,

indicating that there are some similarity in occurrence of

GMs between vessels and fibers. In contrast, strong xylan

labeling was detected in the whole secondary cell wall of

vessels and ray cells, suggesting a different occurrence

between GMs and xylans in vessels and ray cells as shown

in fibers.

In addition to variations in GM distribution among hard-

wood xylem cells, present results also suggest that the same

functional cells between hardwoods and softwoods have

different distributional properties of mannan polymers.

Unlike poplar fibers in this work, GGM labeling was

detected in the whole secondary cell wall of tracheids in

Japanese cedar even though the outer S2 layer showed

weaker labeling than other parts of the cell wall (Kim et al.

2010a). Ray cells in softwood also showed strong GGM

labeling in the whole cell wall (Kim et al. 2011a), while

almost no GM labeling was detected in ray cells of poplar.

These results suggest that there are significant differences in

spatial microdistribution of mannan polymers in addition to

their concentration and chemical structure even in the same

functional cells between hardwoods and softwoods. In con-

trast to mannans, xylans showed in general similar labeling

patterns in tracheids and ray cells of cedar (Kim et al. 2010b,

2011a) to fibers and ray cells of poplar even though the

chemical structure and concentration of xylans differ

between hardwoods and softwoods as in mannan polymers.

GM and xylan distribution in TW

Present results show that tension stress induced a signifi-

cant decrease of GM deposition in fiber cell walls. This

result is in line with chemical data obtained by neutral

sugar analysis of TW (Fujii et al. 1982; Timell 1969) and

recent transcript profiling data reporting decrease of carbon

(C) flux to mannans in TW (Andersson-Gunneras et al.

2006). In contrast, Moon et al. (2011) showed that GM

content in TW is significantly higher at the early stage of

TW formation than OW but became similar during the late

stage of TW formation in yellow poplar, which does not

form a G-layer in TW. Although the variation of GM

content in TW between species and its relationship with

TW formation are not clearly explained by Moon et al.

(2011), present work suggests that GM deposition in

hardwood fiber cell walls is significantly affected by ten-

sion stress. Unlike fibers, GM deposition in vessels was not

changed in TW compared to those in NW and OW, indi-

cating that the GM synthetic pathway in vessels may differ

from that in fibers or may not be affected by tension stress.

This result also suggests that GMs may have important

roles to maintain vessel structure and/or functions even

though their content is very low in vessel cell walls.

In addition to significant decrease of GM deposition in

TW, the present work also shows that the G-layer contains

GMs and its GM content is higher than in the normal

secondary cell wall in G-fibers. The specific functions/roles

of GMs in G-layer are not part of the present work but this

represents the first microscopic report on GM distribution

in G-layer. Additional biochemical and molecular studies

on hemicelluloses including GMs and xylans (see below) in

the G-layer should provide further information on the roles

of hemicelluloses in this cell wall layer.

Fig. 8 Immunofluorescence localization of xylans by LM10 (a, c,

e) and LM11 (b, d, f). Serial sections of Figs. 1a (a, b), 3a (c, d) and

2a (e, f), respectively. Strong labeling was detected in developing (Di)

and mature xylem (Ma) of NW (a, b), TW (c, d) and OW (e, f). Note

the higher variation of labeling in OW compared to NW and TW and

the strong labeling of phloem fiber (Pf) regardless of sample and in

latewood (LW) of OW. g–i Enlargement of squares in b, d, f,respectively. Almost uniform strong labeling was detected in the

whole secondary cell wall of NW (h) and OW (g) fibers, while it was

only detected in the outer secondary cell wall of TW G-fibers (arrowsin i). Some weak labeling in the inner part of the secondary cell wall

in G-fibers (Gf, inset in i; arrows). Bar 100 lm (a–f), 25 lm (g–i)

b

Planta (2012) 236:35–50 47

123

Unlike GMs, strong xylan labeling was detected not

only in NW (and OW), but also the normal secondary cell

wall of TW, indicating that xylan deposition in normal

secondary cell wall of TW may not be much affected by

tension stress even though xylan is the main hemicellulose

in hardwood cell walls. Interestingly, compression wood

(CW), the reaction wood formed in softwoods, showed a

significantly different pattern of xylan labeling in tracheids

compared to NW and almost identical GGM labeling pat-

terns in tracheids between CW and NW (Kim et al. 2011b).

