Expression of smooth muscle myosin light chain 17 and unloaded shortening in single smooth muscle cells

T. J. Eddinger, A. A. Korwek, D. P. Meer, J. J. Sherwood


These experiments were performed to test the hypotheses that myosin light chain 17 (MLC17) a and b isoform expression varies between individual vascular smooth muscle (SM) cells and that their expression correlates with cell unloaded shortening velocity. Single SM cells isolated from rabbit aorta and carotid arteries were used to measure unloaded shortening velocity and subsequently were analyzed via RT-PCR for MLC17 a and b mRNA ratio. The MLC17b/a mRNA and protein ratios from adjacent tissue sections correlate very well (R 2 = 0.68), allowing use of the mRNA ratio to predict the protein ratio. The rabbit MLC17 isoform protein sequence was found to be similar to, but unique from, the swine, mouse, and chicken sequences. Isolated single SM cells from the aorta and carotid have resting lengths of 70–280 μm and shorten to 33–88 μm after contraction. Isolated cell maximum unloaded shortening velocity is highly variable (0.5–7.5 μm/s) but becomes more uniform when normalized to initial cell length (0.01–0.05 cell lengths/s). Carotid cells activated in the presence of okadaic acid (1 μm) have mean maximal unloaded shortening velocities not significantly different from carotid cells activated without okadaic acid (0.016 vs. 0.019 cell lengths/s). Resting cell length before activation is significantly correlated with final cell length after unloaded shortening. Neither initial cell length, final cell length, total cell length change, nor maximum unloaded shortening velocity (absolute or normalized) was significantly correlated with single-cell MLC17b/a mRNA ratio. These studies were performed in isolated single SM cells where unloaded shortening velocity and MLC17b/a mRNA ratios were measured in the same cell. In this preparation, the three-dimensional organization and milieu of the cell is kept intact, but without the intercellular heterogeneity concerns of multicellular preparations. These results suggest the MLC17b/a ratio is variable between individual SM cells from the same tissue, but it is not a determinant of unloaded shortening velocity in single SM cells.

  • arterial muscle
  • aorta
  • carotid
  • reverse transcription-polymerase chain reaction
  • muscle mechanics

smooth muscle (SM) myosin subunit isoforms [myosin heavy chain (MHC), myosin light chain 20 (MLC20), and myosin light chain 17 (MLC17)] have been reported in all SM tissues examined to date. The MLC17 a and b isoforms were first identified by two-dimensional gel electrophoresis (5, 23) and subsequently sequenced (21, 26). These two isoforms are the product of a single gene resulting from tissue-specific alternative splicing. The 151-amino acid polypeptide products differ in five of the nine carboxy-terminal amino acids in bovine (26) and human (28) tissues, but only four differences are found in the chicken (35). These substitutions may be important because crystal structures suggest that the MLC17 subunits associate with the MHC S1 head at the junction between the motor domain and the lever arm (36).

Early kinetic studies reported physiological correlations with expression of the MLC17 isoforms. Purified myosin from porcine stomach (MLC17b = 0) was reported to have a maximal Mg2+-ATPase activity twofold greater than that from porcine aorta (MLC17b = 40%; see Ref. 23). This difference was also apparent when only the S1 head fragment of the myosin molecule was used, suggesting that the difference is not due to differences in the MHC tail isoform (SM1 and SM2) expression between these tissues. Using permeabilized tissue preparations, Malmqvist and Arner (30) reported an inverse relationship between unloaded shortening velocity and relative MLC17b isoform content in SM. Hasegawa and Morita (20) made reconstituted aortic myosin and reported a nonlinear 20-fold decrease in actin-activated ATPase activity as MLC17b increased from 23 to 81%. The lower affinity for MgADP and faster kinetics of phasic SM may be due to its lower expression of MLC17b and/or the higher expression of the MHC head insert (SMB; see Ref. 11). Takeuchi et al. (43) reported unique intra- and intercellular distributions of the MLC17 nonmuscle (nm) and gizzard (gi) isoforms in freshly isolated rat aorta cells. They suggested that their results are consistent with unique physiological roles for these isoforms.

Other recent work suggests that there is only a casual relationship between MLC17a and faster kinetics, indicating another cause. There are two SM MHC head isoforms (SMA, SMB) that differ by the absence or presence, respectively, of a seven-amino acid insert near the ATP-binding site of the MHC head (1, 17, 25, 44). An approximately twofold higher in vitro motility and Mg2+-ATPase activity was reported in phasic (turkey gizzard) vs. tonic (turkey aorta) SM myosin (25). The aorta is reported to express primarily SMA, whereas the gizzard is reported to express primarily SMB SM MHCs. Exchanging the MLC17 isoforms between these two tissues from 0 to 30% MLC17b did not alter the results. Rovner et al. (37) used a baculovirus/insect cell system to express homogeneous heavy meromyosin (HMM) with and without the seven-amino acid head insert and the two MLC17 isoforms. They found a twofold increase in actin-activated ATPase activity and in vitro motility velocity with expression of the MHC head insert (SMB) HMM relative to its absence (SMA). Association of the MLC17 a or b isoforms with the HMM did not affect either measurement. Sweeney et al. (41) expressed chimeric myosin with variable sequences at loop 1 (region of SM MHC head isoforms) and found an approximately twofold increase in in vitro motility and actin-ATPase activity that correlates with increasing loop size/flexibility (from 0 to 28 amino acids). These chimeras were unaffected by the MLC17 isoform type with which they were associated.

Differences in unloaded shortening velocity reported for SM tissue strips show an ∼20-fold range (34). The more recent articles mentioned above report only an approximately twofold difference in in vitro motility attributable to the MHC head isoforms expressed, with no apparent additional changes attributable to MLC17 isoforms expressed. This discrepancy does not appear to be explainable by differences in MLC20 phosphorylation levels. This study was designed to examine the correlation between the MLC17 a and b isoforms and unloaded shortening velocity in single smooth muscle cells (SMCs). This system avoids complications due to cellular heterogeneity in SM tissue but allows the measurements to be made in an intact three-dimensional system where the intracellular milieu is similar to the in vivo situation.



MLC17 isoform nomenclature is confusing due to multiple naming schemes. The original usage of MLC17 a and b isoforms was based on the relative acidic and basic mobilities, respectively, of these two isoforms (5). Subsequent use of MLC17 nm and gi (or SM) for nonmuscle and gizzard (SM) MLC17, respectively (21), while true for those tissues, is confusing because 1) both the nm and gi (SM) are present in most SM tissues (2) and 2) use of the nm/SM nomenclature is not consistent relative to a/b. For example, Nabeshima et al. (35) reported that the chicken fibroblast (MLC17nm) has the inserted sequence that makes it the basic isoform (MLC17b), whereas Lenz et al. (28) reported that the inserted sequence is present in the human “heart aorta” cells (MLC17SM), which makes it the basic isoform (MLC17b). Hasegawa et al. (22) reported the amino acid sequence for the MLC17 a and b isoforms with the more acidic sequence being labeled the MLC17b isoform. Finally, Kelley et al. (25) describe monoclonal antibody production for the MLC17 a and b isoforms using peptides that are basic and acidic relative to each other. In the work reported in this paper, the MLC17 a and b nomenclature used is based on the acidic and basic cDNA sequence generated by the RT-PCR and the acidic and basic mobilities of the protein on two-dimensional gels, respectively.


New Zealand White rabbits were killed by CO2 asphyxiation. Tissue (aorta, bladder, carotid artery, stomach, and uterus) were removed; cleaned of blood and adipose, loose connective, and nonmuscle tissue (endometrium of the uterus, mucosa of the stomach, transitional epithelium of the bladder); and stored in physiological salt solution (in mM: 140 NaCl, 4.7 KCl, 1.2 Na2HPO4, 1.2 MgSO4, 1.6 CaCl2, 5.6 glucose, 0.02 EDTA, and 2.0 MOPS, pH 7.2) at 4°C until further processing.

Protein gel electrophoresis.

Tissues for protein analysis were homogenized in sample buffer [9 M urea, 4% Nonidet P-40, 2% ampholytes, and 20 mM dithiothreitol (DTT)] according to Malmqvist and Arner (30). MLCs were resolved on urea-glycerol-acrylamide gels (40) or by two-dimensional gel electrophoresis (12) and were visualized with Coomassie blue staining. Comparison of the one- and two-dimensional gels for the same tissue samples allowed identification of MLC17 a and b spots on the two-dimensional gels as the faster- and slower-migrating bands on the one-dimensional gel, respectively. For some experiments, the protein on the gels was transferred to nitrocellulose (4), and MLC17 a and b-specific antibodies [Sigma; gift from C. Kelley, National Institutes of Health (NIH)] were used to identify the bands of interest.

Tissue RNA extraction.

Total RNA was extracted from tissues using RNA STAT-60 (Tel-Test “B”). Samples for protein and RNA comparisons were collected from adjacent regions of the same tissues. Extracted RNA (1.0–1.5 μg) was added to the RT reaction mix [2 μl of 5× RT buffer (US Biochemical, Cleveland, OH): 20 μg/ml oligo(dT) (Promega, Madison, WI), 2 U/μl RNasin (Promega), 0.1 μg/μl acetylated BSA, 500 μM 2-deoxynucleotide-5′-triphosphate (dNTP), 1.25 mM DTT, and 10 U/μl Moloney murine leukemia virus-RT (US Biochemical)] to a final volume of 10 μl and was incubated at 37°C for 2 h.

Cell isolation.