These results suggest that minor hemicelluloses, i.e. GMs

in hardwoods and xylans in softwoods may be more

Fig. 9 Immunogold localization of xylans in NW (a–c) and TW

(d–h) by LM10. Both NW (a) and TW (d) showed strong and uniform

labeling in developing fibers. No specific labeling was detected in

cambium cells (Ca). In the mature stage, NW fibers (F) showed

stronger labeling in the outer- than inner S2 layer (b, c), while TW

fibers (Gf) revealed almost uniform labeling in the normal secondary

cell wall (e–h). No notable differences were detected in labeling of

vessels (V) and ray cells (R) between NW (b, c) and TW (g, h). No

labeling of the G-layer with LM10 during G-layer formation (e, f).Bar 500 nm

48 Planta (2012) 236:35–50

123

severely affected by environmental stress compared to the

major hemicelluloses in hardwoods and softwoods.

The present work also shows clearly that G-layer con-

tains xylans even though labeling in the G-layer was much

weaker than that in the normal secondary cell wall of TW.

This result differs from the general idea that G-layer does

not contain xylans (reviewed by Gorshkova et al. 2010).

Based on the binding properties of LM10 and LM11

Fig. 10 Immunogold localization of xylans in NW (a–c) and TW

(d–h) by LM11. No notable differences in labeling were detected in

developing and mature stages of fibers (F or Gf), vessels (V) and ray

cells (R) between NW and TW. Strong and uniform labeling was

detected in all xylem cells of NW and TW, but was not observed in

cambium cells (Ca). In G-fibers (Gf), some specific labeling was

detected in the G-layer from the early stage of G-layer formation

(e, f). Asterisks in e and g show splits between the S2- and G-layer.

Bar 500 nm

Planta (2012) 236:35–50 49

123

(McCartney et al. 2005), we can also expect the G-layer to

contain highly substituted xylans from the labeling of

LM11.

In conclusion, this work indicates that there are signif-

icant variations in GM distribution depending on spatial

location of cell walls and different cell types in poplar.

Present work also indicates that GM distribution differs

significantly from xylan distribution in poplar cells and

distribution of mannans in softwoods. In response to ten-

sion, observations suggest that the distribution of GMs and

xylans are changed in poplar xylem cell walls. Finally, our

immunolocalization work shows clearly the localization of

GMs and xylans in G-layer, which has not been reported

previously. Together this work suggests that each cell type

in hard- or softwoods including reaction wood cells may

have a different hemicellulose composition and reaction

mechanism to environmental stress.

Acknowledgments The authors gratefully acknowledge funding

provided by the Formas FuncFiber Center of Excellence (http://www.

funcfiber.se) and Formas Projects 2008-1399 and 2009-582.

References

Aguayo MG, Quintupill L, Castillo R, Baeza J, Freer J, Mendonca

(2010) Determination of differences in anatomical and chemical

characteristics of tension and opposite wood of 8-year old

Eucalyptus glogulus. Maderas Ciencia y Tecnologia 12:241–251

Andersson-Gunneras S, Hellgren J, Bjorklund S, Regan S, Moritz T,

Sundberg B (2003) Asymmetric expression of a poplar ACC

oxidase controls ethylene production during gravitational induc-

tion of tension wood. Plant J 34:339–349

Andersson-Gunneras S, Mellerowicz EJ, Love J, Segerman B,

Ohmiya Y, Coutinho PM, Nilsson P, Henrissat B, Moritz T,

Sundberg B (2006) Biosynthesis of cellulose-enriched tension

wood in Populus: global analysis of transcripts and metabolites

identifies biochemical and developmental regulators in second-

ary wall biosynthesis. Plant J 45:144–165

Arend M (2008) Immunolocalization of (1,4)-b-galactan in tension

wood fibers of poplar. Tree Physiol 28:1263–1267

Baba K, Sone Y, Kaku H, Misaki A, Shibuya N, Itoh T (1994)

Localization of hemicelluloses in the cell walls of some woody

plants using immno-gold electron microscopy. Holzforschung

48:297–300

Bowling AJ, Vaughn K (2008) Immunocytochemical characterization

of tension wood: gelatinous fibers contain more than just

cellulose. Am J Bot 95:655–663

Clair B, Ruelle J, Beauchene J, Prevost MF, Fournier M (2006) Tension

wood and opposite wood in 21 tropical rain forest species 1.