Individual cells were isolated from carotid or aortic arteries using the method of Driska and Porter (6). Briefly, before digestion, the vessel is incubated in no-Ca2+ PSS (in mM: 140 NaCl, 4.7 KCl, 1.2 Na2HPO4, 2.4 MgSO4, 0.0 CaCl2, 5.6 glucose, 0.02 EDTA, and 2.0 MOPS, pH 7.4) for 15 min before digestion to reduce activation due to mechanical perturbation during isolation. The vessel was ballooned and bathed in a no-Ca2+ PSS solution containing papain (20 U/ml) and DTT (2 mM) at 37°C with 40 cmH2O head pressure applied to balloon the vessel. After 10 min of digestion, the vessel was removed from the enzyme solution and placed in fresh no-Ca2+ PSS. The vessel was bisected longitudinally, and thin strips of the medial layer were teased away, liberating single SMC. To produce cDNA from a single SMC, a single living cell was attached to a glass micropipette (2 μm tip) using a microscope and micromanipulator. The tip of the micropipette was broken into a clean microcentrifuge tube, and the cell was lysed by a freeze-thaw cycle. Ten microliters of RT mix were added to each microcentrifuge tube, vortexed briefly, and incubated at 37°C for 2 h.


Two oligonucleotide primers (corresponding to mouse macrophage nonmuscle myosin, nucleotide positions 471–452 and 195–214; Operon, Alameda, CA; Hailstones and Gunning, GeneBank VO4443) were used in the 50-μl PCR reaction [5 μl of Taq DNA polymerase 10× reaction buffer (Promega): 1.5 mM MgCl2, 200 μM dNTP, 2.0 μM 5′ primer, 2.0 μM 3′ primer, and 1 unitTaq polymerase (Promega)]. Two microliters of the RT reaction were used for cDNA amplification of tissue RNA samples. After a 2-min, 94°C hot start, the PCR temperature protocol was 94°C denaturation (90 s), 55°C annealing (2 min), and 72°C (3 min) primer extension for 35 cycles. An Ericomp thermoelectric thermocycler was used for these studies (USA/Scientific Plastics, Ocala, FL). These primers flank the 44-nucleotide exon that encodes the different region that defines the MLC17a and MLC17b isoforms (28, 35). The PCR products generated with the use of these primers are 231 and 275 bp, corresponding to MLC17a and MLC17b, respectively. Quantitation problems due to differential primer annealing are obviated by using a single pair of primers to generate both MLC17a- and MLC17b-specific PCR products. The MLC17b/a PCR primers flank the intervening sequence in the genomic DNA; therefore, only cDNA products generate appropriately sized PCR products.

For single-cell analysis, the entire RT reaction mixture (10 μl) was routinely used for PCR amplification (60 cycles). Both PCR fragments corresponding to the two MLC17 isoforms derived from rabbit uterus have been cloned into the pCRII vector utilizing the TA cloning protocol (Invitrogen, San Diego, CA). Plasmids were purified by selective adsorption to silica columns (Qiagen, Chatsworth, CA). These clones were sequenced with the use of the dideoxy chain termination method (38).

Single-cell contractions.

Mechanical measurements were performed in a flow-through chamber consisting of a depression slide with three inputs and one output. The concave depression on the slide was surrounded by a wax ring to prevent fluid dispersal during the experiment. The gravity-feed flow through the input tubes was monitored by inclusion of fast green dye (0.002% wt/vol) in the perfusion solutions. Outflow from the perfusion chamber was controlled by a peristaltic pump (Rabbit Miniplus 2; Rainin Instrument; Emeryville, CA). Output was matched to input by adjusting the speed of the peristaltic pump so that the chamber fluid meniscus remained at a constant level.

Single SMCs were isolated from vascular tissue (aorta and carotid) by the methods detailed in Cell isolation. The SMCs were transferred from coated glass isolation dishes (Sigma-cote; Sigma, St. Louis, MO; Slickcoat, Intermountain Scientific, Kaysville, UT) via wide-bore pipettes to the depression slides. Cells were allowed to settle to the bottom of the slide (5 min) before solution exchange from a no-Ca2+ PSS to relaxing buffer [pCa 9.0; relaxing buffer contained (in mM): 14.5 creatine phosphate, 7 EGTA, 20 imidazole, 1 free Mg2+, 4 free MgATP, 5.42 MgCl2, 79.16 KCl, and 4.74 ATP, and 16.33 μM CaCl2, pH 7.0]. Once cells were in the relaxing buffer, permeabilization of the SMC membrane was accomplished by a 7-min incubation in relaxing buffer plus 250 hemolytic U/ml staphylococcal α-toxin (3). After permeabilization, all SMCs were washed with relaxing buffer (2 min).

After the 2-min wash, a permeabilized SMC was selected on the basis of morphology (9) and birefringence of the cell membrane using phase-contrast microscopy. The cell was attached to a small-bore (1- to 3-μm tip) glass micropipette by close apposition of the pipette tip and gentle suction. Once attached, the SMC was lifted off the chamber bottom to ensure absence of adhesion. The cell was stimulated to contract with an activation cocktail consisting of 1 μM phenylephrine and 1 μM histamine in activating buffer [pCa 6.0; activating buffer contained (in mM): 14.5 creatine phosphate, 7 EGTA, 20 imidazole, 1 free Mg2+, 4 free MgATP, 5.31 MgCl2, 68.55 KCl, 4.79 ATP, and 4.91 CaCl2, pH 7.0]. Control experiments included 1 μM okadaic acid [a MLC phosphatase inhibitor (3)] in the activating solution and gave similar results as when it was not present.

Cell length data acquisition.

Before and throughout the single SMC contraction, cell length was continuously monitored by a charge-coupled device (CCD) camera (XC-75; Sony, Park Ridge, NJ) attached to an inverted microscope (Fluovert; Leitz, Overland Park, KS). The output from the CCD camera was recorded by a video recorder (EV-S3000; Sony) on 8-mm videotape. SMC images from the 8-mm tape were captured by a frame grabber (Data Translation, Marlborough, MA) in conjunction with NIH Image software (NIH, Bethesda, MD) at regular intervals encompassing the SMC contraction. Image analyses were performed using NIH Image software (version 1.59). Each individual frame of the cell from successive time points during the contraction was projected on the monitor. The ends of the cell were identified on the frame. With the use of the mouse and starting at one end of the cell, the midpoint of the cell's width was marked along its entire length at 5- to 10-μm intervals. The length of the line connecting these points gave total cell length for each frame. This method gives an accurate measurement of cell length independent of its linearity (33).

Untethered velocity measurements from single SMCs.

Video images from the CCD camera were recorded on a videocassette recorder, digitized, imported into a personal computer, and analyzed with respect to length over a discrete time frame using NIH Image software (version 1.59). Of the 93 cells examined, 67 exhibited the following criteria: 1) characteristic morphology of contracted SMCs (i.e., evaginations in cell membrane postcontraction; see Ref. 9) and 2) a smooth sigmoidal decrease in length over time, indicating unconstrained contraction within the plane of focus of the microscope. Unloaded shortening velocity was taken from the steepest slope of a plot of length vs. time for these 67 cells. After its contraction, each cell from which mechanical measurements were made was transferred to a microcentrifuge tube so that the MLC17b/amRNA ratio could be determined. Both the mechanical measurements and the RT-PCR were performed on the same single cell, thus allowing possible correlations between them to be determined.

Electrophoresis and densitometry.

PCR products were separated by electrophoresis on 8% polyacrylamide gels (13) and were visualized under ultraviolet light after staining with ethidium bromide (1 μg/ml). Band intensity was quantified by image analysis of gel photographs using Ambis 2000 software (Ambis, San Diego, CA). Extensive controls for the validity of these methods for relative quantitation of the RT-PCR products have been reported previously (32).


Comparisons between sample means were tested for significance by a two-tailed t-test, and significance of the correlation coefficient from zero was tested for by a t-test with two degrees of freedom.


To ascertain potential functional roles for the MLC17 a and b isoforms, a set of experiments was carried out to determine the relative MLC17 a and b isoform content in single cells, and the results were compared with some of the mechanical and morphological properties of those same cells. Primers were chosen to span the alternatively spliced exon present in the MLC17 mRNA. The resulting products from RT-PCR, using these primers, were 231 and 275 bp (a and b, respectively). RT-PCR was used to estimate the relative content of the MLC17 a and b proteins at the single-cell level because we are unaware of any procedure to perform direct quantitation of these proteins at the cellular level. To validate the use of the MLC17b/a mRNA relative ratios to predict MLC17b/a protein relative ratios, matched tissue samples from adjacent areas of five tissues (aorta, bladder, carotid, stomach, and uterus) were used to determine the relative mRNA ratios and relative protein ratios. Forty-five pairs of adjacent SM tissue samples from 13 different rabbits were compared. RT-PCR and protein extraction, as described in methods, were performed on adjacent regions of the same tissue. A representative pair of gels is shown in Fig.1. The relative ratios of the MLC17b/a mRNA and MLC17b/a protein, as determined by densitometric analysis of these samples run on standard polyacrylamide or urea-glycerol gels, respectively, were plotted against one another as shown in Fig. 2. TheR 2 value for the least-square fit regression line is 0.68.

Fig. 1.

Myosin light chain (MLC) 17 a and b mRNA and protein profiles from rabbit smooth muscle (SM) tissues. Adjacent regions of SM tissues were analyzed with respect to relative expression levels of MLC17 a and b mRNA and protein. A: polyacrylamide gel of products from RT-PCR (as described in methods) on total RNA extracted from rabbit aorta, bladder, carotid, stomach, and uterus (lanes 3–7, respectively). Lane 1 contains 123-bp DNA ladder; 123- to 492-bp bands are shown. Lane 2 is a negative control (RNA was omitted from the RT-PCR reaction). B: urea-glycerol gel of proteins extracted from adjacent regions of same tissue as in A. Lanes 1–5 correspond to lanes 3–7 in A. LC, light chain.

Fig. 2.

Correlation of MLC17b/a mRNA and protein ratios. A linear regression of 45 pairs of adjacent SM tissue samples from 13 different rabbits is shown (R 2 = 0.68). RT-PCR and protein extraction were performed on adjacent regions of the same tissue. The relative ratios of MLC17b/MLC17amRNA and proteins as determined by densitometric analysis were plotted against one another to determine the correlation between SM MLC17 mRNA and proteins. *, Aorta; ▵, bladder; ⋄, carotid; ○, stomach; +, uterus.