Occurrence and efficiency of the G-layer. IAWA J 27:329–338

Decou R, Lhernould S, Laurans F, Sulpice E, Leple J-C, Dejardin A,

Pilate G, Costa G (2009) Cloning and expression analysis of a

wood-associated xylosidase gene (PtaBXL1) in poplar tension

wood. Phytochemistry 70:163–172

Fisher JB, Stevenson JW (1981) Occurrence of reaction wood in

branches of dicotyledons and its role in tree architecture. Bot

Gaz 142:82–95

Fujii M, Azuma J, Tanaka F, Kato A, Koshijima T (1982) Studies on

hemicelluloses in tension wood: I. Chemical composition of

tension, opposite and side woods of Japanese beech (Faguscrenata Blume). Wood Res 68:8–21

Funada R, Miura T, Shimizu Y, Kinase T, Nakaba S, Kubo T, Sano Y

(2008) Gibberellin-induced formation of tension wood in

angiosperm trees. Planta 227:1409–1414

Gorshkova TA, Gurjanov OP, Mikshina PV, Ibragimova NN, Mokshina

NE, Salnikov VV, Ageeva MV, Amenitskii SI, Chernova TE,

Chemikosova SB (2010) Specific type of secondary cell wall

formed by plant fibers. Russ J Plant Physiol 57:328–341

Herve C, Rogowski A, Gilbert HJ, Knox JP (2009) Enzymatic

treatments reveal differential capacities for xylan recognition

and degradation in primary and secondary plant cell walls. Plant

J 58:413–422

Kaneda M, Rensing K, Samuels L (2010) Secondary cell wall

deposition in developing secondary xylem of poplar. J Integr

Plant Biol 52:234–243

Kim JS, Awano T, Yoshinaga A, Takabe K (2010a) Temporal and

spatial immunolocalization of glucomannans in differentiating

earlywood tracheid cell walls of Cryptomeria japonica. Planta

232:545–554

Kim JS, Awano T, Yoshinaga A, Takabe K (2010b) Immunolocaliza-

tion and structural variations of xylan in differentiating earlywood

tracheid cell walls of Cryptomeria japonica. Planta 232:817–824

Kim JS, Awano T, Yoshinaga A, Takabe K (2011a) Temporal and

spatial diversities of the immunolabeling of mannan and xylan

polysaccharides in differentiating earlywood ray cells and pits of

Cryptomeria japonica. Planta 233:109–122

Kim JS, Awano T, Yoshinaga A, Takabe K (2011b) Occurrence of

xylan and mannan polysaccharides and their spatial relationship

with other cell wall components in differentiating compression

wood tracheids of Cryptomeria japonica. Planta 233:721–735

Marcus S, Blake AW, Benians TAS, Lee KJD, Poyser C, Donaldson

L, Leroux O, Rogowski A, Petersen HL, Boraston A, Gilbert HJ,

Willats WGT, Knox JP (2010) Restricted access of proteins to

mannan polysaccharides in intact plant cell walls. Plant J

64:191–203

McCartney L, Marcus SE, Knox JP (2005) Monoclonal antibodies to

plant cell wall xylans and arabinoxylans. J Histochem Cytochem

53:543–546

Moon D, Shin S-J, Choi JW, Park J-S, Kim W, Kwon M (2011)

Chemical modification of secondary xylem under tensile stress

in the stem of Liriodendron tulipifera. Forest Sci Tech 7:53–59

Moyle R, Schrader J, Stenberg A, Olsson O, Saxena S, Sandberg G,

Bhalerao RP (2002) Environmental and auxin regulation of

wood formation involves members of the Aux/IAA gene family

in hybrid aspen. Plant J 31:675–685

Nishikubo N, Awano T, Banasiak A, Bourquin V, Ibatullin F, Funada

R, Brumer H, Teeri TT, Hayashi T, Sundberg B, Mellerowicz EJ

(2007) Xyloglucan Endo-transglycosylase (XET) functions in

gelatinous layers of tension wood fibers in poplar—a glimpse

into the mechanism of the balancing act of trees. Plant Cell

Physiol 48:843–855

Pettolino FA, Hoogenraad NJ, Ferguson C, Bacic A, Johnson E, Stone

BA (2001) A (1 ? 4)-b-mannan-specific monoclonal antibody

and its use in the immunocytochemical localization of galacto-

mannans. Planta 214:235–242

Pilate G, Dejardin A, Laurans F, Leple J-C (2004) Tension wood as a

model for functional genomics of wood formation. New Phytol

164:63–72

Timell TE (1967) Recent progress in the chemistry of wood

hemicelluloses. Wood Sci Technol 1:45–70

Timell TE (1969) The chemical composition of tension wood. Svensk

Papperstidn 72:173–181

Timell TE (1986) Compression wood in gymnosperms, vol 1.

Springer, Berlin

50 Planta (2012) 236:35–50

123