The rabbit MLC17b amino acid sequence for the carboxy end of the molecule was deduced from the cDNA sequence of the PCR products. The rabbit MLC17b amino acid sequence is unique from the reported swine aortic MLC17b by a change of leucine 116 to an asparagine. The rabbit MLC17b amino acid sequence differs from the reported mouse MLC17nm molecule by a change of aspartate 82 to an alanine. The rabbit MLC17bsequence also differs from the chicken fibroblast MLC17nmby changing valine 77, threonine 85, tyrosine 86, and methionine 127, to isoleucine, cysteine, phenylalanine, and glutamine, respectively. The rabbit MLC17b amino acid sequence differs from the rabbit MLC17a sequence in five of the last nine carboxy-terminal amino acids. The amino acids (alanine 143, phenylalanine 144, histidine 147, isoleucine 148, and serine 150) in the rabbit MLC17b are glutamate, leucine, methionine, valine, and asparagine in the rabbit MLC17a isoform.

Figure 3 A shows the frequency distribution for the MLC17b/a ratio for 46 carotid and 28 aortic SMCs. The mean MLC17b/a ratio for the cells from the carotid is 0.50 ± 0.03 and 0.46 ± 0.06 for the cells from the aorta. The means for these two sets of arterial SMCs are not significantly different, but because they are different arteries, they are analyzed separately. The range of MLC17b/a mRNA ratios for these single cells ranges from ∼0 to 1, whereas this same ratio only ranges from ∼0.5 to 0.7 in the tissues samples. Thus the range of MLC17b/a mRNA ratios is approximately five times greater for single cells from the carotid or aorta than it is for tissue samples from these two arteries.

Fig. 3.

Histogram distribution of isolated rabbit SM aorta (open bars) and carotid (filled bars) cells. A: distribution of MLC17b/a mRNA ratio. B: distribution of initial isolated cell length. C: distribution of isolated cell length after contraction. D: distribution of maximal unloaded shortening velocity of isolated untethered cells.

Mechanical measurements were made on 48 carotid and 19 aorta cells. Figure 3 B shows the frequency distribution of initial cell lengths at the start of each experiment. Cell lengths for both of these arteries ranged from ∼75 to 275 μm, with the mean carotid cell length being significantly longer than the aorta cell length [carotid 168.9 ± 1.2 (SE) μm, aorta 129.0 ± 2.6 μm,P ≤ 0.01]. After contraction of the cells, these cells were all significantly shorter, with their range of lengths being significantly less variable [Fig. 3 C; carotid 58.8 ± 0.3 (SE) μm, aorta 56.2 ± 0.8 μm]. There was no significant difference between final cell length after contraction for cells from the aorta vs. the carotid. Cell unloaded shortening velocity was determined for each cell and ranged from ∼0.5 to 7.5 μm/s (Fig.3 D). The cells from the carotid had unloaded shortening velocities significantly faster than those from the aorta [carotid 3.31 ± 0.04 (SE) μm/s, aorta 2.0 ± 0.08 μm/s,P ≤ 0.01]. Figure4 shows the contraction of a typical cell at four successive time points in response to activation. Final cell length was determined after no significant decrease was observed in cell length for at least 10 s at the end of the contraction.

Fig. 4.

Contraction of a single smooth muscle cell (SMC). A temporal series of digitized images of a permeabilized SMC contracting in response to exposure to an activating solution containing Ca2+ (pCa 6.0), phenylephrine (1 μm), and histamine (1 μm). Cell was suspended from a micropipette (top right in 1–4) and was not in contact with the bottom of the chamber. Frame 1: contraction time 0 s, cell length 198.63 μm; frame 2: contraction time 15 s, cell length 134.31 μm; frame 3: contraction time 30 s, cell length 70.45 μm; frame 4: contraction time 45 s, cell length 58.32 μm.

Figure 5 A shows the relationship between the initial cell length and final cell length recorded before and after cell shortening. Both the carotid and aorta cells show a significant correlation (P ≤ 0.01) for these two values. Figure 5 B shows that there is also a significant correlation (P ≤ 0.01) for both the carotid and aorta cells between initial cell length and unloaded shortening velocity. As initial cell length increases, the unloaded shortening velocity increases. The unloaded shortening velocity does not correlate with final cell length for the carotid cells but does show a weak, but significant (P≤ 0.05), correlation for the aorta cells (Fig. 5 C). Figure5 D shows that there is a significant correlation between total cell length change (initial length − final length) and shortening velocity for both cell types.

Fig. 5.

Velocity and length changes in α-toxin-permeabilized SMCs during unloaded shortening. SMCs were isolated from rabbit vasculature [aorta (open symbols) or carotid (closed symbols)].A: correlation between cell final length after activation and initial length before activation (P ≤ 0.01 for both cell types); carotid R 2 = 0.59; aortaR 2 = 0.68. B: correlation between maximal unloaded shortening velocity and initial cell length before activation (P ≤ 0.01 for both cell types); carotidR 2 = 0.39; aorta R 2 = 0.57.C: correlation between maximal unloaded shortening velocity and final cell length after activation (P ≤ 0.05 for aorta cells); carotid R 2 = 0.06; aorta R 2 = 0.27. D: correlation between total cell length change and maximal unloaded shortening velocity (P ≤ 0.01 for both cell types); carotid R 2 = 0.47; aortaR 2 = 0.60.

For 33 carotid cells and 16 aorta cells, the MLC17b/a ratio is not correlated with the initial cell length (Fig.6 A) or final cell length (Fig.6 B). The total length change that the cell undergoes during an unloaded shortening does not correlate with its MLC17b/aratio (Fig. 6 C). The MLC17b/a ratio is also not correlated with unloaded shortening velocity in the 49 cells that were tested (Fig. 6 D).

Fig. 6.

Velocity and length changes in α-toxin-permeabilized SMCs during unloaded shortening and their correlation with their MLC17b/a ratio. SMCs were isolated from rabbit vasculature [aorta (open symbols) or carotid (closed symbols)].A-C: correlation between single-cell MLC17b/a ratio and initial cell length before contraction (A; NS, carotidR 2 = 0.00, aorta R 2 = 0.01), final cell length after activation (B; NS, carotidR 2 = 0.019, aorta R 2 = 0.00), and total cell length change during contraction (C; NS, carotidR 2 = 0.00, aorta R 2 = 0.01). D: maximal unloaded shortening velocity with the MLC17b/a ratio in each cell; NS, carotidR 2 = 0.01, aorta R 2 = 0.02. All correlations not significant (NS).

Because of the wide range of initial cell lengths (Fig. 3 B), maximum unloaded shortening velocity was also normalized to initial cell length. This would correct for differences that would be attributable to a variable number of contractile units in series. Figure 7 A shows a histogram of shortening velocity (cell lengths/s) vs. cell number. This normalization decreases the range of shortening velocities by a factor of three (0.5–7.5 μm/s, Fig. 3 D; 0.01–0.05 cell lengths/s, Fig. 7 A). There is no correlation between initial cell length (Fig. 7 B), final cell length (Fig. 7 C), total length change (Fig. 7 D), or MLC17b/a ratio (Fig. 7 E) of the single cells examined with normalized unloaded shortening velocity (cell lengths/s).

Fig. 7.

A: histogram of normalized shortening velocity (cell lengths/s). Filled bars, carotid; open bars, aorta. B-D: velocity and length changes in α-toxin-permeabilized SMCs during unloaded shortening. SMCs isolated from rabbit vasculature (aorta, open symbols) or carotid (closed symbols). B: correlation between maximum normalized unloaded shortening velocity (cell lengths/s) and initial cell length before activation [not significant (NS) for both cell types; carotid R 2 = 0.04, aortaR 2 = 0.11]. C: correlation between maximum normalized unloaded shortening velocity (cell lengths/s) and final cell length after activation (NS for both cell types; carotidR 2 = 0.02, aorta R 2 = 0.01).D: correlation between maximal normalized unloaded shortening velocity (cell lengths/s) and total cell length change (NS for both cell types; carotid R 2 = 0.08, aortaR 2 = 0.16). E: correlation between maximum normalized unloaded shortening velocity (cell lengths/s) and MLC17b/a ratio (NS for both cell types; carotidR 2 = 0.006, aorta R 2= 0.007).

To address concerns about possible variable phosphorylation levels between individual cells, 13 additional carotid cells were tested, and okadaic acid (1 μM) was included in the activating solution. The mean shortening velocity of these 13 cells was 0.016 ± 0.001 cell lengths/s (Fig. 8 A), which is not significantly different from the carotid (≤0.019 ± 0.000 cell lengths/s, P ≥ 0.1, Fig. 7 A) or the aortic (≤0.015 ± 0.000 cell lengths/s, P ≥ 0.1) cells tested without the use of okadaic acid. When single-cell unloaded shortening velocity (cell lengths/s, in the presence of okadaic acid) is plotted against the MLC17b/a ratio for each respective cell, there is no correlation (Fig. 8 B). This is consistent with the lack of correlation observed when okadaic acid was not present (Fig.7 E).

Fig. 8.

A: histogram of normalized shortening velocity (cell lengths/s) for 13 carotid cells activated in the presence of 1 μM okadaic acid. Mean = 0.016 ± 0.001. B: correlation between maximal normalized shortening velocity and MLC17b/a ratio for cells activated in the presence of 1 μM okadaic acid (NS;R 2 = 0.00).


SM protein isoforms may have unique functional roles in SMCs. The SM myosin subunit isoforms have been hypothesized to render unique physiological properties to the cell. This study was designed to address the hypothesis that the SM MLC17 a and b isoforms have unique functional roles within SMCs. As mentioned in the introduction, there are numerous studies reported in the literature suggesting that the SM MLC17 a and b isoforms are correlated with specific mechanical properties in SM tissues or in isolated protein preparations (20, 23, 30). In contrast to this, there are also numerous reports suggesting that this correlation is only casual and that the differences observed are due to the SMA and SMB isoforms (11, 25, 37, 41).

Using an experimental protocol to exchange MLC17a for MLC17b in trifluoperazine-treated rabbit bladder strips, Matthew et al. (31) reported that shortening velocity and rate of force development are significantly increased with no MLC20 phosphorylation. They suggest that the MLC17 isoforms may stiffen the lever arm of myosin, modulating MgADP affinity and cooperative nonphosphorylated cross-bridge cycling. However, this is a nonphysiological situation as most, if not all, studies measuring myosin phosphorylation report “basal” levels during relaxed conditions and transient elevated phosphorylation levels during contraction. Our study reports no correlation between the MLC17b/a ratio and unloaded shortening velocity (absolute or normalized) in isolated permeabilized single vascular SMCs. These results would not be consistent with those reported by Matthew et al. (31) if their results also applied to phosphorylated cross bridges, which our experiments are designed to maximize.

As has been suggested (2, 7, 32, 37), the three-dimensional organization of the proteins in the SMC may be critical in determining the functional significance of protein isoforms. Thus in vitro studies may not resolve this question. In addition, numerous studies reporting intercellular variability with regard to various protein isoform expression (7, 8, 10, 14, 24, 27, 29, 32, 39, 45) led to the conclusion that multicellular tissue studies may not be appropriate for protein isoform function studies, as intercellular variability would be averaged across the population, prohibiting resolution of causal relationships. The use of isolated single cells allows measurements to be made with the three-dimensional organization of the cell intact while eliminating the potential problem of intercellular variability. In this study, the possibility of a unique function, with respect to unloaded shortening velocity, for the MLC17 a and b isoforms was tested using isolated single SMCs. Both the analysis of the cellular protein composition (via mRNA measurements) and mechanical properties are made on the same cell. We have shown previously that this is an effective, reliable, and accurate method for testing these hypotheses (32, 33).

RT-PCR is used to provide relative quantitation of the mRNA levels of the MLC17b a and b isoforms. Primers were chosen to span the unique exon present in the pre-mRNA for these proteins that results from alternative splicing of a single gene product (28, 35). The use of a single set of primers that bind to the exact same location on both mRNAs results in two different-sized products whose final concentration is a function of the initial concentration of each of these messages. This allows for relative quantitation of the mRNA at the single-cell level, providing the sensitivity required to detect the small amount of message present at the single-cell level (32). However, the MLC17 a and b proteins and not their mRNAs are what would be responsible for any possible isoform-specific function. Therefore, the predictive power of the MLC17 a and b mRNA levels for the MLC17 a and b protein levels was determined (Figs. 1and 2). The R 2 value for this relationship is 0.68, suggesting that the MLC17b/a mRNA ratio is a good indicator of the MLC17b/a protein ratio. Thus RT-PCR amplification of the MLC17 isoform mRNA allows for quantitation of the MLC17b/a mRNA ratios, which can be used as an indicator of the MLC17b/a protein ratios (32).

Isolated cells were analyzed throughout the experiment for accurate interpretation of the data. Carotid and aortic cells were analyzed separately so that potential differences between cells from these two vessels could be detected. Both the carotid and aortic cells had similar mRNA isoform ratios of 0.50 ± 0.03 and 0.46 ± 0.06 (SE), respectively. For the tissue samples, the carotid and aorta had mRNA isoform ratios of 0.66 ± 0.03 and 0.64 ± 0.03, respectively. The difference between the isolated cell average and the tissue average may be due to the limited number of cells analyzed. It may also, however, be indicative of unique distributions of cells expressing varying amounts of each of these isoforms. Using the “balloon” digestion method of Driska and Porter (6), there may be a disproportionate number of cells from the intimal side of the medial wall. We are currently performing experiments to test this hypothesis.

The carotid cells had significantly longer initial cell lengths than the aorta cells [169 ± 0.15 vs. 129 ± 11.5 (SE) μm]. This significant difference, although small, was not expected. In a separate study of cell lengths, these two tissues had similar-length cells (unpublished data). Some of this difference in cell lengths from these two arteries may be due to cell selection. The balloon digestion method used in this study would give a disproportionate number of cells from the luminal side of the vessel. Isolated individual cell lengths after contraction had similar final cell lengths (59 ± 2.1 vs. 56 ± 3.7 μm). Even though the initial cell lengths ranged from 70 to 279 μm, the final cell lengths only ranged from 33 to 88 μm. The very narrow range of final lengths for both populations of cells may indicate some unique final shortening length and is consistent with our previous work (33). There is a significant (P ≤ 0.01) correlation between initial and final cell lengths for both aorta and carotid cell populations.

Cell length-dependent effects on force generation that are independent of the length-tension relationship due to contractile filament organization have been reported by several research groups. Gunst et al. (15) have proposed that the organization of the cell cytoskeleton (which is dependent on cell length) is more “fluid” in a relaxed tissue than in an activated state and thus can reset itself more rapidly to optimize subsequent contractions. Hai (16) and Szeto and Hai (42) have reported that there are cell-length-sensitive mechanisms of cell depolarization and cytosolic Ca2+concentration that minimally result in length-dependent modulation of myosin phosphorylation. Harris and Warshaw (18) have proposed that the slowing of shortening velocity during an isotonic shortening in single isolated SMCs can be explained by an internal load, which is consistent with the model of Gunst et al. (15). Harris and Warshaw (19) have also reported that stretching isolated toad stomach cells beyond the optimum length shows no “descending limb” as in the traditional length-tension relationship. Our data showing a significant relationship between initial cell length and absolute shortening velocity (Fig. 5 B) are consistent with the idea of length-dependent modulation of activation. The mechanism for this, however, is unknown, as the cells are α-toxin permeabilized and are activated maximally with 1 μM Ca2+, 1 μM histamine, and 1 μM phenylephrine.

The absolute unloaded shortening velocity was significantly faster in the carotid cells than the aortic cells (3.3 ± 0.29 vs. 2.0 ± 0.34 μm/s; P ≤ 0.01, Fig. 3 D), but not when normalized to initial cell lengths (Fig. 7 A). This is consistent with the results that the carotid cell initial lengths were significantly longer and that there is a near-perfect correlation between the absolute cell length change and the initial length of the cell (R 2 > 0.94). Additionally, there is a significant correlation between initial cell length and absolute shortening velocity for both sets of cells (P ≤ 0.02, Fig. 5 B), but this is not significant when using normalized shortening velocity (Fig. 7 B).

These differences between carotid and aortic cell lengths and shortening velocities do not correlate with their MLC17b/aisoform ratio. There are a number of other factors that could affect shortening velocity. Differences in MLC20 phosphorylation levels would cause differences in shortening velocity. This seems unlikely in this study, as micromolar concentrations of two agonists and micromolar free Ca2+ concentration were used to activate the cells, and the presence or absence of okadaic acid did not alter the results in previous work (33). In this study, 13 additional cells were activated similar to all the other cells, but with the addition of 1 μM okadaic acid in the activating solution. Okadaic acid is reported to be an inhibitor of MLC phosphatase (33) and should increase MLC phosphorylation if it is not maximally phosphorylated. This would result in an increased maximum unloaded shortening velocity in these cells. This was not the case. The cells activated in the presence of 1 μM okadaic acid (0.016 ± 0.01 cell lengths/s; Fig.8 A) had maximal unloaded shortening velocities not significantly different from the cells activated without okadaic acid present (carotid: 0.019 ± 0.000 cell lengths/s; aorta: 0.015 muscle lengths/s; Fig. 7 A). In addition, reduced phosphatase activity at room temperature (21°C) is consistent with reduced “latch” cross bridges; thus, velocity should reflect myosin isoform differences much more directly than intact cells or tissue studies at 37°C. The differences in initial cell length may suggest differences in internal loading that could affect shortening velocity. However, when the shortening velocities are normalized to initial cell lengths, there is no correlation with initial cell length, final cell length, or total length change (Fig. 7, B-D). There is no significant correlation between initial cell length and shortening velocity (cell lengths/s; Fig. 7 B) and a very narrow range of shortening velocities for these cells (cell lengths/s; Fig.7 A) with a fourfold range of initial cell lengths (70–280 μm). This is inconsistent with the idea that there is a single narrow “optimum” initial cell length for these cells but indicates that all of these cells are equally able to shorten. Thus internal loading due to partial contraction of cells during isolation or variable phosphorylation is inconsistent with the shortening velocities data reported here.

Neither initial cell length, final cell length, absolute cell length change, nor unloaded shortening velocity correlates with the MLC17b/a mRNA isoform ratio in single cells (Fig. 6). This lack of correlation between unloaded shortening velocity and the MLC17b/a isoform ratio is consistent with those reports suggesting that the MLC17b/a isoform ratio does not affect unloaded shortening velocity (25, 37, 41). In this study, both the MLC17b/a isoform ratio and the unloaded shortening velocity were made in the same single isolated cell, where the three-dimensional organization of the contractile proteins is intact. In addition, because all measurements are made on each individual cell, there is no problem with intercellular variability for each measurement. The lack of correlation reported in this study between the MLC17 a and b isoforms and unloaded shortening velocity implies that there is a different cause for the ∼20-fold variation in tissue shortening velocities (34). We have previously reported that the MHC tail isoforms (SM1 and SM2) also do not correlate with unloaded single-cell shortening. Further work is necessary to identify the proteins and/or pathways that account for these differences in cell shortening velocities.


We thank Barbara DeNoyer for secretarial assistance and Joseph Golden for measuring cells.


  • Address for reprint requests and other correspondence: T. J. Eddinger, Dept. of Biology, Marquette Univ., P.O. Box 1881, WLS 109, Milwaukee, WI 53201-1881 (E-mail: eddingert{at}

  • This work was supported by National Heart, Lung, and Blood Institute Grant HL-62237-01 and National Arthritis and Musculoskeletal and Skin Diseases Institute Grant R15-AR-45294–01 and American Heart Association Grant-in-Aid no. 96066890.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract